A combined genetic, anatomical, and behavioral approach has been undertaken to study the developmental and functional plasticity of identified bristle mechano-sensory neurons in Drosophila. A stereotyped grooming reflex in decapitated flies enabled simple but reliable assessments of the functional output of individual bristle sensory cells to correlate with their axonal projections and terminal arbors revealed by the cobalt backfill technique. Construction of small-patch mosaics that contain only a single mutant bristle allowed functional perturbation of individual neurons within an otherwise normal environment. Mutations that affect nerve excitability and membrane recycling have been used to examine their effects on neuronal pathfinding, arborization, and the initiation and maintenance of functional connections. Previous studies (Burg and Wu, 1986, J. Neurosci. 6:2968-2976; 1989, Dev. Biol. 131:505-514) have demonstrated that para(ts)nap(ts) double-mutant sensory neurons, in which action potentials are unconditionally blocked by defects in sodium currents, and eag Sh double-mutant sensory cells, in which membrane excitability is increased by alterations in potassium currents, can establish and maintain central projections that are indistinguishable from their functionally normal counterparts. Mutations of the shi(ts) gene cause a temperature-sensitive, reversible block of the membrane recycling process, resulting in arrest of neuronal growth in culture (Kim and Wu, 1987, J. Neurosci. 7:3245-3255) and depletion of synaptic vesicles that leads to transmission blockade at established synapses (Ikeda, Ozawa, and Hagiwara, 1976, Nature 259:489-491; Koenig and Ikeda, 1983, J. Neurobiol. 14:411-419; 1989, J. Neurosci. 9:3844-3860). Prolonged heat treatments (up to 16% of total development time) of small-patch shi(ts) mosaics at different pupal stages did not prevent the establishment of central projections characteristic of the various sensory cell types. However, none of the shi(ts) sensory neurons heat-pulsed during the initial or the final 16% of pupal development were able to initiate the reflex behavior, although a proportion of those treated in other periods apparently established functional contacts with appropriate targets to support the characteristic cleaning reflex. The possibility exists that the membrane recycling process blocked in shi(ts) cells provides a crucial mechanism for cell-cell interactions taking place during initial differentiation and final synaptic stabilization, and possibly competition, in the developing sensory neuron. Heat treatments of adult shi(ts) mosaics blocked the reflex initiated by the mutant (but not the surrounding normal) bristles, as expected from the effect of synaptic block. For the majority of bristle types examined, a 2- to 4-h heat pulse blocked the reflex behavior for several days in younger mosaics but caused permanent loss of the reflex in mosaics older than 4 to 5 days of age. This temporal transition correlates well with the age dependence of activity-related modification of nervous system structures (Technan, 1984, J. Neurogenet. 1:1 13-126; Kral and Meinertzhagen, 1989, Philos. Trans. R. Soc. Lond. [Biol.] 323:155-183) and visual discrimination behavior (Mimura, 1986, Science 232:83-85; 1987, Exp. Biol. 46:155-162; Hirsch, Potter, Zawierucha, et al., 1990, Vis. Neurosci. 5:281-289) and may signify a ''critical period'' phenomenon in young adults. The results obtained from small-patch mosaics indicate that electric activity intrinsic to the sensory neuron may not be required for its pathfinding and ramification of terminal arbors but may be important for establishing and maintaining its functional connectivity. Chemical cues either laid down by its neighboring cells or emitting from its target may guide the sensory cell's navigation and sustain its arbor since nonfunctional projections of para(ts)nap(ts) sensory cells and heat-treated shi(ts) neurons persisted for at least 6 days in the mosaics examined.
