Isolation and Characterization of a Novel Cardiac Adenylylcyclase cDNA

7 downloads 0 Views 2MB Size Report
Oct 4, 1991 - A novel adenylylcyclase cDNA (type V) was isolated from a canine heart cDNA library. Northern blotting indicates that the expression of this ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY (0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 267, No. 19, Issue of July 5, pp. 13553-13557,1992 Printed in Ll S.A.

Isolation and Characterizationof a Novel Cardiac Adenylylcyclase cDNA* (Received for publication, October 4, 1991)

Yoshihiro IshikawaSQ, Shuichi KatsushikaQll,Liang Chen(1,Nancy J. Halnon, Jun-ichi KawabeQ,and Charles J. Homcy 11 ** From the Departments of §Pharmacology and **Medicine, College of Physicians and Surgeons of Columbia University, New York. New York 10032 and the IICardiouascularlCNS Research Department, Medical Research Division, American CyanamidLederle Laboratories, Pearl River, New York 10965

A novel adenylylcyclase cDNA (type V) was isolated from a canine heart cDNA library. Northern blotting indicates that the expression of this message is most abundant in heart with a lesser amount in brain but is absent in a variety of other tissues including lung, kidney, skeletal muscle, lymphocyte, and testis. The putative protein product predicted from the cDNA sequence has the motif of tandem six-transmembrane spans separated by a large hydrophilic cytoplasmic loop as seen in other members of the adenylylcyclase family. When this protein is expressed using a CMT cell transient expression system, the adenylylcyclase activity wasstimulated by NaF, GTPrS, and forskolin, but not by calmodulin. The activity was inhibited in a concentration-dependentmanner with either P-site active agents such as adenosine or in the presence of calcium. These data indicate that the protein encoded by this cDNA is adenylylcyclase with the biochemical features characteristic of the cardiac isoform.

several interesting features; the motif of two domains, each containing six predicted transmembrane-spanning regions, separated by a large hydrophilic intracellular loop and terminating in a similarly large intracellular tail. Thetwo intracellular domains are homologous to each other and different adenylylcyclase isoforms share sequence similarity in these regions as well. We first attempted to identify adenylylcyclase mRNA in heart using probes obtained from the conserved intracellular domains of other mammalian adenylylcyclases (types I and 11); however, no mRNA was detected. Consequently,usingthis common domain and a low stringency hybridization protocol we screened a cDNA library prepared from canine heart mRNA. We identified a cDNA clone that is homologous to other adenylylcyclase types, but also possesses certain distinctive features. In this study, we describe thetissuedistribution of this isoform andcertain of its biochemical features. MATERIALSANDMETHODS

Canine Heart cDNA Library; Construction and Screening-Canine left ventriclewas used as a source of mRNA. The library was prepared Activation of the sympathetic nervous system represents a in X g t l O phage with oligo-dT as the primer. Approximately 2 X lo6 major mechanismfor enhancing cardiac contractility in statesplaques were initially screened from the library. Prehybridization was of both volume and pressure overload (1). The release of carried out at least for 2 h in a solution containing 30% formamide, norepinephrine at the synaptic terminaltriggers an increase 5 X SSC, 5 X Denhardt's, 25 mM NaPO, (pH 6.5), 0.25 mg/ml calf in myocardial intracellular cyclic AMP via stimulation of p- thymus DNA, and 0.1% SDS' at 42 "C. Hybridization was then performed in the samesolution at 42 "C. A 970-bp AatI-HincII adrenergic receptors. The agonist-occupied receptor stimufragment from brain (type I) adenylylcyclase cDNA was used as a lates the exchange of G T P for GDP at the a subunit of the probe. The probe was radiolabeled with ["PIdCTP by the multiprimer heterotrimeric G protein, G.. Association of the GTP-liganded random labeling method. After hybridization for 18 h, the blot was a subunit with membrane-bound adenylylcyclasecatalyzes washed in increasingly stringent conditions and then subjected to the conversion of ATP to cyclic AMP. The increased cyclic autoradiography. Sequencing-All positive clones from the canine heart library were AMP levels promote, via activation of protein kinase A, the into pUC18. After restriction mapping, the fragments were phosphorylation of several intracellular proteins responsible subcloned subcloned and sequenced with universal primers or synthesized olifor a calcium-mediated increase in contractility. gonucleotides. For certain fragments, sequencing was performed after The molecular structures of each of the components of the a series of deletions was made by exonuclease 111 digestion. Sequenc@-adrenergic signalingpathway in the heart havebeen eluci- ing was completed bidirectionally at least twice with either or both dated by molecular cloning techniques except for the catalyst (8) Sequenase (9) and Taqpolymerase (10) with or without 7-deazaitself, adenylylcyclase. Studies describing the cloning, struc- dGTP. In certainGC-rich areas, electrophoresis was carried out in a ture, and functionof four mammalian adenylylcyclases (type polyacrylamide gel containing both 8 M urea and 20% formamide in order to eradicate problems with band compression. I-IV) have thus far been reported (2-5). These molecules Northern Blotting-Poly(A)+ RNA was prepared from various together with the multiple drug resistance and cystic fibrosis tissues (brain, heart, lung, kidney, testis, and skeletal muscle) and gene products (6, 7) are members of a superfamily that share cell lines (BAEC,bovine aortic endothelial cell; S49, mouse lymphoma cell; GH,, rat pituitary tumor cell; and PC12, rat adrenal pheochro* This work was supported by National Institutes of Health Grant mocytoma cell) by the method of Chomczynski and Sacchi with HL-38070 and a grant from Lederle Laboratories. Thecosts of oligo(dT) columns (11).Five wg of poly(A)+ RNA were employed for publication of this article were defrayed in part by the payment of Northern blotting. The blot was prehybridized in a solution containing 50% formamide, 5 X SSC, 5 X Denhardt's, 25 mM NaPO, (pH page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The abbreviations used are: SDS, sodium dodecyl sulfate; bp,base $ To whom correspondence should be addressed. pair(s); EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; (I Present address; Dept. of Medicine, National Defense Medical Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid kb, kiloCollege, Tokorozawa 359, Japan. base(s); GTP-yS, guanosine 5'-3-(thio)triphosphate.

