Reverse actin sliding triggers strong myosin binding that moves tropomyosin.

Actin/myosin interactions in vertebrate striated muscles are believed to be regulated by the "steric blocking" mechanism whereby the binding of calcium to the troponin complex allows tropomyosin (TM) to change position on actin, acting as a molecular switch that blocks or allows myosin heads to interact with actin. Movement of TM during activation is initiated by interaction of Ca(2+) with troponin, then completed by further displacement by strong binding cross-bridges. We report x-ray evidence that TM in insect flight muscle (IFM) moves in a manner consistent with the steric blocking mechanism. We find that both isometric contraction, at high [Ca(2+)], and stretch activation, at lower [Ca(2+)], develop similarly high x-ray intensities on the IFM fourth actin layer line because of TM movement, coinciding with x-ray signals of strong-binding cross-bridge attachment to helically favored "actin target zones." Vanadate (Vi), a phosphate analog that inhibits active cross-bridge cycling, abolishes all active force in IFM, allowing high [Ca(2+)] to elicit initial TM movement without cross-bridge attachment or other changes from relaxed structure. However, when stretched in high [Ca(2+)], Vi-"paralyzed" fibers produce force substantially above passive response at pCa approximately 9, concurrent with full conversion from resting to active x-ray pattern, including x-ray signals of cross-bridge strong-binding and TM movement. This argues that myosin heads can be recruited as strong-binding "brakes" by backward-sliding, calcium-activated thin filaments, and are as effective in moving TM as actively force-producing cross-bridges. Such recruitment of myosin as brakes may be the major mechanism resisting extension during lengthening contractions.

An unexpectedly large working stroke from chymotryptic fragments of myosin II.

Recent structural evidence indicates that the light chain domain of the myosin head (LCD) bends on the motor domain (MD) to move actin. Structural models usually assume that the actin-MD interface remains static and the possibility that part of the myosin working stroke might be produced by rotation about the acto-myosin interface has been neglected. We have used an optical trap to measure the movement produced by proteolytically shortened single rabbit skeletal muscle myosin heads (S-1(A1) and S-1(A2)). The working stroke produced by these shortened heads was more than that which the MD-LCD bend mechanism predicts from the full-length (papain) S-1's working stroke obtained under similar conditions. This result indicates that part of the working stroke may be caused by motor action at the actin-MD interface.

Tomographic 3D reconstruction of quick-frozen, Ca2+-activated contracting insect flight muscle.

Motor actions of myosin were directly visualized by electron tomography of insect flight muscle quick-frozen during contraction. In 3D images, active cross-bridges are usually single myosin heads, bound preferentially to actin target zones sited midway between troponins. Active attached bridges (approximately 30% of all heads) depart markedly in axial and azimuthal angles from Rayment's rigor acto-S1 model, one-third requiring motor domain (MD) tilting on actin, and two-thirds keeping rigor contact with actin while the light chain domain (LCD) tilts axially from approximately 105 degrees to approximately 70 degrees. The results suggest the MD tilts and slews on actin from weak to strong binding, followed by swinging of the LCD through an approximately 35 degrees axial angle, giving an approximately 13 nm interaction distance and an approximately 4-6 nm working stroke.

X-ray diffraction indicates that active cross-bridges bind to actin target zones in insect flight muscle.

We report the first time-resolved study of the two-dimensional x-ray diffraction pattern during active contraction in insect flight muscle (IFM). Activation of demembranated Lethocerus IFM was triggered by 1.5-2.5% step stretches (risetime 10 ms; held for 1.5 s) giving delayed active tension that peaked at 100-200 ms. Bundles of 8-12 fibers were stretch-activated on SRS synchrotron x-ray beamline 16.1, and time-resolved changes in diffraction were monitored with a SRS 2-D multiwire detector. As active tension rose, the 14.5- and 7.2-nm meridionals fell, the first row line dropped at the 38.7 nm layer line while gaining a new peak at 19.3 nm, and three outer peaks on the 38.7-nm layer line rose. The first row line changes suggest restricted binding of active myosin heads to the helically preferred region in each actin target zone, where, in rigor, two-headed lead bridges bind, midway between troponin bulges that repeat every 38.7 nm. Halving this troponin repeat by binding of single active heads explains the intensity rise at 19.3 nm being coupled to a loss at 38.7 nm. The meridional changes signal movement of at least 30% of all myosin heads away from their axially ordered positions on the myosin helix. The 38.7- and 19.3-nm layer line changes signal stereoselective attachment of 7-23% of the myosin heads to the actin helix, although with too little ordering at 6-nm resolution to affect the 5.9-nm actin layer line. We conclude that stretch-activated tension of IFM is produced by cross-bridges that bind to rigor's lead-bridge target zones, comprising < or = 1/3 of the 75-80% that attach in rigor.

