Smooth muscle myosin mutants containing a single tryptophan reveal molecular interactions at the actin-binding interface

Abstract

Elucidation of the molecular details of the cyclic actomyosin interaction requires the ability to examine structural changes at specific sites in the actin-binding interface of myosin. To study these changes dynamically, we have expressed two mutants of a truncated fragment of chicken gizzard smooth muscle myosin, which includes the motor domain and essential light chain (MDE). These mutants were engineered to contain a single tryptophan at (Trp-546) or near (Trp-625) the putative actin-binding interface. Both 546- and 625-MDE exhibited actin-activated ATPase and actin-binding activities similar to wild-type MDE. Fluorescence emission spectra and acrylamide quenching of 546- and 625-MDE suggest that Trp-546 is nearly fully exposed to solvent and Trp-625 is less than 50% exposed in the presence and absence of ATP, in good agreement with the available crystal structure data. The spectrum of 625-MDE bound to actin was quite similar to the unbound spectrum indicating that, although Trp-625 is located near the 50/20-kDa loop and the 50-kDa cleft of myosin, its conformation does not change upon actin binding. However, a 10-nm blue shift in the peak emission wavelength of 546-MDE observed in the presence of actin indicates that Trp-546, located in the A-site of the lower 50-kDa subdomain of myosin, exists in a more buried environment and may directly interact with actin in the rigor acto-S1 complex. This change in the spectrum of Trp-546 constitutes direct evidence for a specific molecular interaction between residues in the A-site of myosin and actin.

Myosin acts as a mechanochemical enzyme that converts energy from the hydrolysis of ATP into movement by virtue of its cyclic interaction with actin. The crystal structures of vertebrate myosin subfragment 1 (S1) (1), invertebrate myosin motor domain complexed with ATP analogs (2–4), and actin (5–7) have provided a foundation for examining the molecular details of actomyosin interactions. Based on x-ray crystallography, x-ray diffraction of F-actin gels (8), and image reconstruction analysis of electron micrographs (9–11), several models have been proposed that describe how specific domains or motifs of myosin and actin may interact structurally (reviewed in ref. 12). However, these models are based on static images of myosin in different stages of the contractile cycle and are derived by using techniques that do not allow direct examination of specific molecular interactions at the actomyosin binding interface. Thus, experiments are necessary that can explore specific nucleotide-mediated as well as actin-induced domain movements that occur in myosin, which alter its affinity for actin during the course of the contractile cycle.

One proposed model of the acto-S1 complex describes three specific structural regions within myosin that may interact with actin in a sequential manner during the contractile cycle (9). These regions all are located within or adjacent to the central, most globular portion of S1, known as the 50-kDa domain based on studies using limited tryptic digestion (13). Fig. 1 illustrates the location of these three regions. Two are situated on either side of a large transverse cleft that essentially splits the 50-kDa domain into “upper” and “lower” subdomains, whereas the third is the actual trypsin-sensitive loop that connects the 50-kDa domain to the adjacent 20-kDa tryptic domain. In this model, the highly positively charged 50/20-kDa loop (yellow highlight) initiates binding through electrostatic interactions with the N terminus of actin to form the “weak” binding state. The region on the lower 50-kDa subdomain, referred to as the A-site (highlighted in purple) (14), then forms both ionic and hydrophobic interactions with the N terminus of actin that strengthens the actomyosin complex. The third region of myosin thought to interact with actin, referred to as the R-site (highlighted in green) (14), is recruited last to form more hydrophobic interactions with actin and completes the rigor acto-S1 complex. Finally, the large cleft that separates the upper and lower 50-kDa subdomains is predicted to close during formation of the rigor complex.

Figure 1.

Structure of a fragment of the skeletal muscle myosin molecule (1) corresponding to the smooth muscle construct used in these studies. The positions of the endogenous tryptophans in MDE are highlighted at the corresponding residues in the skeletal structure. The four tryptophan residues mutated to phenylalanine in both of our constructs are highlighted in orange, and the single tryptophans used as fluorescent probes in the 546- and 625-MDE constructs are highlighted in red (Trp-546) and cyan (Trp-625). Additionally, the three putative actin-binding regions described in the text are highlighted: 50/20-kDa loop (yellow), A-site (purple), and R-site (green). This figure was generated with the Swiss pdb program (Glaxo Wellcome Experimental Research, Geneva, Switzerland).

