Reconstitution of Catecholamine-sensitive Adenylate Cyclase RECONSTITUTION OF THE UNCOUPLED VARIANT OF THE S49 LYMPHOMA CELL*

The uncoupled (UNC) variant of the 549 lymphoma possesses receptors and other known regulatory com- ponents necessary for hormone-stimulated adenylate cyclase activity but fails to respond to /I-adrenergic agonists or to prostaglandin E1. A procedure is de- scribed for the reconstitution of responses to hormones in purified plasma membranes of these cells. This tech-nique utilizes cholate extracts prepared from mem- branes of wild type S49 cells or from membranes of other cell lines that lack either /3-adrenergic receptors or the catalytic subunit of adenylate cyclase. The pro- cedure also restores adenylate cyclase activity to membranes of the adenylate cyclase-deficient (cyc-) 549 cell variant. Hormone responses of reconstituted UNC and cyc-membranes resemble those of wild type S49 cell mem- branes in their dependence on agonist concentration for both enzyme activation and binding to P-adrenergic receptors. Effects nucleotides on the binding of agonists, which are lost in UNC and cyc- membranes, are also restored in the reconstituted preparations.

Analysis of the components that comprise a hormone-sensitive adenylate cyclase is at present limited to observation of the various activities that are expressed by the enzyme system. Resolution and reconstitution of some of these activities have led to the identification of at least three distinct components of adenylate cyclase systems that are stimulated by ,&adrenergic agonists. A catalytic entity has been clearly separated from the ligand-binding activity of the P-adrenergic receptor (1,2). More recently, the expression of catalytic activity was shown to depend on the presence of at least two components, the catalyst itself (C)' and a regulatory protein (3,4 Hepes, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic latter appears to be a site of action of guanine nucleotides and of fluoride (4,5). (This regulatory protein has thus been denoted G/F.) The resolution of the components of adenylate cyclase has proceeded by both affinity (6) and conventional (1,2) chromatographic techniques and by the isolation and characterization of genetic variants that are deficient in their expression of certain components of the hormone-sensitive enzyme system (2-5, 7, 8). In particular, a variant of the S49 lymphoma that is essentially devoid of adenylate cyclase activity (when the physiological substrate of the enzyme, MgATP, is present) has been shown to lack detectable activity of the regulatory protein G/F (3,4); a complementary hepatoma clone, HC-1, which is also devoid of adenylate cyclase activity, retains the regulatory protein but is deficient in the activity of the catalyst itself (4).
The existence of the uncoupled variant of the S49 lymphoma suggests a fourth possible candidate for constituency in hormone-sensitive adenylate cyclase (8). The UNC variant possesses an active adenylate cyclase that is responsive to stimulation by both guanine nucleotides and fluoride; in addition, these cells retain an apparently normal complement of /3-adrenergic receptors. By current definition, therefore, these cells possess the identified protein components of the enzyme system. However, neither P-adrenergic agonists nor prostaglandins are capable of stimulating adenylate cyclase activity in either intact UNC cells or in plasma membranes derived therefrom.
To facilitate identification of the defect in the UNC cell membrane, we have developed a method to reconstitute its hormone-sensitive adenylate cyclase activity; this procedure utilizes detergent extracts prepared from membranes of both complementary and fully competent cells. This report provides a description of this method and of some of the characteristics of the factor(s) required for this reconstitution.

EXPERIMENTAL PROCEDURES
The cell lines used in this study have been described previously, and their phenotypic properties are summarized in Table I  basal and fluoride-stimulated activities, but this does not result in the restoration of a hormone response. To eliminate this effect of detergent, a second titration was performed in which the cholate concentration was kept constant. This was accomplished by supplementation of increasing amounts of the reconstituting extract with decreasing amounts of an extract that had been heated at 60°C for 15 min to destroy all reconstituting activity (Fig. 1B). Such titration results in a linear increase in hormone activation and a maintenance of a constant level of activity in the presence of fluoride. Activation of the enzyme by isoproterenol increased to a value that was 70% of that achieved with the halide (comparable to that with wild type membranes), and the effect of isoproterenol was completely blocked by the P-adrenergic antagonist proprano-101. Saturation of the reconstitution of UNC membranes is emphasized in the inset to Fig. 1B. In this depiction fluoridestimulated enzymatic activity is assumed to be a measure of the total adenylate cyclase activity of the UNC membrane, and relative recoupling is estimated as the ratio of stimulation by isoproterenol to that by fluoride. Percentage of stimulation by isoproterenol (compared to the basal rates) shows the same dependence on the addition of extract. A similar plot of the ratio of isoproterenol-to fluoride-stimulated activity from Fig.  1A (not shown) is sigmoidal with increasing amounts of extract, suggesting that an optimal amount of detergent stimulates reconstitution as well as adenylate cyclase activity. Cholate extracts of wild type S49 cell membranes also restore PGE, stimulation of adenylate cyclase activity to UNC membranes (Table II), and reconstitution of activation by PGE, occurs coincidentally with that by isoproterenol. Stimulation by PGEl and isoproterenol was also restored in UNC membranes with extracts from membranes of HC-1, a cell line deficient in the catalytic subunit (C) of adenylate cyclase, and from B82, a cell line lacking /I-adrenergic receptors. However, these responses were not restored when an extract of cycmembranes was used in the reconstitution; this confirms the lack of complementation between the UNC and cyc-variants (11,17).
