Activation of Membrane Guanylate Cyclase by an Invertebrate Peptide Hormone*

Peptide hormones can stimulate cyclic GMP synthe- either of two general mechanisms: a coupling factor the membrane interacting We the metabolism cyclic and further suggests that the membrane cyclase is the most likely target of CHH. Membrane and soluble cyclases can be isolated from homogenates of lobster muscle (in a 3.5: 1 ratio), and both are stimulated by Mn2+ and inhibited by Ca’+. CHH has no effect on the soluble enzyme. Coupling of CHH receptors to the particulate cyclase, however, remains intact in isolated membranes, thus providing a new model system for the study of receptor/ cyclase interactions.

This indicates that the EDRF pathway, although prominent in vertebrate muscle, is not found in crustaceans and further suggests that the membrane cyclase is the most likely target of CHH. Membrane and soluble cyclases can be isolated from homogenates of lobster muscle (in a 3.5: 1 ratio), and both are stimulated by Mn2+ and inhibited by Ca'+. CHH has no effect on the soluble enzyme. Coupling of CHH receptors to the particulate cyclase, however, remains intact in isolated membranes, thus providing a new model system for the study of receptor/ cyclase interactions.
Cyclic GMP is an important second messenger in vertebrate and invertebrate muscle. Mammalian vascular smooth muscle, mammalian cardiac muscle, and the skeletal muscles of invertebrates, like the lobster, are enriched for cyclic GMPdependent kinase and its substrates (Casnellie and Greengard, 1974;Rapoport et al., 1982;Lincoln and Johnson, 1984;KUO and Greengard, 1970;Goy et al., 1984). Furthermore, specific stimuli elevate cyclic GMP levels in these tissues. In vertebrate animals, a variety of vasodilating hormones raise cyclic GMP levels and subsequently induce vascular and cardiac relaxation (reviewed in Waldman and Murad, 1987;Tremblay et al., 1988;and Fischmeister and Hartzell, 1987). In the lobster and other crustaceans, a peptide hormone called crus-* This work was supported by National Institutes of Health Grants NS21290 and NS25915. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
In vertebrates, at least three different enzymes control tissue levels of cyclic GMP. 1) A membrane-specific form of guanylate cyclase is regulated in many tissues, including vascular smooth muscle, by the members of the atria1 natriuretic peptide (ANP) family (Waldman et al., 1985;. The family includes a set of homologous peptides produced by the atrium of the heart and related forms produced in the brain (Seidah et al., 1984;Tanaka et al., 1984;Saper et al., 1985;Sudoh et al., 1988). These peptides selectively turn on the membrane cyclase, probably by interacting with an extracellular binding domain of this membrane-spanning enzyme (Schulz et al., 1989). 2) A Cytoplasmic form of guanylate cyclase is activated in many tissues, again including vascular smooth muscle, by an endogenous regulatory molecule called endothelium-derived relaxation factor (EDRF). In the case of smooth muscle, EDRF is produced in the vascular endothelium in response to a variety of hormones, including the peptides bradykinin, cholecystokinin, substance P, thrombin, and calcitonin gene-related peptide (Griffith et al, 1988). Once produced by the endothelium, EDRF diffuses into the cytoplasm of nearby smooth muscle cells (limited primarily by the relatively short half-life of the EDRF molecule), where it directly stimulates the soluble cyclase. 3) A cyclic GMPspecific phosphodiesterase is regulated by light in vertebrate photoreceptors (reviewed in Stryer, 1986). The capture of a photon of light activates the phosphodiesterase via a GTP binding protein (transducin), which in turn produces a decrease in photoreceptor cyclic GMP levels.
The present study is directed toward identifying the mechanism(s) by which CHH regulates cyclic GMP metabolism in lobster exoskeletal muscle. Because of similarities between this tissue and vertebrate vascular smooth muscle, we were particularly interested in seeing whether CHH would act through the membrane cyclase, as do the peptides in the ANP family, or via EDRF and the soluble cyclase, as do other vertebrate peptides such as bradykinin. We have also investigated the possibility that CHH raises cyclic GMP levels in lobster muscle by inhibiting the cyclic GMP-specific phosphodiesterase (the converse of the effects of light in retinal photoreceptors).

