1: \documentclass[10pt]{article}
2: \usepackage{amsmath}
3: \usepackage{amsfonts}
4: \usepackage{epsfig}
5: \usepackage{color}
6:
7: %\usepackage{time}
8: %\usepackage{eso-pic} %Works with pdflatex
9: %\definecolor{lightgray}{gray}{.80}
10: % \AddToShipoutPicture{
11: % \makebox(650,800){\resizebox{\textwidth}{!}{\rotatebox{60}{\color{lightgray}
12: % Draft:}}}}
13:
14:
15: \input epsf
16: \def\epsf#1#2{\epsfxsize=#1\epsfbox{#2}}
17: \def\deg{^\circ} %
18: \setlength{\unitlength}{1.cm}
19: \renewcommand{\arraystretch}{1.2}
20: \renewcommand{\topfraction}{0.9}
21: \renewcommand{\bottomfraction}{0.9}
22: \renewcommand{\textfraction}{0.1}
23: \renewcommand{\floatpagefraction}{0.8}
24: \renewcommand{\textfloatsep}{10pt}
25: \renewcommand{\floatsep}{10pt}
26: \renewcommand{\intextsep}{9pt}
27: \doublerulesep -1pt
28:
29: %print
30: \textwidth=6.3in
31: \textheight=9in
32: \hoffset=-1.5cm
33: \voffset=0in
34:
35: %screen
36: %\textwidth=7in
37: %\textheight=5in
38: %\hoffset=-4cm
39: %\voffset=-1in
40:
41: \parskip=10pt
42: \parindent=0pt
43: %\pagestyle{headings}
44: \pagenumbering{arabic}
45: \begin{document}
46:
47: \begin{center}
48: {\LARGE Analytic models for mechanotransduction: gating a mechanosensitive channel} \\
49: \vspace{0.5cm}
50: \textsc{ Paul Wiggins${}^*$ and Rob Phillips${}^\dag$ }\\
51: \vspace{0.25cm}
52:
53: ${}^*$Division of Physics, Mathematics \& Astronomy, \\
54: ${}^\dag$Division of Engineering and Applied Science, \\
55: California Institute of Technology, \\
56: 1200 E. California Blvd. \\ Pasadena, California 91125-9500.
57: %\vspace{0.7cm}
58:
59: Correspondence should be addressed to R.P. email: phillips@aero.caltech.edu
60:
61: \end{center}
62:
63: \begin{abstract}
64: \textbf{
65: Analytic estimates for the forces and free energy generated by bilayer
66: deformation reveal a compelling and intuitive model for MscL channel
67: gating analogous to the nucleation of a second phase. We argue that
68: the competition between hydrophobic mismatch and tension
69: results in a surprisingly rich story which can provide both
70: a quantitative comparison to
71: measurements of opening tension for MscL when reconstituted in
72: bilayers
73: of different thickness and qualitative insights into the function of
74: the MscL channel and other transmembrane proteins.
75: }
76: \end{abstract}
77:
78:
79: %\newpage
80:
81:
82:
83:
84: \section{Introduction}
85: The mechanosensitive channel (MscL) is a compelling example of the interaction between a protein and
86: the surrounding bilayer membrane. The channel is gated mechanically by applied tension and
87: is believed to function as an emergency relief valve in bacteria \cite{moe}.
88: MscL is a member of a growing class of proteins which have been determined to be mechanosensitive
89: \cite{Gill}, \cite{hamill}.
90: The dependence of the conductance on applied tension has been studied extensively in patch clamp
91: experiments \cite{suk1}, \cite{perozo1}, \cite{perozo2}.
92: In terms of the observed conductance, these studies have revealed that the channel is very
93: nearly a two state system. MscL spends the vast majority of its life in either a closed state (C)
94: or an open state (O) characterized by a discrete conductance.
95: When the bilayer tension is small, the protein is exclusively in the closed configuration.
96: As the tension grows, the open state
97: becomes ever more prevalent, until it dominates at high tension. The simplest structural
98: interpretation of this conductance data is to assume that each discrete conductance corresponds
99: to a well defined channel conformation. This assumption seems to be compatible with the conductance
100: data\footnote{The area of the states seems roughly independent of applied tension \cite{suk1}.}.
101: Patch clamp experiments have also revealed that there are at least three additional
102: discrete, intermediate conductance levels \cite{suk1}
103: suggesting three additional short lived substates (S1-S3).
104: Rees and coworkers \cite{Rees} have solved the structure for one conformation using X-ray
105: crystallography. This state appears to be the closed state \cite{Rees}, \cite{perozo2}.
106: Perozo {\it et al.} \cite{perozo2} have trapped MscL in the open
107: configuration and used electron paramagnetic resonance spectroscopy (EPR) and site-directed spin
108: labeling (SDSL) to deduce its geometry. Sukharev {\it et al.} \cite{suk2}
109: have also proposed an open state conformation based on structural considerations.
110:
111: The conformational landscape of the MscL channel is extremely complex, depending on huge number
112: of microscopic degrees of freedom which are analytically intractable. Even from the standpoint of
113: numerical calculations, this number is still very large \cite{md1}. As an alternative
114: to a detailed microscopic picture of MscL, we consider a simplified free energy function where we
115: divide the free energy of the system into two contributions, namely,
116: \begin{equation}
117: G = G_P+G_{\cal M},
118: \end{equation}
119: where $G_{P}$ is the free energy associated with the conformation of the protein and $G_{{\cal M}}$ is the
120: deformation free energy from the bulk of the bilayer \cite{huang}. In general, these two terms are
121: coupled. The conformation of the protein depends on the forces applied by the bilayer. The bilayer
122: deformation is induced by the external geometry of the protein. We denote this external geometry with a state
123: vector, $X$, which captures the radius of the channel as well as its orientation relative to the surrounding
124: bilayer as described in more detail below. We calculate the induced bilayer deformation energy, $G_{\cal M}(X)$,
125: by minimizing the free energy of the bilayer and solving the resulting boundary value problem
126: using an analytic model developed for the study of bilayer mechanics \cite{helfrich} and protein-bilayer
127: interactions \cite{huang}, \cite{nielsen}, \cite{goulian}, \cite{ndan4}.
128: We then apply asymptotic approximations to the exact solutions
129: of this model for cylindrically symmetric inclusions, permitting all of the results to be expressed,
130: estimated, and understood with simple scaling relations.
131: The advantage of this model is that it permits us to characterize the protein-bilayer system
132: in a way that is at once analytically tractable and predictive.
133: By understanding the consequences of the simplest models, we develop a framework in which to understand
134: the richer dynamics of the real system. There is a wealth of useful, physical intuition to
135: be gleaned from this model relating to both the function of MscL and
136: that of mechanosensitive transmembrane proteins more generally. In a forthcoming paper,
137: we will show that the mechanics of the bilayer must play an integral role in
138: mechanotransduction and channel function. Specifically, we will present detailed analytic estimates of the
139: free energy generated by bilayer deformation induced by the channel and show that these free energies
140: are of the same order as the free energy differences measured by Sukharev {\it et al.} \cite{suk1}.
