Anumber of observations have lent support to a model in which thermal stress is transduced into a signal at the level of the cellular membranes. Our alternative, but not exclusive, approach is based on the concept that the initial stress-sensing events are associated with the physical state and lipid composition of cellular membranes, i.e., the subtle alteration(s) of membrane fluidity, phase state, and/or microheterogeneity may operate as a cellular thermometer. In fact, various pathological states and aging are associated with typical "membrane defects" and simultaneous dysregulation of heat shock protein synthesis. The discovery of nonproteotoxic membrane-lipid interacting compounds, capable of modulating membrane microdomains engaged in primary stress sensing may be of paramount importance for the design of new drugs with the ability to induce or attenuate the level of particular heat shock proteins.