The Catalytic Unit of Ram Sperm Adenylate Cyclase Can Be Activated through the Guanine Nucleotide Regulatory Component and Prostaglandin Receptors of Human Erythrocyte*

Ram sperm adenylate cyclase is insensitive to fluo- ride, guanine nucleotides, cholera toxin, and hormones, and appears devoid of the guanine nucleotide regulatory component. In this paper, we demonstrate that human erythrocyte membranes are capable of restoring guanine nucleotide regulation and fluoride sensitivity to ram sperm adenylate cyclase. The reconstitution process in the presence of guanyI-5'-yI imidodiphos- phate or NaF is time-dependent, directly proportional to the quantity of erythrocyte protein added to the reconstitution system, and is only observed when Mg- ATP is used as substrate. Furthermore, the guanyl-5'-yl imidodiphosphate-reconstituted activity can be ac- tivated by prostaglandin El or prostaglandin Ez through the binding sites normally present in human erythrocyte membranes. It therefore appears that re- constitution o€ the adenylate cyclase system can be readily performed in normal membranes which are deficient in regulatory component, and not only using defective cultured cell lines.

' The enzyme is soluble when prepared from testis, where it is localized mainly in spermatogonia at various stages of development, while it is entirely membrane-bound in mature spermatozoa (1-4).
The abbreviations used are: N or G/F component, the nucleotidethe catalytic site; Gpp(NH)p, guanyl-5"yl imidodiphosphate; PGE, binding regulatory protein involved in the coupling of receptors with prostaglandin E. type lymphoma cells (13-15), from liver membranes (16), or even more readily, from human erythrocytes (17-21). Indeed, human erythrocytes appear to be relatively enriched in N component (22), as compared to catalytic site and hormonal receptors.
In the present paper, we show that the ram sperm adenylate cyclase can acquire a normal sensitivity to fluoride, Gpp(NH)p, cholera toxin, and Mg2+ by complementation with the guanine nucleotide regulatory component of human erythrocyte. Furthermore, this system can be activated by prostaglandins, most probably by complementation with the prostaglandin receptors from erythrocytes.

EXPERIMENTAL PROCEDURES
Materials-ATP, GTP, creatine phosphate, creatine kinase, and cholera toxin were obtained from Sigma. Guanyl-5'yl imidodiphosphate and NAD were purchased from Boehringer Mannheim. All other chemicals were from E. Merck, Darmstadt.
Ram Sperm Preparation-Semen was diluted in 20 volumes of 150 mM NaCl and washed twice by centrifugation at room temperature at 2,000 rpm for 10 min. The pellets were homogenized in 1 mM NaHC03, centrifuged in a SS34 Sorval rotor for 10 min at 20,000 rpm at 4 "C, resuspended in 1 mM NaHC03, and finally stored in liquid nitrogen.
Human Erythrocyte Membrane Preparation-Human erythrocyte membranes were prepared by the procedure of Steck and Kant (23) as reported previously for the complementation assay (17-22). Adenylate cyclase activity in our preparation was comparable to the activity reported by Johnson et al. (17), Nielsen et al. (20), and Roden et al. (24). It was not detectable when tested in the absence of activator, and attained 0.3 to 0.5 pmol of cyclic AMP formed/mg of protein/min at 30 "C in the presence of 10 FM Gpp(NH)p or 10 mM NaF. In all the complementation experiments described here, the final activities were corrected for the contribution of cyclase activity originating from the erythrocyte membranes.
Adenylate Cyclase Assay-Incubations were performed at 30 "C for the indicated time periods in a final volume of 60 pl, containing 0.5 mM [w3'P]ATP (2.106 cpm), 3 mM MgCh (except when otherwise indicated), 1 mM EDTA, 1 mM cyclic AMP, 50 mM Tris-HC1, pH 7.6, and an ATP-regenerating system consisting of 25 mM phosphocreatine and 1 mg/ml of creating phosphokinase, and the indicated amount of proteins. The reaction was terminated by a modification (25) of the procedure of White (26). Protein was estimated by the procedure of Lowry et al. (27) using bovine serum albumin as a standard.
