Isoform-Specific Regulation of Adenylyl Cyclase Function by Disruption of Membrane Trafficking

    Departments of Medicine (R.D.F.) and Physiology and Pharmacology (R.D.F., S.S.G.F.), University of Western Ontario, Cell Biology Research Group, Robarts Research Institute (Q.D., R.G., J.C., S.S.G.F., R.D.F.), London, Ontario, Canada

    Abstract

    Oligomerization plays an important role in endoplasmic reticulum processing and membrane insertion (and ultimately in regulation of function) of a number of transmembrane spanning proteins. Furthermore, it is known that adenylyl cyclases (ACs), critical regulators of cellular functions, associate into higher order (dimeric) forms. However, the importance of these higher order aggregates in regulating adenylyl cyclase activity or trafficking to the cell membrane is unclear. Therefore, we examined the potential role of oligomerization in the membrane trafficking of adenylyl cyclase. For this purpose, the ability of full-length adenylyl cyclase and various truncation mutants to self-assemble and to be targeted to the cell membrane was assessed. A truncation mutant comprised of the initial six transmembrane spanning domains and half of the C1 catalytic domain coimmunoprecipitated with full-length AC VI. Using both biotinylation assays and assessment of enzyme distribution using sucrose density gradients, we demonstrate that expression of this mutant in human embryonic kidney 293 cells impaired the ability of AC VI to traffic to the plasma membrane. Furthermore, mutant expression resulted in a significant reduction in adenylyl cyclase activity. The decrease in AC VI membrane expression was not caused by alterations in enzyme transcription. The effect of the mutant was specific for the AC V and VI isoforms and expression of the transmembrane M1 domain but not the C1a domain was required for the mutant to affect adenylyl cyclase activity. In aggregate, these data suggest that alterations in the ability of adenylyl cyclases to form higher order forms regulate both enzyme trafficking and enzyme activity.

    Adenylyl cyclases comprise a large family of enzymes that catalyze the synthesis of cyclic AMP from ATP. Adenylyl cyclases constitute the "effector" component of transmembrane signaling pathways that are both positively and negatively regulated by the activity of heterotrimeric G protein and G protein-coupled receptors (GPCRs). The activity of the enzyme adenylyl cyclase is itself regulated by both extracellular stimuli (via GPCR receptor activation) and intracellular stimuli (Cooper et al., 1995; Hanoune and Defer, 2001; Patel et al., 2001). Studies of the regulation of adenylyl cyclases initially focused "upstream" of the enzyme (i.e., on characterization of regulation of enzyme function via G protein and GPCR function/expression) or on the regulation of substrate (Mg2+/ATP) availability. More recently, the importance of regulation of enzyme activity by covalent modifications (i.e., enzyme phosphorylation) has increasingly been appreciated (Hurley, 1998; Tang and Hurley, 1998; Hurley, 1999; Patel et al., 2001; Cooper, 2003). However, regulation of adenylyl cyclase function via regulation of translation and trafficking to the plasma membrane has been largely unexplored. Furthermore, the determinants of adenylyl cyclase trafficking to the cell membrane are unknown.

    Structurally, the adenylyl cyclase enzyme consists of two hydrophobic domains (M1 and M2, containing six transmembrane spans each) and two large cytoplasmic domains (C1 and C2). The C1 and C2 regions are further subdivided in to "a" and "b" regions. The C1a and C2a regions interact to form the catalytic domain of adenylyl cyclase (Tang and Hurley, 1998; Ludwig and Seuwen, 2002). The transmembrane spanning domains have been suggested to be important in the regulation of protein assembly and membrane trafficking of the enzyme (Gu et al., 2001).

    It is known that for a number of other transmembrane spanning proteins, including GPCRs, that oligomerization plays an important role in endoplasmic reticulum processing and ultimately plasma membrane insertion (Balasubramanian et al., 2004; Kaykas et al., 2004; Salahpour et al., 2004). However, whether regulation of the formation of higher order aggregates of adenylyl cyclase might regulate plasma membrane insertion (and ultimately enzyme activity) is unclear.

    Previous studies have identified that some adenylyl cyclase isoforms can be identified as higher order aggregates (Haga et al., 1977; Schlegel et al., 1979; Nielsen et al., 1981; Neer et al., 1984; Yeager et al., 1985; Smigel, 1986). Furthermore, disruption of adenylyl cyclase dimerization has been reported to lead to impairment of enzyme activity. In particular, expression of an inactive truncated form of adenylyl cyclase type VI has been demonstrated to reduce adenylyl cyclase type VIII function (Gu et al., 2002). However, the mechanism underlying this effect and the specific domains that determine this interaction are unclear. Furthermore, the isoform specificity of this effect is unknown.

    Therefore, we tested the hypothesis that membrane expression of adenylyl cyclases could be regulated by disruption of membrane trafficking of the enzyme by using truncation mutants of the enzyme. Data to be presented demonstrate that expression of truncation mutations of AC VI, including the transmembrane M1 (TM1) region, can associate with wild-type adenylyl cyclase and inhibit adenylyl cyclase function on an isoform-specific basis, primarily by interfering with the incorporation of the enzyme in the plasma membrane. These findings support a critical role of association of oligomers of adenylyl cyclase in cytoplasmic compartments in the trafficking of the enzyme and ultimately in the isoform-specific regulation of enzyme function.

