Projectin-Thin Filament Interactions and Modulation of the Sensitivity of the Actomyosin ATPase to Calcium by Projectin Kinase*

The insect muscle protein projectin (900 kDa) belongs to a novel family of cytoskeleton-associated protein kinases (titin, twitchin, and projectin) that are members of the immunoglobulin superfamily. The functions of these kinases are still unknown although recent data suggest a role in modulating muscle activity and generating passive elasticity. An important question is what are the in vivo substrates for these enzymes. We found a thin filament-associated 30 kDa protein that acts as an in vitrosubstrate for projectin kinase from Locusta migratoria. However, we did not find activators for projectin kinase. Neither calcium, calcium with calmodulin, nor cAMP activated the in vitro activity of projectin kinase. Binding studies revealed a strong interaction between projectin and thin filaments comparable with that of the projectin-myosin interaction. That an interaction might be possible in vivo is suggested by immunological studies showing that projectin is attached to the surface of myosin filaments. Since the molecular weights indicate that the 30 kDa protein might be troponin I, which is known to play a central role in modulating cardiac contractile activity, we studied whether phosphorylation of this protein by projectin changes the calcium sensitivity of the actomyosin ATPase. We found a significant increase in the calcium sensitivity. Thus, our results indicate the existence of a novel mechanism of regulation of muscle activity by a cytoskeleton-associated kinase.

The insect muscle protein projectin (900 kDa) belongs to a novel family of cytoskeleton-associated protein kinases (titin, twitchin, and projectin) that are members of the immunoglobulin superfamily. The functions of these kinases are still unknown although recent data suggest a role in modulating muscle activity and generating passive elasticity. An important question is what are the in vivo substrates for these enzymes. We found a thin filament-associated 30 kDa protein that acts as an in vitro substrate for projectin kinase from Locusta migratoria. However, we did not find activators for projectin kinase. Neither calcium, calcium with calmodulin, nor cAMP activated the in vitro activity of projectin kinase. Binding studies revealed a strong interaction between projectin and thin filaments comparable with that of the projectin-myosin interaction. That an interaction might be possible in vivo is suggested by immunological studies showing that projectin is attached to the surface of myosin filaments. Since the molecular weights indicate that the 30 kDa protein might be troponin I, which is known to play a central role in modulating cardiac contractile activity, we studied whether phosphorylation of this protein by projectin changes the calcium sensitivity of the actomyosin ATPase. We found a significant increase in the calcium sensitivity. Thus, our results indicate the existence of a novel mechanism of regulation of muscle activity by a cytoskeleton-associated kinase.
Accessory proteins closely associated with actin filaments, troponin, and tropomyosin, mediate the Ca 2ϩ regulation of skeletal muscle contraction (reviewed in Ref. 1). Moreover, there is evidence for the involvement of a myosin-linked regulatory system in vertebrate skeletal and cardiac muscle during modulation of muscle contraction (2)(3)(4). Most invertebrate muscles are characterized by both myosin and actin-linked regulation (5). The molecular mechanism of myosin filament mediated regulation of skeletal muscle is still little understood. Phosphorylation of the myosin light chains (MLC) 1 is supposed to play a central role in the regulation of skeletal muscle activity as in smooth muscle contraction (3,6).
The discovery of a novel family of myosin filament-associated protein kinases, such as titin, twitchin, and projectin, homologous to MLC kinases of smooth and skeletal muscles (7)(8)(9) raises the question of whether these participate in the regulation of muscle contraction. These extraordinarily large protein kinases are classified as intracellular members of the immunoglobulin protein superfamily on the basis of their molecular structure (7,8,10). Members of this family are the giant mammalian titin (3 MDa, reviewed in Ref. 11) and the invertebrate mini-titins (0.6 -1.2 MDa, reviewed in Ref. 12).
