cAMP is a ubiquitous intracellular signalling molecule that can regulate a wide array of cellular processes. The diversity of action of this second messenger owes much to the localized generation, action and hydrolysis of cAMP within discrete subcellular regions. Further signalling specificity can be achieved by the ability of cells to modulate the frequency or incidence of such cAMP signals. Here, we discuss the use of two cAMP biosensors that measure real-time cAMP changes in the single cell, to address the mechanisms underlying the generation of dynamic cAMP signals. The first method monitors sub-plasmalemmal cAMP changes using mutant cyclic nucleotide-gated channels and identifies an AKAP (A-kinase-anchoring protein)–protein kinase A–PDE4 (phosphodiesterase-4) signalling complex that is central to the generation of dynamic cAMP transients in this region of the cell. The second study uses a fluorescence resonance energy transfer-based cAMP probe, based on Epac1 (exchange protein directly activated by cAMP 1), to examine interplay between Ca2+ and cAMP signals. This study demonstrates real-time oscillations in cAMP driven by a Ca2+-stimulated AC (adenylate cyclase) (AC8) and subsequent PDE4 activity. These studies, using two very different single-cell cAMP probes, broaden our understanding of the specific spatiotemporal characteristics of agonist-evoked cAMP signals in a model cell system.
Compartmentalization of cAMP signals
Numerous hormones and neurotransmitters, acting on GPCRs (G-protein-coupled receptors), mediate a large range of cellular events through the generation of the second messenger, cAMP. The diversity of effects mediated by cAMP is due to both the dynamic regulation and localized action of this highly diffusible messenger. The compartmentalization of cAMP signals to discrete cellular regions is thought to arise from the localization of GPCRs with ACs (adenylate cyclases) that generate cAMP and PDEs (phosphodiesterases) that are responsible for cAMP hydrolysis. The localized actions of cAMP are further established by the presence of signalling complexes based on AKAPs (A-kinase-anchoring proteins) that tether inactive PKA (protein kinase A) molecules to specific sites within the cell where they are primed to phosphorylate local proteins . Direct evidence for the spatiotemporal complexity of cAMP signalling is dependent on the ability to monitor real-time cAMP changes at the cellular, or subcellular, level. In the present study, we discuss the use of several cAMP probes that have been designed to address such issues.
Single-cell cAMP biosensors
Changes in cAMP production are routinely measured in cell populations using radiochemical techniques. However, recent years have seen the introduction of a number of cAMP biosensors to assess single-cell cAMP levels. Many of these sensors are based on various cAMP effectors including PKA [2–5], CNGCs (cyclic nucleotide-gated channels) [6,7] and Epac (exchange protein directly activated by cAMP) [8–10]. The first measurement of cAMP in single cells used a probe based on the ubiquitous cAMP-dependent protein kinase (PKA) . This method involved the microinjection of fluorescently tagged PKA subunits: fluorescein-labelled catalytic (C) subunits and rhodamine-labelled regulatory (R) subunits (termed ‘FlCRhR’ or ‘flicker’). These subunits associate to form a PKA holoenzyme within the cell with FRET (fluorescence resonance energy transfer) occurring between the fluorescein (donor) and rhodamine (acceptor) molecules upon excitation of the fluorescein. Once cAMP is generated, it binds to the regulatory subunits of PKA and the enzyme dissociates with a concurrent decrease in FRET signal. This elegant method has since been used to illustrate the formation of cAMP gradients , to estimate cAMP diffusion rate in invertebrate neurons , to simultaneously monitor the [cAMP] and PKA-mediated regulation of calcium channels in frog ventricular myocytes  and to report dynamic interactions between Ca2+ transients and cAMP in embryonic spinal neurons . A genetically encoded version of this FRET-based cAMP probe has since been developed that avoids the experimental limitations that microinjection entails [3,4]. This probe, which uses CFP (cyan fluorescent protein) (donor) and YFP (yellow fluorescent protein) (acceptor) to label the regulatory and catalytic subunits of PKA respectively, has been used to illustrate discrete microdomains of cAMP in neonatal cardiac myocytes . The green fluorescent protein-tagged PKA probe localizes to T-tubules within the myocytes, a consequence of anchoring to AKAPs, where it can detect a range of dynamic cAMP signals after AC activation.