20220
This is an Open Access article under the CC BY license.

MATERIALS AND METHODS
concentration of methylene blue. After an additional 90 min, the muscle nieces were frozen and processed as above. Preparation of Crude and Purified CHH-CHH is produced by the sinus gland, a well known crustacean neurohemal organ (Keller et al., 1985). In the lobster, as in other crustaceans, two forms of CHH have been identified. ' The major form (called peptide G1 in previous studies) has been purified (Pavloff and Goy, 1990) and its sequence is comparable to that recently established for CHH purified from the crab (Kegel et al., 1989). The minor form, comprising about 10% of the total activity, appears to be closely related to the major form, perhaps differing by-only one or two amino acids (see Newcomb, 1983. andHuberman andAauilar, 1986). These two isoforms of CHH are the only molecules present in the sinus gland with significant stimulatory effects on cyclic GMP metabolism.' For the current studies, sinus glands were identified and dissected as described previously (Goy et al., 1987a).  (Goy and Kravitz, 1989). A reduction in depolarization-evoked muscle contraction is observed, but we believe this is the normal response of the muscle to elevated cyclic GMP levels (Goy et al., 1987b   was adapted. For this assay, either tail muscle or the opener muscle of the dactyl of the walking leg was used. After dissection under cold saline, muscles were rinsed once with cold buffer B (50 mM HEPES, pH 7.4, with 1 mM EDTA, 0.01% bacitracin, and 0.1 mM phenylmethylsulfonyl fluoride) and then homogenized in ice-cold buffer B (1 ml of buffer/O.5 g of tissue wet weight) using a motor-driven Potter-Elvehjem homogenizer with a nylon pestle. After removing a portion of this material (crude cyclase), the remainder was centrifuged for 60 min at 100,000 X g (4 "C). The upper twothirds of the supernatant were carefully removed (to yield the soluble cyclase preparation).
Earlier studies have shown that the cytoplasmic fraction generated by this procedure contains less than 1% of the membrane lipid initially present in the crude homogenate (Goy et al., 1984). The remainder of the supernatant was then discarded and Buffer C (Buffer B + 100 mM KCl) was used to resuspend the pellet (3 ml of buffer/O.5 g of initial tissue wet weight). The resuspended material was again centrifuged for 60 min at 100,000 X g (4 "C), and the new pellet was resuspended by homogenization in Buffer B (3 ml/O.5 g of initial tissue wet weight) and subjected to a third centrifugation for 60 min at 100,000 X g (4 "C  Martell andSmith (1974,1975

Phosphodiesterase Inhibitors Amplify the Effects of CHH in Intact Tissues
A first step toward understanding regulation of cyclic GMP metabolism in lobster tissues is to define the mechanism(s) by which CHH increases cyclic GMP levels. In general, CHH might act either by stimulating the rate of synthesis of cyclic GMP (by activating one of the forms of guanylate cyclase) or by reducing its rate of breakdown (by inhibiting phosphodiesterase). One way to distinguish these two classes of mechanisms is to compare the response of the target tissue in the presence and absence of a phosphodiesterase inhibitor.
If the peptide stimulates cyclase, then the magnitude of the response should be increased by the inhibitor, whereas if it inhibits phosphodiesterase the response should be decreased. As seen in Fig. la, the response to CHH by intact muscle is greatly potentiated by the phosphodiesterase inhibitor IBMX, both at saturating and subsaturating levels of peptide. A similar result (not shown) was obtained with a second phosphodiesterase inhibitor, M&B22948, thought to be more selective for the form of phosphodiesterase that preferentially breaks down cyclic GMP (Bergstrand et al., 1977). Thus, the primary effect of the peptide is probably to stimulate cyclase rather than to inhibit phosphodiesterase.
The initial rate of the response to CHH in intact tissues is roughly linear, both in the presence and absence of IBMX (Fig. lb). Over a period of 20-50 min, however, a plateau is attained. Presumably, this represents the new equilibrium level attained as a result of enhanced synthesis in the presence of the peptide. The plateau level is maintained with only a slight decline for as long as the peptide is present (up to 180 min). This suggests that if the CHH receptor desensitizes, it does so only at a slow rate.
