Phosphorylation-dependent Structural Changes in the Regulatory Light Chain Domain of Smooth Muscle Heavy Meromyosin*

Smooth muscle heavy meromyosin, a double-headed proteolytic fragment of myosin lacking the COOH-terminal two-thirds of the tail, has been shown previously to be regulated by phosphorylation. To examine phosphorylation-dependent structural changes near the head-tail junction, we prepared five well regulated heavy meromyosins containing single-cysteine mutants of the human smooth muscle regulatory light chain labeled with the photocross-linking reagent, benzophenone-iodoacetamide. For those mutants that generated cross-links, only one type of cross-linked species was observed, a regulatory light chain dimer. Irradiated mutants fell into two classes. First, for Q15C, A23C, and wild type (Cys-108), a regulatory light chain dimer was formed for dephosphorylated but not thiophosphorylated heavy meromyosin. These data provide direct chemical evidence that in the dephosphorylated state, Gln-15, Ala-23, and Cys-108 on one head are positioned near (within 8.9 Å) the regulatory light chain of the partner head and that thiophosphorylation abolishes proximity. This behavior was also observed for the Q15C mutant on a truncated heavy meromyosin lacking both catalytic domains. For the actin-heavy meromyosin complex, cross-links were formed in both de- and thiophosphorylated states. S59C and T134C mutants were in a second mutant class, where regulatory light chain dimers were not detected in dephosphorylated or thiophosphorylated heavy meromyosin, suggesting positions outside the region of interaction of the regulatory light chains.

Smooth muscle heavy meromyosin, a double-headed proteolytic fragment of myosin lacking the COOH-terminal two-thirds of the tail, has been shown previously to be regulated by phosphorylation. To examine phosphorylation-dependent structural changes near the head-tail junction, we prepared five well regulated heavy meromyosins containing single-cysteine mutants of the human smooth muscle regulatory light chain labeled with the photocross-linking reagent, benzophenone-iodoacetamide. For those mutants that generated cross-links, only one type of cross-linked species was observed, a regulatory light chain dimer. Irradiated mutants fell into two classes. First, for Q15C, A23C, and wild type (Cys-108), a regulatory light chain dimer was formed for dephosphorylated but not thiophosphorylated heavy meromyosin. These data provide direct chemical evidence that in the dephosphorylated state, Gln-15, Ala-23, and Cys-108 on one head are positioned near (within 8.9 Å) the regulatory light chain of the partner head and that thiophosphorylation abolishes proximity. This behavior was also observed for the Q15C mutant on a truncated heavy meromyosin lacking both catalytic domains. For the actin-heavy meromyosin complex, cross-links were formed in both de-and thiophosphorylated states. S59C and T134C mutants were in a second mutant class, where regulatory light chain dimers were not detected in dephosphorylated or thiophosphorylated heavy meromyosin, suggesting positions outside the region of interaction of the regulatory light chains.
The actin-activated ATPase activity and motor properties of smooth muscle and nonmuscle myosins are regulated by phosphorylation of the regulatory light chain (1)(2)(3). The dephosphorylated forms of these regulated myosins have low ATPase activity and are unable to move actin filaments, whereas phosphorylated forms are activated in both respects. Domain requirements for regulation have been elucidated through studies of various proteolytic and expressed subfragments of SMM. 1 SMM contains two head domains (S1) attached to a long ␣-hel-ical coiled-coil domain (tail). Single-headed myosin (4) and S1 (5,6) are active in both dephosphorylated and phosphorylated states. HMM that lacks the COOH-terminal two-thirds of the tail is double-headed and well regulated (5,7), but expressed HMMs with shorter tails failed to form double-headed structures and were found to be unregulated (8,9) as in S1 and single-headed myosin. Therefore, two heads are critical for down-regulation, suggesting that head-head interaction is an important feature of the dephosphorylated state. However, a double-headed structure may not be sufficient for full regulation, and interactions between heads and rods may also be required (10).