LubrollZA9 extracts of various membranes were previously shown to contain the regulatory component of adenylate cyclase, G/F, and these extracts were capable of restoring adenylate cyclase activity to membranes of the cyc~ variant (4,11); this activity was stimulated by /3-adrenergic agonists. Cholate extracts also have this capability (Table II;  Reconstitution of basal and fluoride-stimulated activities with extracts of membranes from UNC, B82, HC-1, and wild type cells confirms the presence of G/F in these membranes. Recoupling of responses to both P-adrenergic agonists and PGEl was also observed with all but the UNC extract. While G/F and a coupling activity can be reconstituted using appropriate cholate extracts, neither the catalytic subunit, C, nor the P-adrenergic receptor can be incorporated in an active state into deficient acceptor membranes with the procedures utilized to date. This is seen in the failure of a cycextract (which had not been incubated at 23°C) to reconstitute the C-deficient HC-1 membrane or of a wild type extract to reconstitute a response to isoproterenol in membranes from B82 (Table II).

Properties of the Recoupled Hormone
Receptor-Activation of adenylate cyclase in reconstituted UNC membranes shows essentially the same dependence on the concentration of isoproterenol as does that in wild type membranes (Fig. 2). Half-maximal stimulation occurred at 17 and 27 nM isoproterenol in wild type and reconstituted UNC membranes, respectively; reconstitution of cyc-membranes yielded a Knpl for isoproterenol of 19 nM (Table III). When stimulation by PGEl was examined, half-maximal activities were observed at 125 and 190 nM with wild type and reconstituted cyc~ membranes. Slightly greater concentrations of the prostaglandin (K,,, = 350 nM) were required with reconstituted UNC membranes. When membranes from wild type S49 cells are exposed to guanine nucleotides, there is a decrease in the affinity of the  /3-adrenergic receptor for agonists, but not for antagonists (9).
Membranes from the two variants of S49, cyc-and UNC, have lost this capability (8,9). They display only a single affinity that corresponds to the lower affinity of wild type membranes in the presence of guanine nucleotide. Fig. 3A shows the restoration of this nucleotide-mediated change in affinity in reconstituted UNC membranes. No change in agonist competition for ligand-binding sites was seen when UNC membranes were treated with cholate extract that was devoid of reconstituting activity (heated at 60°C for 20 min) (Fig.  3B). Restoration of the nucleotide effect on agonist affinity was also observed when cyc-membranes were reconstituted. The apparent dissociation constants that were observed for isoproterenol with the various membranes are compared in Table III. It is of interest that restoration of the effect of guanine nucleotide on binding was not observed in all reconstituted cyc-and UNC membrane preparations, even when stimulation by hormone was reconstituted and was largely dependent on the addition of GTP. This observation will be discussed below.
Requirements for Reconstitution-Initial attempts to reconstitute responses to isoproterenol in UNC membranes (or adenylate cyclase activity in cyc-membranes) by simple mixture of cholate extracts with suspensions of membranes in Solution A led to preparations that displayed activation during the adenylate cyclase assay and were only partially stable to centrifugation.
This activation was especially prominent when UNC membranes were studied, and more than 10 min was required to achieve maximal stimulation.
