Despite a voluminous knowledge concerning the ultrastructural and molecular composition of active zones, very little is known about how these presynaptic specializations are assembled. Furthermore, with a few exceptions little is known about the roles that active zone molecules play in synaptic function. In this thesis, I describe the identification and characterization of Drosophila mutants for a newly-identified serine-arginine protein kinase, which has been termed serine-argining protein kinase at 79D (SRPK79D).

In Chapter 2 I describe a series of experiments involving quantitative imaging, electron microscopy and electrophysiology. The results of these experiments indicate that SRPK79D is a negative regulator of T-bar assembly and have lead to a new model of active zone assembly. Specifically, the data suggest that in wild type animals active zone constituents are transported from the cell soma to the synapse where they are locally assembled. During transport assembly of these components must be constitutively repressed through the action of SRPK79D to prevent ectopic T-bar formation.

In Chapter 3 I present the initial characterization of a synaptic function deficit at the neuromuscular junction of srpk79D mutants. In otherwise wild type animals, over-expression of the vesicular glutamate transporter (VGlut) leads to an increase in the amplitude of spontaneous fusion events (minis) and an inversely-proportional, compensatory decrease in the average number of vesicles that fuse in response to presynaptic stimulation. Thus, through this process, which we term reverse homeostatic compensation, the average postsynaptic response to presynaptic stimulation in VGlut over-expressing animals is identical to animals expressing normal levels of VGlut. In srpk79D mutants, VGlut over-expression causes an identical increase in the average mini amplitude. Reverse homeostatic compensation also ensues. However, compensation is excessive. Consequently, the average postsynaptic response to presynaptic stimulation is approximately one-half that of wild type animals or animals that over-express VGlut but are wild type for srpk79D.

To date, srpk79D is the first gene shown to be involved in reverse homeostatic compensation. Furthermore, srpk79D is the first gene known to be required to dampen homeostatic compensation. Finally, the data suggest that destabilization of the T-bar may be required for reverse homeostatic compensation, which would represent a novel mechanism in modulation of synaptic efficacy.