Tomographic three-dimensional reconstruction of insect flight muscle partially relaxed by AMPPNP and ethylene glycol.

Rigor insect flight muscle (IFM) can be relaxed without ATP by increasing ethylene glycol concentration in the presence of adenosine 5'-[beta'gamma- imido]triphosphate (AMPPNP). Fibers poised at a critical glycol concentration retain rigor stiffness but support no sustained tension ("glycol-stiff state"). This suggests that many crossbridges are weakly attached to actin, possibly at the beginning of the power stroke. Unaveraged three-dimensional tomograms of "glycol-stiff" sarcomeres show crossbridges large enough to contain only a single myosin head, originating from dense collars every 14.5 nm. Crossbridges with an average 90 degrees axial angle contact actin midway between troponin subunits, which identifies the actin azimuth in each 38.7-nm period, in the same region as the actin target zone of the 45 degrees angled rigor lead bridges. These 90 degrees "target zone" bridges originate from the thick filament and approach actin at azimuthal angles similar to rigor lead bridges. Another class of glycol-PNP crossbridge binds outside the rigor actin target zone. These "nontarget zone" bridges display irregular forms and vary widely in axial and azimuthal attachment angles. Fitting the acto-myosin subfragment 1 atomic structure into the tomogram reveals that 90 degrees target zone bridges share with rigor a similar contact interface with actin, while nontarget crossbridges have variable contact interfaces. This suggests that target zone bridges interact specifically with actin, while nontarget zone bridges may not. Target zone bridges constitute only approximately 25% of the myosin heads, implying that both specific and nonspecific attachments contribute to the high stiffness. The 90 degrees target zone bridges may represent a preforce attachment that produces force by rotation of the motor domain over actin, possibly independent of the regulatory domain movements.

Electron tomography of insect flight muscle in rigor and AMPPNP at 23 degrees C.

Treatment of rigor fibers of insect flight muscle (IFM) with AMPPNP at 23 degrees C causes a 70% drop in tension with little change in stiffness. In order to visualize the changes in crossbridge conformation and distribution that give rise to the mechanical response, we have produced three-dimensional reconstructions by tomography of both rigor and AMPPNP-treated muscle that do not average the repeating motifs of crossbridges, and thereby retain information on variability of crossbridge structure and distribution. Tomograms can be averaged when display of only the regular features is wanted. Tomograms of rigor IFM show double-headed lead and single-headed rear crossbridges. Tomograms of IFM treated with AMPPNP at 23 degrees C reveal many double-headed and some single-headed "lead" bridges but few crossbridges corresponding to the rear bridges of rigor. Instead, new non-rigor forms of variably angled crossbridges are found bound to actin sites not labeled with myosin heads in rigor. This indicates that the rear bridges of rigor have redistributed during the transition from rigor to the AMPPNP state, which could explain the maintenance of rigor stiffness despite the loss of tension. Comparison of in situ crossbridges in tomograms of rigor with atomic model of acto-S1, the complex formed by myosin subfragment 1 and actin, reveals that the regulatory domain of S1 would require significant bending and realignment to fit into both types of rigor crossbridges. The modifications are particularly significant for the rear bridges and suggest that differential strain in the regulatory domain of rear bridges may be the basis for their detachment and redistribution upon binding AMPPNP. Similar comparison using lead-type crossbridges in AMPPNP reveals departures from the rigor acto-S1 atomic model that include azimuthal straightening and a slight M-ward bending in the regulatory domain. Both the motor and regulatory domains of the new non-rigor crossbridges differ from those in the atomic model of acto-S1. A new crossbridge motif identified in AMPPNP-treated muscle consists of paired rigor-like and non-rigor crossbridges and suggests possible transitions in the myosin working stroke.

Three-dimensional structure of nucleotide-bearing crossbridges in situ: oblique section reconstruction of insect flight muscle in AMPPNP at 23 degrees C.