    Materials and Methods

    Cell Culture. Human embryonic kidney (HEK) 293 cells (CRL-1573; American Type Culture Collection, Manassas, VA) were grown in minimal essential medium (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (Hyclone Laboratories, Logan, UT) and 50 e/ml gentamicin (Invitrogen). Cells were maintained at 37°C in a humidified 5% CO2 incubator.

    Construction of Truncation Mutants. We initially generated four adenylyl cyclase type VI truncation mutations, NT (comprising the cytoplasmic N-terminal domain), C1aT (comprising the C1a domain of the enzyme), TM1T (comprising the N-terminal domain and first six transmembrane spanning domains), and 547T (comprising the first 547 amino acids of the enzyme, including the initial transmembrane spanning domain and the C1a domain) (Fig. 1A), to examine the effects of coexpression on wild-type (full-length) adenylyl cyclase VI activity. pIND-AC VI truncation mutants (C1aT, NT, TM1T, and 547T) were constructed by PCR using pIND-FLAG AC VI as template and primers as indicated in Table 1 and Fig. 1A. PCR fragments were digested with HindIII/XbaI and then cloned into pIND vector, which had been digested with identical enzymes. Positive clones were sequenced to verify the right open reading frame and right mutation. All truncation mutants were FLAG-tagged and were all detected in whole cell lysates from transfected cells (Fig. 1B).

    In addition, we subsequently generated a TM2T truncation mutant of AC VI and a TM1 truncation mutant of AC I. The TM2T truncation mutant of AC VI comprised the C1b, the second transmembrane spanning domain and C2 regions of AC VI. FLAG-tagged TM2T was generated by PCR using FLAG-AC VI as template (Table 1). We also generated an AC I truncation mutant, 462T analogous to the 547T mutant of AC VI. The FLAG-tagged AC I 462T was generated by PCR using pcDNA-FLAG-AC I as template (Table 1). PCR fragment was digested with BamI/XbaI and cloned into pIND vector digested with identical enzymes.

    Construction of Myc-Tagged Adenylyl Cyclase VI. Myc-AC VI was generated by PCR using FLAG-AC VI as a template. Using a forward primer containing myc tag (5'-CCC AAG CTT (HindIII) GCC GCC ATG GAG CAG AAA CTC ATC TCT GAA GAG GAT CTG CCC CTG-3') and reverse primer GC TCT AGA CTA ACT GCT GGG GCC CCC ATT GAG-3', PCR fragment was digested with HindII/XbaI and cloned into pIND vector digested with identical enzymes.

    Transfections. Transfections of HEK293 cells were carried out using a modification of the calcium phosphate method as described previously (Cullen, 1987). Cells were transiently cotransfected with 3 e of (pIND) FLAG-AC VI and 3 e of either (pIND) GFP, (pIND) myc-AC VI, or adenylyl cyclase mutants with 6 e of pVgRXR. For isoform experiments, cotransfection of 3 e of (pIND) FLAG-AC isoforms (AC I, AC II, AC V, or AC VI), with 3 e of either (pIND) GFP or truncation mutants plus 6 e of pVgRXR, was used. After 24 h of transfection, expression of inserts was induced by addition of ponasterone A to final concentration to 5 eol/l at 37°C for an additional 24 h.

    Assessment of Adenylyl Cyclase Activity. Adenylyl cyclase activity in response to forskolin (100 e) or forskolin plus vanadate (300 e) was assessed by the conversion of [-32P]ATP to [32P]cAMP as reported previously (Tan et al., 2001). In brief, permeabilized cells were resuspended in Hanks' balanced salt solution with 33 mM HEPES, 0.5 mM EDTA, and 1 mM magnesium sulfate (pH 7.4 at 4°C) added in an aliquot of 40 e to give a final incubation volume of 100 e with 1 e藽i of [-32P]ATP, 0.3 mM ATP, 2 mM MgSO4, 0.1 mM cAMP (used in lieu of a phosphodiesterase inhibitor), 5 mM phosphoenol pyruvate, 40 e/ml pyruvate kinase, and 20 e/ml myokinase. Incubations were carried out at 37°C for 10 min and terminated by addition of 1 ml of a solution containing 100 e of ATP, 50 e of cAMP, and 15,000 cpm of [3H]cAMP. Cells were pelleted by centrifugation at 300g for 5 min. cAMP was isolated from the supernatant by sequential Dowex and alumina chromatography and was corrected for recovery with [3H]cAMP as the internal standard. Adenylyl cyclase activity was linear with time and cell number over the ranges used. Adenylyl cyclase activity was expressed as percentage of control-transfected cells. In experiments examining the effects of various truncation mutants on AC isoforms (ACx), activity was expressed as a percentage of ACx alone-transfected cells [i.e., AC VI plus 547T expressed as a percentage of AC VI alone-transfected cells; (AC VI + 547T)/AC VI)]. In experiments where truncation mutants were transfected alone, the activity was expressed as a percentage of activity in GFP-transfected cells.

