Adenylyl cyclases (ACs) and ATP-pyrophosphate lyases comprise a family of enzymes that catalyze the synthesis of cyclic AMP from ATP. Cyclic AMP (cAMP) was identified in 1957 as the first “second messenger,” relaying signals from hormone-bound receptors to protein kinase A (PKA) and other cAMP-sensitive effectors, including cyclic-nucleotide gated channels, cAMP-activated exchange proteins (EPAC), and a subset of phosphodiesterases that degrade cyclic nucleotides. Catalytic activity of AC is regulated in response to activation of G protein-coupled receptors (GPCRs) by a number of hormones and neurotransmitters. Various studies using biochemical and genetic tools have implicated the importance of cAMP in a variety of physiological functions that include but are not limited to oogenesis, embryogenesis larval development, hormone secretion, glycogen breakdown, smooth muscle relaxation, cardiac contraction, olfaction, water homeostasis, and learning and memory (Sadana and Dessauer 2009). The discovery of cAMP and subsequent studies related to the regulation and physiological functions of AC has given rise to six Nobel Prizes in medicine and chemistry (1971, 1992, 1994, 2000, 2004, and 2012) and remains a highly active area of research (Beavo and Brunton 2002).
Nine mammalian transmembrane ACs have been cloned and characterized since the initial cloning of AC1 in 1989. A tenth “soluble” form of AC (sAC) has also been characterized that lacks transmembrane domains and has distinct regulatory properties. Based upon homology and regulation patterns, membrane-bound ACs are classified into four groups; group I comprises the Ca2+-stimulated AC1, AC3, and AC8; Gβγ-stimulated AC2, AC4, and AC7 belong to group II; group III contains Giα/Ca2+-inhibited AC5 and AC6; while group IV contains the distantly related and largely forskolin-insensitive AC9.
Topology and Structure
The “soluble” form of adenylyl cyclase (sAC) has homology to cyanobacterial adenylyl cyclases with several known splice variants. The overall structure of the catalytic core of sAC is highly conserved with the transmembrane ACs, although the primary sequences differ significantly. Unlike transmembrane ACs, it does not respond to the classic G-protein mode of activation but is regulated by calcium and bicarbonate levels. The sAC isoform has clear roles in male fertility and sperm motility, in addition to potential roles in acid/base homeostasis and metabolic sensing (Levin and Buck 2015).
Regulation of ACs by G Proteins
Select AC isoforms (AC1, AC5, and AC6) are directly inhibited by members of the inhibitory class of G proteins, Giα, Gz, and Go (Beazely and Watts 2006). This regulation can be dependent on the form of activation. For example, calmodulin-stimulated AC1 is potently inhibited by Gi family members (i, z, and o), whereas AC1 is either very weakly inhibited or not at all when stimulated by Gsα or forskolin. AC5 and AC6 are inhibited by Giα and Gzα. Giα interacts in a cleft formed in the C1 domain of AC5, analogous to the Gsα-binding site in C2, and acts in opposition to Gsα.
In addition to the α-subunit, Gβγ-subunits of heterotrimeric G proteins can regulate AC activity. For example, the group I ACs (AC1 and AC8) are inhibited by Gβγ, more potently than the inhibition by Giα. Although Gβγ alone cannot stimulate group II ACs (AC2, AC4, and AC7), it can greatly enhance the activation by Gsα (~five- to tenfold for AC2). This conditional Gβγ stimulation is the hallmark of the group II family of ACs. Gβγ also conditionally stimulates AC5 and AC6 in the presence of Gsα (~1.5–2-fold), but not to the same extent as the AC2 family (Fig. 2).
Additional Modes of AC Regulation
Forskolin, a diterpene derived from the root of the Indian plant Coleus forskohlii, highly activates all the membrane-bound AC isoforms except AC9. This reagent is a highly useful tool for studying AC and the effects of cAMP. It has also been explored for possible medicinal uses and was originally characterized as a vasodilator. Forskolin binds at the interface of the C1 and C2 domains at a site distinct from the catalytic site (Tesmer et al. 1997) and is either additive or synergistic with regulation by Gsα. AC9 is weakly activated by forskolin because it lacks a key amino acid residue within this forskolin-binding pocket. No physiological analogs of forskolin have been identified that can bind at the forskolin-binding pocket to regulate AC activity.