The discovery of Epac as an important cAMP effector  has led to the introduction of several new FRET-based cAMP sensors [8–10]. These Epac-based fluorescent cAMP probes are fast becoming a popular choice for imaging cAMP dynamics as they have a good signal to noise ratio and can be targeted to discrete subcellular compartments. Furthermore, they exhibit faster response times than PKA-based indicators and are not subject to issues arising from subunit reassociation after a stimulus, since both YFP and CFP are tagged to a single Epac-based molecule.
Genetically modified CNGCs provide an alternative method for the visualization of cAMP signals within the vicinity of the plasma membrane [6,15]. These mutant CNGCs can detect changes in local cAMP levels with high temporal resolution, and are ideally suited to monitoring the rapid kinetics of cAMP dynamics at their site of synthesis. The channels are introduced into cells using an adenovirus expression system. Their activation can then be measured using either perforated patch-clamp methods or by imaging with a Ca2+-sensitive dye, such as fura 2, to assess Ca2+ entry into the cell via the activated channel. CNGCs have been used in HEK-293 cells (human embryonic kidney cells) and cardiac myocytes to provide compelling evidence for the existence of cAMP ‘microdomains’ just beneath the plasma membrane where agonist-evoked changes in cAMP are large and more dynamic than those seen in the bulk cytosol. These CNGC studies suggest that cAMP compartmentalization results from restricted diffusion of the second messenger away from the cell periphery due to physical or enzymatic barriers [6,7,15,16].
Use of CNGCs to identify a sub-plasmalemmal cAMP signalling module
Here, we describe the use of recombinant CNGCs to assess local changes in cAMP just beneath the plasma membrane. Previous studies have revealed that cAMP changes in this region of the cell are larger and more rapid than those seen in the bulk cytosol after GPCR stimulation [6,15]. By combining the use of CNGCs with RNAi (RNA interference) and co-immunoprecipitation methods, we can identify the signalling complex that shapes cAMP dynamics in the sub-plasmalemmal cellular microdomain. Assessment of CNGC activity using both calcium imaging and whole-cell patch clamp in the presence of selective pharmacological agents or after knockdown of endogenous signalling proteins (using RNAi) provides evidence for a PDE4–PKA–AKAP signalling complex at the plasma membrane of HEK-293 cells. Using an RNAi approach, we have identified the central anchoring protein as gravin (also known as AKAP250) [16a]. This functional complex is essential for a negative feedback regulation of cAMP that creates rapid transients in local levels of the second messenger after GPCR stimulation. The activation of PDE4 via an anchored PKA to rapidly hydrolyse cAMP at its site of synthesis provides a fine-tuning opportunity in the ability of cAMP to regulate cellular activity. Previous biochemical studies using co-immunoprecipitation and enzyme activity measurements have been vital in suggesting the existence of regulatory modules that localize both PKA and PDE4 activity to certain subcellular regions to tightly control cAMP signals and generate spatial gradients in cAMP [17,18]. However, it is only by looking at intact, live cells and performing molecular dissection using RNAi and real-time measurements of cAMP that the inferred changes in cAMP can be monitored and the existence of functional cAMP signalling complexes can be proven.
