Cool channel subunits reveal their independent interactions with menthol

Menthol is a well-known and widely used compound; it can be found in chewing gums, sweets, toothpaste, mouth wash, after-shave gels, cigarettes, cocktails, body wash, cough drops, etc. It owes its popularity to the cooling feeling it evokes if it reaches the sensory nerve endings in either mucous membranes or the skin. This cool feeling is not due to actual decreased temperatures, but rather to directly activating one of the major molecular sensors of environmental cold temperatures, the transient receptor potential melastatin 8 (TRPM8) ion channel. TRPM8 is activated by cold temperatures in sensory neurons; its opening leads to depolarization and action potentials, which conveys the information of decreased environmental temperatures to higher centres. Menthol also opens TRPM8, and thus activates the same sensory neurons as cold, leading to the well-known cool sensation even if temperature does not change.

TRPM8 is a member of the transient receptor potential (TRP) ion channel superfamily, many members of which are widely considered to be sensory receptors that detect chemical and physical changes in our environment (Voets et al. 2005). Their role is best established in sensing environmental temperatures; TRPM8 is one example. Other sensory TRP channels, such as TRPV1 or TRPV3 are activated by hot or warm temperatures. It is quite intriguing that these heat sensors also have naturally occurring chemical agonists that can evoke the feeling of either painful heat or pleasant warmth, depending on the temperature threshold of the given channel. Capsaicin, the pungent compound in chili peppers is an agonist of the noxious heat sensor TRPV1 channels, while TRPV3 is activated at lower thresholds; its activators carvacrol, eugenol and thymol, from various spices, evoke a warm sensation when applied to the tongue.

Thus most temperature sensitive TRP channels may also be considered ligand gated ion channels. Classical ligand gated channels are activated by small water soluble extracellular molecules, such as the neurotransmitter acetylcholine for nicotinic acetylcholine receptors, or intracellular ligands such as the second messenger cAMP for cyclic nucleotide gated channels. The binding sites for these ligands are usually found in either the extracellular or intracellular hydrophilic regions of the channel protein. The thermo TRP channels described before are different, because their ligands are usually lipophilic compounds, and their binding sites are thought to be in the transmembrane regions of the channel protein.

Most ion channel proteins have a 3-, 4- or 5-fold symmetrical structure, where either identical, or homologous subunits co-assemble to form the conducting pore, and the functional channel. Functional TRP channels are either homo-, or heterotetramers. For a ligand gated channel, this multimeric structure immediately invites the question of how many subunits need to bind the ligand to open the channel.

An article by Janssens & Voets (2011) in this issue of The Journal of Physiology addresses this question on the menthol and cold sensitive TRPM8. To be able to meaningfully study this problem, one needs to be able to do two things. First, one needs to be able to prevent binding of the ligand to one or more subunit(s), while allowing the other(s) to bind. To prevent ligand binding, one needs to know where the ligand binds in the protein sequence; then a mutation can be introduced to prevent binding. The work by Janssens and Voets builds on their own previous work where they did indeed find such a mutant (Voets et al. 2007). Once the right mutant is found, one faces the second problem, the need to be able to tell which subunit(s) one has prevented from ligand binding. TRPM8 is a homotetramer, if one co-expresses a mutant and a wild-type subunit; they will co-assemble randomly, resulting in a mixture of channels with a variable number and position of mutant and wild-type subunits. While the distribution of the various combinations can be predicted with simple statistics, drawing conclusions from such experiments is not straightforward. The alternative approach is to embark on the painstaking task of generating a tandem tetramer molecule, where the four subunits are coupled into one linear sequence using tail to head linkers. In this monster protein, one can introduce the mutations into the subunit of one's choosing. Using this approach, Janssens and Voets measured the menthol concentration dependence of six tandem tertrameric constructs, having the menthol binding site mutation introduced into one, two, three and four subunits. Then plugging the data into various mathematical gating models, they concluded that the channel has four independent and energetically equivalent binding sites that each contributes similarly to the stabilization of the open state.

While a lot of work has been done on various classical ligand gated channels, these results are quite unique, as other ligand gated channels behave differently in various ways as discussed by Janssens and Voets. The question of course remains whether these results can be generalized to all TRP channels, or at least to most sensory channels. TRP channels are famously unruly, and diverse. Nevertheless, most chemical ligands of thermosensory TRPs tend to bind to a similar region to where menthol binds (Jordt & Julius, 2002; Bandell et al. 2006), thus it would not be surprising if other thermo TRP channels showed similar ligand stoichiometry.