AC activity can also be controlled by calcium (reviewed in Halls and Cooper 2011). Most of this regulation is highly isoform specific. For example, although all AC isoforms are inhibited by high, nonphysiological concentrations of calcium, sub-micromolar calcium concentrations inhibit AC5 and AC6 by directly binding to a magnesium ion-binding site within the catalytic core. When bound to calmodulin, calcium can activate AC1 and AC8 and conditionally stimulate AC3 in the presence of Gsα or forskolin. The calcium-/calmodulin-binding sites for AC1 have been mapped to the C1b and C2 domains, whereas the sites on AC8 reside within the N-terminus and C2 domain. Finally, calcium-/calmodulin-bound protein kinases can inhibit ACs. AC3 and AC1, but not AC8, are subject to feedback inhibition by calmodulin kinases (CamK) II and IV, respectively.
Several protein kinases regulate AC activity by directly phosphorylating AC isoforms. Feedback inhibition of AC5 and AC6 occurs via phosphorylation of the C1b domain by cAMP-activated PKA. This type of regulation is facilitated by A-kinase-anchoring proteins (AKAPs) that anchor the regulatory subunit of PKA (Dessauer 2009). AKAPs are scaffolding proteins that bring together a large number of regulatory and downstream effector proteins. They can also bind to a subset of ACs. For example, AKAP79/150 (AKAP5) binds to AC2, AC3, AC5, AC6, AC8, and AC9. The formation of an AKAP79/150-AC-PKA complex facilitates preferential phosphorylation of AC5/AC6 by anchored PKA to inhibit AC activity. Much like calcium/calmodulin regulation of AC1, this sets up an important negative feedback loop to temporally regulate cAMP production and downstream signaling. A similar type of regulation also occurs for mAKAP-AC5-PKA complexes that regulate stress-responsive pathways in the heart (Efendiev and Dessauer 2011).
Regulation by protein kinase C (PKC) isoforms is quite complex. Conventional PKCs (α and β; activated by calcium and diacylglycerol) often stimulate AC2 and AC5 and are synergistic with other forms of regulation. The novel PKCδ isoform (activated by only diacylglycerol) displays synergy with Gsα in activating AC7, while the atypical PKCζ stimulates AC5. Conventional or novel PKCs can also inhibit AC activity. AC4 is inhibited by PKCα when stimulated by Gsα, but not under basal or forskolin-stimulated conditions. Gsα-stimulated AC9 is also inhibited by conventional PKCs, but this may be indirect. The novel PKC δ and ε isoforms inhibit AC6 by phosphorylating several sites within its N-terminus. The sites of PKC phosphorylation on ACs vary between these different isoforms.
Additional kinases that regulate AC activity include Raf1 that can stimulate AC6 activity. The role of phosphatases in regulating ACs should not be overlooked. Although this has not been greatly examined, AC9 is inhibited by the phosphatase calcineurin (also known as protein phosphatase 2B, PP2B); however, it is unknown whether calcineurin acts directly on AC9. Another phosphatase, PP2A, is scaffolded by AC8, although a direct regulatory role is still unclear.
Finally, a wide range of additional proteins have been reported to regulate specific AC isoform activity including regulators of G protein signaling 2 (RGS2; inhibition of AC3, AC5, and AC6), the protein associated with myc (PAM; inhibition of AC1, AC5, AC6, and AC7), the synaptic vesicle protein snapin (prevents PKC inhibition of AC6), and the guanine nucleotide exchange protein Ric8a (inhibits AC5). The functional roles for some of these proteins are further discussed below.
