Mechanisms Restricting Diffusion of Intracellular cAMP

Abstract

Although numerous receptors stimulate cAMP production in a wide array of cells, many elicit distinct, highly localized responses, implying that the subcellular distribution of cAMP is not uniform. One often used explanation is that phosphodiesterases, which breakdown cAMP, act as functional barriers limiting diffusion. However, several studies refute the notion that this is sufficient, suggesting that phosphodiesterase-independent movement of cAMP must occur at rates slower than free diffusion. But, until now this has never been demonstrated. Using Raster Image Correlation Spectroscopy (RICS), we measured the diffusion coefficient of a fluorescently-labeled cAMP derivative (φ450-cAMP) as well as other fluorescent molecules in order to investigate the role that molecular size, cell morphology, and buffering by protein kinase A (PKA) play in restricting cAMP mobility in different cell types. Our results demonstrate that cytosolic movement of cAMP is indeed much slower than the rate of free diffusion and that interactions with PKA, especially type II PKA associated with mitochondria, play a significant role. These findings have important implications with respect to cAMP signaling in all cells.

Figure 1. RICS analysis of fluorescein and EGFP in HEK293 cells.

Representative confocal images of fluorescein (a,b) and EGFP (e,f) in HEK293 cells at low (a,e, scale bar 10 μm) and high (b,f, scale bar 1 μm) magnification. (c,g) Average of the spatial correlation calculated for a region of 64 × 64 pixels (indicated by the red box in b,f) for each of 100 images of fluorescein (c) and EGFP (g) in HEK293 cells. Fit of the correlation functions for fluorescein (d) and EGFP (h). The plot at the top of those panels represents the difference between the autocorrelation and the fit.

Figure 2. RICS analysis of fluorescein and EGFP in adult ventricular cardiac myocytes.

Representative confocal images of fluorescein (a,b) and EGFP (e,f) in myocytes at low (a,e, scale bar 10 μm) and high (b,f, scale bar 1 μm) magnification. (c,g) Average of the spatial correlation calculated for a region of 64 × 64 pixels (indicated by the red box in b,f) for each of 100 images of fluorescein (c) and EGFP (g) in myocytes. Fit of the correlation functions for fluorescein (d) and EGFP (h). The plot at the top of those panels represents difference between the autocorrelation and the fit. (i) Bar plots (average ± s.e.m.) of the diffusion coefficient (D) values obtained for fluorescein and EGFP in HEK293 cells (black bars) and cardiac myocytes (red bars). *p < 0.05, ns = not significant. HEK293 cells: n = 4–8, myocytes: n = 7–22.

Figure 3. RICS analysis of φ450-cAMP and free φ450 in adult ventricular cardiac myocytes.

Representative confocal images of φ450-cAMP (a,b) and free φ450 (e,f) in myocytes at low (a,e, scale bar 10 μm) and high (b,f, scale bar 1 μm) magnification. (c,g) Average of the spatial correlation calculated for a region of 64 × 64 pixels (indicated by the red box in b,f) for each of 100 images of φ450-cAMP (c) and free φ450 (g) in myocytes. (d,h) Fit of the correlation functions. The plot at the top of those panels represents the difference between the autocorrelation and the fit. (i) Bar plots (average ± s.e.m.) of the diffusion coefficient (D) values obtained for φ450-cAMP and free φ450 in HEK293 cells (black bars) and cardiac myocytes (red bars). *p < 0.05, ns = not significant. HEK293 cells: n = 8–12, myocytes: n = 9–37.

Figure 4. Co-localization of φ575-cAMP with the type II regulatory subunit (RII) of PKA in adult ventricular cardiac myocytes.

Confocal images of myocytes expressing RII-ECFP (a,d) or loaded with φ575-cAMP (b,e) at low (left panels, scale bar 10 μm) and high (right panels, scale bar 1 μm) magnification. The merged images (c,f) demonstrate a high degree of co-localization between RII-ECFP and φ575-cAMP. This experiment represents an example of cells in which RII-ECFP exhibited a longitudinal striated pattern. PCC 0.41 ± 0.014, tM1 0.55 ± 0.014, tM2 0.71 ± 0.014, n = 9.

Figure 5. RICS analysis of φ450-cAMP and RII-ECFP in adult ventricular cardiac myocytes treated with St-Ht31 peptide.

Representative confocal images of St-Ht31-treated myocytes expressing RII-ECFP (a,b) or loaded with φ450-cAMP (e,f) at low (a,e, scale bar 10 μm) and high (b,f, scale bar 1 μm) magnification. (c,g) Average of the spatial correlation calculated for a region of 64 × 64 pixels (indicated by the red box in b,f) for each of 100 images of RII-ECFP (c) and φ450-cAMP (g) in St-Ht31-treated myocytes. Fit of the correlation functions for RII-ECFP (d) and φ450-cAMP (h). The plot at the top of those panels represents the difference between the autocorrelation and the fit. (i) Bar plots (average ± s.e.m.) of the diffusion coefficient (D) values obtained for RII-ECFP and φ450-cAMP in control (red bars) or St-Ht31-treated (open bars) myocytes (*p < 0.05. RII-ECFP expressing cells: n = 8–11, φ450-cAMP loaded cells: n = 15–37).

Figure 6. Co-localization of the type II regulatory subunit (RII) of PKA or φ450-cAMP with MitoTracker red in cardiac myocytes.

Representative confocal images of myocytes expressing RII-ECFP (A) or loaded with φ450-cAMP (B) at low (a–c, scale bar 10 μm) and high (d–f, scale bar 1 μm) magnification. Mitochondria labeled with MitoTracker red show a longitudinal striated pattern (b,e). The merged images demonstrate a high degree of co-localization of RII-ECFP (Ac, Af) or φ450-cAMP (Bc, Bf) with MitoTracker red. PCC 0.28 ± 0.017, tM1 0.49 ± 0.018, tM2 0.60 ± 0.017, n = 7 (panel A). PCC 0.22 ± 0.0049, tM1 0.39 ± 0.0064, tM2 0.65 ± 0.0066, n = 13 (panel B).