Mini-titins are distributed throughout the A band of invertebrate muscles, but in insect flight muscles, they are located in the I band (15)(16)(17)(18)(19)(20)(21). First indications of the possible functions of these proteins have been derived from genetic studies on the nematode Caenorhabditis elegans. Mutants of the unc-22 gene encoding twitchin are characterized by a repetetive twitching of the body wall muscle and by an abnormal sarcomere structure (7), suggesting that twitchin-like molecules are directly involved in modulating muscle activity and are also essential for ordered sarcomere assembly. Findings that projectin is able to determine the length of myosin filaments in vitro confirm the idea that these proteins are molecular directives for myosin filament assembly (22). Nevertheless, the function of the kinase domain is still unknown. The recent report on the x-ray crystal structure of the kinase domain of twitchin has given insight into the regulatory power of this kinase and provides evidence for an intrasteric mechanism of protein kinase regulation employing an autoregulatory sequence (23). In vitro autophosphorylation has been described for the bacterially expressed kinase of twitchin from C. elegans (24) and for other members of this family including insect projectin kinase from Drosophila leg and flight muscle (8), crustacean projectin kinase from crayfish tail muscle (25) and mollusc twitchin kinase from Aplysia (26). An endogenous in vitro substrate for mollusc twitchin is the regulatory MLC (27). In contrast, in vitro studies on the kinase catalytic core of twitchin from C. elegans expressed in Escherichia coli, indicate that this twitchin differs in substrate specificity (24). Although synthetic MLC peptides homologous to a part of chicken smooth muscle MLC proved a useful model substrate, MLC peptides derived from C. elegans itself were weakly phosphorylated by twitchin kinase (24). In addition, intact chicken smooth muscle MLC failed as a substrate (24). So far, in vivo substrates for twitchin and other members of the twitchin and titin-like kinases are still unknown. Projectin kinase of insects has been less well studied with respect to potential substrates, even though identification of substrates for mini-titin of different invertebrate species would provide an essential clue for understanding their functions.
In this paper, we demonstrate that projectin from insect muscles is able to autophosphorylate in vitro. Moreover, we identify thin filament-associated 30-kDa protein, not a MLC, as a substrate for projectin. We show that phosphorylation affects the sensitivity of actomyosin ATPase to Ca 2ϩ . Immunoelectron microscopy studies provide evidence that projectin attaches to the surface of myosin filaments. Finally, in vitro binding studies reveal a strong interaction between projectin and thin filaments.

EXPERIMENTAL PROCEDURES
Purification of Projectin and Myosin-Myofibrils were prepared from freshly dissected muscles from the grasshopper Locusta migratoria as described by Etlinger et al. (28). Projectin and myosin were isolated from leg and flight muscle myofibrils as described by Kölsch et al. (22). Washed myofibrils were extracted in 20 mM Na 2 P 4 O 7 , 1 mM CaCl 2, and 10 mM Tris-H 3 PO 4 , pH 8.3, for 1 h by stirring on ice. The insoluble material was removed by centrifugation for 10 min at 20,000 ϫ g. For projectin purification, the supernatant was subjected to chromatography on Mono-Q-FPLC (Amersham Pharmacia Biotech). Subsequently the pellet was used for myosin extraction in 40 mM Na 2 P 4 O 7 , 10 mM ATP, 10 mM MgCl 2 , 5 mM EGTA, and 10 mM Tris-HCl, pH 7.5. Unsoluble proteins were removed by centrifugation for 2 h at 20,000 ϫ g, and the supernatant was applied to Mono-Q-FPLC. Equilibration buffer for projectin chromatofocussing contained 10 mM Na 2 P 4 O 7 , 10 mM Tris-H 3 PO 4 , pH 8.3, and for myosin chromatography, it contained 40 mM Na 2 P 4 O 7 , 10 mM Tris-HCl, pH 7.5. Proteins were eluted from the column with a linear salt gradient (0 -0.5 M NaCl in equilibration buffer). All preparation steps were done at 4°C. Protein concentration was measured according to Lowry et al. (29). The purity of protein preparations was judged by SDS-PAGE using silver and Coomassie Blue staining for protein detection. Proteins were electrophoresed in SDS-polyacrylamide slab gels with the Laemmli buffer system (30) and 2-14% acrylamide gradient.