Treatment with Cholera Toxin-Sperm membrane (1 mg/ml) or human erythrocyte membranes ( 3 mg/ml) were incubated with 50 mM Tris-HC1, pH 7.6, containing 1 mM dithiothreitol, 1 mM NAD, 1 mM ATP, 10 p~ GTP, and 10 pg/ml of cholera toxin previously incubated in the presence of 20 mM dithiothreitol for 30 min at 37 " c (28). After a 20-min incubation at 30 "C, the samples were diluted 25fold in cold 50 mM Tris-HC1, pH 7.6, and centrifuged at 20,000 rpm for 15 min at 4 "C. The pellets were resuspended in 50 mM Tris-HC1, pH 7.6. Controls were incubated in the same conditions but in the absence of cholera toxin.

Reconstitution of Adenylate
Cyclase System 5395

RESULTS
Complementation of Ram Sperm-A titration in which increasing amounts of human erythrocyte membranes were combined with a fixed amount of sperm membranes is shown in Fig. 1. In the absence of erythrocyte membranes, the ram sperm adenylate cyclase activity was 5 pmol of cyclic AMP formed/mg of protein/min and was not increased by addition of either Gpp(NH)p, hormones, or NaF. When erythrocyte membranes were added to the assay mixture, basal activity was not modified; however, stimulation of adenylate cyclase activity by NaF and Gpp(NH)p was clearly restored. It reached a maximum, 55and 24-fold stimulation for NaF and Gpp(NH)p activation, respectively, when 160 pg of human erythrocyte membranes were added to 3 p g of sperm membrane protein.
When the amount of human erythrocyte membranes was kept constant while the amount of ram sperm protein was varied from 0.1 to 10 pg/assay, the reconstituted activity assayed in the presence of 10 p~ Gpp(NH)p was proportional to the amount of ram sperm protein. This activity reached a plateau corresponding to a 23-fold enhancement of cyclase activity (Fig. 2).
Time Human erythrocyte membranes were diluted in 5 mM Na phosphate buffer, pH 8.0, and then mixed with an equal volume of sperm membranes ( 3 pg of protein/assay); 40-pl aliquots were withdrawn in triplicate for determination of reconstituted adenylate cyclase activities with no addition (0) or in the presence of 10 p~ Gpp(NH)p (A) or 10 mM NaF (X) for 40 min at 30 "C. Protein concentrations on the abscissa refer to the concentrations of human erythrocyte membranes in the assay medium. The activity is expressed in picomoles of cyclic AMP formed/mg of sperm protein/min at 30 "C.
were incubated for various periods of time at 30 "C in the presence of 10 PM Gpp(NH)p and all the other assay reagents and in the absence or presence of human erythrocyte membranes (107 pg/asay).   with cholera toxin prior to the reconstitution Sperm membranes and human erythrocyte membranes were treated with cholera toxin as described under "Experimental Procedures.'' Sperm membranes (4 p g of protein/assay) were then assayed for reconstitution with untreated human erythrocytes (144 pg of protein/assay) or with cholera toxin-treated human erythrocytes (130 pg of protein/assay), in the absence or presence of 10 p~ Gpp(NH)p for 35 min at 30 "C. Activities of cholera toxin-treated human erythrocytes were 20 and 50 pmol of cAMP formed/mg of protein in 35 min at 30 "C for basal and Gpp(NH)p activities, respectively. Activities are expressed in picomoles of cAMP formed/mg of sperm protein in 35 min at 30 "C. The fold activation is indicated in parentheses, as compared to the control sperm cyclase with no addition. Cholera toxin-treated sperm lag occurred before Gpp(NH)p started to stimulate cyclase activity; activation was completed in 20 min, and cyclase activity was constant thereafter for at least 40 min. Control sperm adenylate cyclase activity was also linear with time during the same period. The same lag in activation was observed in the presence of Na fluoride. Effect of Cations-It is known that sperm adenylate cyclase can be activated by either Mg2+ or Mn2+ (6,8,9, ll), the ram sperm enzyme being 15-fold more active in the presence of Mn-ATP than in the presence of Mg-ATP. As shown in Fig.  4 A , the enzyme activity