    Immunoprecipitation and Western Blotting. Transfected cells were lysed in buffer A (20 mM Tris, pH 8.0, 1% Nonidet P-40, 0.1% SDS, 140 mM NaCl, 1 mM phenylmethylsulfonyl fluoride). For coimmunoprecipitation experiments, 500 e of cell lysates was incubated with 5 e of an anti-adenylyl cyclase "AC comm" polyclonal antibody that demonstrates cross-reactivity with many adenylyl cyclase isoforms (the epitope was generated against a 14-amino acid peptide of the c-terminal region of adenylyl cyclase, found to be similar in several cloned adenylyl cyclase isoforms; Pieroni et al., 1995) for 1 h at 4°C followed by the addition of 25 e of A/G agarose beads (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and further incubated overnight at 4°C on a rocker platform. The resulting immunoprecipitation complexes were resolved on SDS-PAGE and blotted as described below.

    Sixty micrograms of whole cell extracts were resolved on SDS-PAGE and blotted electrophoretically onto Immun-Blot polyvinylidene difluoride membrane (Bio-Rad, Hercules CA). Membranes were blocked with 5% skim milk and incubated with anti-FLAG (M2) antibody (1:1000; Sigma-Aldrich, St. Louis, MO) overnight at 4°C on a rocker platform. Blots were washed in Tris-buffered saline for 1 h followed by incubation of secondary anti-mouse antibody (1:1000; Sigma-Aldrich) for 1 h at room temperature. Proteins were detected by chemiluminescence as described by the manufacturer's protocol (PerkinElmer Life and Analytical Sciences, Boston, MA).

    Cell Surface Biotinylation. Transfected cells were induced for 24 h with 5 eol/l ponasterone A. Cells were washed with cold PBS three times to remove contaminating proteins and then biotinylated in the presence of 0.5 mg/ml sulfosuccinimidyl-2-(biotinamido) ethyl-1,3-dithiopropionate (EZ-link sulfo-NHS-SS-biotin; Pierce Chemical, Rockford, IL) for 30 min at 37°C. Cells were washed with PBS containing 0.1% bovine serum albumin and then solubilized in lysis buffer. Biotinylated proteins were precipitated by streptavidin-agarose (Sigma-Aldrich) and separated on SDS-PAGE, transferred onto polyvinylidene difluoride membrane, and immunoblotted with anti-FLAG antibody as described above.

    RT-PCR Assessment of AC VI mRNA. Total RNA was extracted from HEK293 cells transfected with AC VI alone or with truncation mutant 547T by TRIzol reagent (Invitrogen Canada Inc., Burlington, ON, Canada). cDNAs were generated by reverse transcription of RNA using SuperScript kit (Invitrogen Canada Inc.) and used as template for PCR reactions with the following oligonucleotide primer pairs: 5'-CCTGGCCAAGGTCATCCATGACAACT-3' and 5'-TGTCATACCAGGAAATGAGCTTG-3' for the internal control GAPDH and 5'-GCGCAACGCGTACCTCAAGGA-3' and 5'-TCTTCTCGAGATCCTCCCTCTGGAAGG-3' for AC VI. Standard PCR was performed with TaqDNA polymerase for 30 cycles. The resulting PCR products were resolved by agarose electrophoresis and visualized by ethidium bromide staining.

    RNA Protection Assays. Total RNA was isolated 24 h after ponasterone A induction from HEK293 cells transfected with AC VI alone or with truncation mutant 547T. RNA protection assay was performed using an Ambion RPAII kit with hybridization at 42°C overnight according to the manufacturer's procedure; 2 x 104 cpm of AC VI (1651eC2010) or GADPH riboprobe was added to 10 e of HEK293 RNA, and the RNA-RNA hybrids that resisted RNase treatment were precipitated, dissolved in loading buffer, and denatured at 95°C for 5 min. Components then were resolved on a 4% polyacrylamide-8 M urea denaturing gel in Tris borate-EDTA buffer. After drying, gels were exposed to X-ray films at eC70°C with intensifying screens.

    Subcellular Fractionation. HEK293 cells expressing AC VI alone or coexpressing AC VI plus the truncation mutation 547T were detached in ice-cold PBS containing 2 mM EDTA. Cells were pelleted and resuspended in homogenization buffer (10 mM HEPES, pH 7.4, 1 mM EDTA, and 0.25 M sucrose and supplemented with protease inhibitor mixture of 1 mM phenylmethylsulfonyl fluoride, with 1 e/ml each of antipain, leupeptin, and pepstatin A. The cells were disrupted using 10 strokes in a Dounce homogenizer followed by four passages through a 25-gauge needle. Nuclei and unbroken cells were pelleted by centrifugation at 3000g for 10 min. The pellet was resuspended in 1 ml of homogenization buffer and centrifuged at 3000g for 10 min. Postnuclear supernatants from two centrifugation steps were combined and centrifuged at 80,000g for 1 h. The vesicle pellet was resuspended in 1 ml of homogenization buffer and loaded on top of 1eC20% iodixanol continuous linear gradients and centrifuged in Beckman SW 41 rotor at 200,000g for 3 h at 4°C. Sequential 0.85-ml fractions were collected from the bottom of the gradient. The subcellular fractions were lysed by the addition of sample buffer, resolved by 8% SDS-PAGE, and immunoblotted with anti-FLAG or a plasma membrane marker antibody (namely, anti-Na+/K+-ATPase; Santa Cruz Biotechnology, Inc.).

