ion selectivity filter within the external pore
of TRPM4 contains a conserved trio of resi-
dues (Phe-Gly-Gln or FGQ) that appear to
form cation-binding sites.
Overall, the TRPM8 structure solved by
Yin et al. is remarkably similar to that of
TRPM4. Differences in regions that connect domains make it difficult to appreciate
their relatedness when overlaying the complete structures, but superposition of individual domains shows that they adopt very
similar folds (see the figure). The TRPM8
structure contains a similar cavity to that
which binds Ca2+ in TRPM4; the channel is
closed and the selectivity filter (not resolved
in the structure) contains the conserved
FGQ sequence. It is unclear why TRPM4
and its closest homolog, TRPM5, are selective for monovalent cations, whereas
TRPM8 and all other TRP channels are also
permeable to divalent cations. The TRPM8
pore may be wider than that in TRPM4, but
higher-resolution structures will be needed
to understand the unique ion selectivity
mechanisms in these channels.
The TRPM4 Ca2+ binding site is particularly
fascinating because it suggests the presence
of a polymodal ligand-binding cavity that
is likely a conserved feature of TRPM channels. Comparison of the Ca2+-binding cavities of the Ca2+-bound and -unbound TRPM4
structures shows local conformational rearrangements that seem to propagate down to
the TRP domain, a conserved region among
TRP channels that is important for opening
these channels. Ca2+ binding to the cavity in
TRPM4 is required for channel opening, and
Ca2+ is also required for TRPM8 activation by
the synthetic compound icilin (11).
Comparison of the TRPM4 and TRPM8
Ca2+-binding cavities shows that they are
remarkably conserved, particularly in the
absence of Ca2+; they are also conserved in
the other Ca2+-modulated TRPM channels
(TRPM2 and TRPM5). Residues required for
activation of TRPM8 by menthol and icilin
(11, 12) are located in the Ca2+-binding cavity,
suggesting that this region is a polymodal
ligand-binding site in different types of TRPM
channels. However, most residues required
for TRPM8 activation by menthol and icilin
are also conserved in TRPM4, raising questions about why that channel is not sensitive
to these compounds or why TRPM8 activation
by menthol does not require Ca2+ binding.
The TRPV1 channel has some overlap-
ping pharmacology with TRPM8, and
both TRPV1 and TRPV2 have a polymodal
ligand-binding cavity in roughly the same
area as TRPM8 (13). However, comparison
of these cavities in TRPV1 and TRPM8 reveals little conservation, suggesting distinct
ligand-binding mechanisms. The overall hydrophobicity of the S1-S4 helices for TRPM4
and TRPM8 channels, which are activated
by positive membrane voltages, suggests a
different voltage-sensing mechanism than
in Kv channels; in the latter, Arg residues
in the S4 helix drive movement of this helix
through the membrane (9).
The TRPM structures (5, 7) provide a
first glimpse of their conserved and unique
amino termini, which contain four mela-
statin homology regions (MHRs) in all TRPM
channels. These MHRs derive their name
from melastatin, a metastasis suppressor in
melanoma cells that is a soluble splice vari-
ant of the amino terminus of TRPM1. The
functions of these amino-terminal domains
are poorly understood, but two other re-
cently reported TRPM4 structures suggest
that the MHRs can regulate the functional
properties of the channel. In Guo et al.’s
structure, the allosteric inhibitor adenosine
59-triphosphate is bound at the interface be-
tween MHR1 and MHR2 of two neighboring
subunits (14). In a second study, Winkler
et al. identified two cavities lined by basic
amino acids where the polyanionic deca-
vanadate binds to prevent channel closure
at negative membrane voltages (15).
Comparison of all available TRPM4 structures (5, 14, 15) reveals only relatively subtle
conformational changes; in each case, the
pore remains closed. In the future, it will be
critical to solve structures of these channels
in an open state to understand the mechanism of channel opening and its regulation
by ligands. That none of these structures
captured an open state might be due to the
absence of phosphatidylinositol 4,5-bisphos-
phate, which is required for opening of both
TRPM4 and TRPM8, as well as the absence
of membrane voltage. It may be worthwhile
to search for mutants that stabilize the open
state for TRPM channels. In the case of
TRPM8, agonists such as menthol and icilin
may help to capture an open state.
These exciting new structures provide a
framework for future studies. It will, for example, be interesting to further investigate
how the MHR domains regulate channel
function, and to explore possible roles in
protein localization, trafficking, or protein-protein interactions. It would also be interesting to reexamine the origin of the voltage
sensitivity of TRPM channels, a feature common to other TRP channels but whose mechanism is unclear. The new structures will also
catalyze experiments to explore the mechanisms underlying temperature sensing in
TRP channels, with the structure of TRPM8
providing a new foundation for investigating
the mechanism by which this channel is activated by cold. j
1. C. Montell, K. Jones, E. Hafen, G. Rubin, Science 230, 1040
2. B.Minke, J.Neurogenet. 24, 216(2010).
3. M.Liao, E.Cao,D.Julius, Y.Cheng, Nature 504,107(2013).
4. E. Cao, M. Liao, Y. Cheng, D. Julius, Nature 504, 113 (2013).
5. H. E. Autzen et al. , Science 359, 228 (2018).
6. I. Mathar et al., Handb. Exp. Pharmacol. 222, 461 (2014).
7. Y. Yin et al., Science 359, 237 (2018).
8. M. J. Pérez de Vega, I. Gómez-Monterrey, A. Ferrer-Montiel,
R. González-Muñiz, J. Med. Chem. 59, 10006 (2016).
9. K. J. Swartz, Nature 456, 891 (2008).
10. J. R. Whicher, R. MacKinnon, Science353, 664 (2016).
11. H. H. Chuang, W. M. Neuhausser, D. Julius,Neuron 43, 859
12. M. Bandell et al., Nat. Neurosci. 9, 493 (2006).
13. F. Zhang et al. , eLife 5, e16409 (2016).
14. J. Guo et al. , Nature 552, 205 (2017).
15. P.A. Winkler, Y.Huang, W.Sun,J.Du, W.Lü, Nature 552, 200
S1-S4 domains S5-S6 domains
For superimposition with TRPM8, the MHR1/2, MHR3,
MHR4, S1-S4, and S5-S6 domains of TRPM4 were
treated as individual domains.
12 JANUARY 2018 • VOL 359 ISSUE 6372 161
TRPM channels come into focus
Structure of the TRPM8 channel, with individual domains of TRPM4 superimposed to illustrate structural
similarities. The enlarged view illustrates the calcium ion–binding cavity within the S1-S4 domain of TRPM4.