To approach this issue we have been studying the formation of connections in the fruit fly brain. The fly is well suited to the identification of mutants in which connection specificity is disrupted. Importantly, we can easily visualize the patterns of connections made by different neurons and manipulate these cells genetically. Three areas of study that are being actively pursued are described below.

The Dscam gene is alternatively spliced to give rise to a large family of highly related proteins that are biochemically distinct

We have studied a number of genes encoding cell surface proteins that are required for normal patterns of connection specificity in the fly. Among these Dscam remains the most intriguing. Remarkably, the Dscam gene encodes more than 38,000 different proteins. This occurs through a process called alternative splicing. All forms or isoforms of Dscam have the same overall domain structure or shape, but have different amino acid sequences within specific regions of the protein. Do these different isoforms have different recognition specificities and, if so, does this property contribute to connection specificity? We have approached this question using a combined biochemical and genetic approach.

Dscam is required for normal connections to form in many regions of the nervous system. In the olfactory system, for instance, it is required for different subclasses of olfactory neurons to connect to the correct group of cells in the brain that processes olfactory information. It is also required for the construction of a central brain structure called the mushroom body (MB) that is important for learning and memory. Using specific genetic manipulations, we have been able to remove all forms of Dscam from single MB neurons and replace them with only one of the 38000 potential isoforms. Surprisingly, different isoforms can provide substantial activity to fulfill the function provided by the entire collection of isoforms indicating that each isoform is likely to share a core biochemical function. What is this biochemical function and is it modulated by the sequence diversity of Dscam?

Using a biochemical approach we have shown that single isoforms of Dscam interact with each other. These findings indicate that one form of Dscam on the surface of a cell will bind to the same form on the surface of an opposing cell. Remarkably, these interactions are very specific. Indeed, two very closely related isoforms do not interact. Hence, Dscam provides a family of closely related proteins with similar functions that shows literally thousands of binding specificities.

While different isoforms of Dscam exhibit different binding specificities single neurons express multiple isoforms. For instance, in collaboration with Andy Chess’ laboratory at Whitehead Institute at MIT, we have shown that each single MB neuron expresses multiple isoforms of Dscam. Furthermore, each MB neuron expresses a combination of isoforms that is distinct from, though perhaps overlapping with, their neighbors. As such, each MB neuron may exhibit a combination of labels that specifies its unique identity. It remains possible that the specific combination of isoforms in a single neuron is not essential, it is only important for each neuron has a cell surface identity that is different from its neighbors. Presumably the nature of Dscam proteins expressed on the surface of a neuron will influence the nature of its interactions with neighboring cells. How these exquisite binding specificities are related to Dscam function in vivo is a critical issue that we are examining using various classical and molecular genetic approaches.

Formation of Neuronal Connections in the Fly Visual System

We have been studying the formation of the connections between photoreceptor neurons (R cells) and their targets in the brain. The compound eye of the fly contains some 800 simple eyes, or ommatidia, and each ommatidium contains eight R cells. These cells fall into three classes based on synaptic specificity. R1-R6 neurons connect to the first optic ganglion, called the lamina, and R7 and R8 neurons extend axons through the lamina and terminate in two distinct layers in the second optic ganglion, the medulla.

During the last year we have focused largely on genes required for R7 specificity. Three different classes of mutations have been identified. One leads to R7 neurons terminating in the R8 layer instead of R7. Another leads to abnormal connections within the appropriate layer . And the third class appears to make connections in multiple layers in the correct ganglion, the medulla. Alternatively, this class of mutations may disrupt the intracellular transport of synaptic vesicle components.

Mutations in N-cadherin disrupt the connections between R7 and its target. N-cadherin is a cell surface protein that promotes interactions between two closely opposed membranes. Molecular analysis revealed that eight isoforms of Dscam are generated by alternative splicing. To our surprise, one mutation in N-cadherin disrupted alternative splicing such that only four of the eight isoforms were generated in the eye. While R7 neurons lacking these isoforms exhibited a defect indistinguishable from R7 lacking all 8 isoforms, some subclasses of olfactory receptor neurons which require N-cadherin for targeting were unaffected by it. This suggested that different isoforms of N-cadherin may be used to regulate connection specificity in different regions of the brain. Further studies are in progress to critically assess why different isoforms of N-cadherin are required in different classes of neurons.

Formation of Connections in the Fly Olfactory System

During the past year we have studying the mechanisms regulating synaptic target specificity in the developing olfactory system. Here different subpopulations of neurons express different odorant receptors. Neurons expressing the same odorant receptor then connect to the same group of neurons in the brain within an anatomical structure called a glomerulus. There are some 43 different odorant receptors and a similar number of glomeruli. Using specific genetic reagents the wiring specificity of different classes of olfactory receptor neurons can be directly visualized and genetically manipulated. We have studied in detail the requirement for N-cadherin and Dscam in elaborating these connections. Dscam plays a key role in promoting targeting to the correct region of the antennal lobe while N-cadherin functions at a later step to promote the formation of a precursorstructure to the glomerulus called a protoglomerulus. We are currently carrying out genetic screens in the olfactory system to identify additional determinants of synaptic specificity.