How neurons acquire a diverse repertoire of structural and functional properties in order to establish working neural circuits remains poorly understood. To address this question, we use the T4/T5 neurons of the Drosophila visual system as a model because of their genetic accessibility and the extensive knowledge we have about their morphology and physiology. The dendrites of T4/T5 neurons represent the first processing stage in the fly visual system in which the direction of image motion is encoded. While T4 neurons respond to motion of brightness increments, T5 neurons detect motion of brightness decrements (Maisak et al., 2013). This specialization results from their dendrites arborizing in different neuropil regions (T4 dendrites in the medulla, T5 dendrites in the lobula) and therefore receiving synaptic input from different neuronal types. As a common attribute of T4 and T5 neurons, their dendrites and axons are confined to single synaptic layers (Fischbach and Dittrich, 1989). We have recently found that two members of the SOX family of transcription factors are required for establishing this layer-specific innervation during the maturation of T4 and T5 neurons (Schilling et al., 2019), which is key for the proper function of the fly motion sensing circuits.
Interestingly, both T4 and T5 neurons can be further subdivided into four subtypes, each responding to motion in one of the four cardinal directions (Maisak et al., 2013). Differences between T4/T5 neuron subtypes in directional tuning are thought to result from differences in the spatial organization of synapses they receive from an identical set of input neurons (Arenz et al., 2017; Takemura et al., 2017). How do these differences emerge during development? Since the orientation of each T4 and T5 dendritic arbor aligns with the neuron´s preferred direction of motion (Takemura et al., 2017), our working hypothesis is that the direction in which a T4/T5 dendrite grows during development determines the arrangement of its synaptic inputs and, thus its direction selectivity. Our current work attempts to understand the cellular rules and underlying molecular mechanisms by which the four T4/T5 neuron subtypes acquire their distinctly oriented dendritic arbors. For this, we first generate tools that allow genetic access to the different T4/T5 neuron subtypes during development. We perform time-lapse imaging experiments of individually labeled T4 and T5 dendrites during development to characterize their growth patterns and to propose cellular rules that account for them. We will test these rules by different manipulations of the system at the cellular level. In parallel, we are transcriptionally profiling the distinct T4/T5 neuron subtypes in order to find molecular candidates that account for their different behaviors during development. In the future, we plan to test these candidates by loss and gain of function analyses.
Understanding how T4/T5 neurons acquire their structural attributes during development will further help to link morphology, function and development in a single neuronal population, as well as to manipulate specific neuronal structural properties to reveal their role in neuronal computation.