The general location of the phosphorylated subunit (RLC) at the junction between the heads and tail (11) suggests that RLC-RLC interactions may be a logical consequence of headhead interactions. Indeed, the RLC has been shown to be critical to the regulatory mechanism. In particular, disruption of the COOH-terminal domain (12,13), the central helix (14), and portions of the NH 2 -terminal charged region (15) of the RLC alters regulation. Smooth muscle myosin with exchanged skeletal RLC was not activated by phosphorylation (16). In contrast to the RLC, the ELC appears to be less important to regulation because it can be replaced by the skeletal isoform or removed entirely with little effect on regulation (17). The requirements for full regulation in smooth muscle myosin are complex, because of a subtle interplay between the light chain domain, the actin binding loop (18), the catalytic domain (19), and the heavy chain sequence within the light chain domain (20).
The exact location of the phosphorylated serine (S19 in chicken gizzard) of the RLC is not known. The analogous residue was found to be within a disordered region of the NH 2 terminus in the crystal structures of both skeletal S1 (21) and the light chain domain of scallop myosin (22). It is possible that this region adopts an ordered structure in a double-headed molecule. Clearly, more information about the structure of the head-tail junction including the RLC is needed to understand the phosphorylation-dependent regulatory mechanism.
Our approach has been to use direct chemical methods to further define this region by looking for conformational differences between the dephosphorylated and phosphorylated forms of smooth muscle HMM. Specifically, we have placed photocrosslinking probes selectively throughout the RLC by use of a set of single cysteine mutants. We have labeled two residues within the putative flexible NH 2 -terminal region (Q15C and A23C), two within the COOH-terminal domain (Cys-108 and T134C), and one within the NH 2 -terminal domain (S59C). After exchange of the labeled RLC into HMM, we found that cross-linked RLC-RLC dimers were formed after irradiation of Q15C, A23C, and WT (Cys-108) but not of S59C or T134C. These RLC-RLC dimers were not formed after irradiation of the respective thiophosphorylated HMMs. Our results suggest that regions of the RLC of one head are proximal to the RLC of the other head in dephosphorylated HMM, and this proximity is abolished in the phosphorylated state. We also show that this conformational change occurs independently of the catalytic domain.
The light chain mutants were prepared by introducing a cysteine residue into (or substituting from) the recombinant human smooth muscle RLC using oligo-site-directed mutagenesis and polymerase chain reaction techniques. Human smooth muscle RLC mutants Q15C, A23C, S59C, and T134C were prepared by substitution of Gln-15, Ala-23, Ser-59, and Thr-134 with cysteine, respectively, and Cys-108 of each was replaced by alanine. The cDNA for the human RLC was subcloned into M13mp19 at the EcoRI restriction site. Site-directed mutagenesis using the Skulptor Mutagenesis kit (Amersham Pharmacia Biotech) was used to substitute the endogenous Cys-108 with alanine and Glu-15, Ala-23, Ser-59, or Thr-134 with cysteine. The oligo for making the Cys-108 to alanine substitution was 5Ј-CGTCGAATGCG-GCAAAGGCG-3Ј. The oligos for A23C, S59C, and T134C mutants were antisense 5Ј-GGTCAAACATGCAGAAGACATTGG-3Ј, 5Ј-GTTCTTCC-CCAGGCAGGCCAGCATG-3Ј, and 5Ј-CTTCCTCATCGCAGAAGCGG-TCAC-3Ј, respectively. Polymerase chain reaction using TAQ polymerase was used to construct the mutant Q15C. The 5Ј-oligo introduced a NdeI restriction site at the start Met-1 as well as introducing the Gln-15 to cysteine. The 5Ј-oligo sequence was 5Ј-CCAGAACGCCC-ATATGTCCAGCAAGCGGGCCAAAGCCAAGGCCACCAAGAAGCGG-CCATGCCGGGCC-3Ј. The 3Ј-oligo used in polymerase chain reaction was complimentary to the 3Ј end of the coding region of the RLC with an introduced EcoRI site at the end of the primer. The 3Ј-oligo was 5Ј-CCGGAATTCTCACGTCTGCCCCGC-3Ј. TA-cloning (Invitrogen) was done directly using the polymerase chain reaction mix with the Invitrogen protocol. Mutant RLC DNAs were confirmed by sequencing and then subcloned into the His-tag vector pET 15b (Novagen) adding GHHHHHHSSGLVPRGSH to the NH 2 terminus of Met-1 of the RLC.