AC Physiological Roles
Tissue distribution and physiological functions of individual mammalian AC isoforms
Sites of expression
Brain, adrenal medulla, SA node
Learning, memory, synaptic plasticity, opiate withdrawal, pain memory, neuronal excitotoxicity
Brain, lung, pancreas, testis, adrenal
Learning, memory, synaptic plasticity, opiate withdrawal, stress anxiety, mood disorders, ethanol consumption, glucose homeostasis
Olfactory epithelium, pancreas, brain, heart, lung, testis, BAT
Olfaction, sperm function, kidney function, diet-induced obesity, depression
Brain, lung, skeletal muscle, heart
Ethanol dependency, immune response, depression
Heart, striatum, kidney, liver, lung, testis, adrenal, BAT
Cardiac contraction, motor coordination, opiate dependency, pain responses, renin secretion, stress responses
Heart, kidney, liver, lung, brain, testis, skeletal muscle, adrenal, BAT
Cardiac contraction and calcium sensitivity, water homeostasis, sympathetic tone
Possible immune responses and cardiac function
Testis and detected in all tissues
Sperm capacitation, fertilization, renal function, acid-base sensing
In Learning and Memory
Nearly all AC isoforms are expressed in the brain, but they display distinct roles. Early studies with Drosophila mutants rutabaga (deficient in Ca2+-/CAM-stimulated AC activity whose sequence is closely related to AC1) demonstrated that mutant flies failed to learn to avoid a neutral odor. The role for AC1 in learning and memory in mammalian systems correlates well with its expression in the brain and the phenotypes of mice deficient in AC1 (AC1 (−/−)). AC8 is also expressed mainly in the brain and shares many regulatory properties of AC1; thus, it is not surprising that it also is involved in learning and memory.
The formation of memories relates to the change in connections between neurons in the brain, often referred to as synaptic plasticity. This change is often measured as long-term potentiation (LTP) which is a long-lasting enhancement of neuronal connections, which enhances synaptic transmission or communication. Synaptic plasticity is dependent on calcium; therefore, it is not surprising that AC1 and AC8 play such a large role. AC1 (−/−) and AC8 (−/−) animals show decreased LTP in the mossy fiber bundles of neurons that form connections with the CA3 region of the hippocampus (Wu et al. 1995; Wang et al. 2003). A complete loss of mossy fiber LTP was observed in the double knockout of AC1 and AC8. In contrast, the overexpression of AC1 actually enhanced recognition memory and LTP (Wang et al. 2004). These and many additional studies suggest that AC1 and AC8 are redundant in learning and memory and the formation of fear-related memories. However, there are some distinct roles for AC1 and AC8. For example, AC1 is important for the stability of neuronal circuits in response to activation deprivation, while AC8 is more involved in anxiety responses. This correlates well with the high expression of AC8, but not AC1, in the thalamus, habenula, and hypothalamus, regions involved in responses to stress. In summary, although AC1 and AC8 are not necessary for survival, they play clear roles related to learning and memory.
Signaling pathways associated with cAMP have long been known to play a crucial role in the processing of painful stimuli (reviewed in Pierre et al. (2009)). Various AC isoforms (AC1, AC2, AC5, AC6, and AC8) are expressed in the spinal cord, but deletion of AC1 and AC5 results in attenuated pain responses. AC1, but not AC8, knockout mice have significantly reduced behavioral responses to acute muscle pain. Chronic muscle inflammatory pain was also significantly reduced in AC1 and the AC1/AC8 double knockouts but could be rescued by activating other ACs using forskolin. Thus AC1 plays an important role in acute and chronic muscle pain, although clearly additional ACs are present that can rescue impaired effects.
Mice deficient in AC5 also have attenuated pain responses in acute thermal and mechanical pain tests. They display decreased sensitivity to inflammatory pain and inflammatory visceral pain. AC5 (−/−) mice also display strongly attenuated mechanical and thermal allodynia (an exaggerated response to normal stimuli) in neuropathic pain models. Although AC1 and AC5 belong to different families in terms of their regulatory properties, they share a common feature in that both are inhibited by the protein PAM, which is upregulated in the spinal cord in response to pain.
The analgesia properties of opiates such as morphine are mediated by Gi-coupled opiate receptors and are related in part to the inhibition of adenylyl cyclase (reviewed in Brust et al. (2015) and Pierre et al. (2009)). Long-term morphine use causes an upregulation of AC signal transduction components (AC1 and AC8, PKA, and CREB) in regions of the brain associated with drug reinforcement and withdrawal. The deletion of AC1 and AC8 causes a reduction in opiate withdrawal behaviors, and the double knockout of AC1 and AC8 displays less morphine-induced hyper-locomotion and no activation of the cAMP-dependent transcription factor, CREB, in the reward response circuitry of the brain (ventral tegmentum). AC1 and AC8 also have some distinct functions during chronic morphine exposure, based upon nonoverlapping patterns of gene expression changes. This difference is also reflected in the addictive properties of ethanol where mice deficient in AC8, but not AC1, had decreased voluntary ethanol consumption.