When the peptide is washed away, the elevated levels of cyclic GMP begin to fall (Fig. lc). The time course of recovery, however, is not simple. Initially, the rate of cyclic GMP breakdown after peptide washout is relatively rapid. This is followed by an extended period (commencing about 50 min after washout) in which the rate of decline is much slower. The biochemical basis for the biphasic nature of the recovery process is not understood. We have noted, however, that during the late phase the tissue is capable of responding normally to a new application of peptide (added at the arrow in Fig. lc).
Because CHH is a very hydrophobic molecule (Pavloff and Goy, 1990), we considered the possibility that the washing procedure might fail to completely remove the peptide from binding sites within the tissue. If so, then subsequent slow release of nonspecifically bound peptide could have produced a low level of free peptide, thereby keeping the level of cyclic GMP elevated above baseline. However, when we subjected tissues to a second washing step 25 min after the initial wash, the subsequent decline of cyclic GMP levels could not be distinguished from that of tissues that had only received the initial wash (compare filled and open circles at 260 min in Fig.  lc). Thus, the late phase of the recovery process probably does not result from incomplete removal of the peptide. In all of the experiments in Fig. 1 we have used both purified CHH (square symbols) and an unpurified CHH-containing extract made from the sinus gland (circular symbols). The results are essentially identical in either case, which suggests that there are no additional hormonal agents produced by the sinus gland whose function is to modify the CHH response.

CHH Activates a Guanylate Cyclase in Muscle Homogenates
Establishing an Assay for Lobster Guanylute Cyclose-Because the results in the preceding section suggest that CHH is likely to act by increasing the rate of cyclic GMP synthesis, we attempted to directly measure guanylate cyclase activity in fractionated lobster tissues. The assay that we used is based on methods that have been successfully applied to homogenates of vertebrate tissues . The ability to quantitatively measure cyclase activity is demonstrated by several features of the assay. 1) Cyclic GMP synthesis is linear for at least 30 min (Figs. 2a and 4). We believe that this approximates the true rate of synthesis, since high levels of IBMX significantly block the enzymatic contribution of phosphodiesterases.
Although it is possible that the lobster preparation contains an IBMX-insensitive phosphodiesterase, we consider this unlikely because no such enzyme has been reported in the literature (Beavo, 1988).
2) The response reflects the rate at which the cyclase enzymes are working rather than limitations imposed by the availability of substrate. As shown in Fig. 2a, cyclic GMP synthesis is not stimulated by doubling the levels of GTP in the reaction mixture. 3) Synthesis can be stimulated, however, by increasing the amount of tissue in the incubation mix. Fig. 2b shows the rate of synthesis obtained as the concentration of purified membranes is varied. The response is proportional to the relative dilution, as expected if the assay accurately reflects the level of enzyme activity. All other experiments reported in this study were performed at rates of synthesis below 0.2 pmol of cyclic GMP/min/40-pl sample, well within the range of demonstrated linearity. 4) As in vertebrate animals, the activity of the lobster enzymes can be enhanced by Mn*+ (Figs. 2c and 3 much protein recovered in the particulate fraction as in the cytoplasmic fraction, the average ratio of membrane to soluble cyclase in intact tissue is calculated to be about 3.5:1. We have noted, however, that the recovery of membrane cyclase is much more variable from preparation to preparation than is the recovery of cytoplasmic cyclase (the standard deviation as a percent of the mean for the former is nearly twice as great as that for the latter). The reason for this is not yet known. It is possible that the membrane cyclase is up-or down-regulated in response to some environmental signal (for example, the molt cycle or the season) and thus varies from animal to animal, while the cytoplasmic cyclase remains relatively insensitive to such forms of regulation.
Alternatively, the membrane cyclase may be more susceptible to losses during handling than the cytoplasmic cyclase, and its recovery therefore more variable. In favor of this latter interpretation, we have observed that cyclic GMP synthesis with the soluble enzyme is linear over the longest intervals tested (up to 210 min), while the membrane enzyme consistently inactivates with time. The rate of inactivation is somewhat variable, but usually complete within 60 to 90 min (see Figs. 2a and 4).
