Molecular Cloning and Characterization of the Type VI1 Isoform of ...

Vol. 269,No.46,Issue of November 18,pp. 28893-28898, 1994 Printed in U.S.A.

JOURNAL OF Blow~rcALCHEMrsTnY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc %E

Molecular Cloning and Characterization of the Type VI1 Isoform of Mammalian Adenylyl Cyclase Expressed Widely in Mouse Tissues and in 549 Mouse Lymphoma Cells* (Received for publication, July 19, 1994, and in revised form, September 6, 1994)

Peter A. Watson$, John Krupinski, Anna M. Kernpinski, and Carole D. Frankenfield From the Weis Center for Research, Geisinger Clinic, Danville, Pennsylvania 17822-2615

We have isolated a5199-nucleotide cDNAfrom a maining isoforms (1, 12). A human clone recently sequenced mouse librarycontaining an open reading frame encod-appears to encodean adenylyl cyclase protein (GenBankTM acing the 1099-amino acid type VI1 adenylyl cyclase pro- cession no. D25538) which is highly homologous to the VI1 type tein. The type VI1 protein is most closely related in pri- protein described in this report. Only the type VI1 isoform mary structure to anunpublishedhumanadenylyl expressed prominently in S49 mouse lymphoma cells (1) and cyclase clone(GenBankm accession no. D25538)and the closely related human AC clone reported in the sequence type I1 adenylyl cyclase. Northern blot analysis demon- data base remain tobe characterized. strates that the type VI1mRNA is most abundant in The type VI1 isoform has been found throughRT-PCR analmouse heart, spleen, and lung. ysis to be expressed at t h e mRNA level in heart, liver, kidney, C A M P content rises rapidly in HEX 293 cells overextestis, skeletal muscle, and brain (1).Type VI1 is the only isopressing type VI1 adenylyl cyclase following treatment form of adenylyl cyclase expressed in theXC rat sarcoma cell with phorbol ester, peaking by 4 min, while cells exline and it is also abundantin S49 mouse lymphoma cells (with pressing the type I1 adenylyl cyclase reach peak accuthe typeVI adenylyl cyclase) and may be responsible for CAMP mulation only after 20 min. Increases in intracellular calcium through treatment of type MI-293 cells with accumulation in these cell lines in response tocell volume exeither ATP or A23187 alone failed to increase intracellu- pansion (13, 14). A cDNA encoding the type VI1 isoform of lar c A ” content. Phorbol ester treatment acted syner- adenylyl cyclase has been isolated from an 549 mouse lymgistically with P-adrenergicstimulation to increase phoma cell library. The cDNA has been expressed inHEK 293 c A ” content in type VII-transformed cells. Pretreat- cells, allowing for biochemical analysisof type VI1 activity. ment of type VII-transformed cells with pertussis toxin EXPERIMENTALPROCEDURES fails to prevent phorbol ester potentiation of isoproterenol stimulation. Thus the ability of phorbol ester to Preparation of S49 Mouse Lymphoma Cell cDNAlA ZAP Libraryisoproterenol-stimulatedtype MIac- S49 mouse lymphoma cell mRNA was isolated utilizing methods and increase basal and tivity appears to be a direct effect on this adenylyl cy- reagents provided in the FastTrack mRNA isolation kit (Invitrogen, clase isoform and not the result of modification of the Inc.). cDNA was synthesized from isolated S49 cellmRNA and coupled to EcoRI-linker arms. cDNAs were subsequently cloned into EcoRIinhibitory G protein, Gi. digested A Z A P vector arms utilizing reagents provided in commercially available A ZAP-cDNA synthesis and packaging kits (Stratagene).The cDNA library was amplified utilizing the XL-1Blue strain of E. coli as A vast array of agonists and olfactory stimuli have been the host. The cDNA library produced by these procedures contained shown to modulate cyclic AMP concentrations in mammalian 1.62 x lo6 independent clones. cells through their interactions with specific receptors coupled Isolation and Characterization of cDNA Clonesfor Qpe VII Adenylyl (G pro- Cyclase-A l-kilobase XbaI fragment of a partial type VI1cDNAobtoguaninenucleotide-bindingregulatoryproteins byDr. teins).’ Like the receptor and G protein families to which they tained from XC rat sarcoma cellclones(graciouslyprovided Masahiro Kawabata and Dr. Ronald Taussig;University of Texas Southa r e coupled, adenylyl cyclase exists as a family of related prowestern Medical Center) was subjected to randomly primed labeling teins in mammalian cells. Eight isoforms of t h e mammalian with [a-32PldCTP.The 32P-labeledproduct was used as a hybridization adenylyl cyclase were identified through RT-PCR analysis of probeto screen approximately 1 x lo6 plaques from the S49mouse cDNAs produced from six different tissues(11, although other lymphoma cell library utilizing standard methods (15). 13 clones were the l-kilobase type VI1 isoforms may exist. Sevenof these isoforms of adenylyl cyclase identified whichreproduciblyhybridizedto have been completely cloned, sequenced, and characterized (1- probe during this library screen. The longest clone,clone 4, was shown to contain a region in the type VI1 cDNA sequence presented in Fig. 1 111, with partial sequences having been published for the re- extending from nucleotide 989 (codonfor amino acid 101) to nucleotide 4521 (3’-untranslated region). A secondary screen identified a 1.5-kilo* This work was supported by National Institutes of Health Grants base clone (clone 4A) which contained sequence encodingthe remainder GM46395 (to J. K.) and HK42365 (to P.A.W.), as well as American of type VI1 adenylyl cyclase. Clone4A spans the region between nucleHeart Association Grant-in Aid 91010670 (to J. K.), and a grant-in-aid otide -678 and 1442 (codon for amino acid 252) in the sequence prefrom the Pennsylvania AfKliate of the American Heart Association (to sented in Fig. 1. Restriction fragments were subcloned and sequenced P. A. W.). The costs of publication of this article were defrayed in partby the payment of page charges. This article must therefore be hereby utilizing dideoxy chain terminators and primers to vector sequencesor synthetic primers to internal sequence in the clones. 7-Deaza-dGTP was marked “aduertisernent”in accordance with 18 U.S.C.Section1734 routinely used to reduce secondary structure during sequencing reacsolely to indicate this fact. $To whom correspondence should be addressed: Weis Center for tions. Either Sequenase (U. S. Biochemical Corp.) or Taq polymerase sequencing reactions for either conResearch, Geisinger Clinic, 100 N. Academy Ave., Danville, PA (Applied Biosystems) was used for ventional [35SldATP/autoradiographybased sequencing or automated 17822-2615.Tel.: 717-271-6671;Fax: 717-271-6701. The abbreviations used are: G-protein, guanine nucleotide-binding sequencing with fluorescently labeled deoxynucleotides (Applied Bioregulatory protein; RT, reverse transcription; PCR, polymerase chain systems). The sequence presented in Fig. 1was determined by sequencreaction; Ro 20-1724,4-(3-butoxy-4-methoxybenzyl)-2-imidazolidinone;ing of both strands of the cDNA clones.A full-length cDNA clone forthe PDBu, phorbol 12.13-dibutyrate. type VI1 adenylyl cyclase wasconstructed utilizing an AatIIrestriction