13553

13554

Cardiac Adenylylcyclase cDNA

6.5), 0.25 mg/ml calf thymus DNA, and 0.1% SDS at 42 "C for 2 h before the addition of probe. Hybridization was performed a t 42 "C for 18 h followed by washing under increasingly stringent conditions. Various restriction fragments from the adenylylcyclase cDNA were used as probes. The probes were radiolabeled by the multiprimer random-labeling method using [32P]dCTP. Expression in CMT Celki-A 4.3-kb, EcoRI-Ssp1 fragment, designated as clone 113-72, was constructed by ligating the EcoRI-SphI (nucleotide 1to 954, from clone 72 in pucls), SphI-EcoRI (nucleotide 954 to 1604, from clone 7), and EcoRI-Ssp1 (nucleotide 1604 to 4303, from clone 8) fragments. The fragment, which contains the entire coding sequence, was subcloned into theEcoRI-EcoRV polylinker site of the plasmid pcDNA I, a mammalian expression vector, and designated pcDNA113-72. Twenty pg of the purified plasmid were transfected into CMT cells (12) by a modification of the method of Golub et al. (13). Briefly, the cells weregrown to 80% confluence. After washing with phosphate-buffered saline twice, 0.5 ml of trypsin solution was added. The cells were incubated for 10 min, and 20 pg of purified plasmid resuspended in 4ml of Dulbecco's modifiedEagle's medium containing 400 pg of DEAE dextran and 0.1 mM chloroquine were added. The cells were incubated for 4 h followed by dimethyl sulfoxide shock for 2 min. After washing with phosphate-buffered saline twice, the induction medium, which contains 1 p~ CdCl, and 0.1 p M ZnC12, wasadded and thecells were incubated at 37 "C for 72 h before harvesting. Control cells were mock-transfected and induced in the same way. Adenylylcyclase Assay-The transfected CMT cells were washed twice with phosphate-buffered saline and then collected into 1 ml of cold buffer containing 50 mM Tris-HC1 (pH 8.0),2 mM EGTA, 10 p~ phenylmethylsulfonyl fluoride, 100 units of leupeptin, and 50 units of egg white trypsin inhibitor (3). Thecells were homogenized with a Polytron (setting 6 for 10 s), and centrifuged at 800 X g for 10 min at 4 "C. The supernatant was further centrifuged at 100,000 X g for 40 min at 4 "C. The resultant pellet was resuspended in 50 mM Tris (pH 8.0), 1 mM EDTA, 1 p~ phenylmethylsulfonyl fluoride, 50 units of leupeptin, and 50 uints of egg white trypsin inhibitor, to a concentration of 5 pg/pl. This crude membrane preparation was used in the adenylylcyclase assay. Adenylylcyclase activity was measured by a modification (14) of the method of Salomon (15). Briefly, the washed membranes from CMT cells (15-30 pg per assay tube) were resuspended in 100 pl of solution containing 1 mM creatine phosphate, 8 uints/ml creatine phosphokinase, 4 mM Hepes (pH 8.0),2 mM MgCl,,0.1 mM cyclic AMP, 0.1 mM ATP, and [32P]ATP(0.2-5 pCi/assay tube). The reaction mixture was incubated at 30 "C for 30 min, and thereaction was stopped by the addition of 100 p1of 2% SDS. To monitor the recovery from the columns, 3H-labeled cyclic AMP was used. Cyclic AMP was separated from ATP by passing through Dowex and alumina columns, and the radioactivity was measured by liquid scintillation counting. Protein concentration was measured by the method of Bradford (16) using bovine serum albumin as a standard.