AC5 also plays an important role in opiate actions. The region of the brain (striatum) that shows the highest μ-opiate receptor levels also contain high levels of AC5. Deletion of AC5 results in a loss of opioid-induced Gi-inhibition of AC activity in the striatum. In addition, the major behavioral effects of morphine including locomotor activation, pain relief, tolerance, reward and physiological dependence, and withdrawal symptoms were attenuated in AC5 (−/−) mice. Thus AC activity plays important roles in opioid responses and addiction, with AC1 and AC8 having roles in withdrawal, hyper-locomotion, and the learned responses to morphine, whereas AC5 is involved in all major behavioral effects of morphine.
In Motor Functions
Motor functions can be divided into voluntary and involuntary (or reflex) movements. Those nonreflex actions require higher cognitive activity in several brain regions, including the striatum. The striatum region of the brain is known to be important for the decision-making of voluntary movements which often requires input from dopamine-releasing neurons. Genetic ablation of AC1 and AC8 did not affect motor coordination. However, AC5 (−/−) mice no longer respond to the dopamine D2 antagonist class of antipsychotic drugs, although general motor control is unaltered (reviewed in Sadana and Dessauer (2009)). This behavior correlates well with the high expression of AC5 in the striatum. Other AC5 (−/−) models exhibit Parkinson’s like motor dysfunction, displaying abnormal coordination, a slowness in the execution of movement (bradykinesia), and locomotor impairment (Iwamoto et al. 2003). Motor coordination can be restored by stimulation of dopamine D2 receptors, while bradykinesia was largely restored by either D1 or D2 stimulation of residual striatal AC activity. Although other Gi-inhibited ACs (AC1 and AC6) are present in the striatum, they cannot fully compensate for AC5 function in dopamine-dependent motor coordination.
In Cardiac Functions
As discussed above, all of the AC isoforms except AC8 are expressed in cardiac myocytes or fibroblasts. AC1 is expressed only in sinoatrial mode, where it may modulate pacemaker activity. Two closely related isoforms, AC5 and AC6 are the major isoforms expressed in cardiac myocytes and have been the focus of several deletion and overexpression studies (reviewed in Sadana and Dessauer (2009)). These ACs appear to exert opposite effects on the heart, since cardiac overexpression of AC6 appears to be protective, whereas disruption of type 5 AC prolongs longevity and protects against cardiac stress (Yan et al. 2007). However, some overlapping functions must exist in the heart, as deletion of AC5 or AC6 does not give rise to a complete loss of sympathetic regulation.
The deletion of AC5 results in ~40% decreased isoproterenol- (an agonist of the beta-adrenergic receptors; βAR) and forskolin-stimulated AC activity in cardiac membranes and isolated myocytes. There is also a loss of acetylcholine-mediated (Gi) inhibition and reduced calcium-mediated inhibition of cAMP production in AC5 (−/−) heart. Although differences in AC5 deletion strains exist, the loss of AC5 results in decreased isoproterenol-stimulated left ventricular (LV) ejection fraction (e.g., the volume of blood pumped out of the heart with each beat). Effects of AC5 deletion are not limited to sympathetic regulation, as loss of AC5 also eliminates parasympathetic control of cAMP levels and attenuates baroreflexes that maintain blood pressure. The deletion of AC6 also decreased AC5 levels (Tang et al. 2008). As expected, βAR-stimulated cAMP levels were reduced ~80% in AC6 (−/−) animals with greatly reduced βAR-stimulated LV contractile function and reduced calcium transients. The latter effect on calcium handling is likely due to the loss of AC6, rather than AC5.
When the heart is stressed by chronic activation of cAMP (often due to prolonged beta-adrenergic receptor activation), a decrease in cardiac function is observed (cardiac myopathy). The levels of AC5 are also known to increase under these conditions. Deletion of AC5 protects the heart against chronic βAR stimulation and chronic pressure overload by attenuating the decline in cardiac function and defending against increased apoptosis. AC5 disruption is also protective against age-related cardiac myopathy and gives rise to an increased lifespan as compared to wild-type animals.