There is limited direct evidence in the literature for the involvement of these regions in actin binding. For instance, the 50/20-kDa loop is protected from proteolysis in the actomyosin complex (15), it can be cross-linked to actin (16), and studies have correlated the kinetic properties of different myosin isoforms to sequence variation in this loop (17, 18). Additionally, there are experiments that suggest the A-site (19) and R-site (20) may interact with actin. However, no direct evidence implicating the role of these regions has been reported because of a lack of suitable techniques for directly examining specific molecular interactions between myosin and actin.

One method of studying domain movements in proteins that may be useful for elucidating mechanisms of actomyosin interactions is fluorescence spectroscopy. Exogenous fluorescent probes that can be chemically attached to specific sites have been used to perform fluorescence quenching, polarization, and energy transfer experiments to examine conformational changes within myosin. These studies have been limited by the availability of chemically reactive side chains and have focused on a small number of sites. Thus, only the nucleotide binding pocket (21–24), the reactive thiol region (25–28), and the light chain domain (29, 30) have been well examined. However, the use of exogenous probes may not be appropriate for studying the actin-binding region because the introduction of a bulky fluorescent molecule may disrupt interactions with actin as well as the structural environment around the probe. Alternatively, the use of intrinsic fluorescent probes such as tryptophan has several advantages because it is site specific, extremely sensitive to local conformational changes, and less disruptive to protein–protein interactions. Nucleotide-sensitive tryptophan fluorescence has been used as a tool to examine general conformational changes in skeletal muscle myosin (31–33), thought to be dominated by Trp-510 (34). Consequently, Park et al. (35) isolated the fluorescence from Trp-510 and demonstrated conformational changes in a cleft surrounding this residue that revealed an important communication route between the active site and the reactive thiol region of myosin. Thus, tryptophan fluorescence can give detailed structural information about the local environment of specific regions of interest in myosin, provided the fluorescence from a single tryptophan can be resolved.

In the present study, site-directed mutagenesis was performed to conservatively replace six of the seven endogenous tryptophans in a truncated version of chicken gizzard S1 containing the motor domain and the essential light chain binding site [motor domain-essential light chain (MDE)], but lacking the regulatory light chain binding site. This MDE construct was used to generate two mutants of smooth muscle myosin containing a single endogenous tryptophan residue located at either Trp-546, in the A-site, or at Trp-625, near the 50/20-kDa loop and at the edge of the 50-kDa cleft (see Fig. 1 for the corresponding locations in the skeletal muscle S1 crystal structure). We determined the fluorescent properties of 546- and 625-MDE by examining their steady-state fluorescence emission, acrylamide quenching, and lifetime decay properties. In addition, we assessed the actin-induced conformational changes in these molecules by examining shifts in the fluorescence emission spectra associated with actin binding. Thus, the single tryptophan-containing mutants we have generated provide a sensitive method of directly examining specific molecular interactions at the actomyosin interface.

MATERIALS AND METHODS

cDNA Construction, Protein Expression, and Purification.

Site-directed mutagenesis was performed on a cDNA clone of chicken gizzard smooth muscle myosin that was truncated at Leu-819, beyond the essential light chain binding site, and referred to as MDE (generously donated by Kathleen Trybus, Brandeis University) (see Fig. 1 for structural representation). To create our two single-tryptophan-containing mutants, we proceeded as follows: first, in both mutants, we performed sequential site-directed mutagenesis to replace five of the endogenous tryptophans with phenylalanine (W29F, W36F, W441F, W512F, and W597F; orange space-filling residues in Fig. 1). In the case of the 546-MDE mutant, we then replaced Trp-625 with phenylalanine to leave a single endogenous tryptophan at position 546 (red space-filling residue in Fig. 1). In the 625-MDE mutant, we replaced Trp-546 with methionine (the residue found at the equivalent position in skeletal muscle myosin), leaving a single endogenous tryptophan at position 625 (cyan space-filling residue in Fig. 1). In addition, a mutant construct, Null-MDE, in which all seven tryptophans were substituted, and a wild-type construct (WT-MDE), with all seven tryptophans remaining, were produced. All MDE constructs contained a FLAG epitope sequence (DYKDDDK) at the C terminus of the heavy chain for purification purposes (36).

