Previous work in our lab used cell sorting and cDNA microarrays to identify
mec-3- dependent genes in embryonic touch receptor neurons (Zhang et al. Nature 418: 331-335, 2002). Those experiments required 2 X 106 cells of which 50% were GFP-positive. We have modified these methods in two ways so that we now make sufficient RNA from only 25,000 cells that are 90-95% GFP-positive. First, by characterizing the fluorescence profile of non GFP-expressing cells, we could exclude autofluorescent cells from our selection. Second, by using Ambion mRNA isolation and amplification kits, we could obtain sufficient RNA for the arrays from much less starting material. The RNA was then used on Affymetrix arrays. Using this approach we identified 201 genes differentially expressed (>4 times upregulated) in the touch receptor neurons. All 25 genes already known to be expressed in these cells were included (usually with very high scores) on our list of upregulated. Genes that were either high on the list or were interesting candidates (GABA and ACHR receptor subunits, Ca++ binding proteins, potassium channel subunits, a paraoxonase homolog) were further characterized by making promoter or protein fusions with GFP. 13 of 16 new genes tested were expressed in these cells, including one, the putative receptor C49G9.1, which co-localizes with the mechanosensory channel. We have used bacterial RNAi and the RNAi-hypersensitive strains
lin-35 and
mec-4ts;
lin-35 to demonstrate that nine of the genes, e.g., C49G9.1 and the Ca++sensor C16H3.1, were needed for optimal touch sensitivity. We used a similar method to identify genes selectively expressed in the FLP neurons, a set of
mec-3 expressing mechanosensory neurons in the head. Because an FLP specific promoter was not known, we isolated the cells from a line expressing P<sub>
mec-3</sub>gfp in which the touch cells were killed using a dominant
mec-4 mutation (in embryos
mec-3 is expressed only in touch neurons and FLP). GFP-positive cells in culture from these embryos were morphologically distinct from touch receptor neurons, suggesting that the touch neurons had been eliminated; the absence of touch neurons was confirmed using an antibody against the touch receptor-specific protein MEC-18. 398 genes were found to be differentially expressed (>4 times upregulated) in the FLP cells. Surprisingly, RNAs for several mec genes, including
mec-18, were expressed in these cells. These results suggest that post-transcriptional control may prevent protein production from these genes in the FLP cells. We are currently confirming the expression pattern of genes of our candidates, one of which is the DEG/ENaC channel gene
asic-1.