Drosophila melanogaster--Cytogenetics.

In the developing CNS, presynaptic neurons often have exuberant overgrowth and
form excess (and overlapping) postsynaptic connections. Importantly, these excess
connections are refined during circuit maturation so that only the appropriate connections
remain. This synaptic rearrangement phenomenon has been studied extensively in
vertebrates but many of those models involve complex neuronal circuits with multiple
presynaptic inputs and postsynaptic outputs. Using a simple escape circuit in Drosophila
melanogaster (the giant fiber circuit), we developed tools that enabled us to study the
molecular development of this circuit; which consists of a bilaterally symmetrical pair of
presynaptic interneurons and postsynaptic motorneurons. In the adult circuit, each
presynaptic interneuron (giant fiber) forms a single connection with the ipsilateral,
postsynaptic motorneuron (TTMn). Using new tools that we developed we labeled both
giant fibers throughout their development and saw that these neurons overgrew their targets and formed overlapping connections. As the circuit matured, giant fibers pruned
their terminals and refined their connectivity such that only a single postsynaptic
connection remained with the ipsilateral target. Furthermore, if we ablated one of the two
giant fibers during development in wildtype animals, the remaining giant fiber often
retained excess connections with the contralateral target that persisted into adulthood.
After demonstrating that the giant fiber circuit was suitable to study synaptic
rearrangement, we investigated two proteins that might mediate this process. First, we
were able to prevent giant fibers from refining their connectivity by knocking out
highwire, a ubiquitin ligase that prevented pruning. Second, we investigated whether
overexpressing Netrin (or Frazzled), part of a canonical axon guidance system, would
affect the refinement of giant fiber connectivity. We found that overexpressing Netrin (or
Frazzled) pre- & postsynaptically resulted in some giant fibers forming or retaining
excess connections, while exclusively presynaptic (or postsynaptic) expression of either
protein had no effect. We further showed that by simultaneously reducing (Slit-Robo)
midline repulsion and elevating Netrin (or Frazzled) pre- & postsynaptically, we
significantly enhanced the proportion of giant fibers that formed excess connections. Our
findings suggest that Netrin-Frazzled and Slit-Robo signaling play a significant role in
refining synaptic circuits and shaping giant fiber circuit connectivity.

Changes in synaptic strength underlie synaptic plasticity, the cellular substrate for learning and memory. Disruptions in the mechanisms that regulate synaptic strength closely link to many developmental, neurodegenerative and neurological disorders. Release site probability (PAZ) and active zone number (N) are two important presynaptic determinants of synaptic strength; yet, little is known about the processes that establish the balance between N and PAZ at any synapse. Furthermore, it is not known how PAZ and N are rebalanced during synaptic homeostasis to accomplish circuit stability. To address this knowledge gap, we adapted a neurophysiological experimental system consisting of two functionally differentiated glutamatergic motor neurons (MNs) innervating the same target. Average PAZ varied between nerve terminals, motivating us to explore benefits for high and low PAZ, respectively. We speculated that high PAZ confers high-energy efficiency. To test the hypothesis, electrophysiological and ultrastructural measurements were made. The terminal with the highest PAZ released more neurotransmitter but it did so with the least total energetic cost. An analytical model was built to further explore functional and structural aspects in optimizing energy efficiency. The model supported that energy efficiency optimization requires high PAZ. However, terminals with low PAZ were better able to sustain neurotransmitter release. We suggest that tension between energy efficiency and stamina sets PAZ and thus determines synaptic strength. To test the hypothesis that nerve terminals regulate PAZ rather than N to maintain synaptic strength, we induced sustained synaptic homeostasis at the nerve terminals. Ca2+ imaging revealed that terminals of the MN innervating only one muscle fiber utilized greater Ca2+ influx to achieve compensatory neurotransmitter release. In contrast, morphological measurements revealed that terminals of the MN inner vating multiple postsynaptic targets utilized an increase in N to achieve compensatory neurotransmitter release, but this only occurred at the terminal of the affected postsynaptic target. In conclusion, this dissertation provides several novel insights into a prominent question in neuroscience: how is synaptic strength established and maintained. The work indicates that tension exists between energy efficiency and stamina in neurotransmitter release likely influences PAZ. Furthermore, PAZ and N are rebalanced differently between terminals during synaptic homeostasis.