view the benefits of inter-kingdom social communication in constantly varying ecological niches. There may also be more practical benefits; generating wine yeasts that are unable to switch to the [GAR+] state would presumably become de rigueur for the wine industry.
Brown, J.C., and Lindquist, S. (2009). Genes Dev. 23, 2320–2332.
Chang, A., Newby, G.A., Weitz, D.A., et al. (2014a). Cell 158, this issue, 1083–1093.
Frey-Klett, P., Burlinson, P., Deveau, A., Barret, M., Tarkka, M., and Sarniguet, A. (2011). Microbiol. Mol. Biol. Rev. 75, 583–609.
Jarosz, D.F., Lancaster, A.K., Brown, J.C.S., and Lindquist, S. (2014b). Cell 158, this issue, 1072– 1082.
REFERENCES
Hall, C., Brachat, S., and Dietrich, F.S. (2005). Eukaryot. Cell 4, 1102–1115.
Riedlinger, J., Schrey, S.D., Tarkka, M.T., Hampp, R., Kapur, M., and Fiedler, H.P. (2006). Appl. Environ. Microbiol. 72, 3550–3557.
Jarosz, D.F., Brown, J.C.S., Walker, G.A., Datta, M.S., Ung, W.L., Lancaster, A.K., Rotem, A.,
Smith, M.G., Des Etages, S.G., and Snyder, M. (2004). Mol. Cell. Biol. 24, 3874–3884.
Alexandre, H., Costello, P.J., Remize, F., Guzzo, J., and Guilloux-Benatier, M. (2004). Int. J. Food Microbiol. 93, 141–154.
Halfman, R., and Lindquist, S. (2010). Science 330, 629–632.
Kunz, B.A., and Ball, A.J. (1977). Mol. Gen. Genet. 153, 169–177.
Building a Temperature-Sensitive Ion Channel Ming-Feng Tsai1 and Christopher Miller1,* 1Department of Biochemistry, Howard Hughes Medical Institute, Brandeis University, Waltham, MA 02453, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.cell.2014.08.008
The biophysical basis of temperature-sensitive ion channel gating has been a tough nut to crack. Chowdhury, et al. use a protein engineering approach to render a temperature-insensitive voltage-gated channel cold- or heat-responsive to reveal principles for temperature-gating and a plausible model for molecularly enabling this mode of environmental responsiveness. Thermodynamics is a funny subject, and arguably the one least loved by cell biology students. The generality of classical equilibrium thermodynamics and its independence of specific molecular models means that its assertions are always true—and often useless. But in this issue of Cell, Chowdhury and colleagues report thermodynamics-inspired experiments that productively attack a long-lingering problem in sensory neurobiology: how certain ion channel proteins achieve the exquisite thermosensitivity that allows the neurons housing them—and us—to detect changes in temperature (Chowdhury et al., 2014). The molecular poster children for neuronal temperature sensing are found in the cation-conducting TRP channel family (Clapham, 2003). Among these are TRPV1, which detects painful heat by activating many-fold for just a few degrees
rise in temperature above 35 C, and TRPM8, which steeply turns on with cooling below 25 . Many studies have sought to identify the ‘‘temperature receptor domain’’ in such channels, and mutants have been found that profoundly alter or even invert their temperature-activation characteristics (Brauchi et al., 2006; Grandl et al., 2008; Yang et al., 2010; Jabba et al., 2014). But since these studies have fingered widely different protein domains as the culprit, no crisp picture or molecular mechanism has yet emerged. A few years ago, David Clapham and one of us suggested that a localized thermoreceptor domain need not exist and that a thermodynamic treatment might provide an alternative pathway into the problem (Clapham and Miller, 2011). We based our suggestion on a commonplace of protein physical chemistry: that if a large number of buried hydrophobic
groups become exposed to water upon a conformational change, the resulting increase in heat capacity (DCp) of the protein, arising from water-ordering near the nonpolar sidechains (Figure 1A), would confer very steep temperature dependence upon the equilibrium constant of that conformational reaction (Schellman et al., 1981). Grossly oversimplifying TRP channel activation as a simple closed % > open equilibrium, we imagined that opening of these channels is accompanied by the exposure to water of hydrophobic residues buried in the closed state—roughly 20 such residues per subunit in the tetrameric channel would do the trick. The residues need not be localized in a particular domain but instead could be distributed throughout the protein. Crucially, this ‘‘DCp’’ idea mathematically entwines enthalpy and entropy components of the conformational free
Cell 158, August 28, 2014 ª2014 Elsevier Inc. 977
