Information about perturbations in the internal and external environment are perceived and processed by sensory systems, which in turn regulate neuronal circuits or central pattern generator(s) to give rise to spontaneous rhythmic behaviors. Here we introduce a paradigm for studying such sensori-motor control: long -term swimming behavior of C.elegans in liquid (M9 media). Initially when shifted from agar to M9 worms continue to swim (S) with 4 body bends/second. They swim continuously for 88+21 (Std. Dev.) minutes before going into quiescent (Q) state, when they lie motionless. After ~5 minutes they spontaneously start to swim when they oscillate for a long time between S and Q bouts. The 2nd S bout is of ~12 min, after which the S bouts gradually decrease to 5 min, while the Q bouts remain invariant at 5 min. These Q periods are not due to fatigue, as there was no correlation of length of first Q and length of first S periods, implying a neural mechanism. This striking ultradian rhythmic behavior in M9 raises several questions: What is the nature of the switch from S to Q? Why length of the first S period is different from subsequent ones? What terminates the nearly invariant Q periods? To answer such questions we noted that the assay was done at 25C, without food and in liquid. Thus the worms, raised at 20C on plates with food, faced at least three altered sensory modalities: food, temperature and internal pressure. We perturbed the genetic and environmental variables to probe into the nature of the S and Q states and the switch. The most critical factors for timing the length of the first S period were difference between rearing and assay temperatures, presence of extra-cellular Ca2+ and food. Consistent with the sensory input requirement, we found mutants affecting cilia development in several sensory neurons, e.g.
osm-5,
osm-6 and
che-2, switch to Q much earlier than wild type and remain in Q state for a long period. In accord with the extracellular Ca2+ requirement, we found that mutations in calcineurin subunits,
tax-6 and
cnb-1, and TRP (Ca2+) channel subunits,
tax-2 and
tax-4, result in an early switch to Q state. In contrast loss of function mutations in
egl-4, a serine-threonine kinase known to target
tax2/4, result in reduction of Q bouts from 5 to 2 minutes. Hence this gene promotes Q. By genetically depleting serotonin (5HT), using
tph-1 and
nss-1 mutants, we found that endogenous 5HT promotes S.
cat-2 mutants, which lack dopamine (DA), have the opposite effect, implying an antagonistic action. Our analysis of sensory mutants identified candidate neurons that regulate the S-Q pattern. We are currently confirming these candidates by laser ablations and epistasis analysis.