141: These analytic calculations reveal a compelling and intuitive model for the gating of the MscL channel
142: which is the subject of this current paper. The competition between hydrophobic mismatch and applied
143: tension, in the presence of radial constraints, generates a bistable system that is implicitly
144: % implicit = involved in the nature or essence of something though not revealed, expressed,
145: % or developed
146: % in this case we mean the last of the three... it functions as a MS channel without all sorts
147: % of magic from G_P
148: a mechanosensitive channel. Furthermore, this simple model provides a picture
149: which is both qualitatively and quantitatively consonant with the measured
150: dependence of the free energy on acyl chain length as observed by Perozo {\it et al.} \cite{perozo1}.
151: In addition, these results may also explain the stabilization of the open state by spontaneous
152: curvature inducing lysophospholipids observed by Perozo {\it et al.} \cite{perozo1}, although more experiments are
153: required to check the consistency of this proposal.
154:
155:
156: \begin{figure}
157: \begin{center}
158: \leavevmode
159: \epsfig{file=basic3_s.eps}
160: \caption{\label{basic} The bilayer-inclusion model. The geometry of the inclusion
161: is described by three parameters: the radius, $R$, the hydrophobic thickness, $W$, and the radial mid-plane
162: slope, $H'$. The hydrophobic mismatch, $2U$, is the difference between the hydrophobic protein thickness, $W$,
163: and the bilayer equilibrium thickness, $2a$. We assume the surfaces of the bilayer are locally normal to
164: the interface of the inclusion, as depicted above, implying that the mid-plane slope is related to the
165: interface angle: $H'=\tan \theta$.}
166: \end{center}
167: \end{figure}
168:
169: \section{The energy landscape of the bilayer}
170: \label{ELB}
171: \begin{figure}
172: \begin{center}
173: \epsfig{file=FMR_s.eps}
174: \caption{\label{FMR} The bilayer deformation energy landscape. The bilayer deformation
175: energy is plotted as a function of the radius
176: for different values of applied tension. The solid curves represent the bilayer deformation
177: energy with a positive line tension, $f$,
178: for various different tensions ($0<\alpha_1<\alpha_2<\alpha_3<\alpha_4$).
179: The competition between interface energy and applied tension naturally gives rise to a
180: bistable potential when the radial domain is limited by steric constraints.
181: The gray regions represent radii inaccessible to the channel due to steric constraints.
182: These constraints are briefly motivated in section \ref{ELB}. If the line tension
183: is negative, depicted by the dotted curve, the potential is never bistable.}
184: \end{center}
185: \end{figure}
186: In the calculations considered here, the geometry of the protein, characterized by
187: the conformational state vector $X$, is described by three geometrical parameters
188: \begin{equation}
189: X = (R, W, H'),
190: \end{equation}
191: where $R$ is the radius of the channel, $W$ is the hydrophobic thickness, and
192: $H'$ is the mid-plane slope. See figure \ref{basic} for details.
193: \label{ELB}
194: Although we have parameterized the conformation space of the protein with these three parameters, in this paper,
195: we will focus on the radial dependence alone, claiming that even in this reduced description,
196: the model provides a rich variety of predictions which are compatible with previous observation and suggest new
197: experiments.
198: The radial dependence of the bilayer deformation energy is particularly
199: important for MscL since the radius undergoes a very large change between the open and closed states \cite{suk2}.
200: The bilayer deformation energy can be written explicitly in terms of the channel radius as
201: \begin{equation}
202: \label{bde}
203: G_{\cal M}=G_0+f\cdot 2\pi R-\alpha\cdot\pi R^2,
204: \end{equation}
205: where $G_0$ and $f$ do not explicitly depend on $R$ and $\alpha$ is the applied tension which triggers
206: channel gating.
207: $G_0$ is a radially independent contribution to the deformation energy which is a function of the other
208: geometrical parameters of the protein. Its importance in gating the channel is most likely secondary since
209: it is independent of $R$ and it will be ignored in the remainder of the discussion.
210: The dependence of bilayer deformation energy on applied tension can be explained intuitively \cite{hamill}.
211: The free energy contribution for a small change in the channel area due to the applied tension can be written
212: $-\alpha dA$ which is the two dimensional analogue of the $-PdV$ term for a gas in three dimensions.
213: At high enough applied tension, the state with the largest inclusion area will have the lowest free energy.
214:
215: The line tension, $f$, contributes an energy proportional to the circumference.
216: Line tensions arise very naturally since the interface area of the inclusion is proportional
217: to the circumference. In what follows, we will discuss the two dominant contributions to this line tension:
218: thickness deformation \cite{huang}, \cite{nielsen}, \cite{goulian} and spontaneous curvature
219: \cite{ndan4}. The thickness deformation free energy is induced by the mismatch between the
220: the equilibrium thickness of the bilayer and the hydrophobic thickness of the protein.
221: The importance of this hydrophobic mismatch in the function of transmembrane proteins has
222: already been established \cite{Killian}.
223: The bilayer deforms locally to reduce the mismatch with the protein as shown in figure~\ref{basic}.
224: Symbolically the thickness deformation energy is \cite{huang},
225: \begin{equation}
226: G_U = f_U\cdot 2\pi R= {\textstyle \frac{1}{2}} {\cal K} U^2 \cdot 2\pi R,
227: \end{equation}
228: where ${\cal K}=2\times 10^{-2}\ kT\ {\rm \AA}^{-3}$ is an effective elastic modulus defined in
229: the appendix which is roughly independent of acyl chain length\footnote{See the appendix for
230: an brief discussion of scaling of the elastic constants.} and $U$ is half the hydrophobic mismatch as defined
231: in figure \ref{basic}.
232: Naturally the energetic penalty for this deformation is
233: proportional to the mismatch squared since the minimum energy
234: state corresponds to zero mismatch. The area of that part of the bilayer
235: which is deformed is roughly equal to the circumference of the channel times an elastic
236: decay length. As a result, the contribution of thickness
237: deformation to the total free energy budget scales with the radial dimension
238: of the channel. We also note that the thickness deformation
239: free energy is always positive.
240:
241: In contrast, the free energy induced by spontaneous curvature
242: can be either negative or positive. Physically, this free energy comes from locally relieving or increasing
243: the curvature stress generated by lipids or surfactants which induce spontaneous
244: curvature \cite{ndan4}, \cite{israelachvili}, \cite{gruner}. Again the radial dependence of this free energy will be linear since
245: the effect is localized around the interface. Since the leaflets of the bilayer can be doped
246: independently \cite{perozo1}, the spontaneous curvatures of the top and bottom leaflets, $C_\pm$,
247: can be different. It is convenient to work in terms of the composite spontaneous curvature of the bilayer,
248: $C \equiv \textstyle \frac{1}{2}(C_+-C_-)$.