Generation of Antibodies against Projectin-Rabbits were immunized subcutaneously with 0.5 mg of purified projectin in complete Freund's adjuvant. They were boosted twice at 3-week intervals with 0.25 mg of projectin in incomplete FreundЈs adjuvant. Serum was collected 2 weeks after the last boost. Immunoglobulins were precipitated with 50% (w/v) ammonium sulfate, resuspended in PBS (1/6 of starting volume) and dialyzed against PBS. Antibodies specific for projectin were isolated by passing the dialyzed probe over a projectin affinity column and eluting the bound antibodies with 0.2 M glycine-HCl, pH 2.5. The affinity column was made by coupling projectin to CNBr-activated Sepharose 4B (Amersham Pharmacia Biotech) according to the manufacturer instructions.
Renaturating Blotting Assay-Autophosphorylating endogenous kinases of a myofibril extract, which was also used for chromatographic purification of projectin, were detected using a modified renaturating blotting assay as described by Kato et al. (31). The extract was mixed with an equal volume of sample buffer which contained 5.5% (w/v) SDS, 12.5% (v/v) 2-mercaptoethanol, 12.5 mM EDTA, 156 mM Tris, pH 6.8, 4 M urea, and subjected to 2-14% SDS-PAGE on mini-gels. Proteins were transferred to polyvinylidene difluoride (Immobilon P; Millipore, 0.45 m pore size) membranes in 192 mM glycine, 25 mM Tris at 150 mA for 30 min (semidry blot). Blots were incubated for 1 h at room temperature with gentle rocking in 7 M guanidine hydrochloride, 50 mM Tris, 50 mM dithiothreitol, 2 mM EDTA, 0.1% (v/v) Triton X-100, pH 8.3. Enzymes were renaturated in 50 mM Tris, 2 mM dithiothreitol, 2 mM EDTA, 0.1% (v/v) Tween 20 for 12 h at 4°C. We did not use casein in the renaturation solution and did not block the membrane after renaturation of enzymes as originally described. To determine unspecific binding of ATP to proteins on the membrane, blots were incubated with [␣-32 P]ATP instead of [␥-32 P]ATP or heat inactivated for 1 min at 80°C. Blot strips were incubated in 50 mM Tris, pH 7.0, 5 mM MgCl 2 , 1 mM DTE, 10 M ATP with 50 Ci [␥-32 P]ATP/ml or 50 Ci [␣-32 P]ATP/ml for 6 h at room temperature. The following substances were added: 5 mM EGTA or 5 mM CaCl 2 or 5 mM CaCl 2 with 5 g/ml calmodulin (Boehringer Mannheim) or 10 g/ml cAMP (Boehringer Mannheim). Kinase reactions were terminated by washing blot strips several times in PBS and then in 1 M KOH, each time for 10 min. Blots were washed again in PBS, stained with India ink, dried, and subjected to autoradiography overnight.
Immunogold Electron Microscopy-Longitudinal flight muscles from Phormia terrae-novae and L. migratoria were isolated from the thorax. The flight muscle from locust was stretched about 20% of resting length before fixation. The leg muscle was stretched and fixed in situ. To do this, the femur was opened and the muscles were dissected with the exception of the flexor tibiae. Femur and tibiae were fixed in the 90 o position. The muscles were then fixed in 4% (w/v) paraformaldehyde in 25 mM phosphate buffer, pH 7.2, for 2 h at 4°C. Muscles were cut into small pieces and postfixed for 15 min at room temperature. The fibers were washed in PBS three times for 10 min and dehydrated in 50 and 70% ethanol each for 20 min at room temperature. Fibers were placed in LR White resin overnight at 4°C and than polymerized in resin at 55°C for 24 h.
Silver sections were mounted on Formvar-coated nickel grids and blocked with 1% (w/v) BSA in PBS, pH 7.4 for 30 min. Sections were exposed for 12 h to primary antibody (10 g/ml anti-projectin antibody in blocking buffer) at 4°C, washed with blocking buffer, exposed to the secondary antibody (10 nm gold-conjugated goat antibody against rabbit IgG (Sigma) diluted 1:10 in blocking buffer) for 1 h at room temperature, rinsed successively in PBS and distilled water and stained with uranyl acetate. Sections were examined in a Phillips EM 301 electron microscope at 80 kV.
