Brain function relies on fast and precisely timed synaptic vesicle (SV) release at active zones (AZs). Efficacy of SV release depends on distance from SV to Ca(2+) channel, but molecular mechanisms controlling this are unknown. Here we found that distances can be defined by targeting two unc-13 (Unc13) isoforms to presynaptic AZ subdomains. Super-resolution and intravital imaging of developing Drosophila melanogaster glutamatergic synapses revealed that the Unc13B isoform was recruited to nascent AZs by the scaffolding proteins Syd-1 and Liprin-α, and Unc13A was positioned by Bruchpilot and Rim-binding protein complexes at maturing AZs. Unc13B localized 120 nm away from Ca(2+) channels, whereas Unc13A localized only 70 nm away and was responsible for docking SVs at this distance. Unc13A(null) mutants suffered from inefficient, delayed and EGTA-supersensitive release. Mathematical modeling suggested that synapses normally operate via two independent release pathways differentially positioned by either isoform. We identified isoform-specific Unc13-AZ scaffold interactions regulating SV-Ca(2+)-channel topology whose developmental tightening optimizes synaptic transmission.

Highlights d Spermidine supplementation age-protects Drosophila brain mitochondria d Brain hypusination levels decay with age but are boosted by spermidine supplementation d Mitochondrial functionality is defective after genetic attenuation of hypusination d Defective hypusination compromises spermidine effects on locomotion and memory

Drosophila is widely used as a genetic model in questions of development, cellular function and disease. Genetic screens in flies have proven to be incredibly powerful in identifying crucial components for synapse formation and function, particularly in the case of the presynaptic release machinery. Although modern biochemical methods can identify individual proteins and lipids (and their binding partners), they have typically been excluded from use in Drosophila for technical reasons. To bridge this essential gap between genetics and biochemistry, we developed a fractionation method to isolate various parts of the synaptic machinery from Drosophila, thus allowing it to be studied in unprecedented biochemical detail. This is only possible because our protocol has unique advantages in terms of enriching and preserving endogenous protein complexes. The procedure involves decapitation of adult flies, homogenization and differential centrifugation of fly heads, which allow subsequent purification of presynaptic (and to a limited degree postsynaptic) components. It is designed to require only a rudimentary knowledge of biochemical fractionation, and it takes ∼3.5 h. The yield is typically 4 mg of synaptic membrane protein per gram of Drosophila heads.

Assembly and maturation of synapses at the Drosophila neuromuscular junction (NMJ) depend on trans-synaptic neurexin/neuroligin signalling, which is promoted by the scaffolding protein Syd-1 binding to neurexin. Here we report that the scaffold protein spinophilin binds to the C-terminal portion of neurexin and is needed to limit neurexin/neuroligin signalling by acting antagonistic to Syd-1. Loss of presynaptic spinophilin results in the formation of excess, but atypically small active zones. Neuroligin-1/neurexin-1/Syd-1 levels are increased at spinophilin mutant NMJs, and removal of single copies of the neurexin-1, Syd-1 or neuroligin-1 genes suppresses the spinophilin-active zone phenotype. Evoked transmission is strongly reduced at spinophilin terminals, owing to a severely reduced release probability at individual active zones. We conclude that presynaptic spinophilin fine-tunes neurexin/neuroligin signalling to control active zone number and functionality, thereby optimizing them for action potential-induced exocytosis.