Recombinant baculoviruses encoding the myosin heavy chain and essential light chain were used to coinfect Sf9 cells followed by the initial stages of MDE purification described previously (37). MDE then was bound to an anti-FLAG antibody column and eluted with the homologous FLAG peptide. For further purification this material was pelleted with actin and released into the supernatant by the addition of 0.5 mM MgATP (37), followed by dialysis into Mops buffer (20 mM Mops/20 mM KCl/2 mM MgCl₂/1 mM EGTA/1 mM DTT/1 mM NaN₃, pH 7.4). The degree of purity was assessed by SDS/PAGE (38) using Coomassie-stained SDS gels. Protein concentrations were determined by the method of Bradford (39).

Actin was purified from chicken pectoralis muscle by using an acetone powder method described previously (40) and was phalloidin-stabilized (41) for the actin pelleting purification and the actin cosedimentation assays. Purified actin concentrations were determined spectrophotometrically by using an extinction coefficient of 0.62 (mg/ml)⁻¹·cm⁻¹ at 290 nm.

Actin-Activated ATPase and Actin Cosedimentation Assays.

Actin-activated ATPase assays were performed at 37°C in Mops buffer. Briefly, 0.05 mg/ml of purified MDE was assayed at a range of actin concentrations (0–90 μM) and inorganic phosphate (P_i) production was determined colorimetrically (42). The ATPase rates [nmol P_i⋅(nmol MDE)⁻¹⋅s⁻¹] of two separate preparations of WT- and 546-MDE and three separate preparations of 625-MDE were plotted as a function of actin concentration and fit with a nonlinear least-squares method. Values of V_max and K_M were determined by assuming Michaelis-Menton kinetics.

Actin cosedimentation assays were performed by incubating 1 μM MDE with 6.5 μM actin in Mops buffer under rigor conditions (absence of ATP) for 30 min on ice, followed by centrifugation at 95,000 rpm for 30 min in a Beckman TLA 120.2 rotor. The supernatant was removed, and the pellet was washed with Mops buffer and resuspended in a volume of SDS-gel sample buffer equivalent to that of the supernatant. Equal amounts of a sample taken before centrifugation as well as samples of the supernatant and pellet after centrifugation were subjected to SDS/PAGE. Assay results were assessed by visual inspection of the amount of MDE in the supernatant and pellet relative to the amount present before centrifugation.

Fluorescence Measurements.

Steady-state fluorescence was measured with a Quantmaster fluorimeter (Photon Technology International, South Brunswick, NJ) using a 75-W Xenon arc-lamp as an excitation source. Emission spectra were measured by exciting the sample at 295 nm through a single-grating excitation monochrometer and monitoring the emitted fluorescence from 305 to 400 nm through a single grating emission monochrometer with a WG-320 cut-off filter. The quantum yield (Φ) was calculated by a comparative method (43) by using l-tryptophan as a standard (Φ = 0.14) (44).

We were able to measure the fluorescence emission spectra from Trp-546 and Trp-625 in our single tryptophan MDE mutants in Mops buffer, both alone and complexed with actin, in rigor or in the presence of MgADP. In the experiments in the presence of actin, the fluorescence emission spectrum of 2–3 μM actin, determined before the addition of 546- and 625-MDE, was subtracted from the fluorescence emission spectrum of 0.25–0.50 μM 546- or 625-MDE bound to actin in a rigor complex or in the presence of 1 mM MgADP. Spectra were corrected for dilution and inner filter effects when necessary. To insure that the fluorescence emission spectrum of actin was not significantly altered by MDE binding, control experiments comparing 2 μM actin alone to 2 μM actin in a rigor complex with the Null-MDE construct (containing no tryptophans) were performed. We also verified that the Null-MDE construct bound normally to actin (data not shown).