    Results

    To test the hypothesis that membrane expression of adenylyl cyclases could be regulated by disruption of membrane trafficking of the enzyme, we sought to initially determine whether expression of the truncation mutants of the enzyme-regulated 1) enzyme activity and 2) membrane insertion of functional enzymes. To examine the functional effects of adenylyl cyclase truncation mutants on adenylyl cyclase activity, we transfected HEK293 cells with either ecdysone-inducible plasmids for GFP, AC VI alone, or AC VI cotransfected with one of the four adenylyl cyclase VI truncation mutants (Fig. 1A). All truncation mutants were detectable by immunoblots in whole cell lysates from transfected cells (Fig. 1B). Expression of wild-type AC VI alone significantly increased forskolin-stimulated adenylyl cyclase activities compared with GFP-transfected cells (163 ± 9% of activity in GFP-transfected cells, n = 13, p < 0.05). Coexpression of either 547T or TM1T truncation mutants with AC VI significantly inhibited forskolin-stimulated adenylyl cyclase activity compared with AC VI alone-transfected cells (Fig. 2). In addition, expression of the 547T mutant alone significantly inhibited endogenous adenylyl cyclase activity by 16 ± 3% (n = 14, p < 0.05) compared with GFP-transfected cells. In contrast, co-expression of the C1aT or NT truncation mutants with wild-type AC VI did not significantly alter forskolin-stimulated adenylyl cyclase activation (Fig. 2).

    To determine whether the effect of these mutants to alter adenylyl cyclase activity might be via disrupting membrane insertion/trafficking, we performed cell surface biotinylation studies. When either expressed individually or together with full-length adenylyl cyclase VI, biotinylated forms of the TM1T and 547T truncation mutants as well as wild-type adenylyl cyclase could be identified (Fig. 3A). However, coexpression of either the 547T or TM1T mutants with wild-type AC VI resulted in a significant reduction in the biotinylated monomeric form of adenylyl cyclase VI (Fig. 3, A and B). In contrast, biotinylated forms of the C1aT or NT were not detectable (data not shown) and did not alter the detection of biotinylated monomeric form of adenylyl cyclase VI (Fig. 3, A and B). Thus, the ability of the TM1T and 547T mutants to inhibit adenylyl cyclase function (but not the C1aT or NT mutants) paralleled their ability to inhibit plasma membrane expression of full-length adenylyl cyclase VI. It is noteworthy that total cellular adenylyl cyclase VI expression in cells (as detected by expression of a 126-kDa protein that migrated identically with recombinant AC VI) was not altered with coexpression of the TM1T mutant (nor with the coexpression of either C1aT and NT mutants; Fig. 4, A and B). However, total cellular AC VI expression was significantly reduced with coexpression of the 547T truncation mutant (32 ± 5% of AC VI alone-transfected cells, n = 6, p < 0.05; Fig. 4, A and B).

    To confirm that 547T expression decreased the proportion of plasma membrane/total cellular AC VI (given that both total cellular and cell surface AC VI expression was decreased), we quantitated the proportion of membrane-associated adenylyl cyclase VI with and without 547T coexpression using sucrose density centrifugation techniques (as determined by that proportion of the expressed adenylyl cyclase that cofractionated with the plasma membrane marker Na+/K+-ATPase). These studies demonstrated that coexpression of the 547T mutant significantly reduced the plasma membrane-associated expression of AC VI (73 ± 3% of AC VI alone, n = 3, p < 0.05).

    To determine whether the apparent reduction in protein expression of wild-type AC VI observed with coexpression of 547T was caused by decreased transcription of wild-type adenylyl cyclase, we performed both RT-PCR and RNase protection assays. RT-PCR revealed comparable detectable fragments of AC VI in RNA isolated from either cells transfected with both AC VI and the 547T truncation mutant or cells transfected with AC VI alone (AC VI + 547T, 107 ± 2% of AC VI alone, n = 3, p = N.S.; Fig. 5A). Likewise, RNase protection assays revealed fragments of adenylyl cyclase VI that were not significantly different when coexpressed with the 547T mutant compared with AC VI alone-expressing cells (AC VI + 547T, 128 ± 18% of AC VI alone, n = 3, p = N.S.; Fig. 5B).

    In aggregate, the reduction in total AC VI protein expression without evidence of alterations in AC VI transcription suggests an enhancement of AC VI degradation with 547T coexpression (beyond its effect to regulate enzyme function by impairing membrane insertion of full-length adenylyl cyclase).

    Previous studies have suggested that the ability of adenylyl cyclase isoforms to form higher order aggregates may be important in regulation of enzyme function (Gu et al., 2001, 2002). Based on these observations, we next sought to more directly determine whether these truncation mutants might interfere with the homo-association of adenylyl cyclase molecules. Therefore, we first attempted to confirm that full-length AC VI molecules could directly associate into higher order aggregates. To do this, cells were transfected with either myc-tagged, FLAG-tagged, or both myc- and FLAG-tagged full-length adenylyl cyclase VI. In coimmunoprecipitation studies, an anti-myc-immunoreactive 126-kDa band was observed after immunoprecipitation with anti-FLAG antibody in cells transfected with both myc- and FLAG-tagged full-length adenylyl cyclase VI (Fig. 6A), but not in cells transfected with either FLAG- or myc-tagged AC VI alone (Fig. 6A). Likewise, immunoprecipitation with anti-myc antibody followed by immunoblotting with anti-FLAG antibody resulted in the identification of a single FLAG-immunoreactive band of 126 kDa (Fig. 6B). These findings are consistent with those in previous reports (Gu et al., 2001, 2002), suggesting that adenylyl cyclase VI molecules can form stable higher order complexes.