The protocol described under "Experimental Procedures" was developed to optimize reconstituted activity in a stable, sedimentable form. Membranes that are reconstituted by this procedure show an enhanced rate of adenylate cyclase activity that is maximal within 2 min of addition of hormone or fluoride and that is linear for at least 30 min. at 3O"C, following the addition of Solution B, results in the best activation by isoproterenol, although reconstitution of fluoride-and Gpp(NH)p-stimulated activity in cyc-membranes shows little effect. A major requirement for Solution B at this stage appears to involve the stabilization of adenylate cyclase activity at 3O"C, since its substitution with Solution A results in large losses of activity. If the initial incubation at 10°C is omitted prior to dilution and incubation in Solution B at 3O"C, recoupling of UNC and reconstitution of cyc-are decreased. Apparently, some time for equilibration at higher protein and detergent concentrations (and lower temperature) aids in reconstitution. Subsequent experiments have shown that a ZO-min incubation on ice is sufficient. The effects of the components of Solution B were also examined ( Table V). ATP plays a major role in stabilization of adenylate cyclase activity, including that already present in the UNC acceptor membranes. The inclusion of GTP appears to cause an enhancement of all reconstituted enzymatic activities. Thus, when UNC membranes are reconstituted in the presence of GTP, hormone-stimulated activity is typically increased to a value that is 60 to 70% of that observed with fluoride. The relative importance of Mg'+ is less clear, since both the membrane suspension and extract contained the cation. Therefore, its removal from Solution B left a concentration of 0.5 to 1.0 rnM Mg2+ in the reconstituting system. However, it should be noted that the concentration of ATP is in excess of that of residual Mg"+; this condition is strongly inhibitory to adenylate cyclase catalytic activity. In addition to the chemical constituents of Solution B, the dilution provided by its addition is important.
If the ingredients in Solution B were added to the same final concentrations but in a smaller volume (such that there was little dilution, 15% U~FSUS lOO%), membrane-bound enzymatic activity suf-   (Table V). Dilution greater than the 2-fold utilized in the described procedure had no further beneficial effect.

Reconstitution of UNC Membranes
Requires a Protein Component-The factor(s) responsible for the reconstitution of UNC membranes exhibits at least three properties that implicate or are consistent with the involvement of a protein in the process. First, the reconstituting activity of the cholate extract is very labile; a 30-min incubation at 37°C is sufficient to inactivate 75% of its recoupling activity in UNC (Table VI). Similar lability is seen for the reconstitution of isoproterenoland Gpp(NH)p-stimulated activity in cyc-membranes, although the response to fluoride appears to be slightly more stable. (A similar or greater difference in the stability of G uersus F activity has been seen previously with Lubrol extracts of G/F (3,4).) More convincing evidence comes from the inactivation of reconstituting activities with the sulfhydryl reagent, N-ethylmaleimide.
Both the hormone-stimulated activity (in UNC and cyc-) and fluoride activation (in cyc-) are inactivated by similar concentrations of the reagent (Table VI). The sensitivity of the extract to digestion by protease is shown in Fig. 4. Reconstitution of stimulation by isoproterenol was about 10 times more sensitive to trypsin digestion than was the restoration of a fluoride response in cyc-. Receptor recoupling in both UNC and cyc-showed identical sensitivity with trypsin. The intermediate sensitivity of stimulation of reconstituted cyc-membranes by Gpp(NH)p is questionable. Reconstitution of responsiveness to hormone in UNC has an apparent stimulatory effect on activation of the enzyme by Gpp(NH)p (Fig. 4A); this effect is trypsin-sensitive, and part of the loss observed in Fig. 4B could be due to this phenomenon, rather than to actual damage to the G/F component. (These two possibilities could, however, be synonymous; see "Discussion.") Effect of Cholera Toxin-treated G/F on the Reconstitution when UNC membranes, which already contain G/F, are reconstituted. This problem was approached by reconstitution of UNC membranes with a cholate extract of membranes obtained from cholera toxin-treated HC-1 cells. Such treatment has been shown to alter the properties of G/F (5,18), and a characteristic reflection of this alteration is marked activation of adenylate cyclase activity by GTP (9,19). Table VII shows that  reconstitution of cyc-membranes with the extract from chol-era toxin-treated HC-1 cells yields preparations that are so activated by GTP; a reduction in the level of response to fluoride is also characteristic. When the extract from toxintreated cells was used to reconstitute UNC membranes, enzymatic activity that was markedly stimulated by GTP was also observed, although fluoride-stimulated activity was not reduced. This probably indicates that a mixture of both toxinmodified G/F and the original UNC G/F are interacting with the endogenous C of the LJNC membrane. DISCUSSION We have established a new method for the reconstitution of membranes of S49 cell variants that are defective in hormonesensitive adenylate cyclase activity; resultant membranes have properties that are essentially indistinguishable from those of wild type cells. UNC and cyc-membranes that have been reconstituted with cholate extracts from various membranes exhibit adenylate cyclase activity that can be stimulated by P-adrenergic agonists, PGE1, guanine nucleotide analogs, and fluoride. All of these activities are recoverable after collection of the reconstituted membranes by centrifugation. The resemblance between reconstituted and wild type membranes extends to their specific activities, to their dependence on the concentration of PGEl or isoproterenol for enzyme activation, and to the KD value for binding of isoproterenol to /?-adrenergic receptors. Thus, the ratio of KD to K,,, for the /3-adrenergic agonist is very high with the reconstituted membranes (>lO ; Table III); such values, which imply the ability to activate adenylate cyclase maximally despite minimal receptor occupancy, have previously been interpreted to be indicative of an efficiently coupled system (20,21).