Host cells used for the expression were Escherichia coli BL21(DE3). 10 ml of overnight culture at 37°C was used to inoculate 1 liter of LB-amp broth. Isopropyl-1-thio-␤-D-galactopyranoside was added to 1 mM when the A 600 reached 0.8. Cultures continued to grow for 6 h. Cells were harvested by centrifugation at 6300 ϫ g for 20 min, and 25 ml of 8 M urea, 20 mM Tris (pH 8.0), 100 mM NaCl was added to each pellet bringing the total volume to 50 ml for 1 liter of culture and stirred on ice for 30 min. The cell lysate was then treated with 100 g of leupeptin (Sigma), 100 g of pepstatin (Sigma), and 200 g of DNase (Sigma)/1 liter of culture and stirred at room temperature for 30 min. The solution was centrifuged at 150,000 ϫ g at 4°C for 20 min. The supernatant was incubated at room temperature with TALON metal affinity resin (CLONTECH) for 30 min and centrifuged at 3,100 ϫ g for 10 min. The resin was resuspended with denaturing buffer, stirred for 10 min, and then centrifuged for 10 min. Resuspension and centrifugation were repeated three more times. The RLC mutant was eluted with 0.2 M imidazole in the denaturing buffer, and the resuspension-centrifugation procedure was repeated four times. The combined supernatant was dialyzed to 1 mM ␤-mercaptoethanol, 50 mM ammonium bicarbonate (pH 7.9), 0.1 mM EGTA, and 0.1 mM EDTA and labeled with benzophenone-4-iodoacetamide (Sigma) as described previously (25). The concentrations of labeled hisRLCs (BP-hisRLCs) were determined using the extinction coefficient of ⑀ 302 M ϭ 22,500 M Ϫ1 cm Ϫ1 (25). Photocross-linking of HMM/BP-RLC-Exchanges of native RLC of HMM for BP-mutant RLC were performed as described previously (26). Excess free RLC in the exchange mixture (typically 20 mg of HMM) was removed by a Superdex 200 gel filtration column (2.6 ϫ 60 cm; 300 mM NaCl, 10 mM MOPS (pH 7.0), 0.1 mM EGTA, 1 mM MgCl 2 , and 0.5 mM DTT). Base-line separation was achieved between the exchanged HMM and the excess free RLC. HMM/BP-hisRLC fractions were pooled and precipitated with 2.5 volumes of saturated ammonium sulfate (pH 8.0) and then dialyzed to 15 mM Tris-Cl (pH 7.5), 150 mM KCl, 2.5 mM CaCl 2 , 10 mM MgCl 2 , 2 mM DTT. Thrombin (Sigma) was added (1 unit/mg HMM/BP-hisRLC), and the mixture was incubated on ice for 4 h. Densitometric scanning of 4 -20% SDS-polyacrylamide gels showed that over 96% of His 6 tags were removed from the RLC by thrombin treatment. HMM/BP-RLC was dialyzed to 10 mM MOPS (pH 7.0), 0.1 mM EGTA, and 1 mM DTT immediately after thrombin treatment. HMM/BP-RLC was phosphorylated by incubating in 50 mM NaCl, 2.5 mM MgCl 2 , 2.5 mM CaCl 2 , 4 g/ml calmodulin, 30 g/ml myosin light chain kinase (27), and 1.5 mM ATP␥S (Roche Molecular Biochemicals) at 25°C and pH 7.0 for 1 h followed by incubating on ice overnight. Complete thiophosphorylation was verified by 8 M urea gel electrophoresis (4). Dephosphorylated control samples were prepared by adding 10 mM EGTA and 5 mM EDTA prior to adding the other components for thiophosphorylation. The buffer was changed to irradiation buffer (10 mM MOPS (pH 7.0), 0.1 mM EGTA, 1 mM DTT) by