Regulation of Ca2+-sensitive adenylyl cyclase in gonadotropin-releasing hormone neurons

In immortalized GnRH neurons, cAMP production is elevated by increased extracellular Ca2+ and the Ca2+ channel agonist, BK-8644, and is diminished by low extracellular Ca2+ and treatment with nifedipine, consistent with the expression of adenylyl cyclase type I (AC I). Potassium-induced depolarization of GT1-7 neurons causes a dose-dependent monotonic increase in [Ca2+]i and elicits a bell-shaped cAMP response. The inhibitory phase of the cAMP response is prevented by pertussis toxin (PTX), consistent with the activation of G(i)-related proteins during depolarization. Agonist activation of the endogenous GnRH receptor in GT1-7 neurons also elicits a bell-shaped change in cAMP production. The inhibitory action of high GnRH concentrations is prevented by PTX, indicating coupling of the GnRH receptors to G(i)-related proteins. The stimulation of cAMP production by activation of endogenous LH receptors is enhanced by low (nanomolar) concentrations of GnRH but is abolished by micromolar concentrations of GnRH, again in a PTX-sensitive manner. These findings indicate that GnRH neuronal cAMP production is maintained by Ca2+ entry through voltage-sensitive calcium channels, leading to activation of Ca2+-stimulated AC I. Furthermore, the Ca2+ influx-dependent activation of AC I acts in conjunction with AC-regulatory G proteins to determine basal and agonist-stimulated levels of cAMP production.     

Dexras1 inhibits adenylyl cyclase

Dexras1 is a steroid hormone-induced Ras family G protein that acts as a receptor-independent activator of signaling by Gi/o family heterotrimeric G proteins. We examined the effects of Dexras1 on the activity of adenylyl cylase, a target of inhibitory regulation by Gialpha x GTP. Constitutively active Gsalpha (Q227L) increased cAMP levels 43-fold above baseline, and Dexras1 expression inhibited cAMP levels by 61% (P < 0.01). Dexras1 mediated inhibition of adenylyl cyclase was blocked by treatment pertussis toxin or by co-expression of RGS4, but was not inhibited by with dominant-interfering (G203T or G204A) mutants of Gi alpha2. Dexras1 decreased forskolin-stimulated CREB activation (P < 0.01) and this activity was also inhibited by co-expression of RGS4. These findings indicate that Dexras1 expression leads to ligand-independent activation of both Gialpha- and G(beta)gamma-dependent arms of the Gi signaling cascade, and suggest that Dexras1 may exert physiologically relevant inhibitory effects on the cAMP-PKA-CREB.     

Evidence for multiple distinctly localized adenylyl cyclase isoforms in mammalian spermatozoa

In addition to a bicarbonate-regulated soluble adenylyl cyclase (sAC), mammalian spermatozoa, like somatic cells, appear to contain receptor/G protein-regulated AC activity that contributes to the modulation of specialized cell processes. This study provides evidence that agents, known to influence somatic membrane-associated AC (mAC) but apparently not germ cell sAC, can modulate cAMP production and functional state in mouse spermatozoa. Specifically, forskolin significantly enhanced cAMP production and capacitation, while inclusion of 2',5'-dideoxyadenosine significantly blocked these responses. Furthermore, GTPgammaS and NaF stimulated cAMP, but GDPbetaS and mastoparan had no apparent effect, consistent with recent evidence that G(s), but not G(i), contributes to AC/cAMP regulation in uncapacitated cells. In addition, intact mouse spermatozoa were screened for all known mAC isoforms by immunolocalization, using commercially available specific antibodies. The most abundant isoforms appeared to be AC2, AC3, and AC8, each with distinct distributions in the acrosomal and flagellar regions; AC1 and AC4 also appeared to be present, although less abundantly, in the midpiece and acrosomal cap regions, respectively. Intriguingly, however, Western blotting revealed that the major immunoreactive proteins in mouse sperm lysates were considerably smaller (approximately 50-60 kDa) than their somatic cell counterparts, suggesting that mature spermatozoa contain multiple mACs which may function in a shortened form. Of particular interest were AC3 and AC8, located in the same regions as, and hence possibly directly associated with, specific cell surface receptors and G proteins that are able to regulate the spermatozoon's acquisition and maintenance of fertilizing ability via changes in AC/cAMP.     

