Spatial organization of adenylyl cyclase and its impact on dopamine signaling in neurons

Abstract

The cAMP cascade is increasingly recognized to transduce physiological effects locally through spatially limited cAMP gradients. However, little is known about how adenylyl cyclase enzymes that initiate cAMP gradients are localized. Here we address this question in physiologically relevant striatal neurons and investigate how AC localization impacts downstream signaling function. We show that the major striatal AC isoforms are differentially sorted between ciliary and extraciliary domains of the plasma membrane, and that one isoform, AC9, is uniquely concentrated in endosomes. We identify key sorting determinants in the N-terminal cytoplasmic domain responsible for isoform-specific localization. We further show that AC9-containing endosomes accumulate activated dopamine receptors and form an elaborately intertwined network with juxtanuclear PKA stores bound to Golgi membranes. Finally, we provide evidence that endosomal localization enables AC9 to selectively elevate PKA activity in the nucleus relative to the cytoplasm. Together, these results reveal a precise spatial landscape of the cAMP cascade in neurons and a key role of AC localization in directing downstream PKA signaling to the nucleus.

Fig. 1. Adenylyl cyclases isoforms localize to primary cilia and endosomes.

a–c Maximum intensity Z-projection of confocal microscopy images of MSNs expressing HA-AC3 (a), HA-AC5 (b), and HA-AC9 (c) and stained for cilia marker Arl13b. d Quantification of cilia localization by measuring the fraction of neurons with AC isoforms localized to cilia, using FLAG-D1R as cilia marker (not shown), and expressed as a percentage of total ciliated cells in the transfected cell population. Data are shown as mean ± s.e.m. from n = 3 independent experiments (total of 68–73 cells/condition). AC3 vs AC9, ***P = 0.0001, AC5 vs. AC9, ***P = 0.0002. e Cilia enrichment index calculated as a ratio of HA-AC fluorescence intensity in cilia (determined by Arl13b and FLAG-D1R) divided by total cell fluorescence. Data are shown as mean ± s.e.m. from n = 4 independent experiments (for AC3 and AC5) and n = 3 independent experiments (for AC9) (31–43 cells total/condition). AC3 vs. AC9, **P = 0.0083, AC3 vs. AC5, **P = 0.0085, AC5 vs. AC9, *P = 0.0175. f–h Representative confocal microscopy images of MSNs expressing AC3-GFP (f), AC5-GFP (g) and AC9-GFP (h) and stained for endosomal marker EEA1. Arrowheads indicate cilia. i Quantification of endosome localization by measuring the fraction of neurons with 10 or more AC-GFP internal puncta, expressed as a percentage of total transfected cells. Data are shown as mean ± s.e.m. from n = 4 independent experiments (34–47 cells total/condition). AC3 vs AC9, ***P < 0.0001, AC5 vs. AC9, ***P = 0.0002. j Endosome enrichment index calculated as a ratio of AC-GFP fluorescence intensity at EEA1-positive endosomes divided by total cell fluorescence. Data are shown as mean ± s.e.m. from n = 5 independent experiments; 38–58 cells total/condition. **P = 0.0012, *P = 0.0351. k Representative iSIM images from n = 3 independent experiments of MSN expressing AC9-GFP and stained for EEA1 and Vps35. Arrowheads indicate AC9-positive subdomains at the endosomal membrane. Scale bars are 10 μm. n.s. not significant, P values are calculated using unpaired two-tailed Student’s t-test. Source data are provided as a Source Data file.

Fig. 2. Dopamine 1 receptor localizes to adenylyl cyclase isoforms positive compartments.

a, b Maximum intensity Z-projection of confocal microscopy images of MSNs expressing FLAG-D1R and HA-AC3 (a) or HA-AC5 (b) from n = 4 independent experiments. Scale bar = 10 μm. c Representative live cell spinning disk confocal images of MSNs transfected with AC9-GFP and FLAG-D1R and treated with 10 μM dopamine (DA) at 0 min. Surface FLAG-D1R was labeled with Alexa Fluor 555-coupled anti-FLAG antibody for 15 min before imaging. Scale bar = 5 μm. d Quantification of surface labeled FLAG-D1R accumulation at segmented AC9-GFP-positive endosomes after vehicle (Veh) or 10 μM DA addition. Data are shown as mean ± s.e.m. from n = 5 independent experiments; (33–35 cells total/condition). Source data are provided as a Source Data file.