Sephadex G-50 (Sigma) centrifugal gel filtration (5-ml column, 0.8-ml load volume). HMM/BP-RLC solutions were centrifuged (350,000 ϫ g) at 4°C for 10 min, and the supernatant fractions were filtered through a 0.45-micron nylon Acrodisc filter (Gelman). The two gel filtration steps, ultracentrifugation and filtration, were sufficient for removing aggregated HMM, as determined by native gel electrophoresis (24). Fresh DTT was added to 1 mM to the HMM/BP-RLC solutions before irradiation. To effect cross-linking, 200 l of 1.2-1.4 mg/ml HMM solution in a 1.5-ml clear microfuge tube was irradiated through a Pyrex filter with a UV lamp (450 Watts, ACE Glass Incorporated) on ice as described previously (25).
Gel Electrophoresis and Western Blots-Gel electrophoresis was performed in the presence of 0.1% SDS with a 4 -20% acrylamide gradient gel (Tris-glycine from Novex, San Diego, CA). Gels were stained with Coomassie Brilliant Blue. Sister gels were transferred to nitrocellulose for subsequent blotting. Western blots were performed by standard procedure (28) for both RLC and ELC antibodies. Polyclonal rabbit anti-RLC antibody and donkey anti-rabbit antibody labeled with horseradish peroxidase (Amersham Pharmacia Biotech) were used as the first and second antibodies, respectively, for RLC blotting. Mouse anti-ELC monoclonal antibody (a gift from Dr. K. Trybus) and sheep antimouse antibody labeled with horseradish peroxidase (Amersham Pharmacia Biotech) were used as the first and second antibodies for ELC blotting. A Renaissance chemiluminescence kit (NEN Life Science Products) was used to visualize the antibody cross-reaction with RLC and ELC. Gels and films were scanned with a laser densitometer (Molecular Dynamics) for presentation. Fig. 1 documents a representative HMM preparation used in this study. HMM (Fig. 1, lane 2) was prepared from chicken gizzard myosin (Fig. 1, lane 1). Labeled mutant RLC (Fig. 1,  (Table I). Excess RLC was removed by gel filtration (Fig. 1, lane 5). Exchange efficiency (typically 75-90%; maximum theoretical value is 95%) could then be determined by densitometry using the ELC to normalize loading amounts. The density of the hisRLC was determined in an independent experiment to be 1.13-fold greater than the RLC. His tags were then removed (Ͼ98%) by thrombin treatment (Fig. 1, lane 7) to avoid potential alteration of regulatory properties of HMM. Thrombin-cleaved mutant RLC migrated slightly above the native RLC because of an additional four amino acids (GSHM). Quantitation of the light chain region by densitometry showed that the ratio of RLC (including both the thrombin-cleaved mutant band and the native band) to ELC for the exchanged samples was within 5% of the value for the unexchanged control HMM (data not shown). This suggests that the gel filtration protocol used here removes essentially all of the excess RLC used during the exchange procedure. The preparation shown in Fig. 1 (lane 7) was similar to all HMM/ BP-RLC mutants prepared, and the respective protein was used for subsequent phosphorylation, ATPase activity, and irradiation studies.