assayed in the presence of 0.5 mM ATP and 1 mM EDTA increased when Mg concentration varied from 0 to 20 mM, reaching a plateau only at high Mg concentration. This dose-response curve was the same whether 10 ~L M Gpp(NH)p was added to the assay medium or not. When 80 pg of human erythrocyte membrane protein were added to the assay medium, the basal activity was not changed, while that assayed in the presence of Gpp(NH)p was The adenylate cyclase activity was measured in the presence of 4 pg/assay of sperm membranes, and in the absence . , basal activities) or in the presence of 10 CM Gpp(NH)p (A, *) for 20 min at 30 "C. Activity was measured in the absence (-) or in the presence of 80 pg of protein/assay of human erythrocyte membranes (---). Activities are expressed in picomoles of cyclic AMP formed/mg of sperm protein/min at 30 "C. dramatically increased. Furthermore, the slope of the doseresponse curve for Mg was modified, the maximal activity being attained at a much lower total Mg concentration (3 mM instead of 20 mM). The dose-response curve for Mn (Fig. 4B) was similar to that of Mg for the basal sperm activity, maximal activity being attained at a concentration of 25 mM total Mn. However, at all Mn concentrations, the activity was not modified when either Gpp(NH)p or sperm membrane or both were added. It therefore appears that reconstitution of a Gpp(NH)p-sensitive adenylate cyclase was only observed when Mg was the co-substrate of the reaction; it is not possible however to decide whether Mg was necessary only for the expression of the reconstituted system or for the reconstitution process per se.
Effect of Cholera Toxin on Reconstitution-Like the AClymphoma cell line system, sperm adenylate cyclase is unresponsive to cholera toxin, whatever the preincubation medium or cholera toxin concentration used (Table I). It was therefore interesting to test whether preactivation of the coupling factor from human erythrocyte membranes with cholera toxin could lead to a permanently activated state of N that might be detectable after the reconstitution process. When the human erythrocyte membranes were preactivated by cholera toxin as described under "Experimental Procedures," the reconstitution process could be observed either for the basal activity or in the presence of Gpp(NH)p, leading to an increase in adenylate cyclase activity of 20-fold and 60-fold, respectively (Table I). The data also show that even after prior treatment with cholera toxin, addition of Gpp(NH)p (10 p~) during the cyclase assay caused a further enhancement (3-4-fold) of the reconstituted enzyme activity. Conversely, preincubation of sperm membrane with cholera toxin did not activate the cyclase system, whether or not the erythrocyte membranes were added afterwards.
Effect of Prostaglandins on Adenylate Cyclase- Fig. 5 shows the effect of PGEl and PGEz on ram sperm adenylate cyclase before and after reconstitution with human erythrocyte membranes. Ram sperm adenylate cyclase activity was not enhanced by either prostaglandin; rather, it was decreased (40% inhibition) at high concentration (0.1 mM) of prostaglandins. This finding is in agreementn with results observed with human, bull, and monkey spermatozoa (6, 8, 9). In contrast,  the reconstituted, Gpp(NH)p-stimulated enzyme was enhanced in a dose-dependent manner by PGEl and PGEz; halfmaximal activation was attained at 0.1 ~L M for PGE, and 0.8 PM for PGE2. Fig. 6 demonstrates that the PGEl response and the Gpp(NH)p "reconstituted" activity were parallel as a function of the concentration of sperm protein. This indicates that prostaglandin activates sperm adenylate cyclase throughout binding sites present on the erythrocyte membrane.

DISCUSSION
Ram sperm adenylate cyclase is similar to that of the AClymphoma cell line mutant described by ROSS et al. (14). Both systems are insensitive to NaF, Gpp(NH)p, and cholera toxin, indicating that they are devoid of a functional guanine nucleotide regulatory component.