Fig. 3. The N-terminus of AC5 is necessary and sufficient for cilia targeting.

a, c, e Maximum intensity Z-projection of confocal images microscopy images of MSNs surface-labeled for FLAG-D1R and expressing AC5-GFP or AC5-AC9-Nter-GFP (a), AC5-GFP or AC5-ΔNter-GFP (c), AC9-GFP or AC9-AC5-Nter-GFP (e). Show above each panel is the schematic representation of each construct showing corresponding mutations and topology. b, d, f Cilia enrichment index of cells coexpressing FLAG-D1R and AC5-GFP or AC5-AC9-Nter-GFP (b) (n = 4 independent experiments; 35-38 cells total/condition; ***P = 0.0003 by unpaired two-tailed Student’s t-test), AC5-GFP or AC5-ΔNter-GFP (d) (n = 3 independent experiments; 39 cells total/condition; ***P = 0.0001 by unpaired two-tailed Student’s t-test), AC9-GFP or AC9-AC5-Nter-GFP (f) (n = 4 independent experiments; 36 cells total/condition; **P = 0.0074 by unpaired two-tailed Student’s t-test). Data are shown as mean ± s.e.m. Scale bars are 10 μm. Source data are provided as a Source Data file.

Fig. 4. A dileucine motif in the N-terminus of AC9 is required for endosome localization.

a Schematic representation of the chimeric mutant AC5-AC9-Nter. The AC5-derived sequence is depicted in orange and the AC9-derived sequence in purple. b Maximum intensity Z-projection of confocal images microscopy images of MSNs expressing AC5-AC9-Nter-GFP and stained for EEA1. c Endosome enrichment index of cells expressing AC5-GFP, AC5-AC9-Nter-GFP or AC9-GFP. Data are shown as mean ± s.e.m. from n = 4 independent experiments; 46-52 cells total/condition; *P = 0.0144 by unpaired two-tailed Student’s t-test. d Schematic of chimeric mutants of AC5 (in orange) containing portions of AC9 N-terminus (in purple). e Endosome enrichment of cells expressing AC5 chimeric mutants. Data are shown as mean ± s.e.m. from n = 5 independent experiments for AC5 and AC5-AC9-1-35 and n = 6 independent experiments for AC5-AC9-N, AC5-AC9-36-70 and AC5-AC9-71-105 (≥26 cells total/condition). ***P = 0.0004, **P = 0.003 for AC5 vs. AC5-AC9-N, **P = 0.0067 for AC5 vs. AC5-AC9-1-35 by unpaired two-tailed Student’s t-test. f Schematic of the three portions of AC9 N-terminus and their respective sequence. g Representative confocal images of MSNs expressing AC5-AC9-1-35-GFP or AC5-AC9-1-35-LL > AA-GFP and stained for EEA1. h Endosome enrichment of cells in g. Data are shown as mean ± s.e.m. from n = 3 independent experiments for AC5, n = 5 for AC5-AC9-1-35 and n = 6 for AC5-AC9-1-35-LL > AA (≥30 cells total/condition). **P = 0.003 by unpaired two-tailed Student’s t-test. i Representative confocal images of MSNs expressing AC5-AC9-1-35-GFP or AC5-AC9-1-35-LL > AA-GFP and stained for EEA1. j Endosome enrichment of cells in i. Data are shown as mean ± s.e.m from n = 8 independent experiments (69–72 cells total/condition). *P = 0.0406 by paired two-tailed Student’s t-test. Scale bars are 10 μm. Source data are provided as a Source Data file.

Fig. 5. Endosomal AC9 modulates PKA activity but not overall cAMP.