Regulation of cAMP responses by the G12/13 pathway converges on adenylyl cyclase VII

Regulation of intracellular cAMP by multiple pathways enables differential function of this ubiquitous second messenger in a context-dependent manner. Modulation of G(s)-stimulated intracellular cAMP has long been known to be modulated by the G(i) and G(q)/Ca(2+) pathways. Recently, the G(13) pathway was also shown to facilitate cAMP responses in murine macrophage cells. We report here that this synergistic regulation of cAMP synthesis by the G(s) and G(13) pathways is mediated by a specific isoform of adenylyl cyclase, AC7. Furthermore, this signaling paradigm exists in several hematopoietic lineages and can be recapitulated by exogenous expression of AC7 in HEK 293 cells. Mechanistic characterization of this synergistic interaction indicates that it occurs downstream of receptor activation and it can be mediated by the alpha subunit of either G(12) or G(13). Our results demonstrate that AC7 is a specific downstream effector of the G(12/13) pathway.     

Alterations of adenylyl cyclase and G proteins in aortocaval shunt-induced heart failure

Unlike most other experimental models of congestive heart failure, the volume overload model induced by aortocaval shunt (AVS) in rats was found to exhibit enhanced beta-adrenoceptor (beta-AR) signaling. To study whether the adenylyl cyclase (AC)-G protein system is involved in such a change, we examined cardiac AC activity and protein content as well as G(s)alpha and G(i)alpha activities, protein contents, and mRNA levels in both left (LV) and right (RV) ventricles at the failing stage (16 wk after surgery). Basal and forskolin-stimulated AC activities were significantly increased in both LV and RV from the failing hearts; this change was associated with an upregulation of type V/VI AC protein. In contrast to 5'-guanylyl imidodiphosphate and NaF, the stimulatory effect of isoproterenol on AC was increased in the failing heart. Although G(s)alpha and G(i)alpha protein contents in the failing hearts were not altered, the mRNA level for G(s)alpha was decreased by 20% and that for G(i)alpha was increased by 20%. In addition, the activity of G(s)alpha, but not G(i)alpha, as assessed by toxin-catalyzed ADP ribosylation, was significantly decreased in the failing heart. Losartan and imidapril treatments improved cardiac function and attenuated alterations in mRNA levels for G(s)alpha and G(i)alpha proteins, as well as G(s)alpha activity, without affecting changes in AC protein content or activities in heart failure due to volume overload. These data suggest that increased AC activity may contribute to the enhanced beta-AR signaling in the AVS model of heart failure, whereas alterations in gene expression for G proteins may be of an adaptive nature at this stage of heart failure.     

Impaired water reabsorption in mice deficient in the type VI adenylyl cyclase (AC6)

Adenylyl cyclase (AC) type VI (AC6) is a calcium-inhibitable enzyme which produces cAMP upon stimulation. Herein, we characterized the specific role of AC6 in the kidneys using two AC6-knockout mouse lines. Immunohistochemical staining revealed that AC6 exists in the tubular parts of the nephron and collecting duct. Activities of AC evoked by forskolin or a selective agonist of the V2 vasopressin receptor were lower in the kidneys of AC6-null mice compared to those of wildtype mice. Results of a metabolic cage assay and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) showed for the first time that AC6 plays a critical role in regulating water homeostasis.     

Nitric oxide modulates Gi-protein expression and adenylyl cyclase signaling in vascular smooth muscle cells