A

i'"

DOIVA

I

4

TAG

4

E

4X 4H4

4

5

I kb

FIG. 1. Partial restrictionmap and thecDNA clones of Type V adenylylcyclase. A, partial restriction map of adenylylcyclase cDNA.Coding portion is boxed andthe hatched box shows the polyadenylation site. E, EcoRI restriction site; H, HincII; S , SphI; X, XhoI; and SS, SspI. B, cDNA clones obtained from the canine heart X g t l O library and thesequencing strategy. Five different clones, 7, 8, 25, 72, and 113, ranging in size from 1.2 to 3.5 kb, constitute a 4.4-kb cDNA, which contains the entire coding region.

obtained clone 72, which extended the sequence an additional 380 bp. The restriction map of these clones is shown in Fig. 1. These clones (7 and 72) contain highly GC-rich regions, and therefore sequencing was performed bidirectionally at least three times using both Sequenase and Taq polymerase with or without7-deaza-dGTP (8). The reaction mixtures were run in a polyacrylamide gel containing both 8 M urea and 20% formamide. A putative ATG translation initiation codon, which exists in thecontext of a reasonable Kozak consensus sequence (17), was identified in an extended open reading frame. We also identified a stop codon (TGA) 81 bp upstream of this ATG in the same open reading frame (Fig. 2). This region is highly GC-rich as seen in other types of adenylylcyclase cDNAs (2RESULTSANDDISCUSSION 5 ) . We concluded that this ATG initiates the translationof The initial screening of the X g t l O library was performed an open reading frame of 3552 bases, encoding a 1184-amino under relatively low stringency conditions as described under acid protein, followed by 634 bp of a 3"untranslated region "Materials and Methods." A 970-bp, Ad-HincII, cDNA frag- upstream of the polyadenylation site. The homology to brain adenylylcyclase (type I) is higher in the cytoplasmic but lower ment from typeI adenylylcyclase cDNA was used as the probe. This portion encodes the cytoplasmic domain and is in the transmembrane portion as shown by the dot matrix known to behomologous among the various isoforms of comparison (Fig. 3B). The first cytoplasmic domain, espeadenylylcyclase (2-5). Among the cDNA clones initially iden- cially the 5' half of the loop, is highly homologous to other tified, the size of the inserts varied from 0.7 to 3.5 kb. Clones types of cyclase including yeast adenylylcyclases (18) and 113 and 8 used the same polyadenylation signal. Three clones, various types of guanylyl cyclase (19, 20). This suggests the 8, 25, and 113, were overlapping and together formed a 3.5- presence of an essential functionwithin this domain, e.g. ATP kb-long fragment. However, this 3.5-kb fragment did not binding. The extracellular loop between the 9th and 10th contain an initiator ATG with an optimal Kozak consensus transmembrane-spanning regions is the largest (Fig. 3A). In sequence in the long open reading frame. In order to obtain comparison to other types of mammalian adenylylcyclases, the 5' end of this cDNA, an 800-bp 5' EcoRI fragment from type V is more similar to types I1 and IV in that it has a clone 8 was used as a probe to screen the X g t l O library again. shorter C-terminal tail, but is unique among the cyclases in After sequencing the entire set of new cDNA clones, it was having a much longer N-terminal tail (type V, 164 amino found that clone 7 overlapped for 800 bases with clone 8 and acids; type I, 63; type 11, 44; type 111, 77; and type IV, 28). The tissuedistribution of this novelgene product was extended the sequence upstream an additional 400 bp. To obtain the additional 5' sequence, fragments consisting of examined by Northern blotting using a probe unique to this either themost upstream 60 or 500 bp of clone 7 were used to cDNA. Tissues known to possess high adenylylcyclase activity rescreen the cDNA library. Out of 82 primary positives we were examined. The message was most abundantly expressed

Cardiac Adenylylcyclase cDNA

13555

FIG. 2. Nucleotide and aminoacid sequence of Type V adenylylcyclase. The entire coding sequence, as well as a portion of the 5"untranslated and complete 3"untranslated regions, are shown. ATG shows the putative translation initiation site in the open readingframe. TGA shows the translation termination a n in the open readingframe. The putative polyadenylation signal is marked (AATAAA).