Figure 1. Temperature Dependence of Ion Channel Activation Driven by Heat Capacity Changes Top: conformational changes occurring upon channel opening for a Shaker-like channel and the consequent transfer of amino acid side chains between hydrophobic and aqueous environments leading to changes in heat capacity (DCp). Bottom: Plots of temperature-dependence of equilibrium constant of opening (Keq, a proxy for open probability) expected for two-state (closed/open) channels, in which the heat capacity is greater in the open conformation than in the closed, illustrate two ways that amino acid substitutions can affect temperature sensitivity. In the left graph, a typical voltage-gated channel (mostly closed at negative voltage) is mutated with hydrophobic substitutions so as to raise DCp from 0.2 kcal/mol-K (dashed curve) to 2 kcal/ mol-K (solid curve), without other changes occurring. This increased DCp produces heat-activated current (red arrow). In the right graph, a hot-activated channel with DCp = 2 kcal/mol-K (red arrow) is converted to cold-activated (blue arrow) by a mutation that via new chemical interactions of the substituted sidechains reduces background chemical enthalpy and entropy without altering DCp. The regions within white boxes represent temperatures accessible to electrophysiological recording and currents observable above typical noise. Curves are drawn according to equation given in Clapham and Miller (2011).
energy change so as to make temperature-dependent activation necessarily nonmonotonic; a plot of the equilibrium constant of opening versus temperature would be U shaped (or rainbow shaped if the hydrophobic residues are more solvated in the closed state). This key feature means that any such channel is in principle both hot activated and cold activated (Figure 1B), with the position of the ‘‘U’’ along the temperature axis determining which activation mode is observed at experimentally accessible temperatures. Point mutations could transform a hotactivating TRP into a cold-activating one, or vice versa, simply by shifting the U plot along the temperature axis (Figure 1B). We offered our analysis only as a theoretical possibility, neither producing nor citing any experiments that would support or refute the idea. Now, Chowdury and coworkers use DCp considerations to design a temper-
ature-sensitive channel. Rather than manipulating TRP channels, which present nearly intractable difficulties in quantifying gating at equilibrium, they seek to build thermosensitivity ‘‘bottomup’’ into Shaker, a robust K+ channel from the same superfamily as TRPs, whose normal voltage-driven activation process is nearly temperature- independent. Shaker-like K+ channels have a voltage-sensing domain (VSD), membrane-embedded but with water-filled clefts, which moves outward upon activation. Since this movement is pretty well understood from crystal structures and modeling (Long et al., 2007; Vargas et al., 2012), the authors were able to pinpoint specific residues throughout the VSD that might change exposure to water when the channel opens. They mutated these systematically to alter their hydrophobicity with the aim of linking opening to a substantial change in Cp, and hence
978 Cell 158, August 28, 2014 ª2014 Elsevier Inc.
to confer temperature sensitivity on the channel. This design strategy appears to work. Shaker channels acquire heat or cold sensitivity by substitution of hydrophobic residues at positions expected to become unburied or desolvated upon opening. The positions at which these oppositely directed shifts occur are rationalized in terms of VSD movement, with activation increasing solvent exposure on one side of the VSD while decreasing it on the other. Moreover, single point mutants, which produce modest thermosensitivity, are combined to obtain stronger temperature-driven activation, in several cases achieving behavior quantitatively comparable to native TRPs. Mutations departing from this picture are also observed and highlighted by the authors, but the overall point is clear: changes of hydrophobic exposure throughout the protein can create strong temperature dependence of channels, seemingly via altered heat capacity changes (which cannot be directly measured here), as thermodynamics demands must occur in such a circumstance. The VSD is a membrane-embedded domain consisting of four helices, S1– S4, with S4 carrying positively charged residues peppering the otherwise hydrophobic helix. It is these charged residues, moving outward upon channel activation, that produce the strong voltage-dependence of Shaker-like K+ channels. TRP channels also have a VSD, but one notably sparse in S4 charges, a deficit that accounts for their much weaker voltage dependence. The authors, wondering why TRPs have lost these charges, substitute S4 arginines with neutral residues in Shaker-like K+ channels engineered as above to be thermosensitive, and show that the thermosensitivity initially present can be enhanced according to the hydrophobicity of the substituted residues, as the channel’s response to voltage becomes more sluggish X. Thus, voltage sensitivity can, in a sense, be traded off for temperature sensitivity. That’s a fascinating idea with intriguing evolutionary implications. Of course, the results here in no way establish that TRPs exploit heat capacity to fulfill their biological purpose, but they do lend that idea some plausibility. The
consequences flowing ineluctably from the First and Second Laws of Thermodynamics applied to a picture of gatingdriven hydrophobic solvation roughly match the temperature behavior of the manipulated Shaker channels. It remains unknown whether evolution transformed voltage-sensitive channels into thermosensitive TRPs in this way, but the basic principle has now found practical application by clever protein engineers. So maybe thermodynamics is worth warming up to after all.