28893

28894

n p e VII Adenylyl Cyclase GGCACGAGCGTGGCTCCCTCCTGCAGCCTCAGGGACACCGGAGATGGCCCAGTGCGGCAGGTAACAGGGTCGGAGGCGGCTACGCGGGGGACCG~GTGGG-588 GACGTTCCCGAGATCCCCAGCCCGCGCCCCGCACGGTCAGTGTGACCAGCTGGGGGCGCCCCGGCTCTCGCGCTCGCGGGCGGGAACTCCAGAGCGCCCG-488 GGCCCTCGGCGGTGGCGGCAGGGAAGAAGATGTCTCTACCTGGAAGCTCCCAGGAGACAGCTGTGGAGCCCACGAGAGCCACAGGGCCTTGACCTCATCT 388 CTTCCCACCATCCAAGAGGGCAAAGTCAGGGGCTTTAAACGGGAGAGGAGAGACGACTCAGCACTCCTTGTGATAAGGCCCTGTGGATGTGTGGCTAGCT -288 TGCTCTGTCCTGTGCGTCAAGGAATGCAGGTGCAAAGAGCTGAGGACCTGGAAGTTTCTGTCAGCTCCTCCTAGACACCCTGAAGACACCAGTGAACAGA 188 ACCCAGTGGCTGGAGCAGGTGTCACATGCTCATGGAGCCACGGGCTCTAGGTGCCTCTGTCCAGTCCTGCCTGATACCCTCAGAAAGGGAAGAAGCCTAG -88 GCCACACACTAAAGCGTCAGCTGACAGATGTATCTCAGACCGTGGGGCTCCT~GCAGGCTGCTGGACCAGGCCACTGGGGAAGAGG

-

-

FIG.1. Translation of type VI1 adenylyl cyclase cDNA sequence. The type VI1 adenylyl cyclase proteinsequenceshown above wasdeduced from two overlapping clones isolated from a n S49 mouse lymphoma cell library as described under "Experimental Procedures." The GenBankTM accession no. for the nucleotide sequence is U12919. The underlined sequences are regions of the form the12 memproteinpredictedto brane-associatedhelices(27)presentin alladenylyl cyclase isoforms cloned at this time.

MPAKGRYFLNEGDEGPDQAALYEKYRLTSLHGPLLLLLLLVAAATCIALISIAFSHEDLRRHPWLGTAFLMLTLFVALYVLVYVECLVQRWLRA~

100

UACLMVLGSVLMWDSLENEAHAUEQVPFFLFVVFVVYALLPLSRRAAIVVGVTSTVSHLLVFGAVTRAFQTSMSSTQLGLQLLANAVlLLGGNFTGAFHK

200

HQLQDASRDLFIYTVKClQIRRKLRVEKRQQENLLLSVLPAHISMG~KLAIIERLKEGGDRHYMPDNNFHSLYVKRHQNVSILYADIVGFTRLASDCSPK

300

E L V W L N E L F G K F D Q l A K A N E C M R l K l L O C ~ I C V S G L W S L P T H A R N C ~ ~ t C E A I ~ ~ ~ l ~ t ~ ~ t H S ~ L400 C~IGLR~~~HD VSLANRMEAAGVPGRVHITEATLNHLDKAYEVEDGHGEQRDPYLKEMNIRTYLVIDPRSQQPPPPSHHLSKPKGDATLKMRASVRVTRYLESWGAARPFA

