Mitochondrial localization of proteins and organelle membrane interactions: two key elements in neurodegeneration explored by new splitGFP tools

Mutations in the mitochondrial serine–threonine kinase PINK1 are associated with familial forms of Parkinson’s disease and mitochondrial Ca2+ overload. The targeting of PKA to mitochondria and its activation rescue functional defects observed in PINK1 deficient neurons and mitochondrial Ca2+ overload due to the loss of PINK1 function. PINK1 and PKA have been proposed to cooperate at the mitochondria level to prevent neurodegeneration, and we have found that PINK1 was able to reduce mitochondrial Ca2+ accumulation. Sustained Ca2+ accumulation into the mitochondrial matrix has been shown to correlate with increases of cAMP levels in the same compartment. If the localization and the action of PKA at the OMM (OMM) are well recognized, its presence and, consequently, its role in the mitochondrial matrix and in the intramembrane space (IMS) is still amply debated. In order to investigate that, we developed a probe based on the splitGFP system and Bimolecular Fluorescence Complementation (BiFC) to monitor PKA distribution at sub-mitochondrial level in living cells. The non-fluorescent GFP1-10 fragment was targeted to the OMM, the IMS and the mitochondrial matrix by the addition of targeting sequences. The β11 fragment, necessary to reconstitute GFP fluorescence, was fused to two PKA regulatory subunits (RIα and RIIβ) and to PKA catalytic subunit (CATα). The co-transfection of the plasmids encoding the targeted GFP1-10 fragments and the β11-CATα or the β11-RIα or the β11-RIIβ in Hela cells revealed the presence of all these subunits at the OMM. Interestingly, strong GFP fluorescence emission in the presence of GFP1-10 fragment targeted to the IMS and mitochondrial matrix was observed in the case of β11-CATα co-expression, but not of β11-RIα and β11-RIIβ, suggesting the presence of PKA CAT-α in these compartments. Then, we evaluated the interference of regulatory subunits with mitochondrial CATα localization, co-transfecting the GFP1-10 fragment targeted to OMM, IMS and mitochondrial matrix with CATα and RIα or RIIβ. In these conditions we still observed fluorescence reconstitution at the OMM and IMS, but not in the mitochondrial matrix when CATα was co-expressed together with RIIβ. In the presence of co-expressed RIα subunit the green fluorescent signal was also detect in the mitochondrial matrix. Then, we analyzed the effect of CATα overexpression on mitochondrial Ca2+ transients, which are strongly decreased compared to the control. This reduction is almost abolished when CATα is co-expressed with RIIβ subunit, but not with RIα suggesting that the modulation of the effects of CATα on Ca2+ transients is dependent on the regulatory subunit in an isoform specific manner. Targeting CATα to the mitochondrial matrix (mtCATα), we have found that the selective expression of mtCATα specifically reduced mitochondrial Ca2+ transients, suggesting the existence of mitochondrial targets for PKA action inside the mitochondria. All together these results reveal that CATα may translocate to this compartment only upon activation and release from RIIβ subunit. In addition to the work on PKA, during my PhD program, I carried out another project based on a different application of the splitGFP tool. The communication between organelles is important to favour different pathways and its dysregulation are present in a number of different diseases, including neurodegenerative disease. In particular, this methodology was used to characterize the ER-plasma membrane (PM) contact sites. The close contacts between the ER and the PM are required for the mechanism of store-operated Ca2+ entry (SOCE), a process induced as a consequence of the Ca2+ depletion of the ER store and dependent on the dynamic interaction between the ER resident protein stromal interaction molecule 1 that acts as Ca2+ sensor (STIM1) and Orai1, the protein forming the channel in the PM that permits Ca2+ entry from the extracellular ambient. To visualize the ER-PM junctions, we generated a YFP1-10 fragment targeted to the PM and β11 strand was targeted to the ER. We generated a construct where the PM-YFP1-10 and the ERS-β11 or ERL-β11 are cloned in the same bicistronic expression vector (SPLICSS/L ER-PM probes). In a first set of experiments these SPLICS probes detected two types of interactions: long and short ER-PM interactions. Then, we investigated whether and how genetic and pharmacological manipulations could impact on ER-PM interface. We analysed the response of SPLICSS/L ER-PM probes to STIM1/Orai1 downregulation and ER Ca2+ depletion. To this purpose, STIM1/Orai1 proteins were silenced by ShRNA or ER Ca2+ depletion was induced by the incubation with 2,5-tert-butylhydroquinone (THBQ) or thapsigargin inhibitors of the sarcoplasmic/endoplasmic reticulum Ca2+-ATPase and with histamine. We have found that SPLICSS signal (monitoring ER-PM short interactions below 10 nm) decreased when we downregulated STIM1 or Orai1 proteins and strongly increased upon ER Ca2+ depletion. We have also detected the ER-PM long interactions under the same conditions and found that upon STIM1 or Orai1 downregulation the number of the long contacts decreased in respect with the control cells. All together, these data reveal that the SPLICS methodology is able to monitor short and long ranges ER-PM interactions and their changes upon pharmacological/genetic manipulations.