    Based on our demonstration that full-length adenylyl cyclase VI molecules can form higher order aggregates, we next determined whether a direct interaction between these truncation mutants and full-length adenylyl cyclase VI might occur in cotransfected cells. We immunoprecipitated wild-type AC VI from whole cell lysates by using an anti-adenylyl cyclase antibody [AC comm, whose antibody epitope was based on the highly homologous carboxyl-terminal domain of adenylyl cyclases (i.e., a domain not common to any of the four truncation mutants)]. When coexpressed with full-length AC VI (Fig. 7), both the TM1T and 547T truncation mutants could be immunoprecipitated by this carboxyl-terminal-specific adenylyl cyclase antibody, consistent with a direct interaction between wild-type AC VI and these truncation mutations. In contrast, the NT and C1aT truncation mutants did not coimmunoprecipitate with full-length AC VI (Fig. 7).

    In aggregate, these studies indicated that 1) adenylyl cyclase VI could form higher order aggregates, and 2) the ability to form heteromeric aggregates between full-length adenylyl cyclase and deletion mutations containing the TM1 domain of the enzyme paralleled their ability to impair plasma membrane insertion of adenylyl cyclase VI and enzyme function.

    To determine whether the effect of the 547T mutant, based on the sequence of adenylyl cyclase VI, was specific for AC VI or was generalized to other isoforms, we examined the effect of coexpressing the 547T-mutant with other adenylyl cyclase isoforms representative of other families of the enzyme (i.e., AC I and AC II as well as AC V; the other member of the ACVI family). Expression of AC I, AC II, and AC V alone significantly increased forskolin-stimulated adenylyl cyclase activity compared with GFP-transfected cells (AC I, 267 ± 37% of GFP; AC II, 267 ± 35% of GFP; and AC V, 294 ± 22% of GFP, n = 9eC13, p < 0.05 for all isoforms) and comparably with that seen with ACVI (163 ± 9% of GFP). The 547T mutant coimmunoprecipitated with all of the isoforms of adenylyl cyclase examined (Fig. 8A). Coexpression of the 547T mutant significantly reduced the extent of adenylyl cyclase activity seen with expression of AC VI and (to a lesser extent) AC V, but it did not alter the extent of adenylyl cyclase activity observed with expression of either AC I or AC II (Fig. 8B). Likewise, biotinylation experiments revealed that AC I cell surface expression was not significantly affected by the coexpression of 547T truncation mutant (98 ± 30% of biotinylation observed in AC I alone-expressing cells, n = 3). Thus, formation of these heteromeric aggregates might occur with all adenylyl cyclase isoforms but is not sufficient to disrupt membrane protein expression and consequently function of adenylyl cyclases.

    Previous studies have demonstrated strong homomeric interactions between the second transmembrane clusters (TM2) of adenylyl cyclase (Gu et al., 2001), suggestive of higher order structures of adenylyl cyclase. Therefore, to determine the effect of TM2 on AC expression, we constructed a truncation mutation of adenylyl cyclase VI incorporating the M2 and C2 domains (the COOH-terminal half) of the enzyme. Coexpression of TM2T with either AC VI or with the other adenylyl cyclase isoforms did not significantly reduce adenylyl cyclase activity for any of the isoforms examined (AC I, 81 ± 6% of control; AC II, 86 ± 4% of control; AC V, 90 ± 2% of control; and AC VI, 82 ± 13% of control, n = 4, p > 0.10 for all isoforms).

    To examine whether the inhibition of adenylyl cyclase VI by a truncation mutant comprising the N-terminal half of adenylyl cyclase VI (the 547T mutant) was AC VI-specific (or would be generalized for truncation mutations constructed from other isoforms of adenylyl cyclase), we generated a similar N-terminal truncation mutant for adenylyl cyclase isoform I (AC I 462T). In contrast to the effect of the AC VI N-terminal mutant (547T), coexpression of the adenylyl cyclase I N-terminal truncation mutant (mutant 462T) did not significantly alter adenylyl cyclase activity for any of the isoforms examined (AC I, 96 ± 5% of control; AC II, 93 ± 6% of control; AC V, 100 ± 3 of control; and AC VI, 85 ± 5% of control, n = 4, p > 0.10 for all isoforms).