When tested separately, each form of the cyclase is stimulated by Mn2+ and inhibited by Ca*+ (Figs. 2c and 3). In contrast to some systems, where Ca*+ has dissimilar actions on the membrane and soluble forms of the enzyme (reviewed in Goldberg and Haddox, 1977), no differential effects are observed in the lobster. Inhibition by Ca*+ may play some part in the physiological regulation of lobster cyclase activity, since the threshold for its effect (1.5 X lo-" M) is within the range that can be attained by cells under normal conditions. It is not known whether inhibition results from a direct interaction of the ion with the cyclase molecules or indirectly via some other enzyme (such as a Ca2+-dependent kinase or protease).
Effect of CHH on Membrane and Cytoplasmic Cychses-When either crude sinus gland extracts or purified CHH are included in the guanylate cyclase assay, they markedly and reproducibly enhance the activity of the membrane form of the enzyme, with little or no effect on the soluble form (Fig.  4). The average effect of crude sinus gland extract on cyclase activity (measured after 30 min of exposure at a concentration of 0.175 gland/ml, approximately the EC&) is 273 f 43% of the control level for isolated membranes (mean f S.E., n = 9) and 101 f 3% of the control level for isolated cytoplasm (mean f S.E., n = 5). Under the same conditions, purified CHH at 0.3 pmol/ml (approximately the EC&) raises membrane cyclase activity to 299 f 53% of the control rate (mean f S.E., n = 3) while again having no effect on isolated cytoplasm (100 f l%, n = 2). The lack of an effect on the soluble enzyme is probably not an artifact resulting from saturation of some step in the pathway, since Mn*+ can stimulate the soluble enzyme more than 3-fold above its basal rate (Fig. 3).
As mentioned above, the effect of the peptide on the membranes is unlikely to result from inhibition of a phosphodiesterase, since the assay mixture contains high levels of IBMX. Data are from experiments with isolated membranes. Similar data (not shown) may be obtained with crude homogenates or with isolated cytoplasm. a, linearity of the response at st.andard (2 mM, w) and elevated (4 mM, W) GTP levels. Each data point for the time course at 2 mM GTP is the mean (*SE.) of 3 to 6 determinations from 3 separate experiments. The data points at 4 mM GTP are the mean (H.E.) for 3 determinations made during one of the 2 mM experiments. The line was fitted by linear regression to the pooled 2 and 4 mM data (for incubation times of 60 min or less), with a correlation coefficient (R' value) of 0.978. When the two sets of data were fitted separately, there was a small decrease (12%) in the rate at 4 mM GTP compared to that at 2 mM. We do not know if this slight apparent inhibition is significant. b, the rate of cyclic GMP synthesis is dependent on the amount of membrane added to the assay. Each point gives the rate of synthesis at a single concentration of protein. Rates were obtained by measuring cyclic GMP levels in duplicate at 0, 10, and 20 min of reaction time, then determining the slope of a line fitted to these data points by linear regression. In all cases, the correlation coefficients for the fitted lines were greater than 0.95. Data were obtained on 2 days, using aliquots frozen from a single batch of membranes with particularly high cyclase activity (approximately 0.4 pmol/min/mg of protein). The line drawn in the figure was determined by linear regression analysis of the rates as a function of protein concentration (correlation coefficient = 0.93). c, the rate of cyclic GMP synthesis is stimulated by Mn*+ (M) and inhibited by Ca2+ (A-A). Each point is the mean (*range) of duplicate determinations made in a single experiment. Lines were fitted by linear regression. These results, therefore, provide strong evidence that the peptide can activate a membrane-bound form of guanylate cyclase. The properties of the response of the purified membranes are consistent with those in the intact cell, except that the efficiency of coupling of receptors to cyclase appears to be reduced in the homogenate.
In intact cells, when tested at the EC&, the peptide can cause 50-to loo-fold increases in cyclic GMP levels (Fig. 1, and Goy et al., 1987a). In contrast, at the EC&,,, the peptide causes only a 2-to 3-fold stimulation of cyclic GMP synthesis in homogenates (see above and Fig.  4). A similar disparity in the magnitude of the response of intact compared to homogenized tissues has been noted for the ANP-sensitive cyclase obtained from vertebrate sources (Hamet et al., 1986).