249: The contribution to the deformation energy arising from spontaneous curvature is given by \cite{ndan4}
250: \begin{equation}
251: G_C = f_C\cdot 2\pi R = K_B C H'\cdot 2\pi R,
252: \end{equation}
253: where $H'$ is the mid-plane slope and $K_B = 20(a/20\ {\rm \AA})^3\ kT$,
254: is the bending modulus which roughly scales as the third power of the bilayer thickness\footnote{See the appendix for
255: an brief discussion of scaling of the elastic constants.}.
256: We will discuss these results in more detail elsewhere.
257: Notice that if $C$ and $H'$ have opposite signs, the deformation
258: energy and the corresponding line tension, $f_C$, will be negative.
259: We note that the elastic theory of membrane deformations associated with proteins like MscL permit
260: other terms (such as mid-plane deformation, for example) which can be treated within the same framework
261: and which give rise to the same radial dependence as that described here. However, for
262: the purposes of characterizing the energetics of MscL, these other terms are
263: less important than the two considered here.
264:
265: Typically, in the absence of large spontaneous curvature, the line tension,
266: $f$, will be dominated by the mismatch and will be positive. A potential of the form described by equation
267: \ref{bde} is depicted schematically in figure \ref{FMR}.
268: In this figure, we have implied that there are steric constraints for the range of radii accessible to the
269: protein. Assuming that there is a lower bound on the radius of the inclusion is very natural. It can
270: be understood as the radius below which the residues begin to overlap. This steric constraint will generate
271: a hard wall in the protein conformation energy, forbidding lower radii. Similar but slightly more elaborate
272: arguments can made for an upper bound. The bilayer deformation energy generates a barrier between small radius
273: and large radius states. The location of the peak of this barrier
274: is the turning point
275: \begin{equation}
276: R_* \equiv \frac{f}{\alpha}.
277: \end{equation}
278: At small tension, the turning point is very large and is irrelevant since it occurs at a radius not
279: attainable by the channel due to the steric constraints, but as the tension increases the position of the
280: turning point decreases. This behavior is reminiscent of the competition between surface tension and
281: energy density for nucleation processes which give rise to a similar barrier ({\it e.g.} \cite{kittel}).
282:
283: \subsection{Induced tension and bilayer induced stabilization}
284: \begin{figure}
285: \begin{center}
286: \epsfig{file=mech_s.eps}
287: \caption{ \label{mech}
288: The free energy of a bistable system. In the background of the plot, we have
289: depicted a mechanical system that is bistable. At the equilibrium length of the
290: springs, the hinge can be at either $R_C$ or $R_O$. To move the hinge between
291: these two stable states it must be compressed, giving rise to a barrier and
292: a transition state at $R_*$. This potential is analogous to that generated by
293: bilayer deformation. This picture also serves as a schematic potential for
294: the gating transitions as described in the text.
295: }
296: \end{center}
297: \end{figure}
298: \label{defforce}
299: Although the conformational landscape of the MscL channel is certainly very complicated,
300: there is an intriguing possibility that the channel harnesses the elastic properties of the bilayer
301: which quite naturally provide the properties we desire in a mechanosensitive channel:
302: a stable closed state at low tension
303: and a stable open state at high tension.
304: In effect, we will treat the bilayer deformation energy as an external
305: potential with respect to the conformational energy landscape of the protein.
306: The physical effects of the radial dependence of the bilayer deformation energy
307: on the inclusion conformation can be recast in a more intuitive form by appealing
308: to the induced tension which accounts not only for the
309: applied far field tension, but also for {\it induced} tension
310: terms due to bilayer deformation. The applied tension is not the whole story!
311: The generalized forces are obtained by differentiating the bilayer deformation energy
312: with respect to bilayer excursions. The net tension induced by the bilayer on the
313: inclusion interface is
314: \begin{equation}
315: \alpha_\Sigma = \alpha - \frac{f}{R},
316: \end{equation}
317: where we have denoted the net tension $\alpha_\Sigma$ since we have already used $\alpha$
318: to denote the applied tension. For radii smaller than the turning point, $R_*$, the
319: bilayer deformation energy is an increasing function of radius and therefore the net
320: tension is negative and acts to compress the channel. For radii larger than that at the turning point,
321: the bilayer deformation energy is a decreasing function of radius and the net tension is positive
322: and acts to expand the channel.
323: The combination of these constraints and the bilayer deformation energy lead to a bistable system where
324: the closed and open states correspond to the constraint-induced radial minimum
325: and maximum, respectively. Recall that the net tension on the closed state will be compressive as long
326: as its radius is smaller than that at the turning point, namely,
327: \begin{equation}
328: R_C<R_*=\frac{f}{\alpha}.
329: \end{equation}
330: This inequality defines the range of applied tension over which the closed state is stabilized by the
331: bilayer deformation energy. The net tension on the open state will be expansive as long as its
332: radius is greater than that at the turning point:
333: \begin{equation}
334: R_O>R_*=\frac{f}{\alpha}.
335: \end{equation}
336: This inequality defines the range of applied tension over which the open state is stabilized by the
337: bilayer deformation energy. There is an intermediate range of tensions for which both states are
338: stabilized by the bilayer,
339: \begin{equation}
340: \frac{f}{R_O} < \alpha < \frac{f}{R_C}.
341: \end{equation}
342: The bilayer deformation energy naturally destabilizes the open state for applied tension
343: below this range while stabilizing the closed state for applied tensions up to the limit of this range.
344: Both effects help to prevent the channel from leaking at low applied tension. This bistability is precisely
345: the desirable behavior for a mechanosensitive channel designed to relieve internal pressure and yet
346: surprisingly little is required from the protein conformational landscape, $G_P$, except for steric
347: constraints which arise very naturally.
348: In figure \ref{mech},
349: we have depicted a mechanical analogue of this bistable system. In figure \ref{FMR2}, we have depicted the
350: way in which MscL mimics this mechanical analogue by using the sum of a schematic protein energy and the
351: bilayer deformation energy to form energy minima corresponding to the open and closed states.
352: \begin{figure}
353: \begin{center}
354: \epsfig{file=FMR2_s.eps}
355: \caption{ \label{FMR2}
356: The total free energy. The total free energy, $G$, of the protein and bilayer are
357: plotted schematically as a function of channel
358: radius. The bilayer deformation energy, $G_{\cal M}$, is represented by the dotted
359: curve. A schematic protein conformation energy is represented by the dashed curve.
360: Their sum gives the total free energy $G$. The protein energy has been chosen to
361: contain a single substate, $S$. There is a conformational energy barrier corresponding
362: to changing the gate conformation of the channel. These transitions occur at $R_{CS}$
363: and $R_{SO}$. $G_P$ also contains steep barriers corresponding to steric
364: constraints. The radii of the conductance states are defined by the free energy minima of $G$.
365: As we have discussed in the text, the
366: bilayer deformation energy can stabilize either or both the closed and open states
367: since they are the states of highest and lowest radius. At the tension depicted above,
368: both have been stabilized. In contrast, the substates are stabilized by protein conformation
369: but never by the bilayer. }
370: \end{center}
371: \end{figure}
372:
373: Our general discussion of the role of the bilayer enables us to make some rather
374: general observations about the nature of the substates. In order to generate a substate,
375: we assume that there is more than one gating transition in the protein conformational energy.