We have previously shown that treatment of rats with the nitric oxide (NO) synthase inhibitor N6-nitro-L-arginine methyl ester for 4 weeks resulted in the augmentation of blood pressure and enhanced levels of Gialpha proteins. The present studies were undertaken to investigate if NO can modulate the expression of Gi proteins and associated adenylyl cyclase signaling. A10 vascular smooth muscle cells (VSMC) and primary cultured cells from aorta of Sprague-Dawley rats were used for these studies. The cells were treated with S-nitroso-N-acetylpenicillamine (SNAP) or sodium nitroprusside (SNP) for 24 h and the expression of Gialpha proteins was determined by immunobloting techniques. Adenylyl cyclase activity was determined by measuring [32P]cAMP formation for [alpha-32P]ATP. Treatment of cells with SNAP (100 microM) or SNP (0.5 mM) decreased the expression of Gialpha-2 and Gialpha-3 by about 25-40% without affecting the levels of Gsalpha proteins. The decreased expression of Gialpha proteins was reflected in decreased Gi functions (receptor-independent and -dependent) as demonstrated by decreased or attenuated forskolin-stimulated adenylyl cyclase activity by GTPgammaS and inhibition of adenylyl cyclase activity by angiotensin II and C-ANP4-23, a ring-deleted analog of atrial natriuretic peptide (ANP) that specifically interacts with natriuretic peptide receptor-C (NPR-C) in SNAP-treated cells. The SNAP-induced decreased expression of Gialpha-2 and Gialpha-3 proteins was not blocked by 1H[1,2,4]oxadiazole[4,3-a]quinoxalin-1-one, an inhibitor of soluble guanylyl cyclase, or KT5823, an inhibitor of protein kinase G, but was restored toward control levels by uric acid, a scavenger of peroxynitrite and Mn(111)tetralis (benzoic acid porphyrin) MnTBAP, a peroxynitrite scavenger and a superoxide dismutase mimetic agent that inhibits the production of peroxynitrite, suggesting that NO-mediated decreased expression of Gialpha protein was cGMP-independent and may be attributed to increased levels of peroxynitrite. In addition, Gsalpha-mediated stimulation of adenylyl cyclase by GTPgammaS, isoproterenol, and forskolin was significantly augmented in SNAP-treated cells. These results indicate that NO decreased the expression of Gialpha protein and associated functions in VSMC by cGMP-independent mechanisms. From these studies, it can be suggested that NO-induced decreased levels of Gi proteins and resultant increased levels of cAMP may be an additional mechanism through which NO regulates blood pressure.     

Mammalian adenylyl cyclase family members are randomly located on different chromosomes

Molecular cloning studies have elucidated the presence of multiple isoforms of mammalian adenylyl cyclase. So far, six different isoforms (I to VI) have been fully characterized. Comparison of their structural and biochemical characteristics suggests that the mammalian adenylyl cyclase family can be classified into four sub-families: type I, type III, type II/IV, and type V/VI. We have determined the chromosomal localization of these genes. Type I gene was assigned to chromosome 7, type III to chromosome 2, types II and IV to chromosomes 5 and 14, and types V and VI to chromosomes 3 and 12. Our results indicate that the different adenylyl cyclase isoforms, even within the same subfamily, are distributed randomly in the genome, in contrast to the chromosomal organization of other components within the same signaling pathway, such as catecholamine receptors and G proteins.     

Identification of RGS2 and type V adenylyl cyclase interaction sites

The production of cAMP is controlled on many levels, notably at the level of cAMP synthesis by the enzyme adenylyl cyclase. We have recently identified a new regulator of adenylyl cyclase activity, RGS2, which decreases cAMP accumulation when overexpressed in HEK293 cells and inhibits the in vitro activity of types III, V, and VI adenylyl cyclase. In addition, RGS2 blocking antibodies lead to elevated cAMP levels in olfactory neurons. Here we examine the nature of the interaction between RGS2 and type V adenylyl cyclase. In HEK293 cells expressing type V adenylyl cyclase, RGS2 inhibited Galpha(s)-Q227L- or beta(2)-adrenergic receptor-stimulated cAMP accumulation. Deletion of the N-terminal 19 amino acids of RGS2 abolished its ability to inhibit cAMP accumulation and to bind adenylyl cyclase. Further mutational analysis indicated that neither the C terminus, RGS GAP activity, nor the RGS box domain is required for inhibition of adenylyl cyclase. Alanine scanning of the N-terminal amino acids of RGS2 identified three residues responsible for the inhibitory function of RGS2. Furthermore, we show that RGS2 interacts directly with the C(1) but not the C(2) domain of type V adenylyl cyclase and that the inhibition by RGS2 is independent of inhibition by Galpha(i). These results provide clear evidence for functional effects of RGS2 on adenylyl cyclase activity that adds a new dimension to an intricate signaling network.     

Opioid tolerance/dependence in neuroblastoma x glioma (NG108-15) hybrid cells is associated with a reduction in spontaneous stimulatory receptor activity

Chronic opioid regulation of stimulatory receptor activity was investigated in neuroblastoma x glioma (NG108-15) hybrid cells stably transfected to express the human beta(2)-adrenoceptor (beta(2)-AR). Expressed beta(2)-ARs are functionally coupled to G proteins and display ligand-independent signalling activity, as demonstrated by the ability of an inverse agonist to attenuate basal adenylyl cyclase (AC) activity. Despite the relative increase in basal AC activity due to the development of tolerance/dependence, chronic morphine treatment was found to completely abolish spontaneous beta(2)-AR activity by reducing basal receptor/G protein precoupling. A similar chronic opioid effect was observed in transiently transfected COS-7 cells. These results indicate that during the state of opioid tolerance/dependence basal levels of AC activity are no longer under the control of spontaneously active stimulatory receptors.     