Myosin and thin filaments were disrupted from myofibrils as described by Huxley (32) and adsorbed to carbon coated grids by floating the grids on drops of the filament suspension for 1 min. Grids were blocked for 15 min with 1% (w/v) BSA in PBS, exposed to primary antibody (1 g/ml anti-projectin antibody diluted in blocking buffer) for 20 min, washed with blocking buffer, exposed to secondary 10 nm gold-conjugated goat antibody against rabbit IgG (diluted 1:10 in blocking buffer), washed with PBS, and negatively stained with uranyl acetate. Control sections or filaments were exposed to nonimmune rabbit IgG.
Protein Kinase Assays and Isolation of Thin Filaments-Purified projectin and crude extracts of flight muscles from Locusta were dialyzed against 200 mM NaCl, 1 mM dithiothreitol in 25 mM Tris, pH 7.0. Purified myosin was dialyzed against 500 mM KCl, 1 mM dithiothreitol in 25 mM Tris, pH 7.0. Thin filaments were prepared as described by Kendrick-Jones et al. (33). Thin filaments of the same preparation were used for binding assays and ATPase measurements. Kinase assays were performed in 200 mM NaCl, 1 mM dithiothreitol, 0.1 mM ATP with 5 Ci [␥-32 P]ATP per 30 l assay in 25 mM Tris-HCl, pH 7.0, and varying Ca 2ϩ concentrations (5 mM CaCl 2 or 5 mM CaCl 2 and 5 mM EGTA or 5 mM EGTA). Projectin was added to a final concentration of 0.2 mg/ml and myosin or filaments or crude extracts to a final concentration of 0.5 mg/ml. Proteins were preincubated for 5 min, and the reactions were started with ATP. Reactions were terminated after 45 min at 30°C by the addition of an equal volume of a 2-fold concentrated sample buffer (4% (w/v) SDS, 12.5% (v/v) 2-mercaptoethanol, 12.5 mM EDTA, 120 mM Tris, pH 6.8, 0.001% (w/v) bromphenol blue and 4 M urea). Proteins were separated by SDS-PAGE. Gels were fixed, stained with Coomassie Brilliant Blue, dried, and used for autoradiography (exposure overnight with intensifying screens).
Solid Phase Binding Assay-Purified projectin from locust flight muscles was biotinylated with D-biotinoyl-⑀-aminocaproic acid N-hydroxysuccinimide ester (Boehringer Mannheim) for 20 min at 4°C. After centrifugation, the supernatant was separated from unbound biotin by gel filtration on a Sephadex G-25. Unlabeled proteins were immobilized on polystyrene ELISA wells at an amount of 300 ng/well for saturation. After blocking with 1% (w/v) BSA, biotinylated projectin was added to each well at different concentrations in high ionic strength buffer containing 0.5 M NaCl in PBS, pH 7.2, and incubated for 1 h at 37°C. The wells were washed five times with PBS, pH 7.2. After reaction with avidin-peroxidase in PBS with 0.1% (w/v) BSA and with 0.05% (v/v) Tween 20, detection was started with a substrate solution using H 2 O 2 and O-phenylenediamine (Sigma) as the chromogen. One molar H 2 SO 4 was used to stop the reaction. The absorbance was measured with a micro-ELISA reader (Dynatech) at 492 nm.
ATPase Assays-Thin filaments (250 g/ml), isolated from locust flight and leg muscles as described previously (33), were preincubated with active or with heat-inactivated (1 min at 80°C) projectin kinase (25 g/ml) in 150 mM NaCl, 10 mM histidine, pH 7.0, 50 mM MgCl 2 , 10 mM ATP, and variable calcium concentrations for 30 min at 30°C. The free Ca 2ϩ concentration was adjusted according to Portzehl et al. (34). ATPase reactions were started by the addition of myosin (250 g/ml) and stopped after 10 min with an equal volume of 10% (v/v) trichloroacetic acid. ATPase activity of myosin was determined by measuring the amount of inorganic phosphate as described previously (35).