The differing roles for AC5 and AC6 may lie in the types of AKAP complexes that they are associated with. For example, AC6 is found in complex with AKAP79/150 in the heart. Deletion of AKAP150 results in loss of βAR-stimulated calcium transients (Navedo et al. 2008), reminiscent of AC6 deletion phenotypes. In contrast, AC5 is found in complex with mAKAP which is associated with the regulation of cardiac hypertrophy and related stress responses (Dessauer 2009). Another AKAP, Yotiao, is important in mediating sympathetic control of cardiac action potential duration. Yotiao is associated with several AC isoforms including AC9 which is present at lower levels in cardiac myocytes. The role for this latter AC isoform is under investigation.
AC2, AC3, and AC4 are expressed in the olfactory system. However, AC3 is the predominant isoform in the olfactory epithelium and is largely responsible for odorant and pheromone detection (reviewed in Wang et al. (2007)). Odorants interact with G protein-coupled receptors to stimulate adenylyl cyclase via Golf. Genetic deletion of AC3 confirms its role in olfaction, as AC3 (−/−) mice suffer from major effects on odorant-induced signaling and are impaired in olfactory-dependent learning and olfaction-based behavioral tests. In addition, AC3 (−/−) mice are unable to detect mouse urine or pheromones. AC3 (−/−) mice also lack intermale aggressiveness and male sexual behavior. AC3 has also been ascribed a role in spermatozoa function and male fertility. In general the vomeronasal organ expressing AC2 is thought to be responsible for pheromone detection, but as discussed, AC3 is also associated with these functions.
In Immune Responses
The role of cAMP in the immune system is complex since cAMP generation can induce apoptosis as well as cell proliferation, differentiation, and activation of various immune cell types. AC7 is the major isoform that regulates cAMP synthesis in both B and T cells. AC 7 (−/−) mice produce fewer leukocytes and have a high mortality rate upon bacterial infections (Duan et al. 2010), which was attributed to changes in the production of certain serum factors required for regulating AC7 activity. AC7 null mice also produce less antigen-specific antibodies to fight infections even though overall immune responses were hyperactive, as measured by the overproduction of cytokines. The unique regulation of AC7 by Gs- and G12/13-coupled receptors makes it well suited to respond to multiple signals and facilitates its multifaceted roles in regulating both innate and adaptive immune responses.
AC6 is localized in the renal tubule and collecting duct of kidneys (Chien et al. 2010). AC6 (−/−) mice had normal glomerular filtration rate but were deficient in water homeostasis (Rieg et al. 2010). AC6-deficient mice drank more water, urinated more, and had low urine solute concentrations. The major protein required for transport of water through the plasma membrane (aquaporin) is also mis-localized, displaying reduced phosphorylation in kidneys of AC6 null mice, consistent with a malfunction in water retention. This phenotype is similar to that of nephrogenic diabetes insipidus and is consistent with a loss of vasopressin-induced cAMP in the inner medullary collecting ducts of AC6 (−/−) kidney. Interestingly, key components of the olfactory signal transduction machinery (olfactory receptor, AC3, and Golf) are expressed in the renal distal nephron. However, roles for AC3 are still unclear.
ACs can be found in every cell type and tissue, consistent with the very large role for cAMP in so many physiological processes. In addition to those outlined above, cAMP has important roles in development and differentiation, cell proliferation, neurodegeneration and neurotoxicity, asthma, diabetes, fertilization, and hormone secretion. Many GPCRs that regulate cAMP levels are currently targeted for treating conditions that include asthma, heart failure, diabetes, pain, migraines, peptic ulcer disease, obesity, schizophrenia, Parkinson’s, and nausea (Pierre et al. 2009). The roles for specific AC isoforms in these pathophysiological conditions have sparked much interest in exploring AC as a drug target. For example, AC5 inhibitors are currently being considered for treatment of heart failure (Ho et al. 2010), while an AC1 activator could be a memory-enhancing or pain-relieving drug (Pierre et al. 2009). Analogs of forskolin that specifically target an AC isoform or selective P-site inhibitors could also prove useful (Iwatsubo et al. 2006). Finally, the local lipid environment of individual AC isoforms or the macromolecular complexes that link ACs to downstream effectors may play a large role in the specificity of downstream signaling. Thus despite the rich history of cAMP research, much work remains to be done.