Acrylamide quenching was monitored by measuring the decrease in fluorescence intensity at the maximum emission wavelength (λ_MAX) as a function of increasing concentrations of acrylamide ([Q]). The fluorescence in the absence of quencher (F₀) divided by the fluorescence in the presence of quencher (F) was used to quantify the relative change in fluorescence from acrylamide quenching (F₀/F). F₀/F was plotted as a function of [Q] and fit to the Stern-Volmer relationship taking into account both dynamic (K_SV) and static (V) quenching constants (45): F₀/F = (1+K_SV[Q])(expV[Q]).

The degree of tryptophan exposure to solvent can be evaluated from the dynamic quenching constant (K_SV), which is a measure of the collisional rate between the tryptophan residue and quencher molecules in the solvent. K_SV was isolated from the static quenching component (V), which arises from interactions of the quencher molecules with tryptophan residues before excitation, by a nonlinear least-squares fit of the data to the Stern-Volmer relationship. The bimolecular quenching constant (K_q), which describes the amount of collisional quenching that occurs during the lifetime of the fluorescent probe, was calculated by dividing K_SV by the fluorescence lifetime (τ).

The fluorescence lifetime decays of our MDE mutants were determined by using a Timemaster time-resolved fluorimeter (Photon Technology International). The sample was excited at 297 nm with a N₂ flash lamp and the time-resolved emission decay collected by a gated photomultiplier tube at λ_MAX (determined from the fluorescence emission spectra). Data acquisition and analysis of the time-resolved emission decays was performed by using proprietary software from Photon Technology International.

RESULTS

Functional Assays.

We successfully expressed two single tryptophan-containing mutants of chicken gizzard smooth muscle myosin, 546- and 625-MDE, that had six of the seven endogenous tryptophans in WT-MDE conservatively replaced (see Materials and Methods). The yields obtained for all purified MDE constructs were in the range of 0.5 to 1 mg per 1 × 10⁹ cultured Sf9 cells. To determine whether our mutagenesis had significantly altered the functional properties of the molecule, we conducted an enzymatic analysis of our mutants. The derived values of V_MAX (0.76 s⁻¹) and K_M (62 μM) which we obtained for the WT-MDE molecule correspond quite well with values that have been reported for biochemically prepared (46) and baculoviral-expressed single headed smooth muscle myosin constructs (47, 48). Furthermore, the V_MAX and K_M values of 625- and 546-MDE are within 50% of those from WT-MDE (see Fig. 2A). As a further verification of the normal function of our mutants, we performed F-actin cosedimentation assays to compare their actin binding properties to WT-MDE under rigor conditions. As shown in Fig. 2B, essentially all of the 546- and 625-MDE preparations pelleted with actin, as was the case for the WT molecule. These results confirm that our mutants have retained their ability to hydrolyze ATP in an actin-dependent manner and to bind actin in a rigor complex.

Figure 2.

Enzymatic and actin-binding properties of 546- and 625-MDE compared with WT-MDE. (A) Actin-activated ATPase rates of 546-, 625-, and WT-MDE are plotted as a function of actin concentration and fit to Michaelis-Menton, yielding the V_MAX and K_M values shown in the Inset. Points with error bars are means of 2–3 different preparations ± SE. (B) Gel electrophoretic analysis of actin cosedimentation of 546-, 625-, and WT-MDE. Actin binding under rigor conditions was examined by comparing the total mixture of MDE (1 μM) and actin (6.5 μM) before sedimentation (T) to the MDE bound to actin in the pellet (P) and unbound MDE in the supernatant (S) after sedimentation (see Materials and Methods).

Fluorescence Properties.

To examine the local environment around Trp-546 and Trp-625 we measured the steady-state fluorescence emission properties of 546- and 625-MDE, respectively. The steady-state fluorescence emission spectra of 546- and 625-MDE compared with l-tryptophan are shown in Fig. 3A, and the derived fluorescence parameters are shown in Table 1. The maximum emission wavelength (λ_MAX) for 625-MDE was 333 nm, and the λ_MAX of 546-MDE was 344 nm. The quantum yields (Φ) of 546- and 625-MDE were determined to be 0.20 and 0.36, respectively. Thus, the Φ and λ_MAX values for 546- and 625-MDE suggest that Trp-546 is almost completely exposed to solvent whereas Trp-625 is quite buried.