RESULTS
To characterize the pattern of protein phosphorylation by endogenous kinases, extracts of purified myofibrils were incubated with [␥-32 P]ATP and various Ca 2ϩ concentrations. Fig. 1 shows the protein composition of myofibrils. Reaction mixtures were analyzed by SDS-PAGE and autoradiography (Fig. 4). Several phosphoproteins were detected in these experiments, and projectin was labeled strongly with 32 P using leg and flight muscle extracts. The pattern of protein phosphorylation was not influenced by calcium, but phosphorylation of projectin decreased slightly in the presence of calcium (results not shown).
In an initial step, we tested whether locust projectin exhibits kinase activity. The kinase activity of projectin was determined with a renaturating blotting assay. Myofibrillar extracts were subjected to SDS-PAGE, and separated proteins were transferred to membranes. After renaturation of proteins, blot strips were incubated with [␥-32 P]ATP. Autoradiography revealed three autophosphorylating kinases in myofibrillar extracts from leg and flight muscles ( Fig. 2A). Control strips that had been incubated with [␣-32 P]ATP showed that the radioactive labeling was not caused by unspecific binding of ATP to proteins ( Fig. 2A). The high molecular weight band on the autoradiogram was identified as projectin by comparison with immunoblots (Fig. 3). To test the influence of several known activators of protein kinases on projectin kinase activity, blot strips were incubated with [␥-32 P]ATP and Ca 2ϩ , Ca 2ϩ and calmodulin, or cAMP. Fig. 2B shows that these kinase activators did not have a significant effect on the kinase activity of renaturated projectin.
We then focussed our interest on substrates for projectin kinase. We purified projectin chromatographically (see Fig. 1). Projectin was incubated in the presence of [␥-32 P]ATP and of a myofibrillar extract from which projectin itself was removed. Fig. 4A shows the SDS-PAGE analyses of the projectin preparations and extracts. The corresponding autoradiogram (Fig.  4B) shows that the phosphorylation of proteins with molecular weights of approximately 36 and 30 kDa was strongly increased in the presence of projectin. From these results, we speculated that MLC might be a substrate for projectin kinase. However, we could not find phosphorylation of MLC by projectin using purified myosin (see Fig. 1) as a potential substrate in kinase assays (results not shown).
In a next step we isolated thin (native actin) filaments to determine whether thin filament-associated proteins are targets for projectin kinase. The protein composition of the thin filaments was analyzed by SDS-PAGE (see Fig. 1). The major proteins of the locust thin filaments were actin (42 kDa) and the regulatory proteins: troponin T (55 kDa), tropomyosin (38 kDa), troponin I (32 and 30 kDa), and troponin C (approximately 20 kDa) (41). The higher molecular mass proteins were not characterized. Immunoblots with anti-myosin antibodies as well as electron microscopy revealed that the preparation was not contaminated with myosin (results not shown). Projectin was incubated with thin filaments, [␥-32 P]ATP and varying calcium concentrations. Reaction mixtures were analyzed by SDS-PAGE (Fig. 5A) and autoradiography (Fig. 5B). Fig. 5B shows that a protein with a molecular weight of 30 kDa (arrow in Fig. 5B) was phosphorylated in the presence of projectin. Multiple bands including a 32-kDa band are phosphorylated in the control, where thin filaments were incubated without projectin, indicating that the thin filaments contain associated kinases. A 30-kDa phosphoprotein, however, is only visible if

FIG. 1. SDS-PAGE analyses of myofibrils, thin filaments, myosin, and projectin preparations. Lanes 1-8 (from left to right) show
Coomassie Blue-stained gels (2-14% gradient), and lanes 9 and 10 show silver-stained gels (2-14% gradient). Note that lanes 9 and 10 are derived from different gels than lanes 6-8, and therefore the mobility of projectin differs slightly. Samples are indicated on top. Position of protein molecular weight standards analyzed in parallel are indicated on the left, and numbers indicate the molecular mass in kDa. Note that projectin and myosin preparations were free of contaminating proteins. Considering the molecular weight, the thin filament preparation consists predominantly of actin (42 kDa), troponin T (55 kDa), tropomyosin (38 kDa), proteins in the range of 30 kDa (probably troponin I isoforms), troponin C (20 kDa), and some protein of higher molecular mass of unknown identity (42 kDa). The 200-kDa band was not a myosin contamination as determined by immunblots with anti-myosin antibodies (results not shown). projectin is added to a thin filament phosphorylation assay. The 30-kDa protein is not a contamination of the projectin preparation as shown in the autophosphorylation assay (Fig.  5). The 55-kDa protein that has been weakly phosphorylated in the autophosphorylation assay (Fig. 5) is possibly a degradation product of projectin. The intensity of this band is slightly enhanced in the presence of calcium, suggesting a calcium-dependent degradation of projectin by proteases (results not shown).