Investigation of a Possible Contribution of the Soluble Cyclase
The results in the preceding section demonstrate that CHH can activate membrane-bound cyclase, but do not rule out the possibility that, in the intact cell, it also activates cytoplasmic cyclase. The failure to observe an effect, on the soluble enzyme in the homogenate studies may simply reflect the fact that the receptors for the peptide are associated with the membrane, and hence cannot interact with soluble cyclase after fractionation.
This idea is somewhat attractive, since it could explain why the response in the intact cell is larger than the response in purified membranes.
In vertebrates, the only definitively characterized route for activation of the soluble cyclase is the EDRF pathway. The identity of EDRF is somewhat controversial. Among the leading contenders are nitric oxide (Palmer et al., 1987;Ignarro et al., 1987) and nitrocysteine (Myers et al., 1990). Once released, EDRF appears to stimulate the activity of the soluble cyclase through the transfer of electrons, perhaps to a heme group associated with the enzyme (Craven and DeRubertis, 1978). Several pharmacological tools have proven useful in the investigation of EDRF-dependent processes (Furchgott and Vanhoutte, 1989 CHH-stimulated cyclic GMP synthesis in intact muscle is Ca'+-independent (Goy et al., 1987a), while EDRF-stimulated cyclase activity in intact vertebrate tissues requires the presence of extracellular Ca'+ (Waldman and Murad, 1987). Taken together, the Ca*+ and methylene blue results strongly suggest that the EDRF pathway does not mediate the effects of CHH.
In further support of this idea, Fig. 5b shows that intact lobster muscle fails to elevate cyclic GMP levels when treated with nitroprusside at a concentration more than 50-fold above that needed to activate the soluble cyclase in vertebrate tissues. We have also tested nitroprusside on intact lobster muscle at a variety of lower concentrations and for times ranging from 1 to 90 min, in all cases without effect (data not shown). As a control for the possibility that nitric oxide was produced by the nitroprusside but was unable to enter the cytoplasm to interact with the soluble cyclase, we applied nitroprusside to a broken cell preparation. Fig. 512 shows that the activator has no effect in a crude homogenate of lobster muscle (which contains both membrane and soluble cyclases).
As positive controls, we have demonstrated that nitroprusside has the expected effect on cyclic GMP synthesis both in intact rat sensory neurons (Fig. 5b) (Gruetter et al., 1979;Garthwaite et al., 1988) (Fig. 5~). These tissues have been shown in other studies to contain a functioning EDRF pathway (Wood et al., 1989;Furchgott and Vanhoutte, 1989). The nitroprusside data suggest strongly that the EDRF pathway not only does not mediate the effects of CHH, but, in fact, is probably not even expressed in lobster muscle. DISCUSSION From previous studies, many lines of evidence suggest that cyclic GMP metabolism plays a significant regulatory role in crustacean muscle. 1) The cytoplasmic levels of cyclic GMP in lobster muscle are much higher than the levels in comparable vertebrate tissues (Goy et al., 1987a).
2) The tissue content of cyclic GMP-dependent protein kinase is among the highest measured in any species (Kuo and Greengard, 1970;Kuo and Shoji, 1982). 3) Endogenous substrates for cyclic GMP-dependent protein kinase are readily demonstrated in lobster muscle (Goy et al., 1984), although they have been difficult to find in vertebrate animals (Nairn and Greengard, 1983). 4) The sinus gland, a prominent lobster neurosecretory organ, contains a peptide (CHH) that dramatically increases the cyclic GMP levels of lobster muscle, and other lobster tissues as well (Sedlmeier and Keller, 1981;Goy et al., 1987a). 5) CHH regulates glycogen metabolism in crustacean muscle (Keller and Andrew, 1973). Preliminary data also suggest that CHH may alter the contractility of lobster muscle, perhaps by modifying excitation-contraction coupling (Goy et al., 1987b).
The experiments in this paper have addressed the mechanism by which CHH mobilizes cyclic GMP metabolism in the neuromuscular target. The observations in Fig. 1 are consistent with the idea that in lobster muscle, as in other tissues, the level of cyclic GMP at any time reflects the relative rates of the cyclase and phosphodiesterase enzymes. The equilibrium level can be raised by inhibiting the phosphodiesterase with IBMX. It can also be raised by treating the tissue with CHH, either in the presence or absence of the phosphodiesterase inhibitor.