a–c cAMP and PKA response in cells expressing cADDis (cAMP) or ExRai-AKAR2 (PKA) sensors alone or with HA-AC3 (a), HA-AC5 (b), or HA-AC9 (c) and treated with 10 μM dopamine (DA). a n = 5 and 3 independent experiments (cAMP, -AC and HA-AC3, respectively; 60-64 cells total/condition), and n = 3 (PKA; 42-61 cells total/condition). b n = 18 (cAMP; 293-310 cells total/condition), n = 3 (PKA; 28-42 cells total/condition). c n = 10 (cAMP; 128-189 cells total/condition), n = 5 and 9 (PKA, -AC and HA-AC9 respectively; 78–126 cells total/condition). a *P = 0.0154 for cAMP and *P = 0.0412 for PKA. b ***P = 0.0002, *P = 0.0432. c **P = 0.0072. d Representative image of MSN expressing the cAMP biosensor cADDis and kinetics of cAMP production over time in MSNs coexpressing cADDis and HA-AC9 (purple, n = 10) or HA-AC9-LL > AA (pink, n = 9) (173–189 cells total/condition) and treated with 10 μM DA. e Integrated cAMP signals of the phases 0-5 min and >5 min (5–30 min) after DA addition (n = 9 for HA-AC9, n = 10 for HA-AC9-LL > AA). f Peak and plateau values were calculated as the maximum ΔF/F₀ (peak) and the average of 20-3₀ min values (plateau) (n = 9 for HA-AC9, n = 10 for HA-AC9-LL > AA). g Representative image of MSN expressing the PKA biosensor ExRai-AKAR2 and kinetics of PKA activity over time in MSNs coexpressing ExRai-AKAR2 and HA-AC9 (purple) or HA-AC9-LL > AA (pink) and treated with 10 μM DA, from n = 9 independent experiments (120–126 cells total/condition). h Integrated PKA signals of the phases 0-5 min and >5 min (5–30 min) after DA addition (n = 9 independent experiments; **P = 0.0082). i Peak and plateau values (n = 9 independent experiments; *P = 0.0483). For all panels, data represent biological replicates and are shown as mean ± s.e.m. Scale bars are 10 μm. P values are calculated by unpaired two-tailed Student’s t-test. Source data are provided as a Source Data file.

Fig. 6. Nuclear PKA activity is dependent on AC9 endosomal concentration.

a, c Representative iSIM images of MSN expressing AC9-GFP and stained for endogenous PKA cat (a, n = 4 independent experiments) or PKA RIIβ (c, n = 3). Nuclei were stained with DAPI. Scale bar = 10 μm. b,d, 3D rendering of cells in a (b) and c (d). Scale bar = 5 μm. e Spinning-disk confocal images from a time series of neurons coexpressing PKAcat-GFP and AC9-HaloTag and treated with 10 μM dopamine (DA) at t = 0 min, from n = 3 independent experiments. Scale bar = 5 μm. f Kymograph of the cell in (e). Scale bar  = 1 μm. g On the left, schematic representation and spinning-disk confocal representative image showing ExRai-AKAR2-NLS localization to the nucleus (Scale bar = 10 μm). On the right, kinetics of nuclear PKA activity over time in MSNs coexpressing ExRai-AKAR2-NLS and HA-AC9 (purple) or HA-AC9-LL > AA (pink) and treated with 10 μM DA, from n = 9 independent experiments (61–69 cells total/condition). h Integrated nuclear PKA signal (n = 9 independent experiments; *P = 0.0409 by unpaired two-tailed Student’s t-test). i Peak and plateau values were calculated as the maximum ∆F/F0 (peak) and the average of 20–30 min values (plateau) (n = 9 independent experiments; *P = 0.0242 by unpaired two-tailed Student’s t-test). j On the left, schematic representation and spinning-disk confocal representative image showing ExRai-AKAR2-NES localization to the cytoplasm and excluded from the nucleus (Scale bar = 10 μm). On the right, kinetics of cytoplasmic PKA activity over time in MSNs coexpressing ExRai-AKAR2-NES and HA-AC9 (purple) or HA-AC9-LL > AA (pink) and treated with 10 μM DA, from n = 6 independent experiments (61–103 cells total/condition). k Integrated cytoplasmic PKA signal (n = 6 independent experiments). l Peak and plateau values (n = 6 independent experiments). For all panels, data represent biological replicates and are shown as mean ± s.e.m. Source data are provided as a Source Data file.

Fig. 7. Spatial organization of adenylyl cyclase isoforms impacts nuclear PKA activity in neurons.

a Diagram of proposed model illustrating the distinct subcellular distribution of AC isoforms in striatal neurons. AC3 is restricted to the primary cilium; AC5 localizes at the plasma membrane, both in cilia and extraciliary surfaces; AC9 localizes at the plasma membrane outside of cilia, and in endosomes. Endosomal localization requires a dileucine motif in the AC9 N-terminus. b Endosomes containing AC9 and D1R are in close proximity with Golgi-associated PKA compartments in the perinuclear region. This enforced proximity ‘in trans’ selectively drives PKA activity in the nucleus (yellow).