The present report demonstrates that human erythrocyte membranes are capable of restoring guanine nucleotide regulation and fluoride sensitivity to the ram sperm membrane adenylate cyclase system. The reconstituted system presents the following characteristics. (i) The reconstitution is timedependent, a lag of 5 to 6 min being observed in the presence of Na fluoride and Gpp(NH)p at 30 "C. (ii) The reconstitution process is proportional to erythrocyte membrane protein added in the presence of Gpp(NH)p or fluoride. The sperm membrane system appears to possess a great excess of catalytic sites/mg of protein as compared to the AC-cell line system, since the reconstitution experiment performed with the same human erythrocyte membrane preparation gave different optimal protein ratios for the different systems: 1 to 100 for sperm membrane to human erythrocyte membranes uersus 1 to 5 for AC-lymphoma cell line (18). This is in agreement with the fact that the specific activity of ACcyclase, assayed in the presence of MgATP, is much lower than that of ram sperm (14). (iii) The reconstitution was only observed when MgATP was used as substrate response of adenylate cyclase activity as a function of MgClz concentration was shifted toward lower concentrations. No reconstitution could be observed in the presence of Mn-ATP. One might hypothesize that the metal binding site is present normally in the ram sperm and that its affinity for Mg2+ is enhanced after the addition of the coupling factor($. Alternatively, Mn2+ ions might somehow uncouple the catalytic site from regulatory components as reported in other systems (29, 30); or else the high affinity for Mg2+ in the reconstituted system might directly reflect the insertion of a saturating amount of functional N from human erythrocyte membranes, while the low affinity apparent before reconstitution would reflect an abnormal, defective, or poorly functional N. (iu) The reconstituted guanine nucleotide-sensitive, ram sperm adenylate cyclase is activated by prostaglandins, probably through the binding sites normally present in the human erythrocyte membranes (22). In contrast, when epinephrine, human chorionic gonadotropin, or follicle-stimulating hormone were added in the final assay system, they failed to further stimulate the activity in the presence of Gpp(NH)p (data not shown). ( u ) Finally, pretreatment of the N component of human erythrocyte membranes by cholera toxin greatly enhanced the ability of the factor to activate ram sperm adenylate cyclase, whether Gpp(NH)p was present or not, in the assay. In the present system, as well as in that described by Nielsen et al. (20) and Lad et al. (21), the mechanism of the reconstitution process was not directly probed. In particular, it is not possible at the present time to decide whether a true membrane fusion occurred or whether components from one membrane migrated to the other. The reconstituted system exhibited a prostaglandin response; this might indicate, but does not prove, that the sperm cyclase itself was transfered to the erythrocyte membrane. The fact that cholera toxin was active only when preincubated with erythrocyte membranes, and not with sperm membranes, is direct evidence that the N component does originate from the erythrocytes.
Inasmuch as the adenylate cyclase activity of human erythrocyte membranes can be considered as negligible (15)(16)(17)(18)(19)(20)(21)(22), the present findings constitute a functional demonstration that N can be activated by cholera toxin in the absence of a fully functional cyclase and that the "activated" N can be transferred to an acceptor system. Noteworthy is the fact that the cyclase reconstituted with cholera toxin-treated N was still sensitive to the further addition of Gpp(NH)p to the assay with membranes from turkey erythrocytes, pigeon erythrocytes, and rat liver. They are compatible with the hypothesis that cholera toxin can exert a stimulatory effect upon adenylate cyclase by a mechanism other than, or in addition to, the classical inhibition of GTPase activity (31).
In this paper we demonstrated that ram sperm membrane adenylate cyclase activity is deficient in guanine nucleotide regulatory protein and can be complemented with an N donor in the absence of detergent. After reconstitution with cholera toxin-treated human erythrocyte, the sperm adenylate cyclase activity was activated 60-fold with respect to basal activity. Ram sperm adenylate cyclase can be used as an acceptor adenylate cyclase deficient in N, and is analogous to the AClymphoma cell line. That it appears to exist as a pure catalytic component embedded in a normal membrane makes it a potentially very useful tool along with other, more sophisticated systems (14,15,35,36) for elucidating many of the open questions concerning the role of the various subunits of the cyclase system.