REFERENCES Brauchi, S., Orta, G., Salazar, M., Rosenmann, E., and Latorre, R. (2006). J. Neurosci. 26, 4835–4840. Chowdhury, S., Jarecki, B.W., and Chanda, B. (2014). Cell 158, this issue, 1148–1158. Clapham, D.E. (2003). Nature 426, 517–524. Clapham, D.E., and Miller, C. (2011). Proc. Natl. Acad. Sci. USA 108, 19492–19497. Grandl, J., Hu, H., Bandell, M., Bursulaya, B., Schmidt, M., Petrus, M., and Patapoutian, A. (2008). Nat. Neurosci. 11, 1007–1013.
Jabba, S., Goyal, R., Sosa-Paga´n, J.O., Moldenhauer, H., Wu, J., Kalmeta, B., Bandell, M., Latorre, R., Patapoutian, A., and Grandl, J. (2014). Neuron 82, 1017–1031. Long, S.B., Tao, X., Campbell, E.B., and MacKinnon, R. (2007). Nature 450, 376–382. Schellman, J.A., Lindorfer, M., Hawkes, R., and Grutter, M. (1981). Biopolymers 20, 1989–1999. Vargas, E., Yarov-Yarovoy, V., Khalili-Araghi, F., Catterall, W.A., Klein, M.L., Tarek, M., Lindahl, E., Schulten, K., Perozo, E., Bezanilla, F., and Roux, B. (2012). J. Gen. Physiol. 140, 587–594. Yang, F., Cui, Y., Wang, K., and Zheng, J. (2010). Proc. Natl. Acad. Sci. USA 107, 7083–7088.
Cell 158, August 28, 2014 ª2014 Elsevier Inc. 979
Brown, J.C., and Lindquist, S. (2009). Genes Dev. 23, 2320–2332.
Chang, A., Newby, G.A., Weitz, D.A., et al. (2014a). Cell 158, this issue, 1083–1093.
Frey-Klett, P., Burlinson, P., Deveau, A., Barret, M., Tarkka, M., and Sarniguet, A. (2011). Microbiol. Mol. Biol. Rev. 75, 583–609.
Jarosz, D.F., Lancaster, A.K., Brown, J.C.S., and Lindquist, S. (2014b). Cell 158, this issue, 1072– 1082.
REFERENCES
Hall, C., Brachat, S., and Dietrich, F.S. (2005). Eukaryot. Cell 4, 1102–1115.
Riedlinger, J., Schrey, S.D., Tarkka, M.T., Hampp, R., Kapur, M., and Fiedler, H.P. (2006). Appl. Environ. Microbiol. 72, 3550–3557.
Jarosz, D.F., Brown, J.C.S., Walker, G.A., Datta, M.S., Ung, W.L., Lancaster, A.K., Rotem, A.,
Smith, M.G., Des Etages, S.G., and Snyder, M. (2004). Mol. Cell. Biol. 24, 3874–3884.
Alexandre, H., Costello, P.J., Remize, F., Guzzo, J., and Guilloux-Benatier, M. (2004). Int. J. Food Microbiol. 93, 141–154.
Halfman, R., and Lindquist, S. (2010). Science 330, 629–632.
Kunz, B.A., and Ball, A.J. (1977). Mol. Gen. Genet. 153, 169–177.