500

HLNHRESVSSSETPISNGRRQKAIPLRRHRAPDRSASPKGRLEDDCDDEMLSAIEGLSSTRPCCSKSDDFHTFGPIFLEKGFEREYRLVPIPRAR~

600

ASLVFVCILLVHLLVMPRMATLGVSFGLVACLLGLVLSFCFATEFSRCFPSRSTLPAISESVETQPLVRLVLVVLTVGSLLTVAIINMPLTLNPGPEQPG

700

DNKTSPLAAQNRVGTPYELLPYYTCSCILGFIACSVFLRMSLELKAMLLTVALVAYLLLFNLSPCUHVSGNSTETNGTQRTRLLLSDAQSMPSHTLAPGA

800

QETAPSPSYLERDLKlMVNFYLILFYATLlLLSRQlDYYCRLDCLUKKKFKKEHEEFETMENVNRLLLENVLPAHVAAHFIGDKAAEDWYHQSYDCVCVH

900

FASVPDFKVFYTECDVNKEGLECLRLLNEIIADFDELLLKPKFSGVEKIKTIGSTYMAAAGLSAPSGHENQDLERKHVHIGVLVEFSMALMSKLDGINRH

1000

SFNSFRLRVGINHGPVlAGVIGARKPQYDlUGNTVNVASRMESTGELGKlQVTEETCTlLQGLGYSCECRGLINVKGKGELRTYFVCTDTAKFQGLGLN

1099

TGAGGTGGCTGGTGGTCAGCCTCCTTCCCCGAGGGAGCCAAGAATGTAGCCCCATGTCTGTTGCAGTGGCTTCTTTGGACTTGCACTACAGGATGGCTTT GACCTGTGCATCAGATTCTGTTTGAAGCAGCTACTGCGTTGTACACAGCGGCTCTGTGCTTCAGCCTCTACAGTTCC~TTAGCTAGACCACTGGTCTA CTACAGGCTGTGTTCATTTCCAGGGTGCTGGGGAAGAGACTTCAGTGCATGACCAAGATAGACATCCACCTTGGTGCCAGTGAACAGCATTCACAGGAGA CAAAAGCTCTACTGGCTACAGGAGGCTCAGCCAGGCTTATTAGCATGGGTTGCTGCTTGCCTTCCTCCCATCAAATCTCCCATGGGATGTTATTCTTTCA ATTAGGCATTCTGGTAAATGGAGTTGAAAACTGTGTATATTGGTGGGTAGTCTC~ACAGCAGAGAAAATGTCTGATCTACACTTGTCACTTTTTTCCA TCTCTGGCTTATGTTTGAAACGGACATGTCATAAACAGAGTTTTAGCTTTACCACTGACTCTTAGATGCTAGACAAAGATCTCCACCTTTCTAGTGTATT TTCTCTTGTTAACCACAGACTACAAGTAAATAGGTCTGCTGTCTAGTGTCCTTTTATAGGATCAGATTGGCTGCAGGGACAGAGTTCTAAGGAATGGGGC TCATAGCAGCAGCAAAGCTTTGAATTTGCAATCAAGCATTTTTTGATGCAAGTCTTTTGGGACAAGGCTCAGGAAGTTTA~GTCTAGAAATGAGGTATG ATGTTTCAGTTTTTCTGTGTGTACTTTATTTATTTTGGAGTCAGGGTCAGCCTAGCTTGGTCTCAAACTTCTAACCTCCTGTTTCGGTCTCCGGAGTGTT TCAATTACAGATGTGCACGACTATCTCCAGCTGTTTCTGTGGGAAAGCCTGTGTAGACAGGCTTGGACAACTTTGTAGCACTTGCCTTTTCTCCAGTCTT CTGAGCTGACGACAGAGCTTCAAGAACAATCCACTTGACAGGAACATGTGTTCAGGACTCTGGCCTGTGAACTGAGCCCCTCAACAAATGCCAAATTGTT CTTATGCAAATGAGTCAAGGCAAAAGCCAGCTTCGTGAGATGGGTGTCTTACTGTGCTTAGCTCCAGAGCATTCCGAGAGAGATGACCAAACACCCCACT CCTTTTTGGAAATGACCTCGTGCCGCTCGTGCC