RESULTS
The effects of thiophosphorylation upon the actin-activated ATPase activity of control HMM and all HMM/BP-RLC mutants (unirradiated) are shown in Table I. All samples behaved similarly to the control (unexchanged) HMM. As expected from previous studies (5), the HMMs lacked activation by thiophosphorylation in the absence of actin but in the presence of actin showed an increase in activity upon thiophosphorylation. The exact fold activation in the presence of actin because of thiophosphorylation was difficult to quantitate from these data because the dephosphorylated activities were low and close to our range of error. However, the activities of the thiophosphorylated samples were all within a factor of 2 of the control. These results show that the HMM/BP-RLC preparations were regulated by thiophosphorylation and therefore were used for subsequent irradiation studies. Fig. 2 shows the effect of irradiation of HMM/BP-Q15C 2 with UV light to activate the benzophenone moiety for cross-linking to proximal residues. In the dephosphorylated HMM ( Fig. 2A,  lane 1), irradiation generated a new band (lanes 2-5) migrating between 34.1 and 51.6 kDa (molecular mass markers not shown) and migrating just below actin (42 kDa) consistent with the expected behavior of a RLC-RLC dimer (RLC ϭ 20 kDa). This band was identified as such after Western blot analysis showed cross-reaction with RLC ( Fig. 2B) but not ELC (Fig. 2C) antibodies. The RLC-RLC dimer band was also generated in the presence of MgADP and MgATP (Fig. 2, lanes 8 and 9, respectively) and in the presence of actin (Fig. 2, lanes 10). The amount of RLC converted to the RLC-RLC dimer was estimated by scanning Coomassie-stained gels in the light chain region using the ELC for normalization of loaded protein. This analysis showed that the cross-linking efficiency was about 12% for all irradiation conditions, including in the presence of 125 and 500 mM NaCl (lanes 6 and 7, respectively). The RLC-RLC dimer band was the only new band observed after irradiation (Fig. 2). Cross-linking was not observed from the RLC to the heavy chain nor to the ELC. In addition intramolecular RLC cross-linking was not detected. To test for this, dilute solutions of free RLC were irradiated to determine the positions on the gels of intramolecularly cross-linked RLC. All new bands migrated below the unirradiated RLC (data not shown). However, these bands were not evident in experiments with the RLC bound to the heavy chain in the HMM (Fig. 2). Therefore, we concluded that the HMM-bound RLC did not cross-link intramolecularly but only intermoleculary to form RLC-RLC dimers.
In contrast to the results for the dephosphorylated HMM/ BP-Q15C, the identical experiment for the thiophosphorylated protein showed no evidence for a RLC-RLC dimer (Fig. 2, D-F), except after irradiation in the presence of actin (Fig. 2, D and  E). Therefore, the results for the HMM/BP-Q15C suggest that there is a clear conformational difference between the dephosphorylated and thiophosphorylated states. The dephosphorylated state apparently positions Cys-15 (Gln-15 in the native RLC) from one head close enough to the other head to allow for cross-linking, whereas this event cannot occur in the thiophosphorylated state.
As evidenced by RLC Western blots, this same general pattern of cross-linking was observed for the HMM/BP-A23C (Fig.  3, A and C) and for the HMM/BP-WT (Cys-108; Fig. 3, B and D). Cross-reaction was not observed with the ELC antibody under any condition (data not shown). Interestingly, these two mutants (both dephosphorylated and thiophosphorylated), like Q15C, also showed generation of an RLC-RLC dimer after irradiation in the presence of actin.