Epac-based cAMP probe used to monitor cAMP and calcium interplay
Several AC isoforms are regulated by sub-micromolar concentrations of Ca2+. AC1 and AC8 are stimulated by Ca2+/calmodulin; AC5 and AC6 are directly inhibited by Ca2+ . The Ca2+-sensitivity of these cAMP-generating enzymes has provided the basis for a number of models (for example ) predicting that dynamic changes in Ca2+ will produce parallel changes in cAMP to enhance the signalling specificity of both second messengers. Regular oscillations of [cAMP] were first detected over 30 years ago in strips of frog ventricular muscle, as a function of the myocardial contraction cycle . However, the first compelling evidence for cAMP oscillations in a single cell came several years later when Gorbunova and Spitzer  provided proof of prolonged, low-frequency cAMP oscillations during bursts of Ca2+ transients in Xenopus spinal neurons using the FRET-based cAMP indicator ‘FlCRhR’. A recent study using a truncated form of the CFP-tagged PKA RII subunit with a targeting motif that directs it to the plasma membrane along with the YFP-tagged catalytic subunit illustrates higher frequency calcium-dependent cAMP oscillations in pancreatic β-cells in response to hormone application . The cAMP rises reported are in phase with increases in intracellular Ca2+ and have an incidence frequency of approx. 0.2–1.5 transients·min−1. Two other recent publications (including our own) have used an Epac-based cAMP sensor  to show rapid cAMP oscillations accompanying periodic changes in Ca2+ [22,23]. In contrast with Dyachok et al. , Landa et al.  report an inverse relationship between depolarization-evoked calcium events and cAMP oscillations in β-cells that is thought to be due to the periodic activation and inactivation of a Ca2+/calmodulin-activated PDE (PDE1). We have used the Epac probe in HEK-293 cells expressing AC8 to show that artificially imposed or agonist-induced Ca2+ rises are accompanied by simultaneous increases in cAMP . In the present study, even higher frequency interplay between Ca2+ and cAMP dynamics was observed with the calcium-mobilizing agonist, carbachol, evoking cAMP events at a rate of up to 3 transients·min−1. In AC8-expressing cells, the dynamic changes in cAMP depend on a combination of Ca2+-stimulated AC8 activity and cAMP-dependent PDE (PDE4) activity (Figure 1). The incidence and amplitude of cAMP oscillations in these cells were increased in the presence of low doses of Gs-coupled agonist (isoprenaline and prostaglandin E1), which were consistent with the AC acting as a coincidence detector, dually regulated by both Gs and Ca2+/calmodulin signals. This dependence of cyclase activity on the convergence of two separate signalling pathways may underlie the proposed role of AC8 (and AC1) in learning and memory [24,25]. This demonstration of frequency coding brought about by the interaction of two discrete signalling pathways provides an important mechanism by which cells can discriminate between various stimuli to mediate a diverse range of downstream events critical to cell survival.
Sequence of events shaping Ca2+-dependent cAMP transients
The monitoring of single-cell and subcellular changes in cAMP is gaining pace rapidly with the development of sensors with improved sensitivity, signal-to-noise ratio, dynamic range, temporal resolution and spatial resolution (aided by specific targeting). Nevertheless, no one sensor is suited to all applications. In order to address different hypotheses, the advantages and disadvantages of each available probe must be considered. CNGCs have the greatest temporal resolution, and have good cAMP-sensitivity, but they are limited to monitoring sub-plasmalemmal cAMP changes [6,7,15,16]. Fluorescent cAMP sensors based on PKA target appropriately to AKAPs , or they can be modified to be targeted to the plasma membrane , and can detect cAMP changes with good sensitivity. However, the PKA-based probes are catalytically active and can buffer cAMP changes and their usefulness for assessing dynamic changes in cAMP can be limited by the time taken for subunits to dissociate/reassociate; in addition, there may be interaction between labelled and unlabelled (endogenous) PKA subunits. The recent introduction of Epac-based cAMP sensors has provided a number of targeted and non-targeted FRET-based probes with a variety of sensitivities and temporal resolutions [8–10]. These probes have the advantage that CFP and YFP are tethered to the same molecule, providing good signal-to-noise ratio, and they have better temporal resolution than the PKA-based FRET probe, although not as good as the CNGCs (Figure 2). Conversely, the full-length Epac probes may have catalytic activity and all of the Epac-based probes have the potential to buffer local cAMP signals. It is likely that the future generation and targeting of high-resolution cAMP biosensors to specific sites within the cell to provide real-time measurements of cAMP changes within subcellular microdomains will provide further insight into the complexities of this finely tuned messenger system and its dynamic interaction with other signalling pathways.
Cartoon for comparison of the temporal characteristics of cAMP signals detected in HEK-293 cells using three different cAMP biosensors
Compartmentalization of Cyclic AMP Signalling: Biochemical Society Focused Meeting held at King's College, Cambridge, U.K., 29–30 March 2006. Organized by D. Cooper (Cambridge, U.K.), M. Houslay (Glasgow) and M. Zaccolo (Padua, Italy). Edited by D. Cooper.
cyan fluorescent protein
cyclic nucleotide-gated channel
exchange protein directly activated by cAMP
fluorescence resonance energy transfer
- HEK-293 cell
human embryonic kidney cell
protein kinase A
yellow fluorescent protein
This work was supported by The Wellcome Trust.