376: One gating transition would correspond to a closed to open transition. An additional gating transition allows
377: three conductance states. We will assume that these transitions are themselves bistable in
378: nature since the conductance data would seem to imply the lifetimes of the transition states
379: are very short compared to the conductance states \cite{suk1}.
380: In other words, the conformational gating transition occurs near a local maximum in the
381: conformational free energy like that depicted in figure \ref{mech}. If we add two such
382: transitions to $G_P$, we generate a substate of intermediate radius between these two transition
383: state radii. A schematic example of this is illustrated in figure \ref{FMR2}.
384: Sukharev {\it et al.} \cite{suk1} have shown that all the substates are short lived and have estimated the
385: areas of each state based on the tension dependence of their free energies\footnote{There is now evidence for
386: additional substates \cite{suk3}.}. Specifically, they
387: have shown that the radii of the substates lie between the open and closed state radii.
388: If the bilayer deformation dominates the free energy of the states,
389: the ephemeral nature of the substates is a natural consequence of their intermediate radii.
390: The compressive tension due to the mismatch stabilizes the state of lowest radius at low applied tension. At high
391: applied tension the bilayer stabilizes the state with highest radius. All the states with
392: intermediate radii are never stabilized by the bilayer and are therefore short lived.
393: Our deceptively simple mismatch model quite naturally leads to short lived substates
394: at intermediate radii.
395:
396: In this section we have shown that some very general assumptions about the form of the
397: bilayer deformation energy lead to channel behavior which is mechanosensitive.
398: Although we have suggested a specific source, thickness deformation, for the line tension, $f$,
399: this linear radial dependence arises very naturally. For instance, if we assume the
400: bilayer is too stiff to deform and the mismatch causes hydrophobic residues to come
401: into contact with the solvent, a larger positive line tension would be induced.
402: The competition between line tension and applied tension is therefore a quite general
403: phenomena for mechanosensitive transmembrane proteins. We shall demonstrate below that, for MscL,
404: this mechanism appears to explain the free energy dependence on bilayer acyl chain length
405: surprisingly well.
406:
407: \section{ Results}
408: \label{mscl_g}
409: The patch clamp experiments of Perozo {\it et al.} \cite{perozo1} go beyond the
410: earlier work of Sukharev {\it et al.} \cite{suk1} by providing experimental values
411: for the free energy difference between the open and closed states for bilayers of
412: several thicknesses. These results can be compared with our predictions. In order to
413: apply our model, we must determine the geometrical parameters of the state vector $X$
414: for the open and closed states and in particular the open and closed radii.
415: The radius of the closed state is known from x-ray
416: crystallography \cite{Rees}: $R_C\sim 23\ \rm \AA$. Structural studies \cite{suk2} and
417: EPR and SDSL \cite{perozo2} experiments have suggested an open state radius of roughly
418: $R_O\sim 35\ \rm \AA$. In order to estimate the line tension and free energy generated by
419: hydrophobic mismatch, we must determine the hydrophobic thickness $W$. (We ignore the
420: difference in hydrophobic thickness between the open and closed states.) In principle,
421: one might have thought this could be deduced from the atomic-level structure of MscL,
422: but in practice, real structures are complicated, often
423: lacking a clear transition from hydrophobic to hydrophilic residues on the interface.
424: However, this width may be deduced from the EPR and SDSL data of Perozo {\it et al.} \cite{perozo1}.
425: EPR and SDSL experiments measure inter-subunit proximity
426: and spin-label mobility, respectively \cite{perozo1}. Compressive tension in the bilayer
427: suppresses the fluctuations of the protein, increasing the subunit proximity and reducing the
428: spin label mobility. In the experiments of Perozo {\it et al.} \cite{perozo1},
429: the applied tension is low implying that
430: the net tension is dominated by the line tension, induced by thickness deformation,
431: \begin{equation}
432: \label{induced_tension}
433: \alpha_{\Sigma} \sim -\frac{{\cal K} U^2}{2R},
434: \end{equation}
435: in the absence of spontaneous curvature.
436: This tension is compressive and proportional to the mismatch squared. Therefore, when the mismatch
437: is zero, the tension reaches a minimum, implying that mobility and subunit separation should reach a
438: maximum. The EPR and SDSL data of Perozo {\it et al.} \cite{perozo1}
439: may turn over for PC12 bilayers, implying that the mismatch is zero which would imply in turn that
440: \begin{equation}
441: W \sim 2a_{n=12},
442: \end{equation}
443: but due to the quadratic dependence on $U$, the slope in the vicinity of the turnover is small.
444: Since PC lipids with acyl chain length shorter than $n=10$ do not form stable bilayers \cite{perozo1},
445: it is difficult to extensively check the quadratic dependence on $U$. The predicted
446: turnover would be more pronounced for PC bilayers with $n<10$. We shall see that
447: this deduced hydrophobic mismatch is compatible with the patch clamp measurements of
448: Perozo {\it et al.} \cite{perozo1}.
449:
450: \subsection{Hydrophobic mismatch}
451: \label{hm}
452: \begin{figure}
453: \begin{center}
454: \epsfig{file=pdata_s.eps}
455: \caption{\label{res} Free energy difference between open
456: and closed states vs lipid acyl chain length.
457: The experimental data of Perozo {\it et al.} \cite{perozo1} for the free energy
458: difference between the open and closed states at zero tension, $\Delta G_0$, is plotted
459: with black circles and error bars. The solid curve represents the theoretical values
460: for the bilayer deformation energy generated by a simple thickness deformation model at
461: zero tension, $\Delta G_{0,\cal M}$.
462: This correspondence is strong evidence for a central role
463: for the hydrophobic mismatch in the function of MscL. The dotted curve represents
464: the translated $\Delta G_{0,\cal M}$ for an engineered MscL channel with a hydrophobic
465: thickness matching a PC14 bilayer.
466: }
467: \end{center}
468: \end{figure}
469: Perozo {\it et al.} \cite{perozo1} have measured $\Delta G_0$, the free energy difference
470: between the open and closed state at zero tension\footnote{The free energy measured by Perozo {\it et al.}
471: \cite{perozo1} is equivalent to the free energy at zero tension modulo several assumptions \cite{hamill}.}
472: for three acyl chain lengths.
473: Using the value we have deduced for $W$, we can now calculate the free energy difference between the
474: open and closed states due to bilayer deformation at zero tension, $\Delta G_{0,\cal M}$, which is given by
475: the line tension contribution alone:
476: \begin{equation}
477: \Delta G_{0,\cal M} = f2\pi\Delta R \rightarrow f_U2\pi \Delta R,
478: \end{equation}
479: where $\Delta R$ is the difference between the open and closed radii.
480: The theoretical result, $\Delta G_{0,\cal M}$, is plotted with the experimental measurements of
481: $\Delta G_0$ in fig \ref{res}. The
482: agreement between experiment and theory is embarrassingly good given the cavalier fashion
483: in which we have chosen the geometrical parameters and that we have neglected the protein conformational energy, $G_P$,
484: entirely.