The two other mutants studied, HMM/BP-S59C and HMM/ BP-T134C, showed a different behavior to those discussed above (Q15C, A23C, and WT (Cys-108)). In the dephosphorylated state no cross-linking was observed by Coomassie-stained gels (data not shown) or Western blot analysis using RLC antibodies (Fig. 4, A and B), except for S59C in the presence of actin. As before, ELC cross-reaction was not observed under any condition (data not shown). Thiophosphorylated mutants (Fig. 4, C and D) also showed no cross-linking, and even the cross-linking observed for dephosphorylated S59C in the presence of actin was abolished. Therefore, except in the presence of actin for S59C, these mutants did not report a difference between the dephosphorylated and thiophosphorylated states. Could the conformational differences between dephosphorylated and thiophosphorylated states that were detected for Q15C, A23C, and WT (Cys-108) be detected for a molecule without catalytic domains? To answer this question BP-Q15C-RLC was exchanged onto a truncated HMM-IIB (nonmuscle brain isoform) that lacked both catalytic domains 3 but contained the full RLC and ELC binding sites and both the RLC and ELC. As found for the full-length HMM/BP-Q15C, irradiation of the truncated HMM generated RLC-RLC dimers in the dephosphorylated state (Fig. 5, lanes 2-4), whereas RLC-RLC dimers were barely detected in the thiophosphorylated state (Fig. 5, lanes 5-7), even with three times more protein loaded per lane. Western blot analysis with ELC antibodies (data not shown) showed no evidence for ELC-RLC cross-linking. Therefore the truncated HMM/BP-Q15C behaved like the full-length HMM/BP-Q15C with respect to the sensitivity of photocrosslinking to the phosphorylation state, suggesting that the conformational change detected does not require the catalytic domains. Table II summarizes the results from this study. We chose these conservative mutations as likely surface residues based upon an alignment with the chicken skeletal RLC sequence and its known three-dimensional structure (21). They also appear as surface residues in our homology model with the scallop regulatory domain (Fig. 6). Residues involved with the metal binding site were avoided because they would likely alter binding affinity of the RLC for the heavy chain. All of the mutants tested were found to exchange to similar levels and to be well regulated by thiophosphorylation ( Table I), suggesting that the structural characteristics of the mutant RLCs were similar to the native protein. Control experiments (not shown) indicate that irradiation of unlabeled HMM generates no new bands and has no effect upon the ATPase activity up until 20 min of irradiation. This is also true for myosin (25,39). For all gels shown (Figs. [2][3][4][5]. Two lanes are designated NaCl; the left lane is 125 mM NaCl, and the right lane is 500 mM NaCl (in other figures where only one lane is designated NaCl, it is 500 mM NaCl). Control experiments with actin alone showed no cross-reactivity to the RLC (see Fig. 4) or ELC antibody (data not shown). All lanes in each gel are from the same actual gel and therefore represent the same exposure of the film. The RLC can appear as two closely spaced bands with the labeled RLC migrating above the unlabeled RLC. The exact appearance of the doublet depends upon the mutant, the extent of exchange and the particular gel and loading. We have no clear evidence that irradiation causes new bands to appear within this RLC doublet region.

DISCUSSION
The labeled RLC mutants tested appeared to fall into two classes. In the first class of mutants, HMMs exchanged with the labeled mutants Q15C, A23C and WT (Cys-108) formed RLC-RLC dimers after irradiation but only in the dephosphorylated state. Upon thiophosphorylation, cross-linking was either not evident (Q15C and A23C) or substantially diminished (WT). The maximum length of the cross-linker from the cysteine sulfur to the photoreactive carbonyl carbon is about 8.9 Å. Therefore we conclude that the dephosphorylated state allows for these regions of the RLC to approach the RLC of the other head to within this distance. The approximate width of the RLC domain is about 25 Å. Our interpretation of the lack of cross-linking in the thiophosphorylated state is that these regions of the RLC are moving away from one another. We do not favor the alternative explanation that thiophosphorylation changes the positioning of selectively reactive amino acids, because of the relatively nonselective behavior of the benzophenone cross-linker used here (30) and because the same general behavior was observed for three mutants in this class. Our results are not consistent with the idea that the phosphorylated state has complete flexibility about the head-tail junction. This is because photocross-linkers may "trap" rare states with short lifetimes. If this mechanism were predominant, we would have expected to detect some cross-linking in the thiophosphorylated state for all mutant positions (also see mutants that show no cross-linking in either state).
It is significant that we observed phosphorylation-dependent cross-linking at increased ionic strength and in the absence of ATP (Figs. 2-4). This would argue that cross-linking was not correlated to a "10 S-like" state because the 10 S state is not stable at high ionic strength and also requires the presence of ATP (25). ATP and ADP did not strongly influence the RLC-RLC cross-linking efficiency. Our photocross-linking efficiency was most strongly influenced by thiophosphorylation and actin binding.