We have not been able to identify the 30-kDa protein up to now because there are few data about the protein composition of thin filaments from Locusta migratoria, and sequence data on locust troponin are lacking. However, for thin filaments of other invertebrates, proteins with a molecular mass of approximately 30 kDa have been identified as troponin I, the inhibitory component of the troponin complex (36 -40). Since the molecular mass of the 30 kDa protein, phosphorylated by projectin, indicated that it might be troponin I, we reasoned that phosphorylation of this protein might affect the actomyosin ATPase activity. To determine whether this is the case, we performed ATPase assays with isolated myosin and the same thin filaments we used for phosphorylation studies in the presence of active or heat-inactivated projectin at varying calcium concentrations. In Fig. 6 the ATPase activity of protein preparations of flight and leg muscles is plotted against the calcium concentration. To rule out that the Ca 2ϩ -sensitivity might be influenced by steric parameters, the experiments were performed with heat-inactivated projectin kinase (80°C for 1 min) before preincubation as control. The half-maximal ATPase ac-tivity (pCa 50 ) was increased from 5.85 to 6.35 when thin filaments were preincubated with active projectin kinase before starting the reaction by the addition of myosin (Fig. 6A). For proteins of leg muscles, pCa 50 was increased from 5.77 to 6.22

FIG. 3. Specifity of anti-projectin antibodies revealed by immunoblots.
Proteins as indicated on top were separated on 4 -14% SDS-polyacrylamide mini-gels, and the semi-dry blot of the gel was exposed to antibodies against projectin (1 g/ml). Bands were visualized with goat antibodies against rabbit IgG, which were conjugated with horseradish peroxidase and diaminobenzidine as chromogenic substrates. The antibody did not show cross-reactions with myosin or other myofibrillar proteins.   , and panel B shows the corresponding autoradiogram. A thin filament-associated protein with a molecular mass of approximately 30 kDa (marked by arrow) was phosphorylated in the presence of projectin (200 g/ml). In addition, 30 and 21 kDa proteins were phosphorylated in the presence of projectin, but weak phosphorylation was also observed if thin filaments were incubated without projectin. The autophosphorylation assay shows a weak additional band with an molecular mass of approximately 55 kDa on the autoradiogram. This is probably a degradation product of projectin. (Fig. 6B). The ATPase activity of myosin in the absence of thin filaments was not affected by projectin (results not shown). Thus, a thin filament-associated protein must be a substrate for projectin and, in addition, is obviously able to regulate the Ca 2ϩ sensitivity. This finding increased our interest on projectin-thin filament interactions.
We performed in vitro binding studies to study the projectinthin filament interaction. Thin filaments, containing troponin and tropomyosin, or purified myosin from flight muscles were immobilized on microtiter wells and exposed to increasing projectin amounts. BSA was used as a negative control protein.
The binding capacity of the polystyrene wells was 100 ng of protein/well according to the manufacturer instructions. We found that for myosin and for thin filaments, the maximum binding of projectin was achieved when 30 nM projectin was added per well (Fig. 7). Higher projectin concentrations resulted in decreased binding of projectin to thin filaments or to myosin, possibly due to steric hindrance. The amount of projectin bound to thin filaments or to myosin, represented as absorbance, is within the same range.