485: There is a very important point to be made about these results. Perozo {\it et al.} \cite{perozo1}
486: have measured three data
487: points and our model is quadratic,
488: implying that we could have chosen the parameters of our model to fit the data
489: points perfectly since any three points lie on a parabola, but our parameters have in fact been deduced
490: independently rather than fit, which is why
491: this correspondence with the data is remarkable. This model corresponds to a channel where the free
492: energy difference between the open and closed states is dominated by the bilayer deformation
493: rather than protein conformation.
494: Our model implies that $\Delta G_0$ for PC10, PC12, and PC14 should be very small. Unfortunately these
495: bilayers have proved too weak for patch-clamp measurements of $\Delta G_0$ \cite{perozo1}.
496: Certainly none of these bilayers trap the channel in the open state \cite{perozo1}.
497:
498: The opening tension is defined as the tension at which the open and closed state probabilities
499: are equal or, analogously, the tension at which the free energies of the open and closed states are equal.
500: The opening tension is
501: \begin{equation}
502: \label{opening_tension}
503: \alpha_{1/2} = \frac{f}{\overline R}+\frac{\Delta G_P}{\Delta A},
504: \end{equation}
505: where ${\overline R}\equiv {\textstyle \frac{1}{2}}(R_C+R_O)$ is the mean radius, $\Delta G_P$ is the difference in
506: the open and closed state protein conformation energy, and $\Delta A$ is the difference in open
507: and closed state area. When the bilayer deformation energy dominates, the opening tension is determined by the
508: first term alone. Changing the sensitivity of the channel is straightforward from
509: this perspective. Changes in the length of the hydrophobic region of the protein can increase or
510: decrease the opening tension of the channel. For example, MscL channels might be engineered with
511: an expanded hydrophobic region which matches PC14 bilayers. Our mismatch based
512: theory would predict that the free energy versus acyl chain length curve would simply be translated
513: to higher $n$ so that the minimum $\Delta G_{0,\cal M}$ is realized for a PC14 bilayer. This shift should be
514: measurable, reducing $\Delta G_0$ for PC16, PC18, and PC20 bilayers.
515: The reduction in mismatch may also allow MscL to be reconstituted into PC22 bilayers, allowing
516: an additional data point. The proposed shift should also be measurable in EPR and SDSL
517: measurements of residue proximity and mobility. The maximum mobility and separation should now be
518: centered around $n=14$, perhaps permitting a clear measurement of the rise in induced tension for a PC10 bilayer
519: predicted by the quadratic dependence of the line tension on the mismatch.
520:
521:
522: \subsection{Spontaneous curvature}
523: \label{sc}
524: Perozo {\it et al.} \cite{perozo1} have proposed that asymmetric bilayer stresses play a
525: central role in MscL gating. They have proposed this model based on patch clamp,
526: EPR, and SDSL experiments showing that spontaneous curvature can induce MscL channel
527: opening. Specifically, Perozo {\it et al.} find that MscL reconstituted into PC vesicles with high
528: enough concentrations of asymmetrically incorporated LPC stabilizes the open state of the channel,
529: while MscL reconstituted into PC vesicles with symmetrically incorporated LPC, does not stabilize the
530: open state. Unlike Keller {\it et al.} \cite{keller}, Perozo {\it et al.}
531: have measured neither the spontaneous curvature for the mixed bilayer nor
532: the free energy difference between open and closed states as a function of LPC concentration.
533: In the absence of these quantitative experimental results, it is difficult to make concrete comparisons between
534: our model and the experimental data. For large spontaneous curvature \cite{keller} but
535: a relatively modest complementary mid-plane slope, the free energy difference between the two states due
536: to spontaneous curvature is
537: \begin{equation}
538: \label{spont}
539: \Delta G_{C} \sim 2\pi (R_O-R_C) K_B CH' =
540: -16 \left(\frac{20\ {\rm \AA}}{C^{-1}}\right)\left( \frac{H'}{-0.2}\right) \ kT,
541: \end{equation}
542: an energy typically large enough to stabilize the open state.
543: In section \ref{ELB}, we have made some rather general arguments about the shape of the
544: bilayer deformation energy landscape. We now return to this picture briefly to discuss the
545: consequences of spontaneous curvature. In our discussion, we assumed that the line tension,
546: $f$, was typically positive, but we remarked
547: that this need not be the case in the presence of large spontaneous curvature.
548: If $f$ is negative, as depicted by the dotted curve in
549: figure \ref{FMR}, the only state stabilized by the bilayer is the open state, which very naturally
550: gives rise to the open state stabilization observed by Perozo {\it et al.} \cite{perozo1}.
551: Alternatively, this result can be understood from the predicted opening tension in equation \ref{opening_tension}.
552: When $\Delta G_0$ is bilayer deformation dominated, a negative $f$ implies that the opening tension is
553: itself negative! A compressive force is required to stabilize the closed state. This
554: argument gives a tantalizingly simple explanation for the open state stabilization
555: but in the absence of measured values for the spontaneous curvature induced by LPC,
556: we can only conclude that spontaneous curvature could stabilize the open states for
557: rather generic values of the parameters. An experimental consistency check of these results
558: is fairly simple. Perozo {\it et al.} have incorporated LPC asymmetrically. The same
559: experiment might be repeated with $H_{II}$ phase inducing lipids which can also be used to
560: generate spontaneous curvature but of the opposite sign. For the DOPC/DOPE system of Keller
561: {\it et al.} \cite{keller}, the spontaneous curvature is known and tunable as a function of
562: concentration. Our results predict that $\Delta G_0$ should be linear in $C$ \cite{ndan4}
563: and compatible with equation \ref{spont}.
564:
565: \subsection{Transition Rates}
566: \label{tr}
567: Although we have not discussed transition states in detail, the proposed model
568: also predicts that the transition rates between states depends on the line tension, $f$.
569: If we assume the transition rates to be given by the standard Arrhenius form:
570: \begin{equation}
571: k = k_0\exp( -G_B/kT ),
572: \end{equation}
573: where $G_B$ is the free energy barrier height for the transition and $k_0$ is roughly constant,
574: the results can be understood intuitively by consulting figure \ref{FMR2}. At fixed tension, forward transition
575: rates (transitions to increasing radii) are reduced by line tension while backward rates are increased.
576: At any fixed tension, we would therefore expect the forward rates for MscL reconstituted into a PC16
577: bilayer to be faster than for a PC18 bilayer while the backward rates are slower since the PC18 bilayer has
578: a greater line tension. The ratio of these rates can be calculated but depends on the radial location
579: of the transition state. At the opening tension for each bilayer, it can be shown that if the
580: bilayer deformation energy dominates, the predicted rates for the transitions $C\rightleftharpoons O$ will be faster
581: for PC16 bilayer than for PC18 bilayer. Again, the ratio of these rates can be calculated in terms of the radii of
582: the states and transition states. These two sets of qualitative rate predictions are straightforward
583: to check experimentally.