The Q15C mutant was also exchanged into a truncated nonmuscle HMM IIB lacking both catalytic domains. As found for the native full-length HMM, RLC-RLC cross-linking was observed in the dephosphorylated state exclusively, suggesting that the catalytic domain is not necessary for the molecule to adopt the close positioning between the RLC domains. This result is generally consistent with the lack of effect of nucleotides upon our cross-linking results. Our findings also support fluorescence anisotropy measurements on a similar construct showing an increase in mobility of the RLC domain upon phosphorylation (31). Therefore, the catalytic domains are not required for phosphorylation-dependent changes in positioning between the two RLC domains.
In the second class of mutants, S59C and T134C, crosslinking was not observed in either the dephosphorylated or thiophosphorylated states (Fig. 4). Therefore we conclude that these regions of the RLC are faced away from the interfaces of the RLC that appear to approach closely as defined by the first class of mutants.
Ser-59 and Thr-134 (Class II) are located in the B-helix and at the end of the G-helix, respectively (Fig. 6A). Interestingly, these two residues are on opposite faces of the RLC from the residues in Class I (Fig. 6A). Before this study, little information had been available about the positions of the NH 2 -terminal region of the RLC, including Gln-15 and Ala-23, flanking the thiophosphorylated residue Ser-19, because the homologous residues from the skeletal myosin S1 (21) and the scallop regulatory domain (22) were found to be disordered. The equivalent region of the smooth muscle RLC is 24 residues in length. From our results with the Q15C and A23C mutants, it is clear that at least one of the interactions of this 24-residue region is  Fig. 2. For A, nothing was loaded in lane 4, whereas in C the load was the same as in Fig. 2; for D, lane 5 is 500 mM NaCl. The smearing observed above the RLC-RLC dimer was assigned as nonspecfic cross-reaction because it did not consistently appear and was not associated with Coomassie-stained bands.
in close approach to the other head. Fig. 6B shows the relative positions of Phe-25 (the first visualized NH 2 -terminal residue) and Cys-108, which is located at the top of the E-helix within the COOH-terminal domain of the RLC as defined by the break between the D and E Helix (21). It is interesting that Phe-25, which must be close to A23C, and Cys-108 are on the same face of the RLC and that both have the same cross-linking behavior (Class I).
An energy-minimized model of the head-tail junction with two scallop myosin regulatory domains (22) coupled to a model of the scallop tail, shows specific interactions between the two RLC (32) in a pseudo-symmetrical relationship. This model shows Gln-46 (scallop; equivalent to our S59C mutant) close to three residues in the COOH-terminal domain of the RLC of the other head. In our study S59C shows no RLC-RLC cross-linking, suggesting that at least in this myosin, the contact is not within 8.9 Å. The scallop model also does not predict the crosslinking we observe from Cys-108. Therefore in general our results are not in accordance with the scallop model.