Because the arrangement of projectin within the myosin filaments of insect muscles is unknown, we studied the ultrastructural localization of projectin in asynchronous flight muscle of P. terrae novae and in synchronous flight muscle and leg muscles of L. migratoria using immunogold labeling. Projectin was restricted to the I band (Fig. 8B) in asynchronous flight muscle. This has also been described previously for flight muscles of other diptera (16, 17, 21). However, in synchronous locust flight muscle and leg muscle, projectin is distributed throughout the A band (Fig. 8, D and F). Immunogold labeling of isolated filaments confirmed these results and, in addition, revealed that projectin is attached to the surface of myosin filaments (Fig. 8G). However, we did not find projectin bound to thin filaments, indicating that specific conditions must be met for a strong interaction between projectin and thin filaments to occur. This result conflicts with earlier data from immunofluorescence studies, which suggested that projectin connects the myosin filaments to the Z disc in synchronous insect flight muscles (11). DISCUSSION The data presented here demonstrate the kinase activity of projectin, a twitchin-like muscle protein of insects. In addition to autophosphorylation, which has already been described for Drosophila projectin (8), we identified an in vitro substrate for projectin kinase activity. Using isolated thin filaments as a substrate in kinase assays, we showed that projectin phosphorylates thin filament-associated 30 kDa protein. On the basis of its molecular mass, we tentatively assume this protein to be troponin I. Furthermore, we have shown that phosphorylation of the 30-kDa protein increases the Ca 2ϩ sensitivity of the actomyosin ATPase activity. We did not find a significant effect of Ca 2ϩ -calmodulin, cAMP, and calcium on autophosphorylating activity.
Earlier studies on the molecular structure of the kinase domain have indicated that MLC may be an in vivo substrate for projectin because the kinase domains resemble strongly the catalytic domain of MLC kinase (7,8). However, in our studies, we failed to phosphorylate MLC with projectin. In contrast, mollusc twitchin does phosphorylate MLC (27). So far it is unclear whether this different substrate specificity in vitro reflects physiological differences in the activation of muscle contraction in mollusc and insect muscles. Mollusc muscles contain only a myosin filament-associated regulatory system (41), whereas in most other invertebrates, regulatory mecha- nisms exist that are mediated by both actin and myosin filaments (5).
For the understanding of the regulation of twitchin-like kinases and the signaling pathways involved, it is important to obtain insight into the functions of these enzymes. It has been suggested that calmodulin could activate twitchin and titin because the sequence of the autoinhibitory region of twitchin kinase is similar to that of the calmodulin binding domain and because binding of calmodulin to twitchin and titin has been shown (42). However, although calmodulin binds to Aplysia twitchin, it does not activate it (26). We likewise did not find an activation of projectin kinase by Ca 2ϩ -calmodulin. Heierhorst et al. (43) have recently reported that Ca 2ϩ -S100A1, a member of the large Ca 2ϩ -binding protein family, greatly enhances the enzyme activity of the autoinhibited Aplysia twitchin fragment TWK-43 and C. elegans twitchin. These authors also demon-strated that the S100A1 2 -binding site is a part of the autoregulated sequence positioned in the active site responsible for intrasteric autoinhibition of twitchin kinase. Twitchin is the first enzyme to be described that is activated by a member of the S100 family. Because twitchin and projectin are closely related proteins, it will be interesting to determine whether S100A1 protein also activates insect projectin.
cAMP and Ca 2ϩ intracellular signaling pathways interact at several levels in the hierarchy of control (44). Regarding this aspect, it is noteworthy that mollusc twitchin has been identified as the major substrate for neuropeptide-activated cAMPdependent protein kinase and that the phosphorylation state was shown to correlate with the extent of neuropeptidergic modulation of muscle relaxation in vivo (45). We have not yet determined whether locust projectin is a target for cAMP-dependent kinases. In vertebrate cardiac muscles, cAMP pathways are known to be involved in the regulation of muscle contraction. cAMP-dependent phosphorylation of troponin I decreases the Ca 2ϩ -sensitivity of the ATPase of cardiac actomyosin (46). Despite the evolutionary divergence between vertebrates and invertebrates, a similar mechanism, with troponin as a target for protein kinases, might exist in invertebrate muscles. Our result that phosphorylation of a 30-kDa protein associated with thin filaments increases the sensivity of actomyosin to Ca 2ϩ points to this new aspect of invertebrate muscle regulation.