-

in the heart, a tolesser degreein the brain, while no expression other types of adenylylcyclase, an RNA species of approxiwas detected in other tissues (testis, skeletal muscle, kidney, mately 6 kb was also seen in poly(A)+ RNA prepared from and lung). Both heart and brain contained messages of -5 heart and brain if the hybridization was carried out under and 7 kb (Fig. 4). The ratio(3:2) of these mRNAswas similar relatively less stringent conditions. We have also detected a between thetwo tissues. When different probes from differentmuch larger species in skeletal muscle under similar condiportions of the cDNA were used, i.e. thetransmembrane tions. The size of this message was approximately 9.5 kb. portion, the entire cDNA, or3"untranslatedportion,the These additional mRNAs might represent othermembers of Northernblottingresults were similar. Thus far we have the adenylylcyclase family not yet identified. obtained six type V adenylylcyclase clones which contained a Biochemical characterization of the protein product encomplete 3' end. There was no divergence in their sequence coded by this cDNA was obtained using a CMT cell-expresand all used the same polyadenylation site. Taken together, sion system (12). The CMT cell is a derivative of the COS these findings suggest that the two messages are most likely cell in which expression of the T-antigen is under the control products of the same gene, probably the resultof alternative of a metallothionein promoter. Thus theaccumulation of Tsplicing of a single precursor RNA. The message sizes predict antigen in the cell is further enhanced by the addition of that each containsat least 1-3 kb of 5"untranslated sequence. heavy metal ion in the medium. The adenylylcyclase cDNA We havealso examined the expression of this message in construct (113-72) was cloned into pcDNA 1, a cytomegaloviseveral cell lines. Messenger RNA of similar size was detected rus promoter-driven expressionvectorwith an SV40 enin GH, and PC12 cells but not in S49 or BAEC (data not hancer-origin of replication element. A crude membrane prepshown). It is of interest that when a probe from the second aration was prepared from the transfected CMT cells and a cytoplasmic portion of this cDNA was employed, i e . a 1.6-kb variety of agents known to stimulate adenylylcyclase were XhoI fragment from clone 113, which shares homology with examined (TableI). There was a dose-dependent relationship

Cardiac Adenylylcyclase cDNA

13556 5.00

A

4.00 3.00 2.00

1.53 0.00

-1.00 -2.00

-3.00 -4.w -5.w 1030

I 200

400

600

BOO

I

1000

TYP V

FIG. 3. Hydropathy plot of typeV adenylylcyclase and protein dot matrix comparison with type I adenylylcyclase. A, MacVector 3.5 software was used to analyze the structure of type V adenylylcyclase. The method of Kyte and Doolittle (25) was used with a window size of 7. Twelve peaks are numbered, which represent putative transmembrane spanning regions. B, MacVector 3.0 was used for the amino acid dot matrix comparison with the stringency set at60% and thewindow size a t 8.

FIG. 5. Effect ofcalcium on type V adenylylcyclase activity. Adenylylcyclase activity (CMT cells transfected with pcDNA113-72 or cardiac sarcolemma) was measured in the presence of increasing concentrations of calcium (0-1 mM). Cardiac sarcolemma was prepared as described (26). Both membrane preparations were first washed with EGTA prior to assay (27). Similar results were obtained in three independent experiments. The efficiency of transfection was confirmed by a t least a 4-fold increase in both basal and forskolinstimulated adenylylcyclase activities over control. Transfected membrane (O), transfected membrane with calmodulin (200 nM) (W), cardiac sarcolemma (O), cardiac sarcolemma with calmodulin (200 nM) (0).