3397 3497 3597 3697 3797 3897 3w7 4097 4197 4297 4397 4497 4530

endonucleasesite a t nucleotide12Ol/amino acid 172 of the cDNN protein sequence (Fig. 1) in the region of overlap between clone 4 and 4A. This full-length cDNA was subcloned into pBluescript in the multicloning site utilizing EcoRI restriction sites present in the original linker added during library construction. Northern Blot Analysis-The presence of type VI1 mRNA in mouse tissues was assessedby hybridization analysis of a commercially available mouse multiple tissue Northern blot (Clontech Laboratories). A cDNA hybridization probe was produced by randomly-primed labeling of the type VI1 cDNA in the presence of [ C X - ~ ~ P I ~Blots C T P .of fractionated RNA were prehybridizedfor 4 hfollowed by hybridization with lo6 c p d m l of randomly primed cDNA probe at 42 "C for 20 h in 50% formamide, 5 x SSC, 50 mM Na' phosphate (pH 6.5),1% sodium dodecyl sulfate, 0.5% powdered milk, and 0.5 mg/ml sonicated salmon sperm DNA. Blots were rinsed with 1 x SSC, washed for 15 min in 1 x SSC + 0.1% SDS at room temperature, and then 0.1% SSC and 1% SDS a t 65 "C for 20 min (20x SSC = 0.3 M sodium citrate,3 M sodium chloride). RNA loading in lanesof gels was standardized by rehybridization to a p-actin cDNA probe (Clontech Laboratories). Expression of Type VII and Type II Adenylyl Cyclase cDNAs in HEK 293 Cells (ATCC CRL 1573)"The full-lengthcDNAs for type I1 adenylyl cyclase (provided graciously by Dr. Randall Reed, Johns Hopkins University) and typeVI1 adenylyl cyclase were subcloned into the expression vector pCMV-Neo describedpreviously (1). HEK 293 cells (IO6 cells/60-mm dish) were transfected with either the type VII-pCMV-Neo construct ora pCMV-Neo construct with typeVI1 cDNA in an antisense orientation by calcium-phosphate co-precipitation as described previously (16). Geneticin (0.5 mg/ml active compound; Life Technologies, Inc.) was used t o select HEK 293 cells which had stably incorporated the expression-construct DNA and thus gained resistance to the antibiotic through expressionof the neomycin-resistance gene in the pCMVNeo vector. This selection favored HEK293 cells expressing the type VI1 cDNA which was subcloned into the same vector. Cell colonies surviving antibiotic selection from a single transfection were pooled into a polyclonal population for subsequent analysis. Cyclic AMP Accumulation in DansfectedCells-24 h prior to experimental analysis, geneticin-resistant populations of HEK 293 cells originally transfected with either the type VI1 or type I1 adenylyl cyclase expression-constructs were passaged to10-cm tissue culture dishes(1x lo6 cells/dish). Cells were washed twice prior to preincubation for 20 min at37 "C in defined medium mimickingDulbecco's modified Eagle's medium in salt and nutrient content and buffered with 15 mM HEPES. Where indicated, staurosporine was added to a final concentration of 100 ILM in dimethylsulfoxide (0.1%final concentration) during both the

preincubationandduringtheexperiment.Dimethyl sulfoxide was added toa final concentrationof 0.1%in controls for these experiments. Cells were pretreated with pertussis toxin (final concentration 100 ng/ ml) for 24 h where indicated. Experimental treatments were initiated by the exchange of preincubation culture medium for culture medium containing agentsas detailed in the legends of the appropriate figures. After incubationa t 37 "C for the times indicated, medium was aspirated and cells were fixed and CAMP extracted with 1.0 ml of ice cold6% trichloroacetic acid. Cells were scraped from culture dishes, the dishes washed with a n additional 0.5 ml of 6% trichloroacetic acid and the combined trichloroacetic acid extracts centrifuged to pellet denatured proteins. Supernatants were ether-washed to remove trichloroacetic acid, lyophilized, and resuspendedfor subsequent acetylation and analysis for CAMPcontent by automated radioimmunoassay (17). Precipitated proteins were resuspended in 1 N NaOH and protein content was assessed by dye binding (18). RESULTS AND DISCUSSION

Results from a variety of experiments indicate that a proportion of the diversity in the regulation and responsiveness of adenylyl cyclase in different cells and tissues canbe accounted for by differential expression of isoforms of adenylyl cyclase therein. Eight distinct gene products have been identified in the family of mammalian adenylyl cyclases (1-11). Despite the high degree of conservation in the topographical structures of the adenylyl cyclases thus far cloned, these isoforms demonstrate significant sequence diversity. This diversity in the primary structure between adenylyl cyclase isoforms reflects a high degree of functional diversity in terms of the activity and regulation of these isoforms. In most tissues examined, multiple adenylyl cyclase isoforms are expressed at the mRNA level, with as many as seven isoforms being expressed in a single tissue as determined by RT-PCR (1).Multiple isoforms of adenylyl cyclase also appear to be expressed in clonal lines of cultured cells. RT-PCR analysis indicates four isoforms of adenylyl cyclase are expressed at the mRNA level in HEK 293 sequence D25538) and two cells (types I, 111, VI, and the human isoforms expressed in S49 mouse lymphoma cells (type VI and VII) (data not shown). To our knowledge only one cell line has been identified which expresses a single isoform of adenylyl