The alteration in the affinity of binding of /3-adrenergic agonists that is characteristically observed in wild type membranes in the presence of guanine nucleotide is also observable in reconstituted UNC and cyc-membranes. However, the reconstitution of this phenomenon is not completely reproducible, despite the fact that reconstitution of responses to hormones is always observed. The reason for this variability is not known. Endogenous GTP (from membranes or extract) does not appear to be the problem, since reconstituted responses to hormones always show a similar dependence on added nucleotide (-50%). This is lower than the 90 to 95% dependency of wild type membranes but is similar to the value of 70% seen in the same membranes after they are subjected to treatment with cholate as in a reconstitution protocol. Furthermore, this treatment with cholate does not interfere with the observation of nucleotide-induced changes in affinity of wild type membranes for /3-adrenergic agonists. Inconsistency in the ability to observe guanine nucleotidemediated changes in affinity could be indicative of relatively inefficient recoupling of receptors to a regulatory component of adenylate cyclase, despite a high value of KU/K,,, and the possibility of efficient coupling of adenylate cyclase to receptors. One could envision a stimulation where all adenylate cyclase is subject to regulation by receptor, but where all receptor is not recoupled to the crucial component of the enzyme.
The method described above contrasts sharply with the previously reported reconstitution of cyc-membranes with extracts prepared with Lubrol 12A9 (4,5,11). In this case stable reconstitution was achieved only in the presence of certain activators of the enzyme system (i.e. Gpp(NH)p, fluoride, or GTP when extracts of cholera toxin-treated cells were used) (5). The essential irreversibility of these activating ligands prevented the subsequent examination of the effects of other ligands after stable reconstitution had been achieved. There was little stable reconstitution of responsiveness to /?adrenergic agonists, although isoproterenol was effective in mixtures of membranes and extract. The effects of prostaglandins were minimal or nonexistent in membranes or in mixtures of membranes and extract. The latter result was probably due to interference from Lubrol that remained in the suspension of reconstituted membranes. The reasons for the markedly different results obtained with the two different detergents are not clear. Lubrol may have a greater affinity for factors that are necessary for reconstitution, such that a more specific interaction with acceptor membranes (presumably promoted by the activating ligands) and higher temperature are required to facilitate the movement of G/F into the membrane.
Also poorly understood are the improvements in the fidelity of reconstitution that are imparted by the specific procedures and conditions of the protocol (Tables IV and V). It is readily admitted that these evolved largely empirically; however, their existence provokes speculation.
The protocol may be divided, simplistically, into three phases: (a) incubation at low temperature and high concentrations of membranes and extract-containing detergent; (b) dilution; and (c) incubation at higher temperature.
We suggest that the first incubation at high concentrations of soluble factors and membranes facilitates their initial interaction.
An optimal concentration of cholate at this stage might provide beneficial perturbation of the acceptor bilayer. The data of Fig. 1  while solubilization is achieved at concentrations of cholate that are above the critical micellar concentration, optimal reconstitution is obtained when the detergent concentration in Stage 1 is near to or slightly below this value (22,23).) Reconstitution at this stage is stimulated greatly by the use of Solution B. Presumably the addition of GTP and ATP promotes the interaction of factors with acceptor membranes. Elevation of temperature, the third stage, appears to facilitate the stable incorporation of factors into the membrane, perhaps in part because of necessary alteration of the lipid structure. Inclusion of ATP at this stage stabilizes adenylate cyclase against detergent-facilitated denaturation. Further speculation is limited by our current necessary reliance on determinations of enzymatic activity to assess the status of individual components of the system during their reconstitution.
