Mechanosensation underlies multiple senses, such as touch, pain, hearing, and proprioception. The molecules that mediate most of the mechanical senses have not been identified. Genetic and molecular methods have identified several putative mechanosensitive proteins. However, how the mechanotransduction machineries organize and function remains largely unknown. To understand the organization of the mechanotransduction complex, I studied the DEG/ENaC mechanosensory channel that detects gentle touch in the six touch receptor neurons (TRNs) of C. elegans. Previous studies from our lab have suggested that this channel complex contains two pore-forming subunits MEC-4 and MEC-10 (DEG/ENaC proteins) and two auxiliary subunits MEC-6 (paraoxonase-like protein) and MEC-2 (stomatin-like protein). However, questions remain about what molecules really constitute this mechanosensory channel complex. Studying this particular DEG/ENaC channel in C. elegans will not only elucidate the organization of one major mechanosensory complex, but also improve our knowledge of other DEG/ENaC proteins, which are found in both vertebrates and invertebrates, and involved in various functions, e.g. mechanosensation, sodium taste, acid sensation, synaptic plasticity, and sodium homeostasis.

My thesis research investigated the molecular organization and formation of the DEG/ENaC mechanosensory channel in C. elegans. In collaboration with Ehud Isacoff's lab, I analyzed the stoichiometry and co-localization of the potential channel subunits using single molecule optical imaging. In Xenopus oocytes, MEC-4 and MEC-10 form trimers, either of MEC-4 alone or of MEC-4 and MEC-10 in a ratio of 2:1. MEC-2 and MEC-6 do not seem to colocalize with the MEC-43 or MEC-42MEC-10 trimers at the single molecule level, and thus, may not be part of the channel complex. To study the role of MEC-6, I characterized its homologous protein POML-1. Compared to MEC-6, POML-1 appears to play a similar but relatively minor role in the TRNs. As with mec-6, loss of poml-1, completely suppressed mec-4(d) induced neuronal degeneration. [mec-4(d) encodes a hyperactive channel and causes neuronal degeneration in vivo]. Loss of poml-1 alone had no effect, but in sensitized background, it completely abolished touch sensitivity.

Surprisingly, most of MEC-6 and POML-1 proteins were found in the endoplasmic reticulum (ER), rather than on the plasma membrane, consistent with the finding in Xenopus oocytes that MEC-6 is not part of the MEC-4 mechanosensory channel. I provided several lines of compelling evidence to demonstrate that MEC-6 and POML-1 are required for MEC-4 folding and transport, and likely function as ER chaperones. First, loss of these proteins dramatically reduced MEC-4 protein level, eliminated the punctate distribution of MEC-4 in the neuronal process, and altered the MEC-4 folding status in the TRNs. These phenotypes are also shared by calreticulin (CRT-1), a chaperone in the ER. Second, MEC-6 also substantially increased MEC-4 surface expression in Xenopus oocytes, though POML-1 and CRT-1 did not have the same effect in oocytes. Third, overexpressing a transport protein, SEC-24, partially rescued the transport defects caused the poml-1 and crt-1 mutations. Based on the finding that loss of poml-1 reduces MEC-4 protein levels and suppresses neurodegeneration caused by the hyperactive MEC-4(d) channel, I used the poml-1 deletion as a sensitized background to identify genes that normally inhibit MEC-4(d) neurotoxicity through a genetic screen.

I found that the loss of two genes, mec-10 and C49G9.1, makes mec-4(d) more toxic. The proteins encoded by these genes affect mec-4(d) neurotoxicity through different mechanisms. MEC-10 inhibits MEC-4(d) without affecting MEC-4 surface expression. In contrast, both in vivo and in vitro data suggested that C49G9.1, a membrane protein specific to nematodes, can reduce MEC-4 surface expression, which contributes to, at least in part, its inhibitory effect on MEC-4(d). C49G9.1 does not incorporate into the MEC-4/MEC-10 channel, though they may transiently interact, because C49G9.1 did not appear to co-localize with MEC-4 either in vivo or in vitro, but co-immunoprecipitated with MEC-4. In summary, my doctoral research has refined the model of the MEC-4/MEC-10 complex. In particular, my studies resolved the subunits composition of DEG/ENaC channel at the single molecule level, by showing that they form MEC-42MEC-10 trimers in Xenopus oocytes. Notably, MEC-2 and MEC-6 may not be part of the complex.

Indeed, I provided compelling evidence to demonstrate that MEC-6 and POML-1 are needed for MEC-4 folding and transport, and likely function as chaperones and/or assembly factors. In addition, I identified a novel membrane protein, C49G9.1, which negatively regulates MEC-4 surface expression and/or activities. This work has revised our understanding of a major mechanosensory complex and described a new class of chaperone proteins as well as a new inhibitor protein for DEG/ENaC proteins.