28895

Type VII Adenylyl Cyclase cyclase exclusively, the ratXC sarcoma cell line which appears to express only the type VI1 isoform."." Thus, an acute understanding of the pattern of regulation and response of specific tissues and cells to stimuli which influence the CAMP-dependent signal transduction pathways mustinvolve understanding of the biochemical nature of the specific isoforms of adenylyl cyclase expressed within that tissue orcell. Type VI1 adenylyl cyclase was found to be expressed significantly in both XC rat sarcoma and S49 mouse lymphoma cell lines by PCR analysis. Therefore, a 1-kilobase restriction fragment from a partial clone isolated from XC rat sarcoma cells (graciously provided by Drs. Masahiro Kawabata and Ronald Taussig) wasused to screen an S49cell cDNAlibrary generated in this laboratory. Two overlapping clones were isolated from the S49 cell library which,when combined, cover the 5199 nucleotides of the cDNA sequence for type V I 1 (Fig. 1).The type VI1 clone depicted in Fig. 1, constructed from the 2 partial clones isolated from the mouse library, encodes a protein of 1099 amino acids. The sequence of a type VI1 PCR product, which was originally used to identify type VI1 a s a unique adenylyl cyclase isoform (1)is identicalto the region highlighted in the type VI1 sequence in Fig. 1. The nucleotide sequence which surrounds the codon for the initiator methionine is in fair agreement with the consensus for a site of eukaryotic ribosome binding (19). This initiator methionine codon is in frame with an upstream stop codon starting at position -66. The initial methionine codon in the open reading frame ispreceded by six additional ATG sequences in the cDNA, the importance of which is not known. The predicted secondary and tertiary structure of the type VI1 protein resembles that of the previously cloned adenylyl cyclase isoforms (1-11).A 33-amino acid amino-terminal cytoplasmic domainis followed by alternating setsof six hydrophobic transmembranespansand hydrophilic cytoplasmic domains. Each of these cytoplasmic domains contains a region which bears stronghomology to nucleotide binding regionsin a number of proteins, such as the catalytic domain of the guanylyl cyclases (20). Potential extracellular sites for N-linked glycosylation are predicted a t amino acids 702 (between predicted transmembrane spans 9 and 10) and 771(between predicted transmembrane spans 11 and 12). The N-linked glycosylation sitepredicted between transmembrane spans9 and 10 is found in allisoforms of adenylyl cyclase cloned thus far, while the site between transmembrane spans 11 and 12 ispredicted in several isoforms, including types V, VI, VI11 (1, 5, 7, 9-11), and a recently isolated human adenylyl cyclase clone (GenBankT" accession no. D25538). The typeVI1 protein appears to be most closely related to the human clone D25538 with predicted amino acid identity of 84% between the mouse type VI1 protein and the proteinsencoded by the humanclone. !l'ype VI1 and the humanD25538 clone would represent themost closely related adenylyl cyclase sequences. Alternatively, it ispossible that thesetwo protein represent species-specific homologs with sequence differences reflecting the fact that the cDNAs were isolated from mouse versus human cDNA libraries. Distinguishing between these possible explanations hasnot yet been possible. The type VI1 protein is also closely related to the type I1 adenylyl cyclase in structure. The type VI1 protein shows significant homology with the type I1 protein with amino acid identity of 53% between the two isoforms. However, the expression of these isoforms of adenylyl cyclase appears to differ, at

* M. Kawabata, R. Taussig, and A. G. Gilman, personal communication. '' P.A. Watson, J. Krupinski, A. M. Kempinski, and C. D. Frankenfield, unpublished observations.

-4.4 - 2.4 - 1.35

FIG.2. Northern blot analysis of tissue distribution of mRNA for type VI1 adenylyl cyclase in mouse tissues. Distribution of type VI1 mRNA was assessed using a Mouse Multiple Tissue Northern blot (Clontech Laboratories) containing3 pg of poly(A*)-enrichedRNA from tissuesindicated.Themembranewasscreened a s described under "Experimental Procedures" using a randomlyprimed "P-labeled cDNA probe generated from the full-length typeVI1 cDNA depicted in Fig. 1. Membranes were subjected to autoradiography for 6 days. Migration of RNA size standards is shown on the left (in kilodaltons).

least at the mRNA level. PCR results demonstrate that type VI1 is expressed at the mRNA level in heart, liver, kidney, testis, skeletal muscle, and brain (1).Northern blot analysis of mouse tissues indicates high level expression of a single 7.25kilobase message in heart,lung, spleen and brain withexpression to a lesser extent inkidney, skeletal muscle, and liver (Fig. 2). A message of identical size was detected by Northern blot analysis of poly(A+)RNA isolated from S49 mouse lymphoma and XC rat sarcoma cell lines (data not shown). Northern analysisreveals the mRNA for type I1 adenylyl cyclase is also expressed in the lung, as well as in the brain and olfactory bulb (4). Comparison of results from a qualitative screen by RT-PCR (1)and the Northernblot analysis presented inFig. 2 reinforce that caution mustbe exercised when interpreting RT-PCR results with regardt o the relative abundance of either mRNA or protein in a cell or tissue. While both methods identified expression of type VI1 adenylyl cyclase at the mRNA level in heart, brain,kidney, skeletal muscle, liver, and testis; the latter two tissues which gave relatively weak signals in the Northern blot analysis for type VI1 mRNA demonstrated predominance for type VI1 sequence in the RT-PCR results (liver, 33% and testis, 46% of sequenced fragments). Experiments involving overexpression of adenylyl cyclases in HEK 293 cells were performed to ascertain if the significant homology which is observed between the type VI1 protein and the type I1 adenylyl cyclase at the structural level is reflected in qualitative similarities with regard to the regulatory interactions which the two proteins have with other signaling pathways. It should be noted that quantitative comparisons between the activities of the type I1 andtype VI1 adenylyl cyclases in transformed HEK 293 cells were not considered prudent due toseveral factors. Cellsstably expressing the type I1 adenylyl cyclase utilized in comparative experiments were originally derived for a previous series of experiments. Previous experience has shown that both the host HEK 293 cells as well a s polyclonal-transformed populations of these cells dem-

28896

o p e VII Adenylyl Cyclase

TABLE I Increases in intracellular calcium or protein kinase C activation augment CAMPaccumulation following isoproterenol stimulation in type VII-transformed HEK 293cells while attenuating isoproterenol-stimulated CAMP accumulation in control HEK 293cells Cells were plated for experiments asdescribed under “Experimental Procedures” and preincubated for 20 min in defined medium containing Ro 20-1724. Fresh medium containingisoproterenol (ISO; 1 VM), A23187 (1PM),ATP(50 pd,or PDBu (100 m)in theindicated combinations was added, and the cells were incubated for 2 min. Experiments were terminated, and CAMP was extracted and analyzed as described under “Experimental Procedures.” Results are expressed as mean 2 S.E. and represent results from a minimum n = 3. cDNA construct

