Advances in measuring real-time cAMP dynamics have provided direct evidence for spatial and temporal heterogeneity of cAMP signals in both excitable and non-excitable cells (Zaccolo & Pozzan, 2002; Rich et al. 2000). Such compartmentalization of cAMP signals is thought to be due to physical or enzymatic barriers limiting diffusion of cAMP into the bulk cytosol. These cAMP microdomains may function to differentiate the downstream effects of the numerous agonists acting via adenylyl cyclase (AC). In excitable cells further differentiation may result from the frequency-encoding of cAMP transients, or oscillations (Cooper et al. 1995; Gorbunova & Spitzer, 2002).

We have used an adenovirus expressed cyclic nucleotide-gated (CNG) channel to determine whether cAMP oscillations can be detected in a non-excitable cell line and identify some of the signalling components involved. Human embryonic kidney cells (HEK-293) were infected with mutant olfactory CNG channel α subunit as described previously (Rich et al. 2001). Channel activation was monitored using either patch-clamp to detect changes in whole-cell current or fluorescent imaging of Fura-2 loaded cells to detect Ca²⁺ influx.

Application of 100 nM prostaglandin E₁ (PGE₁) produced a large cAMP transient that in ~9 % of cells (n = 134) was followed by one or more smaller cAMP transients. Pharmacological inhibitors of type 4 phosphodiesterase (PDE4, 10 µM Rolipram), protein kinase A (PKA, 10 µM H-89) and A-kinase anchoring protein (AKAP, 20 µM Ht31) were used to assess the role of each signalling component in the initial return of cAMP levels to baseline. Our results suggest that a PDE4/PKA/AKAP complex provides local feedback regulation of cAMP levels within the near membrane ‘microdomain’. Furthermore, low doses of PDE4 inhibitor enhanced detection of single cell cAMP ‘oscillations’. Following pretreatment with 0.01 µM rolipram or 0.01 µM RO-20-1724 ~22 % of cells (n = 171) produced more than one cAMP transient in response to 100 nM PGE₁. It is possible that low doses of rolipram or RO-20-1724 inhibit high affinity PDE4 activity induced by local increases in cAMP and unveil near-membrane cAMP oscillations. However, basal PDE activity and CNG channel expression levels (probe sensitivity) are likely to be key factors in determining whether cAMP oscillations will be observed in non-excitable cells.

This work was funded by the Wellcome Trust