Building a Temperature-Sensitive Ion Channel Ming-Feng Tsai1 and Christopher Miller1,* 1Department of Biochemistry, Howard Hughes Medical Institute, Brandeis University, Waltham, MA 02453, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.cell.2014.08.008
The biophysical basis of temperature-sensitive ion channel gating has been a tough nut to crack. Chowdhury, et al. use a protein engineering approach to render a temperature-insensitive voltage-gated channel cold- or heat-responsive to reveal principles for temperature-gating and a plausible model for molecularly enabling this mode of environmental responsiveness. Thermodynamics is a funny subject, and arguably the one least loved by cell biology students. The generality of classical equilibrium thermodynamics and its independence of specific molecular models means that its assertions are always true—and often useless. But in this issue of Cell, Chowdhury and colleagues report thermodynamics-inspired experiments that productively attack a long-lingering problem in sensory neurobiology: how certain ion channel proteins achieve the exquisite thermosensitivity that allows the neurons housing them—and us—to detect changes in temperature (Chowdhury et al., 2014). The molecular poster children for neuronal temperature sensing are found in the cation-conducting TRP channel family (Clapham, 2003). Among these are TRPV1, which detects painful heat by activating many-fold for just a few degrees
rise in temperature above 35 C, and TRPM8, which steeply turns on with cooling below 25 . Many studies have sought to identify the ‘‘temperature receptor domain’’ in such channels, and mutants have been found that profoundly alter or even invert their temperature-activation characteristics (Brauchi et al., 2006; Grandl et al., 2008; Yang et al., 2010; Jabba et al., 2014). But since these studies have fingered widely different protein domains as the culprit, no crisp picture or molecular mechanism has yet emerged. A few years ago, David Clapham and one of us suggested that a localized thermoreceptor domain need not exist and that a thermodynamic treatment might provide an alternative pathway into the problem (Clapham and Miller, 2011). We based our suggestion on a commonplace of protein physical chemistry: that if a large number of buried hydrophobic
groups become exposed to water upon a conformational change, the resulting increase in heat capacity (DCp) of the protein, arising from water-ordering near the nonpolar sidechains (Figure 1A), would confer very steep temperature dependence upon the equilibrium constant of that conformational reaction (Schellman et al., 1981). Grossly oversimplifying TRP channel activation as a simple closed % > open equilibrium, we imagined that opening of these channels is accompanied by the exposure to water of hydrophobic residues buried in the closed state—roughly 20 such residues per subunit in the tetrameric channel would do the trick. The residues need not be localized in a particular domain but instead could be distributed throughout the protein. Crucially, this ‘‘DCp’’ idea mathematically entwines enthalpy and entropy components of the conformational free
Cell 158, August 28, 2014 ª2014 Elsevier Inc. 977
Figure 1. Temperature Dependence of Ion Channel Activation Driven by Heat Capacity Changes Top: conformational changes occurring upon channel opening for a Shaker-like channel and the consequent transfer of amino acid side chains between hydrophobic and aqueous environments leading to changes in heat capacity (DCp). Bottom: Plots of temperature-dependence of equilibrium constant of opening (Keq, a proxy for open probability) expected for two-state (closed/open) channels, in which the heat capacity is greater in the open conformation than in the closed, illustrate two ways that amino acid substitutions can affect temperature sensitivity. In the left graph, a typical voltage-gated channel (mostly closed at negative voltage) is mutated with hydrophobic substitutions so as to raise DCp from 0.2 kcal/mol-K (dashed curve) to 2 kcal/ mol-K (solid curve), without other changes occurring. This increased DCp produces heat-activated current (red arrow). In the right graph, a hot-activated channel with DCp = 2 kcal/mol-K (red arrow) is converted to cold-activated (blue arrow) by a mutation that via new chemical interactions of the substituted sidechains reduces background chemical enthalpy and entropy without altering DCp. The regions within white boxes represent temperatures accessible to electrophysiological recording and currents observable above typical noise. Curves are drawn according to equation given in Clapham and Miller (2011).
energy change so as to make temperature-dependent activation necessarily nonmonotonic; a plot of the equilibrium constant of opening versus temperature would be U shaped (or rainbow shaped if the hydrophobic residues are more solvated in the closed state). This key feature means that any such channel is in principle both hot activated and cold activated (Figure 1B), with the position of the ‘‘U’’ along the temperature axis determining which activation mode is observed at experimentally accessible temperatures. Point mutations could transform a hotactivating TRP into a cold-activating one, or vice versa, simply by shifting the U plot along the temperature axis (Figure 1B). We offered our analysis only as a theoretical possibility, neither producing nor citing any experiments that would support or refute the idea. Now, Chowdury and coworkers use DCp considerations to design a temper-
ature-sensitive channel. Rather than manipulating TRP channels, which present nearly intractable difficulties in quantifying gating at equilibrium, they seek to build thermosensitivity ‘‘bottomup’’ into Shaker, a robust K+ channel from the same superfamily as TRPs, whose normal voltage-driven activation process is nearly temperature- independent. Shaker-like K+ channels have a voltage-sensing domain (VSD), membrane-embedded but with water-filled clefts, which moves outward upon activation. Since this movement is pretty well understood from crystal structures and modeling (Long et al., 2007; Vargas et al., 2012), the authors were able to pinpoint specific residues throughout the VSD that might change exposure to water when the channel opens. They mutated these systematically to alter their hydrophobicity with the aim of linking opening to a substantial change in Cp, and hence