Figure 3.

Steady-state fluorescence properties of 546- and 625-MDE. (A) Emission spectra of 546- and 625-MDE compared with l-tryptophan (excitation wavelength of 295 nm). Emission spectra were normalized for differences in concentration. The quantum yields (Φ) and emission peak maxima (λ_MAX) are indicated in Table 1. (B) Acrylamide quenching of 546- and 625-MDE. Relative fluorescence intensity changes from quenching (F₀/F) are plotted as a function of acrylamide concentration (average of three trials ± SE). The data was fit to the Stern-Volmer relationship as described in Materials and Methods. The calculated dynamic (K_SV), static (V), and bimolecular (K_q) quenching constants are shown in Table 1.

Table 1.

Summary of the fluorescence properties of 546- and 625-MDE compared to l-tryptophan

Mutant	λMAX, nm	Φ	τ, ns	KSV, M−1	V, M−1	Kq, M−1⋅ns−1
546-MDE	344	0.20	2.6†	11.5	3.0	4.4†
625-MDE	333	0.36	3.0	5.3	2.5	1.7
l-Tryptophan	351	0.14*	2.6*	—	—	—

The emission peak maximum (λ_MAX), quantum yield (Φ), and lifetime (τ) values for the mutants are compared with l-tryptophan. Also, the results from acrylamide quenching of 546- and 625-MDE are compared, including the calculated dynamic (K_SV), static (V), and bimolecular (K_q) quenching constants. 

^*Values taken from the literature. 

^†Estimated values. 

The time-resolved fluorescence decay measured for 625-MDE was best-fit by a single exponential term yielding a lifetime of 3.0 ns (data not shown). We were unable to directly determine the lifetime decay of 546-MDE because of its low quantum yield. However, because 546-MDE demonstrated characteristics of an almost fully exposed tryptophan (Φ, λ_MAX, and K_SV), we assumed a lifetime decay value of 2.6 ns, the lifetime determined for l-tryptophan in solution (49), to allow us to estimate the bimolecular quenching constant (see below).

To quantitatively determine the degree of exposure to solvent of Trp-546 and Trp-625, we used acrylamide to quench the tryptophan fluorescence from 546- and 625-MDE (Fig. 3B). Acrylamide quenching was evaluated by using the Stern-Volmer relationship (see Materials and Methods). The calculated values of K_SV and V for 546-MDE (11.5 M⁻¹ and 3.0 M⁻¹, respectively) and 625-MDE (5.3 M⁻¹ and 2.5 M⁻¹, respectively) are shown in Table 1. The bimolecular quenching constant (K_q) also was evaluated for both mutants. The K_q was calculated to be 1.76 M⁻¹⋅ns⁻¹ for 625-MDE and was estimated to be 4.42 M⁻¹⋅ns⁻¹ for 546-MDE (Table 1). The estimated value of K_q for 546-MDE is similar to that of a fully exposed Trp in proteins (4.0 M⁻¹⋅ns⁻¹) (46), and the calculated K_q value for Trp-625 indicates it is more than 50% protected from solvent quenching.

Actin-Induced Changes in Spectra of 546- and 625-MDE.

Finally, we examined the actin-induced changes in the environment of Trp-546 and Trp-625 to determine whether specific regions of the putative actin-binding region of myosin are involved in or affected by actin binding. The steady-state fluorescence emission spectra of 546- and 625-MDE were compared in the presence and absence of actin. This analysis was accomplished by subtracting the fluorescence of actin alone from the fluorescence when 546- and 625-MDE were bound to actin in a rigor complex (MDE:actin) (see Materials and Methods). The spectrum of 625-MDE was not significantly altered in the actin-bound conformation (Fig. 4A). However, there was a large 10-nm blue shift in the λ_MAX (344 → 334 nm) of 546-MDE upon actin binding in rigor (Fig. 4B) and in the presence of MgADP (data not shown). Thus, although the environment around Trp-625 is unchanged, Trp-546 appears to become more buried upon actin binding in rigor or in the presence of MgADP.