The general effect of adding actin (Table II) during irradiation for the Class I mutants was to permit RLC-RLC crosslinking in the thiophosphorylated states. For the second class of mutants, actin either had no effect (T134C) or allowed for cross-linking in the dephosphorylated state (S59C). These experiments were done with a 2-fold excess of actin over HMM heads. We have not determined whether the cross-linking observed in the actoHMM complex is intra-or intermolecular; however, by the efficiency of cross-linking it is likely that it is intramolecular as in HMM alone. If our interpretation of the cross-linking results in HMM alone (that thiophosphorylation allows for RLC domain separation) is correct, then in the acto- HMM complex, the head-head separation is diminished. This result seems counterintuitive. Binding of HMM heads to adjacent actin monomers would be expected under these conditions. Using the skeletal actoS1 structure as a model (21), two heads attached to adjacent monomers would most certainly involve some distortion in the coiled-coil tail to allow for separation of the RLC domains. Our results suggest that for smooth muscle HMM, the conformations of the heads when bound to actin must differ from this model in the region of the RLC, allowing for proximity of the two RLCs. In contrast, the acto S1 model (21) is consistent with cross-linking between the catalytic domains observed in the smooth actoHMM complex (33). Further experiments are required to establish the molecular basis for our findings. It has been proposed that full regulation of SMM requires interaction of the heads with the S2 portion of the tail (20) in the dephosphorylated state. However, results from deletion mutants of the S2 portion of the rod suggest that a specific amino acid sequence at the head-rod junction may not be critical for full regulation (34). It is interesting to note that, at least for the mutants tested, we did not observe RLC-heavy chain cross-linking, suggesting that the mutant residues did not approach the heavy chain within 8.9 Å. Atomic force images of smooth muscle myosin also did not show clear evidence for interaction of the RLC with the tail region (37).
It is interesting to place our results into the context of previous studies of the effects of phosphorylation on the structure of regulated HMMs. It is difficult to compare results with whole myosin because of its ability to adopt the filamentous and folded (10 S) forms that are not available to the HMM fragment. To date, phosphorylation of smooth muscle HMM has been correlated to four effects. Phosphorylation increases the sedimentation coefficient (35), stabilizes a conformation that appears visually to have the heads extended away from the tails (in contrast to heads bent back toward the tail) (35), increases proteolytic susceptibility of the head-tail junction (36), and increases the mobility of the active site region and the RLC (31). A cryo-atomic force microscopy study showed that the range of separation between the two heads was increased by thiophosphorylation of 6 S smooth muscle myosin molecules at high ionic strength (37). Together these studies show an immobilized dephosphorylated head-tail junction, with closer heads, and a more compact overall structure than the phosphorylated state. Studies with scallop HMM (calcium-regulated) show similar trends induced by calcium binding (38). However, from these relatively low resolution studies it is not possible to compare detailed molecular interactions controlled by phosphorylation.
Our study provides novel information through the use of a site-directed chemical technique that provides higher resolution information about many sites within the RLC subunit. The labeled mutants report a conformation specific to dephosphorylated smooth HMM that involves a close juxtaposition of the two RLC subunits and a separation of the RLC subunits upon phosphorylation. Class I mutations appear to be at the interface between the two RLCs (approaching within 8.9 Å). Our method does not distinguish between a close-fit interaction (tight) and one that is weak but within distance range of the photocross-linking reagent. We found that Class II mutations are unable to cross-link in HMM, which is expected behavior for probes that do not have access to the interface between the RLCs. We are currently extending the power of the photocrosslinking technique to map the sites of cross-linking and to further examine the positioning of the RLC in actoHMM conformations, which are difficult to study with other methods.
FIG. 6. Space-filling homology model of the smooth muscle myosin regulatory domain. This is a comparative protein model that represents the most probable structure (40) for the smooth muscle myosin (gizzard) regulatory domain (RLC, ELC, and heavy chain) generated from a sequence alignment with the scallop regulatory domain and the respective structural coordinates (22). Graphics were drawn with RasMol (Mac Vers. 2.6). The heavy chain is the darkest shade. A, this orientation shows the positions of the Class II mutants (Ser-59 (S59) and Thr-134 (T134); these mutants do not cross-link in either the de-or thiophosphorylated states. The NH 2 terminus of the heavy chain extends to the left toward the catalytic domain. The three-dimensional position is such that the actin filament (left) and the myosin filament (right) are approximately vertical using the chicken actosubfragmentone model (41). The M-line is at the top, and the direction of the powerstroke moves the RLC from the top to the bottom. B, this orientation is identical to that in A, except that the image has been rotated 180°about the vertical axis to look at the back-side of the RLC. Cys-108 is a Class I mutant (only cross-links in the dephosphorylated state). Also shown is Phe-25 (F25), which is the first residue at the NH 2 terminus for which there is structural information.