Modeling of Galpha(s) and Galpha(i) regulation of human type V and VI adenylyl cyclase

We examined the kinetics of Galpha(s) and Galpha(i) regulation of human type V and type VI adenylyl cyclase (AC V and AC VI) in order to better model interactions between AC and its regulators. Activation of AC VI by Galpha(s) displayed classical Michaelis-Menten kinetics, whereas AC V activation by Galpha(s) was cooperative with a Hill coefficient of 1.4. The basal activity of human AC V, but not that of AC VI, was inhibited by Galpha(i). Both enzymes showed greater inhibition by Galpha(i) at low Galpha(s) concentrations; however, human AC V was activated by Galpha(i) at high Galpha(s) concentrations. Neither regulator had an effect on the K(m) for Mg-ATP. Mutations made within the Galpha(s) binding pocket of AC V (N1090D) and VI (F1078S) displayed 6- and 14-fold greater EC(50) values for Galpha(s) activation but had no effect on Galpha(i) inhibition of basal activity or K(m) for Mg-ATP. Galpha(s)-stimulated AC VI-F1078S was not significantly inhibited by Galpha(i), despite normal inhibition by Galpha(i) upon forskolin stimulation. Mechanistic models for Galpha(s) and Galpha(i) regulation of AC V and VI were derived to describe these results. Our models are consistent with previous studies, predicting a decrease in affinity of Galpha(i) in the presence of Galpha(s). For AC VI, Galpha(s) is required for inhibition but not binding by Galpha(i). For AC V, binding of two molecules of Galpha(s) and Galpha(i) to an AC dimer are required to fully describe the data. These models highlight the differences between AC V and VI and the complex interactions with two important regulators.     

Different roles for Gi and Go proteins in modulation of adenylyl cyclase type-2 activity

The effect of Gi/o protein-coupled receptors on adenylyl cyclase type 2 (AC2) has been studied in Sf9 insect cells. Stimulation of cells expressing AC2 with the phorbol ester 12-O-tetradecanoyl phorbol-13-acetate (TPA) led to a twofold stimulation of cAMP synthesis that could be blocked with the protein kinase C inhibitor GF109203X. Activation of a coexpressed alpha2A-adrenoceptor or muscarinic M4 receptor inhibited the stimulation by TPA almost completely in a pertussis toxin-sensitive manner. Activation of Gs proteins switched the response of the alpha2A-adrenoceptor to potentiation of prestimulated AC2 activity. The potentiation, but not the inhibition, could be blocked by a Gbetagamma scavenger. A novel methodological approach, whereby signalling through endogenous G proteins was ablated, was used to assess specific G protein species in the signal pathway. Expression of Go proteins (alphao1 + beta1gamma2) restored both the inhibition and the potentiation, whereas expression of Gi proteins (alphai1 + beta1gamma2) resulted in a potentiation of both the TPA- and the Gs-stimulated AC2 activity. The data presented supports the view of AC2 as a molecular switch and implicates this isoform as a target for Go protein-linked signalling.     

Rodent oocytes express an active adenylyl cyclase required for meiotic arrest

The intracellular levels of cAMP play a critical role in the meiotic arrest of mammalian oocytes. However, it is debated whether this second messenger is produced endogenously by the oocytes or is maintained at levels inhibitory to meiotic resumption via diffusion from somatic cells. Here, we demonstrate that adenylyl cyclase genes and corresponding proteins are expressed in rodent oocytes. The mRNA coding for the AC3 isoform of adenylyl cyclase was detected in rat and mouse oocytes by RT-PCR and by in situ hybridization. The expression of AC3 protein was confirmed by immunocytochemistry and immunofluorescence analysis in oocytes in situ. Cyclic AMP accumulation in denuded oocytes was increased by incubation with forskolin, and this stimulation was abolished by increasing intraoocyte Ca(2+) with the ionophore A23187. The Ca(2+) effects were reversed by an inhibitor of Ca(2+), calmodulin-dependent kinase II. These regulations of cAMP levels indicate that the major cyclase that produces cAMP in the rat oocyte has properties identical to those of recombinant or endogenous AC3 expressed in somatic cells. Furthermore, mouse oocytes deficient in AC3 show signs of a defect in meiotic arrest in vivo and accelerated spontaneous maturation in vitro. Collectively, these data provide evidence that an adenylyl cyclase is functional in rodent oocytes and that its activity is involved in the control of oocyte meiotic arrest.     