We have explored the three-dimensional structure of myosin crossbridges in situ in order to define the structural changes that occur when nucleotide binds to the myosin motor. When AMPPNP binds to rigor insect flight muscle, each half sarcomere lengthens by approximately 2.0 nm and tension is reduced by approximately 70% without a reduction in stiffness, suggesting partial reversal of the power stroke. We have obtained averaged oblique section three-dimensional reconstructions of mechanically monitored insect flight muscle in AMPPNP that permit simultaneous examination of all myosin crossbridges within the unit cell and direct comparison of calculated transforms with X-ray diagrams of the native fibers. Transforms calculated from the oblique section reconstruction of AMPPNP insect flight muscle at 23 degrees C show good agreement with native X-ray diagrams, suggesting that the average crossbridge forms in the reconstruction reflect the native structure. In contrast to the rigor lead and rear crossbridges in the double chevrons, the averaged reconstruction of AMPPNP fibers show only one crossbridge class, in the position of the rigor lead bridge. The portion of the crossbridge close to the thick filament appears broader than in rigor, and shows a small 0.5 to 1.0 nm M-ward shift of the regulatory domain region of myosin. In transverse view, AMPPNP "lead" crossbridges are less azimuthally bent than rigor. Fitting the atomic model of actomyosin subfragment 1 to the averaged crossbridges shows that the detectable differences between rigor bridges and between rigor and AMPPNP bridges occur in the alignment and angles of the regulatory domains and suggests that rear bridges are more strained than lead crossbridges. The apparent absence of rear bridges in AMPPNP in averaged reconstructions indicates detachment of a number of force-bearing bridges, which conflicts with the maintained stiffness of the fibers used for the reconstruction. This conflict may be explained if rigor rear bridges become distributed irregularly over more actin sites in AMPPNP, so that their average density is too low to appear in the averaged reconstructions. The reconstructions indicate that in insect flight muscle the response of in situ rigor crossbridges to AMPPNP binding is not uniform. Lead bridges persist but have altered structure in the light chain domain, whereas rear bridges detach and possibly redistribute. Shape changes in attached myosin heads within the myofibrillar lattice are in the appropriate direction and of the appropriate magnitude needed to explain the sarcomere lengthening. This could be a direct response to nucleotide binding, a passive response to rear bridge detachment, or a combination of both.

Movement and force produced by a single myosin head.

Muscle contraction is driven by the cyclical interaction of myosin with actin, coupled to the breakdown of ATP. Studies of the interaction of filamentous myosin and of a double-headed proteolytic fragment, heavy meromyosin (HMM), with actin have demonstrated discrete mechanical events, arising from stochastic interaction of single myosin molecules with actin. Here we show, using an optical-tweezers transducer, that a single myosin subfragment-1 (S1), which is a single myosin head, can act as an independent generator of force and movement. Our analysis accounts for the broad distribution of displacement amplitudes observed, and indicates that the underlying movement (working stroke) produced by a single acto-S1 interaction is approximately 4 nm, considerably shorter than previous estimates but consistent with structural data. We measure the average force generated by S1 or HMM to be at least 1.7 pN under isometric conditions.

Single-molecule mechanics of heavy meromyosin and S1 interacting with rabbit or Drosophila actins using optical tweezers.

Single-molecule mechanical interactions between rabbit heavy meromyosin (HMM) or subfragment 1 (S1) and rabbit actin were measured with an optical tweezers piconewton, nanometer transducer. Similar intermittent interactions were observed with HMM and S1. The mean magnitude of the single interaction isotonic displacements was 20 nm for HMM and 15 nm with S1. The mean value of the force of single-molecule interactions was 1.8 pN for HMM and 1.7 pN with S1. The stiffness of myosin S1 was determined by applying a sinusoidal length change to the thin filament and measuring the corresponding force; the mean stiffness was 0.13 pN nm-1. By moving an actin filament over a long distance past an isolated S1 head, we found that cross-bridge attachment occurred preferentially at a periodicity of about 40 nm, similar to that of the actin helical repeat. Rate constants for the probability of detachment of HMM from actin were determined from histograms of the lifetime of the attached state. This gave a value of 8 s-1 or 0.8 x 10(6) M-1 s-1 for binding of ATP to the rigor complex. We conclude (1) that our HMM-actin interactions involve just one head, (2) that compliance of the cross-bridge is not in myosin subfragment 2, although we cannot say to what extent contributions arise from myosin S1 or actin, and (3) that the elemental movement can be caused by a change of shape of the S1 head, but that this would have to be much greater than the movements suggested from structural studies of S1 (Rayment et al., 1993).

Relation between magnetically-applied force and velocity in beads coated with rabbit myosin, sliding on actin cables in Nitellopsis cells.