584:
585:
586: \label{caveats}
587: In the argument above, we focused on the radial dependence of bilayer deformation free energy,
588: and fixed the other components of the state vector, $X$.
589: In principle, there is a potentially important piece of radial dependence we are missing. The
590: internal conformation may effectively couple the radius to the other parameters
591: in the state vector $X$, adding additional implicit radial dependence. For example,
592: the thickness of the inclusion, $W$, is almost certainly a function of radius.
593: It is also very natural to couple the mid-plane slope to the radius.
594: We have ignored these dependences in order to develop an
595: intuitive and simple one dimensional picture with as few undetermined constants and couplings as possible.
596: Provided that the bilayer deformation energy change is dominated by the
597: radial change, this simplified model is a useful tool for understanding the bilayer-inclusion
598: interaction. More elaborate models might easily be built from the general analytic framework
599: we have constructed. This framework will be described in a forthcoming paper.
600:
601: \section{Discussion}
602:
603: %\subsection{Overall picture}
604: We have argued that the bilayer deformation energy is harnessed by MscL to govern channel gating.
605: Indeed, we have shown that a model which attributes the entire free energy difference between the
606: open and closed states to the bilayer deformation energy is compatible with the experimental data.
607: These results are somewhat surprising since it has been shown experimentally that the mutation of
608: a single residue in the vicinity of the channel gate, can significantly effect channel gating
609: \cite{ou}, \cite{yoshimura}. The protein conformational energy cannot be neglected in general.
610: In fact, we have assumed that the protein
611: conformational energy is large enough to constrain the channel geometry since we have assumed it is
612: the bilayer which deforms rather than the protein. In principle, the closed state could have been
613: stabilized by protein conformation alone, rather than mismatch, but exploiting bilayer deformation
614: provides a robust mechanism for mechanotransduction, a design principle which functions in spite of
615: the enormous number of nearly degenerate microstates endemic to proteins. Even for proteins as simple as
616: myoglobin, Frauenfelder {\it et al.} \cite{frauen} have shown that the macroscopic conformation
617: corresponds to an enormous number of structurally distinct microstates. These ideas have already been
618: exploited for channel proteins. Goychuk and H\"anggi \cite{hanggi} have used this degenerate
619: landscape to derive the empirical rate law for voltage gated channels.
620: In light of these results, it is very natural to suppose that the protein conformational energy of
621: the MscL protein gives rise to a vast number of nearly degenerate states as well.
622: The bilayer deformation energy naturally breaks this degeneracy
623: and forms a mechanotransducing channel. The ensemble of microstates we observe as the closed states is
624: stabilized at low applied tension by the line tension, while the ensemble of microstates we observe as the
625: open state is stabilized by high applied tension.
626: The importance of bilayer deformation in mechanotransduction may help to explain why there are no obvious
627: sequence motifs associated with mechanosensitivity \cite{Bass} since a mismatch requirement does not imply
628: sequence specificity. Harnessing bilayer elasticity does have one noted disadvantage. The gating of a channel
629: will be affected
630: by the membrane environment which surrounds it. This is precisely what the experiments of Perozo {\it et al.}
631: \cite{perozo1} have shown. In realistic cell membranes there is an enormous diversity of
632: proteins and lipids which would imply that the free energy, and therefore opening tension, in these membranes
633: would be heterogeneous. Sukharev and co workers \cite{suk3} have evidence for exactly this variability for
634: MscL in giant spheroplasts.
635:
636: The correspondence between our simple theoretical
637: model for the gating of the MscL channel and experiment is at least strongly suggestive that
638: this mechanism is exploited by the MscL channel. Our model can also naturally explain the
639: stabilization of the open state by LPC \cite{perozo1} as well as the ephemeral nature of
640: the substates \cite{suk1}. The tractable nature and simplicity of the model allow extensive
641: analytic calculations to be made, which have in turn lead to numerous experimental predictions,
642: discussed in section \ref{mscl_g}. Specifically we have predicted (i) a shift in the
643: curve relating the free energy difference and acyl chain length when the hydrophobic thickness of
644: the channel is altered, (ii) the dependence of the free energy on spontaneous curvature and,
645: in particular, on the concentration of spontaneous curvature-inducing molecules, and (iii)
646: the qualitative dependence of the transition rates on acyl chain length and spontaneous
647: curvature.
648:
649: We have developed an extensive framework for studying bilayer-inclusion interactions in the
650: MscL system. The model we have discussed here is the simplest implementation of these results
651: and a more thorough description of the model will appear in a future paper. There are several very
652: natural extensions to the current work. For example, we have focused here on the radial
653: dependence only, but, as we have briefly alluded to in section \ref{caveats},
654: there are two additional geometric parameters which may also play important
655: roles in the function of the MscL channel. More detailed measurements of the rates and free energies
656: of the various states will no doubt prove our simplified model incomplete and provide motivation and
657: insight into a more detailed model of channel gating.
658: The simplicity and generality of the competition between applied tension and line tension suggests that
659: it may be a quite general phenomena for mechanotransduction. We hope to apply similar ideas to other
660: mechanosensitive systems. More generally, we are also intrigued by the possibility of finding
661: analogous bilayer deformation driven conformational changes for other transmembrane proteins which do
662: not exhibit mechanosensitive function, perhaps illuminating a more general qualitative design
663: principle for the function of transmembrane proteins.
664:
665:
666:
667:
668: \section{Methods}
669: \subsection{Lipids}
670: When discussing the lipids used by other authors, we have used the same naming convention they employed:
671: 10:0 dicaproyl-phosphatidylcholine (PC10),
672: 12:0 dilauroyl-phosphatidylcholine (PC12),
673: 14:1 dimirstoyl-phosphatidylcholine (PC14),
674: 16:1 dipalmitoleoyl-phosphatidylcholine (PC16),
675: 18:1 dioleoyl-phosphatidylcholine (PC18,DOPC),
676: 20:1 Eicossenoyl-phosphatidylcholine (PC20),
677: lysophospholipid (LPL),
678: lysophosphatidylcholine (LPC),
679: dioleoyl-phosphatidylethanolamine (DOPE).
680: We have preformed calculations based on elastic moduli measured by Rawicz {\it et al.} \cite{length}.
681: To extrapolate the effects of changes in the acyl chain length, we use the moduli for a typical 18:1
682: phosopholipid (1,2-dioleoyl-sn-glycero-3-phosphocholine) and the scaling relations listed below
683: \begin{equation}
684: \begin{array}{cclcc}
685: 2a & = & 37\ {\rm \AA} & \propto & a^1 \\
686: K_A & = & 0.64\ kT\, {\rm \AA}^{-2} & \propto & a^1 \\
687: K_B & = & 21\ kT & \propto & a^3 \\
688: {\cal K} & = & 2.1 \times 10^{-3}\ kT\, {\rm \AA}^{-3} & \propto & a^0
689: \end{array}
690: \end{equation}
691: where $2a$ is the equilibrium thickness of the bilayer, $K_A$ is the area expansion modulus,
692: $K_B$ is the bending modulus, and
693: \begin{equation}
694: \label{calk}
695: {\cal K} \equiv \sqrt{2}\left(\frac{K_A^3K_B}{a^6}\right)^{1/4},
696: \end{equation}
697: is the composite elastic modulus appearing in the line tension, $f_U$. The proportionality to
698: bilayer thickness is approximate only and is deduced from treating the bilayer as a thin shell.