Transgenic replacement of type V adenylyl cyclase identifies a critical mechanism of beta-adrenergic receptor dysfunction in the G alpha q overexpressing mouse.

Chronic activation of Gq coupled receptors, or overexpression of G alpha q, in cardiomyocytes results in hypertrophy, enhanced expression of fetal genes, decreased basal and beta-adrenergic receptor (beta AR) stimulated adenylyl cyclase (AC) activities, and depressed cardiac contractility in vivo. Among several abnormalities of the beta AR-Gs-AC pathway that occur in G alpha q overexpressing transgenic mice, we have investigated whether the observed approximately 45% decrease in type V AC expression and function compared to non-transgenic (NTG) is the basis of the above phenotype. Transgenic mice were generated that overexpressed by approximately 50% the rat type V AC in the heart using the alpha-myosin heavy chain promoter. These mice were mated with the G alpha q transgenics resulting in animals (ACV/G alpha q) that had restored levels of forskolin stimulated AC activities in cardiac membranes. In addition, basal cardiac AC activities were normalized in the ACV/G alpha q mice (NTG=23+/-4.4, G alpha q=14+/-3.6, ACV/G alpha q=29+/-5.3 pmol/min/mg) as were maximal isoproterenol stimulated activities (59+/-8.9, 34+/-4.6, 52+/-6.7 pmol/min/mg respectively). Cardiac contractility was also improved by ACV replacement, with increased fractional shortening (51+/-2%, 36+/-6%, 46+/-3% respectively). In contrast, hypertrophy and expression of hypertrophy associated fetal genes were not affected. Thus the observed decrease in type V AC that accompanies the development of the cardiac phenotype in the G alpha q model is the dominant mechanism of dysfunctional beta AR signalling and contractility. In contrast, the decrease in type V AC or beta AR signalling to cAMP is not the basis of the hypertrophic response.     

RGS2 interacts with Gs and adenylyl cyclase in living cells

Regulator of G Protein Signalling (RGS) proteins impede heterotrimeric G protein signalling. RGS2 decreases cAMP production and appears to interact with both adenylyl cyclase (AC) and its stimulatory G protein Gs. We showed previously that Green Fluorescent Protein-tagged RGS2 (GFP-RGS2) localizes to the nucleus in HEK 293 cells and is recruited to the plasma membrane when co-expressed with Gsalpha, or the Gs-coupled beta2-adrenergic receptor (beta2AR). Here, using confocal microscopy we show that co-expression of various AC isoforms (ACI, ACII, ACV, ACVI) also leads to GFP-RGS2 recruitment to the plasma membrane. Bioluminescence Resonance Energy Transfer (BRET) was also used to examine physical interactions between RGS2 and components of the Gs-signalling pathway. A BRET signal was detected between fusion constructs of RGS2-Renilla luciferase (energy donor) and Gsalpha-GFP (energy acceptor) co-expressed in HEK 293 cells. BRET was also observed between GFP-RGS2 and ACII or ACVI fused to Renilla luciferase. Additionally, RGS2 was found to interact with the beta2AR. Purified RGS2 selectively bound to the third intracellular loop of the beta2AR in GST pulldown assays, and a BRET signal was observed between GFP-RGS2 and beta2AR fused to Renilla luciferase when these two proteins were co-expressed together with either ACIV or ACVI. This interaction was below the limit of detection in the absence of co-expressed AC, suggesting that the effector enzyme stabilized or promoted binding between the receptor and the RGS protein inside the cell. Taken together, these results suggest the possibility that RGS2 might bind to a receptor-G protein-effector signalling complex to regulate Gs-dependent cAMP production.     

Mediation of adenylyl cyclase sensitization by PTX-insensitive GalphaoA, Galphai1, Galphai2 or Galphai3