Treatments Control

A23187

ATP

PDBu

CAMP (pmol lmg proteid Type VII-sense -IS0 +IS0 Type VII-antisense -IS0 +IS0

7.0k0.4 10.820.5 13.72 0.1 5 5 . 6 2 1.3 1 9 1 k2151552473 *73227 2 5 7 3.7 2 0.2

188 2 10

3.8 2 0.2 1232 7

0.4 4.6 92 2 6

3.420.1 168 2 8

onstrate passage-dependent variations in theirbiochemical behavior. For example, the host HEK 293 cell populations used for these transfection experiments demonstrate markeddifferences in the activation of endogenous adenylyl cyclases by isoproterenol. However, the data inTable I indicate that the isoproterenol-stimulated activity in these type VII-transformed HEK 293 cells appears to be accounted for by activation of endogenous adenylyl cyclases as reflected by the valuesfor the antisense-transformed control cells. Similar results were obtained in experimentson HEK 293 cells expressing the typeI1 enzyme (data notshown) relative to thecontrol cells from that transfection, although the relative accumulation of CAMP in these two cell populations following stimulation with isoproterenol is different (22-fold increase in CAMP content in host HEK 293 cells used in typeI1 transfection experiments; 51-fold increase in CAMPcontent in host HEK 293 cells used for type VI1 transformations). It was also impossible to ascertain or control for potential differences in the level of expression of these adenylyl cyclases at the protein level in these transfection experiments. It was felt that these limitations precluded useful comparisons regarding quantitative differences in the activities and sensitivities of the type I1 and type VI1 isoforms to various stimuli in this cell system. Such comparisons will require experiments isolated in membranes preparedfrom cells expressing the type I1 and type VI1 adenylyl cyclases at quantitativelyequal levels asdeterminedwith isoform-specific antibodies. Concern must be raised regarding the apparent lack of a type VII-dependent component inthe isoproterenol-stimulated CAMPaccumulation in transformed HEK 293 cells (Table I). This data would appear to preclude potentiation of isoproterenol stimulation of type VI1 by activation of protein kinase C. However, treatment of control cells with either isoproterenol, PDBu, or both alsocannot account for the increasesobserved in type VII-transformed cells (Table I). Two potential explanations may account for this apparent discrepancy. The first is that GB-ais in some manner limiting to the stimulation of adenylyl cyclase isoforms in HEK 293 cells. Thus, stimulation of the totalpool of adenylyl cyclase molecules in HEK 293 cells by isoproterenol will remain at a constant level, regardless of whether this pool is made upof endogenous adenylyl cyclases or of a combination of exogenously and endogenously expressed cyclase isoforms. Data obtained in HEK 293 cells expressing the type VI adenylyl cyclase would appear to contradict this possibility as type VI expression does increase with isoproter-

enol stimulation relative to control cells (1).It is also possible that the type VI1 adenylyl cyclase is not activated by GB-ain HEK 293 cells unless it isphosphorylated by protein kinase C. Both of these potential explanations may also involve differential localization of the exogenously expressed isoforms relative to these signal transductioncomponents. Qualitative similarities and differences in the regulation of type I1 and type VI1 adenylyl cyclases were still apparent in experiments performed utilizing these transformed polyclonal populations of HEK 293 cells. HEK 293 cells overexpressing either the typeI1 or type VI1 isoforms demonstrate no increase in CAMPaccumulation concurrent with increases in intracellular calcium following treatment with the calcium ionophore A23187 (Figs. 3a and 3b) or the purinergic-agonist ATP (11) (Fig. 3, C and D).However, cells overexpressing either thetype I1 (21, 22) (Fig.3E) or type VI1 isoforms (Fig. 3F) show significant increases in activity following treatment with the phorbol ester PDBu (phorbol 12,13-dibutyrate). Treatment with PDBu synergistically increases sensitivity to P-adrenergic stimulation in cells overexpressing either type I1 or type VI1 adenylyl cyclase (Fig. 3, E and F ) . These responses to phorbol ester treatment are also seen with adenylyl cyclase activity in S49 mouse lymphoma cells (23, 241, cells shown previously to express the type VI1 adenylyl cyclase isoform (1). Increasing intracellular calcium by treatment of transformed HEK293 cells with the calcium ionophore A23187 or treatment with the purinergic agonist ATP (11)also acted synergistically to increase beta-adrenergic stimulationof CAMPproduction (Fig. 3, A-D). Inhibition of protein kinase C activity in typeVII-transformed HEK 293 cells by staurosporine treatmentsignificantly diminished the synergistic effects of increased intracellular calcium or ATP treatment and P-adrenergic stimulation (Fig. 4) indicating that a significant proportion of this synergism is the result of protein kinase C activation. In contrast, control HEK 293 cells transfected with the pCMV-Neo plasmid containing the type VI1 cDNA in an antisense orientation demonstrated unaltered P-adrenergicresponsiveness following treatment with A23187, ATP, or PDBu (Table I). Two potential sites of action exist for the ability of protein kinase C to increase basal adenylyl cyclase activity and to potentiate isoproterenol-stimulated activity in type VII-transformed HEK 293 cells. The first involves the well documented ability of protein kinase C to phosphorylate the alpha-subunit of the inhibitory guanine nucleotide-binding regulatory protein, G,, and attenuate the “tonic inhibition” of adenylyl cyclase. This results in a net increase in the stimulatory drive on adenylyl cyclase. Such an effect has been described in experiments on membranes prepared from S49 mouse lymphoma cells which express typeVI1 adenylyl cyclase (23,24). A second possible site of action for protein kinase C with regard to its effects on type VI1 activity is direct modification of the type VI1 enzyme itself. Experiments were performed to distinguish between these two possibilities. Type VII-transformed HEK 293 cells were pretreated for 24 h with pertussis toxin (100 ng/ml) to ADP-ribosylate G,-cy and prevent the inhibition of adenylyl cyclase activity (25). If modification of the G,-a subunit by protein kinase C was responsible for the enhanced type VI1 adenylyl cyclase activity following phorbol ester treatment, pertussis toxin treatment should prevent phorbol ester potentiation of this activity. Results of experiments inHEK 293cells overexpressing type VI1 adenylyl cyclase presented in Fig. 5 indicate that pretreatment with pertussis toxin failst o prevent phorbol ester-dependent potentiation of isoproterenol stimulation. In identical experiments performed on HEK 293 cells transformed with the type VI1 cDNA cloned into pCMV-Neo in an antisense orientation, pertussis toxin had no effect on either