    Discussion

    Previous studies have suggested that the ability of adenylyl cyclase isoforms to form higher order aggregates were important in regulation of enzyme function (Gu et al., 2001, 2002). The present studies demonstrate that disruption of the ability of the enzyme to form effective higher level aggregates is associated with impairment of the membrane expression of the enzyme and ultimately activity. We demonstrate that expression of a truncation mutation of AC VI impairs adenylyl cyclase function in both native cells and wild-type AC VI expression models. This adenylyl cyclase VI truncation mutation coimmunoprecipitated with wild-type (full-length) AC VI and decreased the plasma membrane expression of the enzyme (as detected in biotinylation experiments). This effect was specific for truncation mutations containing the M1 domain of the enzyme (i.e., 547T and TM1T) and was not replicated by smaller truncated forms of the enzyme that did not contain the M1 region or by expression of a truncation mutant comprising the carboxyl-terminal half of the enzyme. The effect of this mutant was adenylyl cyclase isoform-selective, i.e., the mutant did not affect the function or expression of representative isoforms of other adenylyl cyclase subfamilies. Additionally expression of an analogous truncation mutation of AC I did not replicate the effect of the truncation mutation of AC VI.

    Inhibition of adenylyl cyclase activity has been previously been related to alterations in concentrations of ionic determinants of activity (i.e., either decreasing magnesium concentrations or increasing calcium concentrations), activation of inhibitory G proteins, or by covalent modification of the receptor (e.g., by protein kinase A-mediated serine phosphorylation of the receptor) (Hurley, 1998; Tang and Hurley, 1998; Hurley, 1999; Patel et al., 2001; Cooper, 2003). However, the effect on the enzyme's activity of disrupting its functional assembly, and the mechanism of this effect, has been largely unexplored. It is notable that the importance of the cytoplasmic domains in regulating adenylyl cyclase activity has been extensively studied. Using small synthetic peptides based on the sequence of the C1a region of adenylyl cyclase V sequence data, Kawabe et al. (1994) demonstrated inhibitory effects on adenylyl cyclase II and V. However, whether these effects were caused by direct inhibition of enzyme activity or disruption of enzyme trafficking was not addressed.

    The importance of the TM1 domains in impairing enzyme function by disrupting enzyme trafficking has been largely unexplored. Largely based on the study of adenylyl cyclase VIII, interactions between the transmembrane spanning domains have been suggested to be a critical determinant in the regulation of the enzyme. Gu and coworkers demonstrated that with adenylyl cyclase VIII expression, enzyme activity could be suppressed by coexpression of either an inactive adenylyl cyclase type VIII deletion mutant (deletion of amino acids 582eC594) or an inactive adenylyl cyclase VI deletion mutant (deletion of amino acids 553eC666). It is notable that these suppressive effects of the adenylyl cyclase type VIII deletion mutants were also apparent with expression of adenylyl cyclase isoforms V and VI (Gu et al., 2002). In contrast, the current study indicates that the TM1 domain is the critical determinant for efficient AC VI trafficking and that this effect is specific in the context of 1) the adenylyl cyclase domain, 2) the adenylyl cyclase isoform from which the truncation protein was constructed, and 3) the adenylyl cyclase isoform(s) that are affected by expression of this mutant protein. The reason for the apparent variance of the present findings with those of Gu et al. (2002) is unclear. However, it is notable that in their studies, the apparent expression of adenylyl cyclase (as judged by the proportional increase in activity over sham-transfected cells) was much greater than in the current studies (500% versus 170%). Thus, whether the lack of isoform specificity of their constructs reflected the difference in expression levels remains to be determined.

    To date, nine membrane-bound isoforms of adenylyl cyclase have been cloned and characterized in mammalian tissues and/or cells (Defer et al., 2000). The overall structure is similar between all membrane-bound isoforms. However, despite the similarities in structure and function of the adenylyl cyclase isoforms, it has been increasingly appreciated that their regulation and tissue-specific expression varies considerably (Defer et al., 2000; Hanoune and Defer, 2001; Cooper, 2003). Previously differences between isoforms have been demonstrated in the context of regulation by G proteins, ionic concentrations, and via differences in the patterns of covalent modification of the enzyme (primarily via phosphorylation) (Hurley, 1998, 1999; Tang and Hurley, 1998; Patel et al., 2001; Cooper, 2003). The current studies identify another potentially important isoform-specific mechanism of regulation. The effect of the 547T-truncation mutant of AC VI did have effects on AC V function. However, this effect was not seen for representative isoforms from the two other major subfamilies. There is very strong sequence homology between AC V and AC VI (>80%; Hanoune et al., 1997); thus, a common effect would not have been unexpected. However, somewhat surprisingly, the other AC isoforms studied coimmunoprecipitated with the mutant, although the mutant did not alter their functions. Thus, heteromeric association (as detected by this approach) might occur, but it is clearly not sufficient to account for the disruption of function. Whether this reflects quantitative differences in affinity for heteromeric association or the importance of additional interaction sites not common to these other isoforms remains to be determined.