MWS(kb) 9.49 7.46

-

4.4

-

1.35 -

H

B

T

S

K

L

FIG. 4. Northern blot analysis of mRNA from different tissues. Five pgof poly(A) RNA were employed for each assay. An EcoRI-Him11 0.9-kb fragment from the cardiac adenylylcyclase brain ( B ) ,testis ( T ) skeletal , cDNA was used as the probe. Heart (H), and lung ( L ) . muscle ( S ) , kidney (K),

TABLE I Adenylylcyclase activity in transfected CMT cells Adenylylcyclase activity (control cells versus cells transfected with the type V adenylylcyclase cDNA) was measured in the presence of the following activators: NaF (10 mM), GTPyS (100 p ~ )forskolin , (100 p M ) . Values are means f S. E.

I 0

20

40

60

80

100

P-rite lnhibltofs VU)

FIG. 6. Effect of adenosine and its analogues on type V adenylylcyclase activity. Membranes (transfected with pcDNA113-72 or control) were preincubated with 5 mM Mn*+and 100 p~ forskolin for 10 min prior tothe assay. Adenosine (W), 2’3’-AMP (*), and 2’-deoxy-3’-AMP (0).Similar deoxyadenosine (O), results were obtained in three independent experiments. Each point is the average of triplicate determinations.

was transfected, the resultant adenylylcyclase activity was slightly (-15%) lower than thatof mock-transfectedcells. All p m l f m i n mg.protein activities were enhanced several-fold in the transfected cells, as compared to controls, with the forskolin-stimulatedactivity Control 15 f 3.1 30 f 4.5 52 f 7.2 2.5 & 0.5 Transfected 11 f 1.6 41 f 4.7 84 f 12.5 321 f 28 showing the greatest increase in activity (>6-fold inceaseover ’Control < transfected, p < 0.001. Similar results were obtained control). in three independent experiments. For each experiment, the results The calcium/calmodulin sensitivity of this protein was also represent the average obtained from five independently transfected assessed. Adenylylcyclase purified from heart has previously plates of CMT cells. been shown to be inhibited by the addition of calcium (21). Thirty pg of the crude CMT cell membranepreparation were between the amount of plasmid transfected and the resultant incubated in the presence of increasing concentrations of adenylylcyclase activity, both basal and forskolin-stimulated CaC12. The transfected adenylylcyclase inthe CMT cells was (0-30 pg per transfection). When the plasmid without insert inhibited in a concentration-dependent manner by the addiBasal”

Nap

GTPyS”

Forskolin”

Cardiac Adenylylcyclase cDNA tion of 0-1 mM calcium. Addition of calmodulin (200 nM) did not alter this inhibitory effect. We also assessed the adenylylcyclase activity in canine cardiac sarcolemma and found degree of inhibition in the that it too exhibited a similar presence of calcium (Fig. 5). Another featureof adenylylcyclase is its inhibition through an allosteric purine binding or P-site (22,23). Adenylylcyclase activity in the transfected CMT membranes exhibited a concentration-dependent inhibition in the presence of adenosine and its analogues (0-100 PM). This was more apparent when the enzyme was firstactivated by the addition of Mn2+forskolin. The order of potency was as follow: 2'-deoxy-3'AMP > 3'-AMP > 2'-deoxyadenosine > adenosine (Fig. 6). The above data indicate that the protein encoded by this cDNA is adenylylcyclase with the biochemical features of the cardiac isoform. As described above, this adenylylcyclase isoform demonstrates several unique features. Although the type V isoform described in this study and the recently reported type IV are both expressed in heart, the distribution of the type V appears to be relatively limited (heart and neural tissue), while type IV is widely expressed in a variety of different tissues (5). In both heart and brain, mRNA species of -5 and 7 kb were observed for the type V isoform. Our data suggest that these are likely produced via an alternative splicing mechanism although the use of different polyadenylation signals cannot be excluded. Finally the enzymatic activity of this isoform is not stimulated with calmodulin and is inhibited by the addition of calcium. Defects in catecholamine release, /+adrenergic receptor content, receptor-G. coupling, and adenylylcyclase catalytic activity have been identified either in tissue removed from failing human hearts (1,24) or in the hearts of animals with experimentally induced heart failure. Judging the significance of these changes as to their cause and effect relationship, however, is hampered by the very nature of the fact that the process of heart failure can only be studied over a relatively long time frame and in the intact animals. Nevertheless, a recentreport suggests that an abnormality in myocardial cyclic AMP production may be a fundamental defect present in patients with end-stage heart failure (24). The cloning of the cDNA for this cardiac isoform of adenylylcyclase permits the assessment of whether the decrease in adenylylcyclase enzymatic activity, which occurs during the progression of heart failure, is mediated via apre- orpost-translational mechanism. Acknowledgments-We thank Drs. A.G. Gilman(University of Texas Southwestern Medical Center) and R. Reed (Johns Hopkins University) for providing us with adenylylcyclase cDNAs, Dr. Brian

13557

F. Hoffman (Columbia University) for his advice and encouragement, and Dr. S. Wrenn (Medical Research Division, American CyanamidLederle Laboratories) for helpful discussions. 1.

