Identifying olfaction's ‘other channels’

The molecular identity of a calcium-activated chloride channel involved in olfactory transduction has been elusive. New evidence shows that an initial candidate, bestrophin-2, cannot be the olfactory channel. Additional studies now identify anoctamin 2 (TMEM16B) as a strong candidate. Olfactory sensory neurons (OSNs) transduce odorous chemical stimuli to electrical changes that can initiate the perception of a smell. In vertebrates, the transduction occurs in the membranes of chemosensory cilia that extend from the OSNs into the nasal mucus (reviewed in Kleene, 2008). Odor molecules contacting a cilium can activate a G-protein-coupled cascade that produces cAMP within the cilium. cAMP can open membrane channels that allow Ca2+ to enter the cilium and depolarize the neuron. Although this would seem sufficient to excite the OSN, another significant step follows. As Ca2+ accumulates within the cilium, it can activate a second population of channels. These channels allow Cl− to exit the cilium, further depolarizing the neuron. In mammalian OSNs during strong stimulation, 80 to 90% of the receptor current flows through the Cl− channels. The first of the two transduction channels, the olfactory cyclic-nucleotide-gated (CNG) channel, was discovered in 1987 (Nakamura & Gold, 1987). Soon the classical biophysical properties of the channel had been thoroughly described. Channels activated directly by cAMP or indirectly by odor stimulation had very similar properties (Firestein et al. 1991). On that basis, it was reasonable to believe that the channel gated by cAMP accounted for the odor-induced depolarization of the OSN. However, much more compelling evidence followed the molecular identification of a CNG channel subunit protein (Dhallan et al. 1990). The olfactory epithelia of knockout mice lacking this protein were almost totally insensitive to most odors tested (Brunet et al. 1996). The second channel in olfactory transduction, a Ca2+-activated Cl− channel, was discovered in 1991 (Kleene & Gesteland, 1991). Again, the basic properties of the channel were promptly described. Unlike the CNG channel, though, the olfactory Cl− channel has remained one of unknown molecular identity. Progress in the molecular genetics of Cl− channels has been hard won. (As recently as 1993, even the Biophysical Society herded posters on Cl− channels into a session called ‘Other Channels’.) Attempts to clone the olfactory Cl− channel by homology with Ca2+-activated Cl− channels of the CLCA family were unsuccessful. However, in 2006 Pifferi et al. presented a case that bestrophin-2 (Best2) might be the olfactory Cl− channel. Best2 can function as a Ca2+-activated Cl− channel when exogenously expressed, although its role in vivo has been controversial (Hartzell et al. 2009). Pifferi et al. (2006) showed that Best2 is expressed in the proximal sections of the olfactory cilia. Furthermore, the olfactory Cl− channel and expressed Best2 channels have similar biophysical properties (anion selectivities, unitary conductances, voltage dependences, sensitivities to blockers). The authors also highlighted one difference: Best2 is about 12 times more sensitive to Ca2+ than is the olfactory channel. Such differences can simply be due to the absence of additional subunits or modulators in the expression system. On the whole, it was reasonable to propose that Best2 might be the olfactory Cl− channel. In this issue of The Journal of Physiology, Pifferi et al. present new evidence that refutes their former hypothesis. When knockout mice lacking Best2 became available, the authors compared them to wild-type mice and found no significant differences in their olfactory physiologies. In both intact epithelia and isolated OSNs, the electrical responses to odors and olfactory second messengers have no requirement for Best2. The sensory endings of OSNs lacking Best2 have Ca2+-activated Cl− currents like those in the wild-type. Despite its similar properties, Best2 cannot account for the odor-induced Cl− current. As it happens, this important negative finding has developed concurrently with a positive one. New evidence suggests that the olfactory Cl− channel is anoctamin 2 (ANO2, also known as TMEM16B). In 2005, TMEM16B topped a list of proteins preferentially expressed in mouse OSNs (Yu et al. 2005; see also Stephan et al. 2009). Within the olfactory epithelium, its transcript was shown to be specific to OSNs. At that time, though, TMEM16B was a putative membrane protein of unknown function. Only after it was reported that anoctamin/TMEM16 proteins form Ca2+-activated Cl− channels (reviewed in Hartzell et al. 2009) did ANO2/TMEM16B emerge as a likely olfactory transduction channel. TMEM16B was identified in the proteome of rat olfactory cilia (Mayer et al. 2009). An ANO2::EGFP construct traffics to olfactory cilia when delivered via an adenovirus (Stephan et al. 2009). When expressed in a mammalian cell line, ANO2 has properties that resemble those of the olfactory Cl− channel (Pifferi et al. 2009a; Stephan et al. 2009). This may sound like the Best2 story all over again: a ciliary protein with properties resembling the olfactory Cl− channel. However, ANO2 is an even closer match. Unlike Best2, ANO2 in an expression system has about the same Ca2+ sensitivity as the olfactory Cl− channel (Pifferi et al. 2009a; Stephan et al. 2009). If ANO2 can be reliably knocked out or knocked down, the ANO2 hypothesis will probably survive the test that sank Best2. Better still, a mouse lacking the olfactory Cl− current can tell us whether that current is required for normal olfactory behaviour. Although the Cl− current is typically the great majority of the receptor current, the current through the CNG channels alone seems more than sufficient to lead to action potentials. For now, the functional importance of the Cl− channel remains a matter of speculation.