Figure 4.

Actin-induced shifts in the steady-state fluorescence emission spectra of 625- and 546-MDE. Normalized emission spectra of 0.25–0.5 μM MDE in the presence and absence of 1 μM actin are compared. (A) The fluorescence emission spectrum of 625-MDE bound to actin (625-MDE:actin) is quite similar to 625-MDE alone, with no change in the emission peak maximum (333 nm). (B) The fluorescence emission spectrum of 546-MDE demonstrates a 10-nm blue shift in the emission peak maximum (344 → 334 nm) upon actin binding (546-MDE:actin).

DISCUSSION

To understand the molecular basis of muscle contraction it is important to examine how conformational changes in the actin-binding region of myosin result in alterations in its affinity for actin throughout the contractile cycle. In an effort to study the dynamic structural and functional properties of the putative actin-binding domain of myosin, we generated two mutants of smooth muscle myosin containing a single tryptophan at different positions in this interface. We examined the fluorescence from Trp-546, which is located in the lower 50-kDa subdomain A-site and predicted to be involved with hydrophobic interactions with the C terminus of actin (9). We also examined the fluorescence from Trp-625, which is located in the upper 50-kDa subdomain at the edge of the 50-kDa cleft and just before the 50/20-kDa loop. This location is important because the 50/20-kDa loop is thought to participate in electrostatic interactions with the N terminus of actin (15–18) and the 50-kDa cleft was predicted to close in the rigor acto-S1 complex (9, 10). Thus, we were able to use two naturally occurring tryptophan residues as intrinsic fluorescent probes to directly examine the role of specific sites within the upper and lower 50-kDa subdomains of myosin in forming a rigor complex with actin.

We conservatively replaced six of the seven endogenous tryptophans to produce two single tryptophan-containing mutants (546- and 625-MDE) that maintained their actin-activated catalytic activity and ability to bind actin in a rigor complex. Our mutations may not have been disruptive because four tryptophans were replaced with residues occurring at the homologous position in other myosin isoforms. In addition, six of the seven tryptophans were conservatively replaced with phenylalanine, which has similar properties, including hydrophobicity and van der Waals volume. Previously it was shown that mutations that decrease the hydrophobicity of the A-site of smooth muscle heavy meromyosin (HMM), Trp-546–Ser and Phe-547–His, decreased actin-activated ATPase and actin binding activities of the construct (19). However, the substitution of tryptophan to methionine at residue 546 (the corresponding residue in skeletal muscle myosin) in the 625-MDE mutant is probably benign because Trp-546 is mostly exposed and methionine is a more hydrophobic residue than serine. Thus, despite the potential sensitivity of the A-site to structural changes, our conservative substitutions have allowed our mutants to maintain nearly normal actin-dependent ATPase and actin binding properties.

The fluorescence properties of 546- and 625-MDE measured in the absence of actin are in good agreement with the conformation of the corresponding residues in the crystal structure of chicken skeletal muscle myosin S1 (1). The acrylamide quenching and fluorescence emission results demonstrate that Trp-546 is a nearly fully exposed tryptophan, whereas Trp-625 is less than 50% exposed, based on comparison of these results with previous studies of tryptophans in proteins (46). The low value of K_q and the single lifetime of Trp-625 suggest that, although Trp-625 is not completely buried, it may be restricted to a single conformation, perhaps because of side-chain packing interactions. Interestingly, although the lifetime of 546-MDE was not measured directly, the estimated K_q is similar to that observed for an exposed tryptophan in proteins and fits well with the measured fluorescence values for Trp-546 (λ_MAX, Φ, and K_SV). Furthermore, the fluorescence profiles of both 546- and 625-MDE are not sensitive to ATP. These results are consistent with the published crystal structures of the myosin motor domain of Dictyostelium discoideum, which show that the conformation of residues corresponding to Trp-546 and 625 are unaltered in the presence of different ATP analogs (2–4). Thus, our fluorescence data from 546- and 625-MDE support the available structural data and provide a unique opportunity to study dynamic conformational changes in these two endogenous tryptophan residues in the presence of actin.

