C. elegans can sense a range of temperatures and migrates toward the temperature at which it was cultivated with food for several hours on a temperature gradient. Laser ablation experiments revealed that thermosensory neuron AFD and its downstream interneurons, AIY and AIZ, are critical for this thermotaxis behavior (1). To date, several genes required for thermotaxis have been identified. The
tax-4 and
tax-2 genes encoding alpha and beta subunits of cyclic nucleotide-gated channel, respectively, are expressed in several sensory neurons including AFD, and both
tax-4 and
tax-2 mutants exhibit athermotactic (non-temperature-responsive) phenotype (2,3). The
ttx-3 gene encodes a LIM homeodomain protein that is specifically expressed in AIY, and
ttx-3 mutants show cryophilic (cold-seeking) phenotype (4). Like
ttx-3 mutants,
ttx-1 and
ttx-2 mutants are cryophilic and are nearly, if not completely, normal in other behaviors such as chemotaxis. Since AFD and AIY are likely to be involved specifically in thermotaxis, we tried to observe any defects in AFD and AIY neurons of
ttx-1 and
ttx-2 mutants, by introducing AFD-specific marker
gcy-8::GFP and AIY-specific
ttx-3::GFP. The same constructs were also introduced to
ttx-3,
tax-2 and
tax-4 mutants for comparison. We found that the expression of
gcy-8:GFP in AFD was significantly down-regulated in
ttx-1,
tax-2 and
tax-4 mutants. Noticeably, we have also observed that the microvillus-like structures at sensory ending of AFD were often abnormal in
ttx-1 mutants, which is consistent with the previous report based on the EM analysis (5). By contrast,
ttx-2 and
ttx-3 mutants expressed
gcy-8:GFP at the same level as wild type animals. The expressions of
ttx-3::GFP in AIY of
ttx-1,
ttx-2,
tax-2 and
tax-4 mutants were all indistinguishable when compared with that of wild type, but as already reported, it was much reduced in
ttx-3 mutants, probably due to self-regulation (4). We are intrigued by the low expression of
gcy-8::GFP in AFD of
tax-2 and
tax-4 mutants, since we think it highly unlikely that a cyclic nucleotide-gated channel plays a direct role in transcription of
gcy-8 encoding a guanylyl cyclase (6). Thus, the present study suggests that the
ttx-1 gene functions in AFD, but does not necessarily support the possibility that it encodes a transcription regulator over other possibilities. We are now testing the model that disruption of a gene participating in signal transduction pathways for thermotaxis in AFD, AIY or AIZ, leads to down-regulations of other genes involved in the same or related signal transduction cascades. If so, appropriate GFP markers will then be used not only to identify the cellular focus, but also to help estimate the function of the genes affected in thermotaxis-defective mutants. We thank Barbara Wedel and David Garbers for
gcy-8::GFP ; Oliver Hobert for
ttx-3::GFP; Kazuyoshi Ishii, Megumi Shirakawa and Yasumi Ohshima for ttx strains. We are indebted to Yasumi Ohshima for help in his laboratory at the initial stage of this work.