Our finding that projectin does not phosphorylate MLC raises the question how does projectin interact with thin filaments in vivo. To determine whether such an interaction as we have observed in in vitro binding studies is possible with projectin and thin filaments in vivo, we studied the arrangement of projectin within myosin filaments in more detail. Previous information on the localization of projectin in synchronous insect flight muscles is based on immunofluorescence studies (12). However, interpretation of fluorescence staining may be complicated since immunoreactivity may reflect either abundance or accessibility of the labeled species. We have tried to overcome this problem using two different approaches, labeling of isolated filaments and postembedding immunogold labeling. The results of these studies indicate that projectin binds to the surface of the myosin filaments along their whole length. Isolated thin filaments were not labeled with antibodies against projectin. This may be because of the isolation conditions for myosin and thin filaments used in this study. Confirmatory results have recently arisen from experiments which show that titin inhibits thin filament motility in vitro in a calcium-dependent manner (47). The authors also demonstrated that titin interacts with native thin filaments and F-actin and that increasing calcium concentration results in enhanced binding.
Another aspect of the possible physiological function of the thin filament-projectin interaction is the proposed elasticity of this class of molecules. The molecular structure of these proteins makes them a plausible candidate for generating passive elasticity in muscles, and for titin, new data corroborating this idea exist (13,14). The sticky proteins consist predominantly of fibronectin type III-like domains and IgG-like domains (7,8,10). The fibronectin type III-like domains of this protein class are thought to constitute the molecular basis for elasticity as one domain can reversibly unfold from 4 to 29 nm (48). Because projectin in many invertebrate muscles obviously does not anchor the myosin filaments in the Z disc as titin does, it is difficult to explain how projectin should generate passive tension. An elastic connection would be possible, however, if projectin binds to thin filaments.
If troponin I is an in vivo substrate for projectin, both enzyme and substrate would be fixed within the cell. Interaction could FIG. 8. Localization of projectin in situ revealed by immunoelectron microscopy (indirect immunogold labeling). A and B, asynchronous flight muscles of Phormia terrea novea. Panel B shows the labeling of an ultrathin section of LR White-embedded flight muscles with anti-projectin IgG. Panel A shows a serial section that was exposed to nonimmune rabbit IgG. Note the deposition of gold particles at both sides of the Z disc. C-H, synchronous flight muscle and leg muscle of L. migratoria. Ultrathin sections of LR White resin embedded leg (D) and flight (F) muscles were exposed to anti-projectin antibodies. Gold particles are clearly restricted to the A band in both muscle types. Control sections were exposed to nonimmune IgG (C and E). Immunogold labeling of isolated locust flight muscle thick (arrows) and thin (arrowheads) filaments with anti-projectin IgG results in deposition of gold particles along the thick filaments (G). Panel H shows control. No gold particles were found on thin filaments. The average diameter of the gold particles is 10 nm. Scale bars ϭ 500 nm be controlled by the dynamics of the cytoskeleton, which therefore would be involved in the regulation of the enzyme activity. We postulate that the projectin-thin filament interaction could represent an intracellular sensor that could provide dynamic feedback to the cytoskeleton by transforming mechanical into biochemical signals.
Our results on projectin and thin filament interactions may give new insight into the function of these cytoskeleton-associated protein kinases. Studies with specific inhibitors like antibodies which inhibit projectin kinase activity in muscles will be a powerful tool for investigating whether a correlation exists between the phosphorylation of actin filament proteins by projectin kinase and muscle activity.