We have succeeded in controlling the sliding movement of myosin-coated magnetizable beads on actin cables in Nitellopsis cells by the inhomogeneous magnetic field adjacent to a small, strong permanent magnet. The relation between magnetic force acting on the bead and the bead velocity was, in many respects, similar to that obtained from the same system by the use of centrifugal force (Oiwa et al., 1990). In particular, force favouring the motion (negative load) had little effect on the velocity until it was sufficient to pull the bead off the actin, whereas a relatively small positive load caused a reduction in velocity to a plateau value. Although the present method does not allow a good control of force direction, it demonstrates the promise of magnetic force in studying in vitro motility.

Isoproterenol and GTP gamma S inhibit L-type calcium channels of differentiating rat skeletal muscle cells.

In adult skeletal muscle, G-proteins have been shown to modulate the calcium channels both directly and through a cAMP-dependent phosphorylating mechanism. We have investigated the action of G-proteins on the L-type calcium current in cultured rat muscle cells (myoballs) under voltage clamp in whole cell or perforated patch modes. Intracellular photolytic release of 200 microM GTP gamma S inhibited the L-type calcium current. Inclusion of 500 microM uncaged GTP gamma S in the patch pipette in the whole cell configuration reduced the calcium current by a similar amount. Under perforated patch conditions external application of 10 microM of the beta-adrenergic agonist isoproterenol also reduced the calcium current. Pretreatment of the cells with pertussis toxin reversed the effect of GTP gamma S and removed that of isoproterenol. We conclude that rat myoballs contain beta-adrenergic receptors that inhibit the L-type calcium current, and that this inhibition is mediated by a pertussis toxin-sensitive G-protein.

Inferences concerning crossbridges from work on insect muscle.

This paper presents a number of separate results concerning crossbridge attachment: [1] X-ray diffraction from live bumble bee flight muscle shows a set of layer lines distinct from that of relaxed Lethocerus, in which the apparent myosin helix is shorter than that of the actin. [2] Rigor crossbridges of Lethocerus are not rotatable by stretch. [3] Rabbit and Lethocerus fibres in rigor relaxed by ATP at -35 degrees C show evidence of non-rigor crossbridge attachment.

Calcium signals in single T cells on activation by lectin.

Succinylation of concanavalin A (Con A) reduces its oligomer size while retaining its mitogenicity, and provides a probe of T cell activation. We have observed responses of cytosolic ionized calcium to succinyl Con A in suspensions of Jurkat and rat lymph node (LN) cells, using a fluorimeter, and in single cells settled on glass, using a dual wavelength video imaging system. In the fluorimeter a mitogenic level of succinyl Con A (30 micrograms/ml) produced only a 15-30 nM rise in average cell calcium in the suspended Jurkat or rat cells whereas the use of quantitative video imaging produced asynchronous 250-1000 nM pulses of free calcium in 35% of Jurkat cells and 300-850 nM pulses in 45% of rat LN cells. In Jurkat cells these pulses were sometimes repetitive, giving rise to apparent oscillations. In the fluorimeter 30 micrograms/ml of native Con A (a supra-mitogenic concentration) produced a 300 nM rise in average cell calcium in suspended Jurkat cells, and a 100 nM rise in rat LN cells; when major histocompatibility complex class II-bearing cells were removed the response rose. Mitogenic Con A (3 micrograms/ml) produced a much lower rise in calcium. With video imaging the response seen was greater. Levels greater than 30 micrograms/ml Con A caused 700-5000 nM pulses synchronously in 94% of Jurkat cells and 250-1000 nM pulses in 73% of rat LN cells. At 3 micrograms/ml Con A produced asynchronous 300-1100 nM pulses in 36% of rat LN cells. We conclude that the absence of a calcium signal in the fluorimeter can conceal asynchronous calcium responses in individual cells and that brief asynchronous cytosolic calcium pulses are sufficient for lectin to activate rat T cells.

Quantal release of Ca2+ from intracellular stores by InsP3: tests of the concept of control of Ca2+ release by intraluminal Ca2+.