699: For more elaborate models and scaling arguments, see Rawicz {\it et al.} \cite{length}. By
700: fitting the Peak-to-Peak head group thickness \cite{length} for saturated and mono-unsaturated lipids,
701: we find that the relation between bilayer thickness and acyl chain length is roughly
702: \begin{equation}
703: 2a = 1.3n+17\ {\rm \AA}.
704: \end{equation}
705: Again, more elaborate models and scaling arguments are found in Rawicz {\it et al.} \cite{length}.
706:
707: \subsection{Line tension}
708: In terms of $h(r)$, the bilayer mid-plane position, and $u(r)$, half the difference of the bilayer
709: thickness and the equilibrium thickness,
710: the mean curvature contributions to the free energy density are \cite{helfrich}
711: \begin{equation}
712: {\cal G}_{\rm B} = {\textstyle\frac{K_B}{2}}[\underbrace{\left(\nabla^2 h\right)^2+
713: \left(\nabla^2 u\right)^2}_{\cal M}-\underbrace{C_+\nabla^2[h+u]-C_-\nabla^2[u-h]}_{\partial \cal M}],
714: \end{equation}
715: where the variation of the ${\cal M}$ terms contribute to the action in the bulk (bilayer),
716: the $\partial{\cal M}$ terms are total derivatives and can be evaluated at the
717: interface, the constant terms are dropped, $K_B$ is the bending modulus for the bilayer,
718: and $C_\pm$ are the spontaneous curvatures of the upper and lower
719: leaflets, respectively. The tension contribution to the free energy density is \cite{goulian}
720: \begin{equation}
721: {\cal G}_{\alpha} = \underbrace{{\textstyle\frac{\alpha}{2}}\left(\nabla h\right)^2}_{\cal M},
722: \end{equation}
723: where the thickness deformation term can be ignored.
724: The interaction free energy density between the two layers is \cite{huang}
725: \begin{equation}
726: {\cal G}_{\rm I}= \underbrace{\frac{K_A}{2a^2}u^2}_{\cal M},
727: \end{equation}
728: where $K_A$ is the area expansion modulus and $a$ is the equilibrium thickness of a leaflet.
729:
730: %\subsection{Equilibrium equations and solutions}
731: The equilibrium equations that result from the minimization of $u(r)$ and $h(r)$ are
732: \begin{eqnarray}
733: 0 = \frac{\delta G[u, h]}{\delta u} &=& \left[K_B \nabla^4+\frac{K_A}{a^2}\right]u(r), \\
734: 0 = \frac{\delta G[u, h]}{\delta h} &=& \left[K_B \nabla^4-\alpha \nabla^2\right]h(r).
735: \end{eqnarray}
736: In the asymptotic regime, large $R$, we can ignore the curvature of the interface and replace the Laplacians
737: with $d^2/dr^2$. For $u(r)$, the radially decreasing solutions are exponentials with complex wave numbers
738: \begin{equation}
739: \beta_\pm = -\left(\frac{K_A}{a^2K_B}\right)^{1/4}e^{\pm i\pi/4}.
740: \end{equation}
741: For $h(r)$, there is only a single radially decreasing solution with decay length:
742: \begin{equation}
743: \beta_h = \sqrt\frac{\alpha}{K_B}.
744: \end{equation}
745: At the interface, the bilayer deforms to match the hydrophobic thickness of the protein:
746: \begin{equation}
747: u(R) = U = {\textstyle\frac{1}{2}}W-a.
748: \end{equation}
749: If the inclusion interface is relatively flat and the top and bottom surfaces of the bilayer are normal
750: to the inclusion interface as depicted in figure \ref{basic}, the radial slope of $u(r)$ is zero at the interface:
751: \begin{equation}
752: u'(R) = 0,
753: \end{equation}
754: although this too may be used as a geometrical parameter \cite{ndan4}.
755: For $h(r)$, we set the radial slope at the interface:
756: \begin{equation}
757: h'(R) = H'.
758: \end{equation}
759: At infinity, the bilayer is unperturbed, implying
760: \begin{eqnarray}
761: u(\infty) &=& 0, \\
762: u'(\infty) &=& 0,\\
763: h(\infty) &=& 0.
764: \end{eqnarray}
765: The equilibrium solutions can be found by matching the boundary conditions above.
766:
767: %\subsection{Bilayer Deformation Line Tensions}
768: The line tension is
769: \begin{equation}
770: f = \int_R^\infty dr \left({\cal G}_{\rm B}+{\cal G}_{\alpha}+{\cal G}_{\rm I}\right).
771: \end{equation}
772: Integrating this equation by parts twice yields a bulk term proportional to the
773: equilibrium equations, which is zero since the equations are satisfied, and surface terms.
774: The total line tension is:
775: \begin{equation}
776: f = \underbrace{{\textstyle \frac{1}{2}} {\cal K} U^2}_{f_U} + \underbrace{{\textstyle \frac{1}{2}}
777: \sqrt{K_B\alpha} H'^2}_{f_H} + \underbrace{K_BCH'}_{f_C},
778: \end{equation}
779: where $\cal K$ has been defined in equation \ref{calk}, $C\equiv {\textstyle \frac{1}{2}}(C_+-C_-)$ is the
780: composite spontaneous curvature, and $f_H$ is the mid-plane deformation contribution to the line tension
781: which is typically small enough to ignore.
782:
783: \begin{thebibliography}{99}
784:
785: \bibitem{moe}
786: Blount, P. \& Moe, P. C.
787: Bacterial mechanosensitive channels: integrating physiology, structure and function.
788: {\it Trends. Microbiol.} {\bf 7}, 420-424 (1999).
789:
790: \bibitem{Gill}
791: Gillespie, P.G. \& Walker, R.G.
792: Molecular basis of mechanosensory transduction.
793: {\it Nature} {\bf 413}, 194-202 (2001).
794:
795: \bibitem{hamill}
796: Hamill, O. \& Martinac, B.
797: Molecular Basis of Mechanotransduction in Living Cells.
798: {\it Physiol. Rev.} {\bf 81}, 685-740 (2001).
799:
800:
801:
802: \bibitem{suk1}
803: Sukharev, S.I., Sigurdson, W.J., Kung, C. \& Sachs, F.
804: Energetic and Spatial Parameters for Gating of the Bacterial Large Conductance
805: Mechanosensitive Channel, {\it MscL}.
806: {\it J. Gen Physiol.} {\bf 113}, 525-539 (1999).
807:
808:
809:
810: \bibitem{perozo1}
811: Perozo, E., Kloda, A., Marien Cortes, D. \& Martinac, B.
812: Physical principles underlying the transduction of bilayer deformation forces during
813: mechanosensitive channel gating.
