Tito A. Serafini
Synaptic Specificity in the CNS
We seek to understand the molecular basis of syna ptic specificity and neuronal identity in the mammalian central nervous system. We use a combination of biochemistry, molecular biology, embryology, and transgenic mouse technology to identify and study molecules mediating postsynaptic target recognition and the initiation of synaptogenesis in the developing brain and spinal cord. Furthermore, because target specificity is intimately linked with neuronal identity, we are identifying expressed genes serving as markers for the many different neuronal types found in the cerebral cortex. Through this endeavor and our in vitro analysis of neuronal recognition and targeting, we are assembling a molecular Utoolbox^ that will eventually allow us to manipulate neuronal connectivity in the mammalian brain, thereby permitting us to determine how specific neuronal connections contribute to circuit development, information processing, and systems level phenomena such as learning and behavior.
In our earlier identification of the netrins, endogenous chemoattractants for developing axons, we isolated and characterized the proteins using functional, in vitro assays to quantitate netrin activity. Several new projects in the laboratory employ the same basic strategy of in vitro reconstitution to characterize and identify molecules mediating cell-cell recognition and the initiation of synaptogenesis in several different yet complementary systems.
Identification of motoneuron recognition molecules. During development, sensory axons, extending from cell bodies located within dorsal root ganglia, grow into the spinal cord, entering on its dorsal side. The axons relaying nociceptive and thermoceptive information do not grow beyond the dorsal region of the spinal cord; however, axons relaying proprioceptive information grow to reach the motor columns in the ventral region of the spinal cord, where they synapse directly with motoneurons. Because we can purify motoneurons and elicit proprioceptive sensory axon growth selectively in culture, we have been able to reconstitute aspects of this targeting event in vitro. We are now using this reconstituted system to assay for molecules important in synapse formation when motoneurons are the postsynaptic partner.
Identification of a cerebellar granule cell recognition signal. The cerebellar cortex contains only five or six neuronal cell types, several of which can be isolated as relatively pure populations. Granule cells, the most abundant neuronal cell type in the cerebellar cortex (comprising approximately 90% of the neurons), normally receive synaptic input from mossy fibers projecting from the pons. In vitro, mossy fiber axons grow out from explants of basilar pons. If these pontine explants are cultured on a field of purified granule cells, however, mossy fiber growth is arrested upon contact with the granule cells, providing evidence for a growth-arrest signal on the granule cells that is recognized by their presynaptic partners. We are in the process of isolating the molecule or molecules on granule cells responsible for delivering this target recognition signal.
Identification of molecules mediating subcellular specificity in synaptic site selection. Not only are specific postsynaptic partners chosen by particular presynaptic neurons, but often specific places on those particular partners are chosen for the site of synapse formation. For example, the basket cells of the cerebellar cortex form synapses predominantly on the cell bodies and initial axon segment, rather than the dendrites, of Purkinje cells, the major output cell of the cerebellar cortex. We are developing a basket cell purification procedure in order to reconstitute this subcellular site selectivity in vitro and to identify the Purkinje cell molecules that mediate such exquisite selectivity.
A functional genomics of neuronal identity in the cerebral cortex. The aforementioned targeting phenomena yield to an in vitro analysis at present because the different neurons involved include some of the few mammalian central nervous system neurons for which molecular markers, usually expressed genes, have been identified and characterized. Unfortunately, a significant impediment to such studies in the cerebral cortex is the paucity of such markers useful for uniquely identifying and isolating different types of cortical neurons. Not only are such markers important for the eventual isolation of different classes of cortical neurons, but they are also necessary for a description of cortical architecture and development that does not rely upon extrinsic descriptive criteria. In addition, the identification of such uniquely expressed genes will allow for selective, cell-type specific manipulations of cortical architecture through transgenic mouse technology. In order to obtain these markers, we are screening arrayed cortical cDNA libraries in a highly multiplex fashion with amplified mRNA representations obtained from single or a few identified cortical neurons.