A possible mechanism for the generation of 'quantal' release of intracellular Ca2+ by InsP3 (Muallem et al., J. biol. Chem. 264, 205-212 (1989)) has been put forward in which intraluminal Ca2+ levels modulate InsP3 receptor structure (Irvine, FEBS Lett. 263, 5-9 (1990)). Here we have modelled such a steady-state mechanism, with an InsP3-sensitive store plus an InsP3-insensitive one, to test its ability to mimic published data. We have also performed experiments on InsP3-stimulated rat liver microsomes to test whether the model is consistent with one-way Ca2+ fluxes at a steady state. The model can simulate quantal release, in that InsP3 produces a release of part of the stored Ca2+ which is initially rapid relative to the one-way flux. In the original form of the model, in which InsP3-modulated Ca2+ binding to the intraluminal site opens the Ca2+ channel, the range of InsP3 concentrations needed to release Ca2+ is greater than that observed. When the model is changed so that Ca2(+)-modulated InsP3 binding opens the channels, the effective InsP3 range is shortened, but the quantal release effect is reduced. Other published data on one-way fluxes, and our own data on microsomes, can be simulated when leakage from the InsP3-insensitive store is adjusted to fit the observations; these data therefore do not test the existence of a steady state in the InsP3-sensitive store. We conclude that sensitivity of Ca2+ release to intraluminal Ca2+ provides a steady-state explanation of most, but not all, current quantal release observations.(ABSTRACT TRUNCATED AT 250 WORDS)

GTP gamma S causes contraction of skinned frog skeletal muscle via the DHP-sensitive Ca2+ channels of sealed T-tubules.

We have investigated the involvement of G-proteins in excitation-contraction coupling of fast-twitch skeletal muscle, using a fibre preparation designed to retain intact T-tubules and sarcoplasmic reticulum. The nonhydrolysable analogue of guanosine triphosphate, GTP gamma S (50-500 microM) caused a strong, transient isometric contraction in this preparation. Reduction of ethylene-bis(oxonitrilo)tetraacete (EGTA) in the sealed T-tubules from 5 mM to 0.1 mM lowered the threshold to GTP gamma S and removal of sodium reversibly raised it. The dihydropyridine (DHP) calcium channel antagonists nicardipine and nifedipine allowed a first contraction and then blocked subsequent GTP gamma S action. The phenylalkylamine methoxyverapamil (D-600) did likewise, reversibly, at 10 degrees C. The guanosine diphosphate analogue, GDP beta S, and procaine reversibly blocked the action of GTP gamma S; pertussis toxin also blocked it. Photolytic release of 40-100 microM GTP gamma S within 0.1 s from S-caged GTP gamma S caused contraction after a latent period of 0.3-20 s. We conclude that GTP gamma S can activate contraction in frog skeletal muscle via a route requiring both the integrity of the T-tubular DHP-sensitive calcium channel (DHPr) and the presence of sodium in the sealed T-tubules. We propose that in this preparation GTP gamma S activates a G-protein, which in turn activates the DHPr as a calcium channel and releases stored calcium from within the sealed T-tubule. Implications of these results for the excitation-contraction coupling mechanism in skeletal muscle are discussed.

X-ray diffraction and electron microscopy from Lethocerus flight muscle partially relaxed by adenylylimidodiphosphate and ethylene glycol.

The low-angle X-ray diffraction pattern from Lethocerus flight muscle fibres was recorded in rigor or under two conditions that modify crossbridge structure and behaviour, aqueous adenylylimidodiphosphate (AMPPNP) and AMPPNP + calcium in an ethylene glycol-water mixture. The effects on the 38.7 nm layer-line peaks (hk.6) of the diffraction patterns were studied in detail. In aqueous AMPPNP at room temperature, a condition in which rigor tension drops to half without loss of stiffness, the peaks remained nearly as intense as in rigor except for the 10.6, which dropped to half. In 20% (v/v) ethylene glycol-AMPPNP + 100 microM-Ca2+ at 23 degrees C (gly + pnp + Ca), a condition which removed muscle tension but left stiffness close to the rigor value, the 10.6 and 11.6 peaks greatly decreased but the 31.6 remained relatively high. The 14.5 nm meridional peak (00.16) became stronger on addition of AMPPNP and again on adding glycol + calcium. Considered in terms of constructively interfering filaments and crossbridges, the X-ray data indicated a transfer of diffracting crossbridge mass towards the thick filament as relaxation proceeds. We compared the X-ray diffraction patterns and crossbridge structure seen with electron microscopy (EM) under the same chemical conditions. EM and X-ray observations were mutually quite consistent overall. However, X-ray data indicated that more crossbridge mass was stereospecifically related to actin before fixation in the partially relaxed state (gly + pnp + Ca) than was suggested by the disordered crossbridge profiles seen by EM. We conclude that myosin heads at the start of the power stroke may both be closely related to their thick filament origins and form actin-determined attachments to the thin filament.

Slip of rabbit striated muscle in rigor or AMPPNP.