28897

Type VII Adenylyl Cyclase

-

+~ 2 3a7 1

+4

0,

75

ATP

ISO+AZ~I~~

n

I

40

25

0

2

4

6

8

E A 500{

I

0

4

2

Time (min.)

Time (min.)

Time (min.)

5

/

a IS0

-eISO+PDBu 0IS0 +PDEu

+ATP +IS0 +IS0

+ ATP + ATP

/

“ .

2

8

l i m e (min.)

Time (rnin.)

Time (min.)

FIG.3. Regulatory interactions between 8-adrenergic stimulation, intracellular calcium, purinergic agonists, and phorbol ester treatment in HEK 293 cells expressing typeI1 or type VI1 adenylyl cyclases.HEK 293 cells overexpressing the type I1 (Panels A, C, and E )or type VI1 (Panels B , D, and F ) adenylyl cyclase isoforms were treated with combinations of isoproterenol (1p ~ and ) the calcium ionophore A23187 (1~ M ; and A B ) ,ATP (50 p ~C;and D), or PDBu (100 n ~E ;and F ) . All experiments were performed in the presence of the phosphodiesterase inhibitor Ro 20-1724. CyclicAMP accumulation was terminated at theindicated time by addition of cold trichloroacetic acid (final concentration 6%). Samples were analyzed as described under “Experimental Procedures.” Data represent means of at least two experiments ( n = 3 in each experiment) with S.E. less than 5% of the mean.

ISense+PTox

,” 500

R H Antisense+PTox

E a \

5

-0

250

E None ATP A23187

PDBu

Additional Agents FIG.4. Inhibition of protein kinase C with staurosporine attenuates potentiation of 8-adrenergic stimulationof type VI1 adenylyl cyclase activity by A23187,ATP,or phorbol ester treatment. Experiments were performed as described in Fig. 2 with the exception that where indicated, cells were pretreated with 100 n~ staurosporine (Stauro.)for 20 min prior to agonist addition. Ro 20-1724 and isoproterenol were present in all cases a t 100 and 1 p ~ respectively. , Additions to experimental groups are indicated under the abscissa. CAMPconcentrations in control cells were 7.0 2 0.5 pmol/mg of protein when F&I 20-1724 was the only agent added, and 5.3 2 0.6 pmoUmg protein following staurosporine pretreatment. Data represent means+. S.E. of at least two experiments ( n = 3 in each experiment).

basal or agonist-stimulated activities of the endogenous HEK 293 cell adenylyl cyclases isoform(s) (Fig.5). We therefore conclude that a significant proportion of the ability of protein kinase C to enhance isoproterenol-stimulated type VI1 adenylyl cyclase activity is the resultof an effect of protein kinase C on the enzyme which is independent of actions on Gi-a. In this manner, type VI1 appears t o be similar to the type V isoform which has recently beenshown to be directly phosphorylatedby protein kinaseC isoforms alpha and zeta(26). Phosphorylation by either protein kinaseC isoform results in potentactivation of type V activity and potentiation of activation of type V by forskolin. Additionally, phosphorylation of type V adenylyl cyclase by protein kinase C-a potentiates G,-a-mediated activation of type V activity, similar t o the potentiation of isoproterenol-stimulated activity by phorbol ester treatment with the type VI1 isoform described here. While the type I1 and type VI1 isoforms of adenylyl cyclase are similar in certain aspectsof their structure and function,