2. 3. 4. 5. 6. 7.

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

REFERENCES Homcy, C. J., Vatner, S. F., and Vatner, D. E. (1991) Annu. Reu. Physiol. 5 3 , 137-159 Krupinski, J., Coussen, F., Bakalyar, H. A., Tang, W.-J., Feinstein, P. G., Orth, K., Slaughter, C., Reed, R. R., and Gilman, A. G. (1989) Science 244, 1558-1564 Bakalyar, H. A., and Reed, R. R. (1990) Science 250, 1403-1406 Feinstein, P. G., Schrader, K. A., Bakalyar, H. A,, Tang, W.-J., Krupinski, J., Gilman, A. G., and Reed, R. R. (1991) Proc. Natl. Acad. Sci. U. 5'. A. 88,10173-10177 Gao, B., and Gilman, A. G. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,10178-10182 Gros, P., Croop, J., and Housman, D. (1986) Cell 47, 371-380 Rommens, J. M., Iannuzzi, M. C., Kerem, B.,Drumm, M.L., Melmer, G., Dean, M., Rozmahel, R., Cole, J. L., Kennedy, D., Hidaka, N., Zsiga, M., Buchwald, M., Riordan, J. R., Tsui, L. C., and Collins, F. S. (1989) Science 2 4 5 , 1059-1065 Khambaty, F. M., and Ely, B. (1990) BioTechniques 9 , 714-715 Tabor, S., and Richardson, C. C. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,4767-4771 Innis, M. A., Myambo, K.B., Gelfand, D. H., and Brow, M. A. D. (1988) Proc. Natl. Acad. Sci. U. S. A. 85,9436-9440 Chomczynski, P., and Sacchi, N. (1987) Anal. Biochem. 1 6 2 , 156-159 Gerard, R. D., and Gluzman, Y. (1985) Mol. Cell. Biol. 5, 32313240 Golub, E. I., Kim, H., and Volsky, D. J . (1989) Nucleic Acids Res. 17,4902 Bockaert, J., Hunzicker-Dunn, M., and Birnbaumer, L. (1976) J. Biol. Chem. 251,2653-2663 Salomon, Y., Londos, C., and Rodbell, M. (1974) Anal. Biochem. 58,541-548 Bradford, M. M. (1976) Anal. Biochem. 7 2 , 248-254 Kozak, M. (1989) J. Cell Biol. 1 0 8 , 229-241 Kataoka, T., Broek, D., and Wigler, M. (1985) Cell 4 3 , 493-505 Chinkers, M., Garbers, D. L., Chang, M. S., Lowe, D. G., Chin, H., Goeddel, D. V., and Schulz, S. (1989) Nature 3 3 8 , 78-83 Singh, S., G, L. D., Thorpe, D. S., Rodriguez, H., Kuang, W. J., Dangott, L. J., Chinkers, M., Goeddel, D. V., and Garbers, D. L. (1988) Nature 3 3 4 , 708-712 Potter, J. D., Piascik, M. T., Wisler, P. L., Robertson, S. P., and Johnson, C. J. (1980) Ann. N. Y. Acad. Sci. 3 5 6 , 220-231 Bushfield, M., and Johnson, R. A. (1990) Biochem. SOC.Trans. 18,150-151 Tang, W.-J., Krupinski, J., and Gilman, A.G. (1991) J. Biol. Chem. 266,8595-8603 Feldman, M. D., Copelas, L., Gwathmey, J . K., Philips, P., Warren, S. E., Schoen, F. J., Grossman, W., and Morgan, J. P. (1987) Circulation 7 5 , 331-339 Kyte, J., and Doolittle, R. F. (1982) J. Mol. Biol. 1 5 7 , 105-132 Vatner, D.E., Homcy, C. J., Sit, S. P., Manders, W. T., and Vatner, S. F. (1984) J. Clin. Znuest. 7 3 , 1473-1482 Lynch, T. J., Tallant, E. A., and Cheung, W. Y. (1976) Biochem. Biophys. Res. Commun. 68,616-625