Abstract
The chemical synapse is the central point of communication in neuronal systems. The strength of individual synapses is constantly adapted in response to different levels of neuronal activity. This plastic adaptation happens in time scales of milliseconds to hours and even days. So far, the knowledge about the involved molecular processes is limited. Potential mechanisms that could contribute are post-translational modifications, such as phosphorylation.
A key player in synaptic transmission and synaptic plasticity is RIM1α, a multi-domain protein located in the cytomatrix at the active zone (CAZ). RIM1α is fundamentally involved in calcium channel clustering, vesicle to calcium channel coupling, vesicle docking, priming and synaptic plasticity. Additionally, it was proposed that RIM1α and protein kinase A (PKA) centrally participate in the generation of presynaptically mediated long-term potentiation. Here, we applied mass spectrometry (MS) together with phosphoenrichment, biochemical methods and live cell imaging to decipher the importance of RIM1α phosphorylation for synaptic transmission.
We used different imaging tools (FM dyes and iGluSnFR) and developed screening assays for RIM1α phosphorylation sites with functional relevance for synaptic transmission. To this end we verified in FM dye imaging experiments that the knock-out of RIM1α or the ablation of all large isoforms of RIM lead to a reduced release probability that can be rescued by expression of a N-terminally GFP-tagged RIM1α (GFP-RIM1α). Using bioinformatics and phospho-proteomics of stimulated hippocampal neurons we identified a set of potential phosphorylation sites in RIM1α and mutated these to phospho-deficient and phospho-mimetic GFP-RIM1α variants. The mutated GFP-RIM1α variants were expressed in RIM1α knock-out or RIM1/2 conditional double knock-out (cDKO) neurons and rescue efficacy of synaptic release was investigated.
Three of the identified sites failed to rescue the reduced release probability when rendered phospho-deficient, while one site increased the synaptic release. Furthermore, we show that GFP-RIM1α variants that carry a release relevant mutation, are located in putative synaptic structures, but that their persistence in the cytomatrix at the active zone is changed. This could point to altered protein-protein interactions at the active zone.
One protein that preferentially bound to phosphorylated RIM1α was Serine/Arginine-rich protein specific kinase 2 (SRPK2). A more detailed investigation of SRPK2 function revealed that this kinase is involved in modification of neurotransmitter release in a RIM dependent manner. We propose that the strength of synaptic transmission correlates with the level of SRPK2 in the synapse. We identified three phosphorylation sites in RIM1α that could be necessary to act as phospho-switches to set the SRPK2 dependent synaptic release probability.
Taken together, our data suggest an essential function of RIM1α phosphorylation for synaptic vesicle release. We could identify several functionally relevant phosphorylation sites in RIM1α and we have evidence that these potentially affect the dwell time of RIM1α in the CAZ, probably by changing protein-protein interactions. Finally, we identified SRPK2 as novel kinase in the presynapse that interacts with RIM1α and is involved in synaptic transmission.