Is0

PDBu

Iso+PDBu

FIG.5. Pretreatment of type VII-transformed HEK 293 cells with pertussis toxin augments isoproterenol-stimulated activity butfails to block phorbol ester potentiationof isoproterenolstimulated activity.HEK 293 cells overexpressing type V I 1 adenylyl cyclase (Sense)and HEK 293 cellstransformed with the type VI1 cDNA subcloned into pCMV-Neo in an antisense orientation (Antisense) were pretreated with 100 ng/ml pertussis toxin (PTox) for 24 h. Cells were treated with isoproterenol (1 PM), PDBu (50 nM), or both for 2 min. Controls for PDBu-treated cells had dimethyl sulfoxide added to a final concentration of 0.1%, similar to that of PDBu treated cells. The basal CAMP concentration was 5.12 f 0.5 pmoVmg of protein for control cells and was unchanged in response to pretreatment with pertussis toxin. Experiments were terminated and samples extracted and analyzed for CAMPas indicated under “Experimental Procedures.” Data represent mean & S.E. of at least three samples.

they differ in severalsignificant ways. The activity of both type I1 andtype VI1 isoforms of adenylyl cyclase is increased through activation of protein kinase C activity in response to treatment of cells with phorbol esters (Fig. 3, E and F).However, the rateof activation and subsequentinactivation of these adenylyl cyclase isoforms differs significantly in response to activation of protein kinase C in transformed HEK 293 cells. rapid activation of type VI1 Phorbol ester treatment results ain activity, peaking by 4 min. This activity, as well as the synergistic P-adrenergiclphorbol ester stimulated activity, undergo type VII-transformed HEK 293 cells (Fig. rapid inactivation in 33’). Activity of the typeI1 isoform is increased more gradually in response to protein kinase C activation, with peak activity occurring at 20 min (additional data not shown) (22). These results emphasize subtle differences in the regulatoryproperties of the adenylyl cyclases that aremanifested through crosstalk with other signal transduction pathways.

28898

n p e VII Adenylyl Cyclase REFERENCES

1. Krupinski, J., Lehman, T., Frankenfield, C., Zwaagstra, J., and Watson, P. (1992) J. Biol. Chem. 267,2485%24862 2. Krupinski, J., Coussen, E , Bakalyar, H., Tang, W., Feinstein, P., Orth, K., Slaughter, C., Reed, R., and Gilman, A. (1989) Science 244, 1558-1564 3. Bakalyar, H., and Reed, R. (1990) Science 250, 1403-1406 4. Feinstein, P., Schrader, K., Bakalyar, H., Tang, W., Krupinski, J., Gilman, A,, and Reed, R. (1991) Proc. Natl. Acad. Sei. U. S. A. 88, 10173-10177 5. Ishikawa, Y., Katsushika, S., Chen, L., Halnon, N., Kawabe, J.-I., and Homcy, C. (1992) J. Biol. Chem. 267, 13553-13557 6. Gao, B., and Gilman,A.(1991)Proc. Natl. Aead. Sci. U. S. A. 88,10178-10182 7. Premont, R., Chen,J., Ma, H.-W., Pommapalli, M., and Iyengar,R. (1992)Proc. Nut. Acad. Sci. U. S. A. 89,9809-9813 8. Yoshimura, M., and Cooper, D. (1992)Proc. Nut. Acad. Sci. U. S. A. 89,671G 6720 9. Wallach, J.,Droste, M., Kluxen, F., PfeutTer, T.,andFrank, R. (1994)FEBSLet. 338,257-263 10. Katsushika, S., Chen, L., Kawabe, J.-I., Nilakantan, R., Halnon,N., Homcy, C., and Ishikawa, Y. (1992) Proc. Natl. Acad. Sei. U. S. A. 89, 8774-8778 11. Cali, J., Zwaagstra, J., Mons, N., Cooper, D., and Krupinski, J. (1994) J. Biol. Chem. 269, 12190-12195 12. Hellevuo, K., Yoshimura, M., Kao, M., Hoffman, P., Cooper, D., and Tabakoff,

B. (1993) Biochem. Biophys. Res. Commun. 192, 311-318 13. Watson, P. (1989) J. Bid. Chem. 264, 14735-14740 14. Watson, F? (1990) J. Bid. Chem. 266, 656943575 T. (1989) Molecular Cloning: 15. Sambrook, J., Fritsch,E.,andManiatis, A Laboratory Manual, ed. 2, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 16. Gorman, C., Gies, D., and McCray, G. (1990)DNA Protein Eng. Techn. 2.3-10 17. Brooker, G., Terasaki, W., and Price, M. (1976) Science 194, 270-276 18. Lowry, O., Rosebrough, N.,Fan; A., and Randall,R. (1951)J. Biol. Chem. 193, 265-275 19. Kozak, M. (1991) J. Biol. Chem. 266, 19867-19870 20. Chinkers, M., and Garbers, D.(1989) Science 245,1392-1394 21. Jacobowitz, O., Chen, J., Premont, R., and Iyengar, R. (1993) J. Biol. Chem. 268,3829-3832 22. Yoshimura, M., and Cooper, D. (1993) J. Biol. Chem. 268, 4604-4607 23. Bell, J., and Brunton, L. (1986) J. Biol. Chem. 261, 1203G.12041 24. Bell, J., and Brunton, L.(1987) Am. J. Physiol. 262, E783-E789 25. Katada, T., Bokoch, G., Northup, J., Ui, M., and Gilman, A. (1984) J. Biol. Chem. 259,356&3577 26. Kawabe, J-I., Iwani, G., Ebina, T., Ohno, S., Katada, T., Ueda, Y., Homcy, C. J., and Ishikawa, Y. (1994)J. Biol. Chem. 269, 16554-19558 27. Eisenberg, D.,Schwartz, E., Komaromy, M., and Wall, R. (1984) J . Mol. Biol. 179, 125-142