A major question involves the nature of the lesion in the UNC variant. Previous analysis has demonstrated that the gross lipid composition of the UNC membrane is identical with that of wild type and that there is an identical electrophoretic pattern of the major membrane proteins (24). The reconstitution procedure described herein was thus designed to help answer this question. To date, we can say that the lesion is correctable by the appropriate addition of one or more factors and that a crucial factor is temperature-sensitive and susceptible to inactivation by A'-ethylmaleimide and by trypsin. Furthermore, initial data (not shown), obtained by ultracentrifugation of cholate extracts through sucrose gradients, are consistent with the residence of UNC-reconstituting activity in a fraction with a molecular weight of approximately lo5 (similar to that of G/F). These data seem to indicate the involvement of at least one protein in the reconstitution of UNC membranes. This protein might be either a new component of the adenylate cyclase system or one of the three components already identified, receptor, G/F, or C. Considerable evidence indicates that the lesion does not involve the receptor. The defect in UNC membranes results in the loss of responses to both PGE, and ,&adrenergic agonists, and, therefore, it presumably involves a common link through which both types of receptors interact with adenylate cyclase. Also consistent with this reasoning is the fact that reconstitution of UNC membranes with cholate extract restores both responses in parallel. Furthermore, the procedures utilized to date have failed to reconstitute responses to padrenergic agonists in receptor-deficient membranes (Table  II). Finally, examination of responses to isoproterenol in heterokaryons formed by the fusion of UNC or cyc-S49 cells with B82 (in the presence of cycloheximide) indicates that both S49 cell variants contain functional P-adrenergic receptors (17). The remaining receptor-related hypothesis is that the receptor for PGE, is reconstituted and that it serves as an essential coupling factor for the P-adrenergic receptor. There is no reason to consider this as a likely possibility.
The catalytic subunit of adenylate cyclase, C, is also not a good candidate for the site of the UNC defect. In addition to the fact that UNC has normal basal and adenylate cyclase activity, cholate extracts containing C do not restore enzymatic activity to C-deficient membranes (Table II). Furthermore, the reconstitution of UNC acceptor membranes is routinely accomplished with extracts in which C activity has been destroyed by warming or with extracts from C-deficient cells (Table II), and, if UNC acceptor membranes are incubated at 37°C prior to reconstitution, the restoration of response to hormone is reduced in parallel with the loss of UNC catalytic activity (not shown).
A defect in the regulatory component of adenylate cyclase, G/F, remains as an excellent possibility to explain the deficiency of UNC, despite the existence of Gpp(NH)p-and fluoride-stimulated cyclase activity in these membranes. The following observations are consistent with this hypothesis: (a) recoupling activity is as stable as or is more labile than G/F activity when both are inactivated by temperature, N-ethylmaleimide, or proteases; (b) initial experiments indicate cofractionation of UNC and cyc~ reconstituting activities by gel filtration or sucrose density gradient centrifugation; (c) UNC and cyc-(G/F-deficient) variants are not complementary in vitro (Table II; Refs. 11 and 17); and (d) G/F activity is probably incorporated into UNC membranes during the procedure utilized to reconstitute them, and this G/F may function in tandem with endogenous C (Table VII). These observations are consistent with the notion that there exists a domain of G/F that is crucial for responsiveness to hormones. It is reasonable to speculate that this is a region of the protein that is subject to post-translational modification and that the system necessary for such modification is altered in the UNC variant.
(The high frequency of the UNC lesion and the constancy of the phenotype (8) perhaps argue against a specific alteration in primary amino acid sequence of G/F.) However, the observations just stated are also completely consistent with the deficiency in UNC of an unidentified central component of the adenylate cyclase system. While lack of complementation between UNC and cyc-and cofractionation and co-inactivation of recoupling and G/F activities tend to suggest linkage of function, this linkage could be noncovalent.
The existence of a coupling factor that is normally tightly bound to G/F would explain the coincidence of behavior and the lack of complementation with cyc~ if G/F is necessary, for example, for the insertion of the factor into the membrane.
We cannot yet distinguish between these two possibilities or the somewhat less likely possibility of a com-