    Previous studies have focused on disrupting enzyme activity by impairing activity of the membrane-bound protein. The current studies suggest a novel approach for regulating function by disrupting enzyme association in cytoplasmic compartment, probably by disrupting the formation of higher order aggregates of adenylyl cyclase. Intramolecular dimerization of adenylyl cyclase between the two transmembrane spanning domains has been suggested to be important for the association of C1 and C2 domains necessary for catalytic activity and G protein regulation of the enzyme. Older literature had suggested the possibility of adenylyl cyclase dimerization. In particular, earlier reports had indicated that detergent-solubilized adenylyl cyclase preparations displayed molecular masses more than 200 kDa, suggestive of higher order complexes or dimerization of adenylyl cyclase (Haga et al., 1977; Schlegel et al., 1979; Nielsen et al., 1981; Neer et al., 1984; Yeager et al., 1985; Smigel, 1986). More recently, the ability of adenylyl cyclase isoforms to dimerize has been demonstrated by coimmunoprecipitation and fluorescence resonance energy transfer analysis (Gu et al., 2001, 2002). However, the importance of this association for either enzyme activity or for trafficking was unclear. Based on the current studies, we would suggest that formation of homologous dimers is critical in regulation of enzyme trafficking. Furthermore, this might be a mechanism for ensuring the fidelity of expression of these membrane proteins, if the hetero-oligomerization of "defectively" translated adenylyl cyclases failed to be processed appropriately for trafficking to and insertion in the cell membrane.

    It is important to note that our data do not rule out a role for more complex aggregates in either the membrane processing of adenylyl cyclases or ultimately in regulating enzyme function. Adenylyl cyclase isoforms have been shown to be a component of lipid rafts (Ostrom et al., 2004). Thus, whether our findings are best accounted for by disruption of a more "simple" higher order protein aggregate or of a more complex lipid raft structure has not been elucidated.

    It is notable that in GPCR signal transduction systems, the concept that protein dimerization in the endoplasmic reticulum might be an important determinant in regulating their trafficking and ultimately their expression and function in the plasma membrane is not novel. A similar mechanism has been implicated as a means of regulating fidelity of expression of G protein-coupled receptors such as the 2-adrenoceptor (Salahpour et al., 2004), the Frizzled receptor (Kaykas et al., 2004), and the GABAb receptor (Balasubramanian et al., 2004). Furthermore, an expanding list of plasma membrane proteins, including the insulin receptor (Bass et al., 1998), transforming growth factor  receptors (Gilboa et al., 1998), -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid calcium channels (Greger et al., 2002), and shaker potassium channels (Papazian, 1999), have been shown to oligomerize in endoplasmic reticulum compartments. For these proteins, oligomerization has been suggested to be an important step in cell surface trafficking, perhaps relating to enhancing their association with molecular chaperones critical in protein processing (Yuan et al., 2003). Our findings suggest that adenylyl cyclase membrane insertion may be regulated similarly.

    Whether regulation of adenylyl cyclase trafficking is an important determinant of adenylyl cyclase function either physiologically (or pathologically) is unknown. In addition, what role truncated adenylyl cyclase molecules may play in regulating adenylyl cyclase trafficking and hence adenylyl cyclase function is unclear. A truncated form of adenylyl cyclase VI, consisting of the first half of the molecule, was previously cloned from a canine cardiac cDNA library (Katsushika et al., 1993). However, what role this truncated adenylyl cyclase VI mutant plays in vivo is unknown. In addition, a novel isoform of adenylyl cyclase (DAC78C) was recently identified in Drosophila melanogaster (Cann and Levin, 2000). This adenylyl cyclase was similar to mammalian transmembrane adenylyl cyclase isoforms. It is interesting that this adenylyl cyclase was expressed as two structurally distinct proteins (one full-length and one truncated) in a temporal and tissue/cellular-specific pattern (Cann and Levin, 2000). What role this truncated adenylyl cyclase plays in the embryogenesis of D. melanogaster is unknown. Thus, whether similar truncated adenylyl cyclases occur and/or play a role in mammalian cellular regulation of adenylyl cyclase is unknown but could represent an unappreciated mechanism for regulation of adenylyl cyclase action.

    These studies were supported by grants-in-aid from the Heart and Stroke Foundation of Ontario (to R.D.F.) and Canadian Institutes of Health Research (to R.D.F.). R.D.F. and S.S.G.F. are Career Investigators of the Heart and Stroke Foundation of Ontario.

    Q.D. and R.G. contributed equally to this work.

    doi:10.1124/mol.104.006817.

    References

    Balasubramanian S, Teissere JA, Raju DV, and Hall RA (2004) Heterooligomerization between GABAA and GABAB receptors regulates GABAB receptor trafficking. J Biol Chem 279: 18840eC18850.

    Bass J, Chiu G, Argon Y, and Steiner DF (1998) Folding of insulin receptor monomers is facilitated by the molecular chaperones calnexin and calreticulin and impaired by rapid dimerization. J Cell Biol 141: 637eC646.

    Cann MJ and Levin LR (2000) Restricted expression of a truncated adenylyl cyclase in the cephalic furrow of Drosophila melanogaster. Dev Genes Evol 210: 34eC40.

    Cooper DM (2003) Regulation and organization of adenylyl cyclases and cAMP. Biochem J 375: 517eC529.

    Cooper DM, Mons N, and Karpen JW (1995) Adenylyl cyclases and the interaction between calcium and cAMP signalling. Nature (Lond) 374: 421eC424.

    Cullen BR (1987) Use of eukaryotic expression technology in the functional analysis of cloned genes. Methods Enzymol 152: 684eC704.

    Defer N, Best-Belpomme M, and Hanoune J (2000) Tissue specificity and physiological relevance of various isoforms of adenylyl cyclase. Am J Physiol 279: F400eCF416.