Single glycerol-extracted rabbit psoas muscle fibres have been slowly extended either in rigor or in the unhydrolysable ATP analogue, AMPPNP, and their sarcomere length, sarcomere structure and tension measured. The length of regularly arrayed sarcomeres, measured by optical diffraction, increased continuously as the muscle was stretched; the maximum sarcomere extension seen was approximately 6%. In the electron microscope sarcomeres from extended muscle fixed in rigor or AMPPNP remained regular in their internal structure, without rupture or obvious lengthening around the Z line. During steady extension at 0.024% per min the tension in the muscle fibre rose until it reached a limiting value [Tm] when the sarcomeres had stretched by 0.8-1.6% and then remained constant with continued extension, while the sarcomeres continued to stretch. Provided that a novel form of preparation of the glycerol-extracted fibres was employed, Tm in rigor was a large fraction of the tension expected from an active isometric muscle fibre. In the presence of AMPPNP Tm was reduced by a factor of 2 to 3. Step extension by 0.08% at 5-min intervals gave the same pattern of mechanical response with similar values of Tm. The isometric tension decay in the interval between the steps was very rapid at first and slowed continuously until the next step. The average speed of tension fall between 30 and 300 s after stretch was measured at each step and plotted relative to the tension in the muscle. The relationship approached linearity, although with a significant upward curvature at high tension.(ABSTRACT TRUNCATED AT 250 WORDS)

Two attached non-rigor crossbridge forms in insect flight muscle.

We have performed thin-section electron microscopy on muscle fibers fixed in different mechanically monitored states, in order to identify structural changes in myosin crossbridges associated with force production and maintenance. Tension and stiffness of fibers from glycerinated Lethocerus flight muscle were monitored during a sequence of conditions using AMPPNP and then AMPPNP plus increasing concentrations of ethylene glycol, which brought fibers through a graded sequence from rigor relaxation. Two intermediate crossbridge forms distinct from the rigor or relaxed forms were observed. The first was produced by AMPPNP at 20 degrees C, which reduced isometric tension 60 to 70% below rigor level without reducing rigor stiffness. Electron microscopy of these fibers showed that, in spite of the drop in tension, no obvious change from the 45 degrees crossbridge angle characteristic of rigor occurred. However, the thick filament ends of the crossbridges were altered from their rigor positions, so that they now marked a 14.5 nm repeat, and formed four separate origins at each crossbridge level. The bridges were also less slewed and bent than rigor bridges, as seen in transverse sections. The second crossbridge form was seen in glycol-AMPPNP at 4 degrees C, just below the glycol concentration that produced mechanical relaxation. These fibers retained 90% of rigor stiffness at 40 Hz oscillation, but would not bear sustained tension. Stiffness was also high in the presence of calcium at room temperature under similar conditions. Electron microscopy showed crossbridges projecting from the thick filaments at an angle that centered around 90 degrees, rather than the 45 degree angle familiar from rigor. This coupling of relaxed appearance with persistent stiffness suggests that the 90 degree form may represent a weakly attached crossbridge state like that proposed to precede force development in current models of the crossbridge power stroke.

Two attached non-rigor crossbridge forms.

It is possible to produce a graded progression from rigor toward relaxation using MgAMPPNP and substituting ethylene glycol for part of the solvent water. Fibers have been brought through this progression to various stages while measuring isometric force and stiffness, then fixed for thin-section electron microscopy. Distinct state-dependent crossbridge forms were observed in thin cross and longitudinal sections. When MgAMPPNP was added to rigor fibers at 23 degrees C, the tension dropped to about one-third of its original value, but crossbridge angle remained at 45 degrees. Distinct changes were seen in crossbridge shape and angle close to the thick filament, presumably in the S2 region of myosin. Adding 30% glycol in the presence of AMPPNP reduced tension to nearly zero while stiffness remained high, provided either calcium was present or the muscle was kept cold. Under these conditions, the crossbridges were oriented at approximately 90 degrees to the filaments, and in cross-section appeared straight and joined the thick filament at separate azimuths. Raising the glycol concentration to 40% or the temperature to 23 degrees C in the absence of calcium lowered the stiffness to a value slightly above that of MgATP relaxed muscle. The 90 degrees crossbridge forms seen in stiff versus relaxed fibers were closely similar but the distribution of bridges and the optical transforms suggested more bridge attachments when stiffness was high. The 90 degrees crossbridges appear structurally distinct from the rigor form and may in the stiff fibers represent a stable but weak binding state of the actomyosin contact.