The observed actin-induced changes in the spectra of 546- and 625-MDE can be evaluated in terms of proposed models of the acto-S1 complex (see Introduction). The fluorescence spectrum of 625-MDE bound to actin is quite similar to the unbound spectrum, indicating that the conformation of Trp-625 is unaltered by actin binding. Because tryptophan fluorescence is very sensitive to its local environment, the lack of change in Trp-625 fluorescence indicates that the region around this residue does not change significantly upon actin binding. This insensitivity of Trp-625 to actin binding seems somewhat surprising given the evidence that the 50/20-kDa loop interacts with actin, and suggests the flexible loop does not communicate conformational changes to the upper 50-kDa subdomain of myosin. The lack of change in the environment around Trp-625 upon actin binding suggests that the movements in the cleft separating the upper and lower 50-kDa subdomain may not be as pronounced as predicted by acto-S1 docking models (9, 10). Also, the residues located in a conserved helix lining the cleft on the lower 50-kDa subdomain appear to be vital for the normal function of the molecule (50), implying that conformational changes may be more prominent in the lower 50-kDa subdomain of the cleft. Therefore, by demonstrating that fluorescence from Trp-625 does not change in the presence of actin, we have provided evidence that actin-induced conformational changes that may be associated with the 50-kDa cleft and 50/20-kDa loop are not transmitted to the upper 50-kDa subdomain.

The large 10-nm blue shift in the λ_MAX of Trp-546 induced by actin binding indicates that this residue undergoes a substantial change in environment and becomes more buried in the acto-S1 rigor complex. This change in environment of Trp-546 could indicate that Trp-546 directly interacts with actin and/or that an actin-induced conformational change occurs in myosin, which causes it to become less exposed to the solvent. We favor specific interactions of Trp-546 with actin for two reasons. First, it is located near several exposed hydrophobic residues in the A-site (skeletal residues Pro-529, Met-530, Ile-535, Phe-542, and Pro-543) that are orientated in the crystal structure with their side chains pointing away from the myosin molecule (9). Second, when two residues in this hydrophobic patch are changed to more polar amino acids (Trp-546–Ser and Phe-547–His), a significant decrease in actin-dependent function is observed (19). This region of the A-site may interact with C-terminal hydrophobic actin residues (Ala-144, Ile-341, Ile-345, Leu-349, and Phe-352) and form a hydrophobic pocket in the rigor complex (12). The large blue shift in Trp-546 fluorescence upon actin binding indicates that it may be buried within a highly ordered complex of hydrophobic residues between actin and myosin that represent the primary binding site between the two molecules. Electron micrographs of smooth muscle myosin decorated actin filaments suggest a structural change at the actomyosin interface in the presence of MgADP compared with rigor (11). However, our results indicate that the environment around Trp-546 in the MgADP bound complex is similar to rigor and thus other regions of the actomyosin interface may be responsible for the MgADP-dependent changes observed previously. Our data constitutes direct evidence that the A-site in the lower 50-kDa subdomain of myosin is indeed an important component of the strongly bound actomyosin complex.

We have designed an extremely sensitive method, which uses endogenous fluorescence from single tryptophan-containing mutants of smooth muscle myosin, to examine specific molecular interactions at the actomyosin interface. We have determined that Trp-625 is not sensitive to actin binding, suggesting that actin-induced conformational changes that may occur in the 50/20-kDa loop and 50-kDa cleft are not communicated to the upper 50-kDa subdomain of myosin. We also have demonstrated that Trp-546 becomes more buried upon actin binding. Thus, these data provide direct evidence of a specific molecular interaction of the A-site of myosin with actin in the strongly bound actomyosin complex. The successful use of these single tryptophan-containing mutants in the current work indicates that this method will be a powerful one for examining myosin structure/function relationships in future studies.

ABBREVIATIONS

MDE

motor domain essential light chain construct

S1

myosin subfragment 1

WT

wild type