The enhancement of the peptide's actions by IBMX suggests that the peptide acts primarily by stimulating one or both of the cyclases rather than by inhibiting a phosphodiesterase enzyme. The data do not, however, rigorously exclude the phosphodiesterases from all aspects of the response. Parallel regulation of the phosphodiesterase could, for example, serve to amplify the cyclase response (or act in a negative feedback loop to limit it).
One of the primary contributions of this work is the demonstration that CHH is capable of stimulating guanylate cyclase activity in isolated membranes (Fig. 4). This provides a plausible mechanism by which the peptide might carry out its biochemical actions in intact cells. It also places the peptide in the company of a relatively small number of biological activators of the cyclic GMP cascade. To date, only the mutually homologous members of the ANP family Song et al., 1988), ACTH (Nambi and Sharma, 1981), and a pair of sea urchin peptides called speract and resact (Hansbrough and Garbers, 1981;Suzuki et al., 1984), have been shown to activate the membrane cyclase. A variety of cross-linking, purification, and cloning studies suggest that the receptors for these peptides in fact comprise an extracellular domain of the cyclase enzyme itself (reviewed in Schulz et al., 1989). This ligand binding domain appears to communicate with an intracellular catalytic domain by means of a single membrane-spanning helix. It will be important to determine whether the same is true of the CHH receptor in the lobster. Since there is no sequence homology between CHH and the peptides of the ANP, ACTH, or speract/resact families, it is apparent that the peptide recognition component associated with the lobster cyclase cannot be identical with those in vertebrates or sea urchins. However, the fact that the sequence of the catalytic component of the cyclase is maintained across all species tested suggests that the conserved regions of the molecule could be used to clone the lobster cyclase gene(s). This would establish a novel member of the guanylate cyclase family for comparative structure/ function analysis and would provide an additional vantage point for evaluating evolutionary constraints on the structure of the membrane cyclase.
We have tried to address the question of whether CHH might, in addition, make use of the cytoplasmic form of guanylate cyclase, which is reasonably abundant in lobster muscle. Although we cannot entirely exclude a contribution of the soluble cyclase to the receptor-mediated response, Fig.  5 strongly suggests that if coupling occurs between peptide receptors and the soluble cyclase then it does not occur by means of the EDRF pathway. In fact, the EDRF pathway appears not to function at all in lobster muscle, since a dose of nitroprusside that causes more than an 80-fold increase in the cyclic GMP level of cultured rat sensory neurons has no effect on isolated lobster muscle. This is somewhat unexpected, since the ability to respond to nitroprusside is one of the hallmarks of the vertebrate cytoplasmic cyclase. It remains to be seen whether the lobster-soluble enzyme displays other significant differences from the vertebrate enzyme.
If the EDRF pathway is not active in the lobster, then the existence of significant levels of cytoplasmic cyclase is somewhat puzzling. The cytoplasmic enzyme is clearly distinct from the membrane cyclase, since it does not sediment at 100,000 X g, is not coupled to CHH receptors, and does not show the time-dependent inactivation that is characteristic of the membrane form. In this regard, we do not know whether the lobster soluble cyclase is the product of a unique gene, or a degradation product resulting, for example, from proteolytic cleavage of the membrane cyclase. However, in either case, it seems likely that the soluble form is a natural component of intact tissue, rather than an artifact of homogenization, since the ratio of membrane to soluble cyclase remains constant even if we greatly prolong the interval between homogenization and centrifugation.' This raises the question of the functional significance of the soluble cyclase in intact tissue. Several possibilities should be considered. 1) Perhaps CHH (or some other lobster hormone) stimulates the soluble cyclase by means of a coupling factor distinct from EDRF. 2) Perhaps the soluble cyclase is tonically active in a resting cell, but is inhibited following exposure to agents that raise intracellular Ca*+ (as in Figs. 2c and 3), thus leading to a decrease in cyclic GMP levels. 3) Perhaps the cytoplasmic form is an intermediate in the degradation of the membrane cyclase and serves no other biological role.
Whatever explanation proves to be correct, it is clear that describing the properties of the cyclic GMP system is essential to our basic understanding of metabolic regulation in the lobster. It is also apparent that structural and functional comparison of the guanylate cyclases of the lobster (both soluble and membrane) to those of vertebrate animals remains a fruitful area for further enquiry.