    Gilboa L, Wells RG, Lodish HF, and Henis YI (1998) Oligomeric structure of type I and type II transforming growth factor beta receptors: homodimers form in the ER and persist at the plasma membrane. J Cell Biol 140: 767eC777.

    Greger IH, Khatri L, and Ziff EB (2002) RNA editing at arg607 controls AMPA receptor exit from the endoplasmic reticulum. Neuron 34: 759eC572.

    Gu C, Cali JJ, and Cooper DM (2002) Dimerization of mammalian adenylate cyclases. Eur J Biochem 269: 413eC421.

    Gu C, Sorkin A, and Cooper DM (2001) Persistent interactions between the two transmembrane clusters dictate the targeting and functional assembly of adenylyl cyclase. Curr Biol 11: 185eC190.

    Haga T, Haga K, and Gilman AG (1977) Hydrodynamic properties of the -adrenergic receptor and adenylate cyclase from wild type and variant S49 lymphoma cells. J Biol Chem 252: 5776eC5782.

    Hanoune J and Defer N (2001) Regulation and role of adenylyl cyclase isoforms. Annu Rev Pharmacol Toxicol 41: 145eC174.

    Hanoune J, Pouille Y, Tzavara E, Shen T, Lipskaya L, Miyamoto N, Suzuki Y, and Defer N (1997) Adenylyl cyclases: structure, regulation and function in an enzyme superfamily. Mol Cell Endocrinol 128: 179eC194.

    Hurley JH (1998) The adenylyl and guanylyl cyclase superfamily. Curr Opin Struct Biol 8: 770eC777.

    Hurley JH (1999) Structure, mechanism and regulation of mammalian adenylyl cyclase. J Biol Chem 274: 7599eC7602.

    Katsushika S, Kawabe J, Homcy CJ, and Ishikawa Y (1993) In vivo generation of an adenylylcyclase isoform with a half-molecule motif. J Biol Chem 268: 2273eC2276.

    Kawabe J, Ebina T, Ismail S, Kitchen DB, Homcy CJ, and Ishikawa Y (1994) A novel peptide inhibitor of adenylyl cyclase (AC). A peptide from type V AC directly inhibits AC catalytic activity. J Biol Chem 269: 24906eC24911.

    Kaykas A, Yang-Snyder J, Heroux M, Shah KV, Bouvier M, and Moon RT (2004) Mutant Frizzled 4 associated with vitreoretinopathy traps wild-type Frizzled in the endoplasmic reticulum by oligomerization. Nat Cell Biol 6: 52eC58.

    Ludwig MG and Seuwen K (2002) Characterization of the human adenylyl cyclase gene family: cDNA, gene structure and tissue distribution of the nine isoforms. J Recept Signal Transduct Res 22: 79eC110.

    Neer EJ, Lok JM, and Wolf LG (1984) Purification and properties of the inhibitory guanine nucleotide regulatory unit of brain adenylate cyclase. J Biol Chem 259: 14222eC14229.

    Nielsen TB, Lad PM, Preston MS, Kempner E, Schlegel W, and Rodbell M (1981) Structure of the turkey erythrocyte adenylate cyclase system. Proc Natl Acad Sci USA 78: 722eC726.

    Ostrom RS, Bundey RA, and Insel PA (2004) Nitric oxide inhibition of adenylyl cyclase type 6 activity is dependent upon lipid rafts and caveolin signaling complexes. J Biol Chem 279: 19846eC19853.

    Papazian DM (1999) Potassium channels: some assembly required. Neuron 23: 7eC10.

    Patel TB, Du Z, Pierre S, Cartin L, and Scholich K (2001) Molecular biological approaches to unravel adenylyl cyclase signaling and function. Gene 269: 13eC25.

    Pieroni JP, Harry A, Chen J, Jacobowitz O, Magnusson RP, and Iyengar R (1995) Distinct characteristics of the basal activities of adenylyl cyclases 2 and 6. J Biol Chem 270: 21368eC21373.

    Salahpour A, Angers S, Mercier JF, Lagace M, Marullo S, and Bouvier M (2004) Homodimerization of the 2-adrenergic receptor as a pre-requisite for cell surface targeting. J Biol Chem 279: 33390eC33397.

    Schlegel W, Kempner ES, and Rodbell M (1979) Activation of adenylate cyclase in hepatic membranes involves interactions of the catalytic unit with multimeric complexes of regulatory proteins. J Biol Chem 254: 5168eC5176.

    Smigel MD (1986) Purification of the catalyst of adenylate cyclase. J Biol Chem 261: 1976eC1982.

    Tan CM, Kelvin DJ, Litchfield DW, Ferguson SS, and Feldman RD (2001) Tyrosine kinase-mediated serine phosphorylation of adenylyl cyclase. Biochemistry 40: 1702eC1709.

    Tang WJ and Hurley JH (1998) Catalytic mechanism and regulation of mammalian adenylyl cyclases. Mol Pharmacol 54: 231eC240.

    Yeager RE, Heideman W, Rosenberg GB, and Storm DR (1985) Purification of the calmodulin-sensitive adenylate cyclase from bovine cerebral cortex. Biochemistry 24: 3776eC3783.

    Yuan H, Michelsen K, and Schwappach B (2003) 14-3-3 dimers probe the assembly status of multimeric membrane proteins. Curr Biol 13: 638eC646.