978 Cell 158, August 28, 2014 ª2014 Elsevier Inc.
to confer temperature sensitivity on the channel. This design strategy appears to work. Shaker channels acquire heat or cold sensitivity by substitution of hydrophobic residues at positions expected to become unburied or desolvated upon opening. The positions at which these oppositely directed shifts occur are rationalized in terms of VSD movement, with activation increasing solvent exposure on one side of the VSD while decreasing it on the other. Moreover, single point mutants, which produce modest thermosensitivity, are combined to obtain stronger temperature-driven activation, in several cases achieving behavior quantitatively comparable to native TRPs. Mutations departing from this picture are also observed and highlighted by the authors, but the overall point is clear: changes of hydrophobic exposure throughout the protein can create strong temperature dependence of channels, seemingly via altered heat capacity changes (which cannot be directly measured here), as thermodynamics demands must occur in such a circumstance. The VSD is a membrane-embedded domain consisting of four helices, S1– S4, with S4 carrying positively charged residues peppering the otherwise hydrophobic helix. It is these charged residues, moving outward upon channel activation, that produce the strong voltage-dependence of Shaker-like K+ channels. TRP channels also have a VSD, but one notably sparse in S4 charges, a deficit that accounts for their much weaker voltage dependence. The authors, wondering why TRPs have lost these charges, substitute S4 arginines with neutral residues in Shaker-like K+ channels engineered as above to be thermosensitive, and show that the thermosensitivity initially present can be enhanced according to the hydrophobicity of the substituted residues, as the channel’s response to voltage becomes more sluggish X. Thus, voltage sensitivity can, in a sense, be traded off for temperature sensitivity. That’s a fascinating idea with intriguing evolutionary implications. Of course, the results here in no way establish that TRPs exploit heat capacity to fulfill their biological purpose, but they do lend that idea some plausibility. The
consequences flowing ineluctably from the First and Second Laws of Thermodynamics applied to a picture of gatingdriven hydrophobic solvation roughly match the temperature behavior of the manipulated Shaker channels. It remains unknown whether evolution transformed voltage-sensitive channels into thermosensitive TRPs in this way, but the basic principle has now found practical application by clever protein engineers. So maybe thermodynamics is worth warming up to after all.
REFERENCES Brauchi, S., Orta, G., Salazar, M., Rosenmann, E., and Latorre, R. (2006). J. Neurosci. 26, 4835–4840. Chowdhury, S., Jarecki, B.W., and Chanda, B. (2014). Cell 158, this issue, 1148–1158. Clapham, D.E. (2003). Nature 426, 517–524. Clapham, D.E., and Miller, C. (2011). Proc. Natl. Acad. Sci. USA 108, 19492–19497. Grandl, J., Hu, H., Bandell, M., Bursulaya, B., Schmidt, M., Petrus, M., and Patapoutian, A. (2008). Nat. Neurosci. 11, 1007–1013.
Jabba, S., Goyal, R., Sosa-Paga´n, J.O., Moldenhauer, H., Wu, J., Kalmeta, B., Bandell, M., Latorre, R., Patapoutian, A., and Grandl, J. (2014). Neuron 82, 1017–1031. Long, S.B., Tao, X., Campbell, E.B., and MacKinnon, R. (2007). Nature 450, 376–382. Schellman, J.A., Lindorfer, M., Hawkes, R., and Grutter, M. (1981). Biopolymers 20, 1989–1999. Vargas, E., Yarov-Yarovoy, V., Khalili-Araghi, F., Catterall, W.A., Klein, M.L., Tarek, M., Lindahl, E., Schulten, K., Perozo, E., Bezanilla, F., and Roux, B. (2012). J. Gen. Physiol. 140, 587–594. Yang, F., Cui, Y., Wang, K., and Zheng, J. (2010). Proc. Natl. Acad. Sci. USA 107, 7083–7088.
Cell 158, August 28, 2014 ª2014 Elsevier Inc. 979