Chronic activation of mu-opioid receptors, which couple to pertussis toxin-sensitive Galphai/o proteins to inhibit adenylyl cyclase (AC), leads to a compensatory sensitization of AC. Pertussis toxin-insensitive mutations of Galphai/o subtypes, in which the pertussis toxin-sensitive cysteine is mutated to isoleucine (Galpha ), were used to determine whether each of the Galphai/o subtypes is able to mediate sensitization of AC. Galpha , G , G or G were individually transiently transfected into C6 glioma cells stably expressing the mu-opioid receptor, or transiently co-expressed with the mu-opioid receptor into human embryonic kidney (HEK)293T cells. Cells were treated with pertussis toxin to uncouple endogenous Galphai/o proteins, followed by acute or chronic treatment with the mu-opioid agonist, [D-Ala2,N-Me-Phe4,Gly5-ol]enkephalin (DAMGO). Each Galphai/o subtype mediated acute DAMGO inhibition of AC and DAMGO-induced sensitization of AC. The potency for DAMGO to stimulate sensitization was independent of the Galphai/o subtype, but the level of sensitization was increased in clones expressing higher levels of Galphai/o subunits. Sensitization of AC mediated by a component of fetal bovine serum, which was also dependent on the level of functional Galphai/o subunits in the cell, was observed. This serum-mediated sensitization partially masked mu-opioid-mediated sensitization when expressed as percentage overshoot due to an apparent increase in AC activity.     

The type 8 adenylyl cyclase is critical for Ca2+ stimulation of cAMP accumulation in mouse parotid acini

Capacitative Ca(2+) entry stimulates cAMP synthesis in mouse parotid acini, suggesting that one of the Ca(2+)-sensitive adenylyl cyclases (AC1 or AC8) may play an important role in the regulation of parotid function (Watson, E. L., Wu, Z., Jacobson, K. L., Storm, D. R., Singh, J. C., and Ott, S. M. (1998) Am. J. Physiol. 274, C557-C565). To evaluate the role of AC1 and AC8 in Ca(2+) stimulation of cAMP synthesis in parotid cells, acini were isolated from AC1 mutant (AC1-KO) and AC8 mutant (AC8-KO) mice and analyzed for Ca(2+) stimulation of intracellular cAMP levels. Although Ca(2+) stimulation of intracellular cAMP levels in acini from AC1-KO mice was indistinguishable from wild type mice, acini from AC8-KO mice showed no Ca(2+)-stimulated cAMP accumulation. This indicates that AC8, but not AC1, plays a major role in coupling Ca(2+) signals to cAMP synthesis in parotid acini. Interestingly, treatment of acini from AC8-KO mice with agents, i.e. carbachol and thapsigargin that increase intracellular Ca(2+), lowered cAMP levels. This decrease was dependent upon Ca(2+) influx and independent of phosphodiesterase activation. Immunoblot analysis revealed that AC5/6 and AC3 are expressed in parotid glands. Inhibition of calmodulin (CaM) kinase II with KN-62, or inclusion of the CaM inhibitor, calmidazolium, did not prevent agonist-induced inhibition of stimulated cAMP accumulation. In vitro studies revealed that Ca(2+), independently of CaM, inhibited isoproterenol-stimulated AC. Data suggest that agonist augmentation of stimulated cAMP levels is due to activation of AC8 in mouse parotid acini, and strongly support a role for AC5/6 in the inhibition of stimulated cAMP levels.     

Inactivation of the mouse adenylyl cyclase 3 gene disrupts male fertility and spermatozoon function

Mammalian spermatids and spermatozoa express functional G protein-coupled receptors. However, bicarbonate-regulated soluble adenylyl cyclase (AC), the major AC present in these cells, is not directly coupled to G proteins. To understand how G protein-coupled receptors signal in spermatozoa, we investigated whether a conventional transmembrane cyclase is present and biologically active in these cells. Here, we provide evidence for expression of type 3 AC (AC3) in male germ cells and describe the effects of disruption of the AC3 gene on fertility and function of mouse spermatozoa. As previously reported in rat, AC3 mRNA is expressed in mouse testes and localized, together with soluble AC mRNA, mainly in postmeiotic germ cells. AC3 protein was detected by immunolocalization in round and elongating spermatids in a region corresponding to the developing acrosome and was retained in the mature spermatozoa of the epididymis. Forskolin caused a small increase in cAMP production in mouse spermatozoa, but this increase could not be detected in the AC3(-/-) mice. Inactivation of the AC3 gene did not have overt effects on spermatogenesis; however, AC3(-/-) males were subfertile with only three litters generated by 11 males over a period of 6 months. When used in in vitro fertilization, spermatozoa from these AC3(-/-) mice produced few embryos, but their fertilizing ability was restored after removal of the zona pellucida. Despite an apparently normal structure, these spermatozoa had decreased motility and showed an increase in spontaneous acrosome reactions. These data support the hypothesis that AC3 is required for normal spermatid or spermatozoa function and male fertility.