814: {\it Nat. Struct. Bio.} {\bf 9}, 696-703 (2002).
815:
816:
817: \bibitem{perozo2}
818: Perozo, E., Kloda, A., Marien Cortes, D. \& Martinac, B.
819: Open channel structure of MscL and the gating mechanism of mechanosensitive channels.
820: {\it Nature} {\bf 418}, 942-948 (2002).
821:
822:
823: \bibitem{Rees}
824: Chang, G., Spencer, R.H., Lee, A.T., Barclay, M.T. \& Rees, D.C.
825: Structure of the MscL homologue from Mycobacterium tuberculosis: A gated mechanosensitive ion channel.
826: {\it Science} {\bf 282}, 2220-2226 (1998).
827:
828:
829: \bibitem{suk2}
830: Sukharev, S.I., Durell, S.R. \& Guy, H.R.
831: Structural Models of the MscL Gating Mechanism.
832: {\it Biophys. J.} {\bf 81}, 917-936 (2001).
833:
834:
835: \bibitem{md1}
836: Gullingsrud, J.R., Kosztin, D. \& Schulten, K.
837: MscL gating studied by molecular dynamics simulations.
838: {\it Biophys. J. } {\bf 80}, 2074-2081 (2001).
839:
840:
841: \bibitem{huang}
842: Huang, H.W.
843: Deformation Free Energy of Bilayer Membrane and its Effect on Gramicidin Channel Lifetime.
844: {\it Biophys. J.} {\bf 50}, 1061-1070 (1986).
845:
846: \bibitem{helfrich}
847: Helfrich, W.
848: Elastic Properties of Lipid Bilayers.
849: {\it Z. Naturforsch. C.} {\bf 28}, 693-703 (1973).
850:
851: \bibitem{nielsen}
852: Nielsen, C., Goulian, M. \& Andersen, O.
853: Energetics of Inclusion-Induced Bilayer Deformations.
854: {\it Biophys. J.} {\bf 74}, 1966-1983 (1998).
855:
856: \bibitem{goulian}
857: Goulian, M., Mesquita, O.N., Fygenson, D.K., Nielsen, C., Andersen, O.S. \& Libshaber, A.
858: Gramicidin Channel Kinetics under Tension.
859: {\it Biophys. J.} {\bf 74}, 328-337 (1998).
860:
861:
862: \bibitem{ndan4}
863: Dan, N. \& Safran, S.A.
864: Effect of Lipid Characteristics on the Structure of Transmembrane Proteins.
865: {\it Biophys. J.} {\bf 75}, 1410-1414 (1998).
866:
867: \bibitem{Killian}
868: Killian, J. A.
869: Hydrophobic mismatch between proteins and lipids in membranes.
870: {\it Biochimica et Biophysica Acta} {\bf 1376}, 401-416 (1998).
871:
872: \bibitem{israelachvili}
873: Israelachvili, J. N.
874: {\it Intermolecular and Surface Forces.} 2nd edn. (Academic Press,London, 1991.)
875:
876: \bibitem{gruner}
877: Gruner S.M.
878: Stability of Lyotropic Phases with Curved Interfaces.
879: {\it J. Phys. Chem.} {\bf 93}, 7562-7570 (1989).
880:
881: \bibitem{kittel}
882: Kittel, C. \& Kroemer, H.
883: {\it Thermal Physics.} 2nd ed. (W.H. Freeman and Company, New York, 1980.)
884:
885:
886: \bibitem{keller}
887: Keller, S.L., Bezrukov, S. M., Gruner, S.M., Tate, M.W. \& Vodyanoy, I.
888: Probability of Alamethicin Conductance States Varies with NonLamellar Tendency of Bilayer Phospholipids.
889: {\it Biophys. J.} {\bf 65}, 23-27 (1993).
890:
891: \bibitem{ou}
892: Ou, X., Blount, P., Hoffman, R. J. \& Kung, C.
893: One face of a transmembrane helix is crucial in mechanosensitive channel gating.
894: {\it PNAS} {\bf 95}, 11471-11475 (1998).
895:
896: \bibitem{yoshimura}
897: Yoshimura, K., Batiza, A., Schroeder, M., Blount, P. \& Kung, C.
898: Hydrophilicity of a Single Residue within {\it MscL} Correlates with Increased Channel Mechanosensitivity.
899: {\it Biophys. J.} {\bf 77}, 1960-1972 (1999).
900:
901: \bibitem{frauen}
902: Frauenfelder, H., Sligar, S.G. \& Wolynes, P.G.
903: The energy landscapes and motions of proteins.
904: {\it Science} {\bf 254}, 1598-1603 (1991).
905:
906: \bibitem{hanggi}
907: Goychuk, I. \& H\"anggi, P.
908: Ion channel gating: A first-passage time analysis of the Kramers type.
909: {\it PNAS} {\bf 99}, 3552-3556 (2002).
910:
911: \bibitem{Bass}
912: Bass, R. B., Strop, P., Barclay, M. \& Reese, D. C.
913: Crystal Structure of Escherichia col MscS, a Voltage-Modulated and Mechanosensitive Channel.
914: {\it Science} {\bf 298}, 1582-1587 (2002).
915:
916: Strop, P., Bass, R. \& Rees, D.C.
917: Prokaryotic mechanosensitive channels. {\it Adv. Prot. Chem.}
918: {\bf 63}, 177-209 (2003).
919:
920:
921:
922: \bibitem{suk3}
923: Chiang, C.S., Anishkin, A. \& Sukharev, S.
924: {\it Biophys. J.} in press (2003).
925:
926: \bibitem{length}
927: Rawicz, W., Olbrich, K.C., McIntosh, T., Needham, D. \& Evans, E.A.
928: Effect of Chain Length and Unsaturation on Elasticity of Lipid Bilayers.
929: {\it Biophys. J.} {\bf 79}, 328-339 (2000).
930:
931: \bibitem{dekruijff}
932: de Kruijff, B.
933: Lipids beyond the bilayer.
934: {\it Nature} {\bf 386}, 129-130 (1997).
935:
936: \bibitem{JGIII}
937: Gullingsrud, J. \& Schulten, K.
938: Gating of MscL Studied by Steered Molecular Dynamics.
939: {\it Biophys. J.} {\bf 85}, 2087-2099 (2003).
940: \end{thebibliography}
941:
942:
943: \section{Acknowledgments}
944: We are grateful to Doug Rees, Tom Powers, Jan\'e Kondev, Klaus Schulten, Evan Evans,
945: Sergei Sukharev, Eduardo Perozo, Olaf Andersen, Sylvio May, Ben Freund, and Mandar Inamdar
946: for useful discussions, suggestions, and corrections. We also acknowledge support from
947: the Keck Foundation, NSF and the NSF funded Center for Integrative Multiscale
948: Modeling and Simulation. PAW acknowledges support through an NSF fellowship.
949:
950:
951:
952:
953: \end{document}
954:
955:
956:
957:
958:
959:
960:
961:
962:
963:
964:
965:
966:
967:
968: