cond-mat0108362/bio.tex
1: %\documentstyle[preprint,psfig,pre,multicol,aps]{revtex}
2: \documentstyle[graphicx,multicol,pre,aps]{revtex}
3: 
4: %\def\sech{\mbox{sech}}
5: %\def\ds{\displaystyle}
6: 
7: \begin{document}
8: 
9: \draft
10: 
11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
12: \title{Nonlinearity-induced conformational instability and dynamics 
13: of biopolymers}
14: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
15: 
16: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
17: \author{Serge F. Mingaleev$^{1,2,3}$,
18: Yuri B. Gaididei$^{2,3}$, Peter L. Christiansen$^{2}$,
19: and Yuri S. Kivshar$^{1}$}
20: \address{$^{1}$ Nonlinear Physics Group,  
21: Research School of Physical Sciences and Engineering, \\
22: Australian National University, Canberra ACT 0200, Australia \\
23: $^{2}$ Department of Mathematical Modeling,
24: The Technical University of Denmark, DK-2800 Lyngby, Denmark \\
25: $^{3}$ Bogolyubov Institute for Theoretical Physics, 14-B Metrologichna
26: St., 03143 Kiev, Ukraine}
27: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28: 
29: %\date{\today}
30: \maketitle
31: 
32: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33: \begin{abstract}
34: We propose a simple phenomenological model for describing the
35: conformational dynamics of biopolymers via the nonlinearity-induced
36: buckling and collapse (i.e. coiling up) instabilities. 
37: We describe the buckling instability analytically, and then
38: demonstrate the role of both instabilities 
39: in the folding of biopolymers through the 
40: numerical simulations of the three-dimensional dynamics of 
41: a long semiflexible chain in the aqueous environment. 
42: \end{abstract}
43: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
44: 
45: \pacs{87.15.-v; 36.20.-r}
46: 
47: \begin{multicols}{2}
48: \narrowtext
49: 
50: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
51: %%\section{Introduction}
52: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
53: 
54: Conformational flexibility is a fundamental property of polymers 
55: which differentiates them from small molecules and gives rise to
56: their remarkable properties. 
57: However, even the properties of polymers seem meagre when compared 
58: to the functionality of biopolymers. 
59: A distinctive feature of biopolymers is that, on the one hand, 
60: they are heterogeneous and, on the other, their elementary sub-units 
61: have a complex structure and can carry long-lived nonlinear 
62: excitations \cite{myoglobin,peyrard,davydov,dna,scott,cruzeiro}. 
63: While the significance of the heterogeneity gained 
64: wide recognition as being of fundamental importance 
65: for protein folding \cite{Taylor:2001-apr:RPP}, 
66: the function of nonlinear excitations remains obscured, 
67: in spite of a large body of theoretical research 
68: \cite{peyrard,davydov,dna,scott}. 
69: 
70: Here we suggest {\em a new role of nonlinear 
71: excitations} in one of the most important functionalities of 
72: biopolymers, their {\em conformational dynamics}. 
73: For the first time to our knowledge, we demonstrate 
74: that nonlinear excitations 
75: can facilitate bending or folding of a semiflexible 
76: molecular chain via the buckling or collapse 
77: instability, and thus provide a 
78: possible scenario for the conformational dynamics of 
79: biopolymers.
80: We consider a semiflexible chain that models a biopolymer and 
81: assume that its mechanical degrees of freedom are coupled to the
82: internal degrees of freedom such as amide-I vibrations 
83: in proteins \cite{myoglobin,peyrard,davydov}, base-pair 
84: vibrations in DNA \cite{peyrard,dna}, or polarons in 
85: proteins \cite{peyrard,davydov}. 
86: This coupling provides the energy needed to
87: overcome the potential barriers for folding, thus avoiding the need
88: for stochastic thermal activation. 
89: We reveal that such a simple model describes the conformational
90: dynamics as {\em a deterministic controlled process}: when the
91: amplitude of the internal excitation becomes large, the
92: chain displays the buckling or collapse instability that
93: subsequently initiates the chain folding.
94: We believe that such simple physics can also account for the
95: kinetics of conformational phase transitions of semiflexible
96: polymers in solutions \cite{geom}.
97: 
98: The importance of nonlinear excitations has already been 
99: emphasized in the modeling of biopolymers and conjugated 
100: polymers. In particular,
101: it was suggested that solitons may provide a possible physical mechanism
102: for the energy (or charge) transport and storage in proteins 
103: \cite{peyrard,davydov}
104: and conducting  polymers \cite{polymers}. The concept of nonlinear
105: localized modes was also employed to explain some
106: specific features of the DNA dynamics \cite{peyrard,dna}. 
107: However, most of those
108: studies were dealing with nonlinear properties of  {\em straight} 
109: molecular chains and never discussed changes in their geometrical
110: structure. The study of nonlinear properties of {\em curved chains} has been
111: initiated only recently \cite{curv} and, in particular,
112: it was shown that the bending in a curved chain manifests 
113: itself as an effective trap for nonlinear localized modes. 
114: In this Letter, we make an important step forward and, 
115: taking into account {\em flexibility} of biopolymer chains, 
116: demonstrate that nonlinear excitations may act as drivers 
117: giving impetus to {\em conformational dynamics} of 
118: biopolymers.
119: 
120: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
121: %%\section{Model}
122: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
123: 
124: Let us consider a simple phenomenological model of a biopolymer 
125: consisting of particles with mass $M$, indexed by $n$, and located at the
126: points 
127: $\bbox{r}_n=\{x_n, \, y_n, \, z_n\}$.  We assume that each particle
128: represents a complex sub-unit of the biopolymer which, 
129: additionally to its position $\bbox{r}_n(t)$, 
130: carries an internal excitation which can 
131: be characterized, in some approximation 
132: \cite{peyrard,davydov,dna,scott,polymers}, 
133: by the complex amplitude, 
134: $\psi_n(t)$. Such an internal mode
135: can represent, for instance, a polaron state 
136: (an excess electron accompanied by phonons) or vibrational state
137: (amide-I vibrations in proteins or base-pair vibrations in DNA).
138: The Hamiltonian of such a chain is written as
139: $H= T + U + V$, where 
140: $T= (M/2) \sum_{n} (d \bbox{r}_n/dt)^2$ is the kinetic energy, 
141: $U(\bbox{r}_n)$ is the potential energy
142: of inter-particle interactions, and 
143: \begin{displaymath}
144: V(\psi_n, \bbox{r}_n)= \sum_{n} \left\{ 2|\psi_n|^2 -
145: \sum_{m \neq n} J_{n m} \psi^*_n \psi_{m} -
146: \frac{1}{2} \,\chi\, |\psi_n|^4 \right\} 
147: \end{displaymath}
148: is the energy of the internal excitation.
149: Here $\chi$ characterizes on-site self-trapping 
150: nonlinearity of the internal excitation, and 
151: $J_{nm}$ are the excitation transfer coefficients that 
152: depend on the distance in the embedding space  between the 
153: particles $n$ and $m$: 
154: $J_{n m} \equiv J(|\bbox{r}_n -\bbox{r}_m|)= (e^{\alpha}-1) 
155: \exp\left(-\alpha |\bbox{r}_n-\bbox{r}_{m}|\right)$ 
156: \cite{note0}. 
157: From the Hamiltonian, we obtain the equations of motion 
158: \begin{eqnarray}
159: %%\begin{array}{l}
160: M \frac{d^2 \bbox{r}_n}{dt^2} + \nu \frac{d \bbox{r}_n}{dt} + 
161: \frac{d U}{d \bbox{r}_n} - \sum_{k} \sum_{m \neq k}
162: \frac{d J_{k m}}{d \bbox{r}_n} \psi^*_k \psi_{m} = 0 \; , 
163: \nonumber \\
164: i \frac{\partial \psi_n}{\partial t} - 2 \psi_n +
165: \sum_{m \neq n} J_{nm} \psi_m + \chi \, |\psi_n|^2 \psi_n =0 \; ,
166: %%\end{array} 
167: \label{eq-mot}
168: \end{eqnarray}
169: where we allowed additionally for a viscous damping $\nu$ of the 
170: aqueous environment. We assume that the number of excitations, 
171: $\sum_{n} |\psi_n|^2=1$, is conserved by Eq. (\ref{eq-mot}). 
172: This approximation is fully justified for polaron states 
173: and should be fairly good for vibrational states, 
174: taking into account that in the case under study the lifetime 
175: of amide I vibrations is greatly inceased as a result of their 
176: self-trapping localization \cite{myoglobin,cruzeiro}.
177: 
178: We consider the potential energy of inter-particle interactions as a sum, 
179: $U(\bbox{r}_n)=U_{S}+U_{B}+U_{R}$, of the stretching energy 
180: %%\begin{displaymath}
181: $U_{S} = (\sigma/2) \sum_{n} (|\bbox{r}_n-\bbox{r}_{n-1}|-a)^2$, 
182: %%\label{poten-US}
183: %%\end{displaymath}
184: bending energy \cite{note2} 
185: \begin{equation}
186: U_{B} = \frac{\kappa}{2} \sum_{n}
187: \frac{\theta_n^2}{1-(\theta_n / \theta_{\rm max})^2} \; , 
188: \label{poten-UB}
189: \end{equation}
190: and (to describe correctly coiling up of the chain) the 
191: energy of short-range repulsive interactions between particles 
192: (considered as elastic balls of the diameter $d$): 
193: \begin{equation}
194: U_{R} = \frac{\delta}{2} \sum_{n} \sum_{m \neq n}
195: (d-|\bbox{r}_n-\bbox{r}_{m}|)^2 \; ,
196: \label{poten-UR}
197: \end{equation}
198: for $|\bbox{r}_n-\bbox{r}_{m}|<d$, and $U_{R} = 0$,
199: otherwise.
200: Here $a$ is the equilibrium lattice spacing (we take $a=1$), 
201: $\sigma$ is the elastic modulus of the stretching 
202: rigidity of the chain, $\theta_n$ is the angle 
203: between neighboring bond vectors, 
204: $(\bbox{r}_{n} - \bbox{r}_{n-1})$ and $(\bbox{r}_{n+1} - \bbox{r}_{n})$, 
205: that meet at the $n$-th particle [see Fig. \ref{fig:chain}(a)],
206: $\theta_{\rm max}$ is a maximum bending angle, and 
207: $\kappa$ is the elastic modulus of the bending rigidity of the chain. 
208: 
209: First of all, we recall the properties 
210: of nonlinear localized modes in a curved chain, when the 
211: chain geometry is ``frozen''. Such an analysis has been 
212: recently carried out in Ref. \onlinecite{curv}, where it was shown 
213: that the chain curvature affects strongly the properties of 
214: nonlinear modes, creating an effective double-well potential around 
215: the bending point, which 
216: is responsible for the symmetry-breaking effect: 
217: as the nonlinearity $\chi$ increases above some threshold, 
218: a symmetric 
219: stationary mode becomes unstable, and it transforms into
220: an energetically more favorable asymmetric stationary mode.
221: In that study, it was also found that the energy of a 
222: nonlinear mode decreases as a square-law of the curvature 
223: when the curvature is small. 
224: Thus, since the bending energy of the chain increases
225: also as the square of the curvature, one can expect that
226: there should exist a converse influence of the internal nonlinear
227: modes on the conformational dynamics of the chain. 
228: Below, we confirm this conjecture by straightforward 
229: numerical calculations and find that not only the chain tends
230: to bend, for small enough $\kappa$, but it also 
231: collapses (i.e. coils up), for smaller $\kappa$.
232: 
233: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
234: %%\section{Ground State}
235: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
236: 
237: As the first step of our study in this Letter, we analyze 
238: the stationary ground-state configurations of 
239: the chain described by the model introduced above. 
240: We assume that the chain is infinitely long, 
241: planar ($z_n = 0$), and inextensible ($\sigma \to \infty$),
242: so that the distance between  the neighboring sites of the chain remains
243: constant. In this case,  a  spatial conformation of the chain is 
244: completely defined by the bending angles $\theta_n$.
245: 
246: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
247: %\vspace{2mm}
248: \begin{figure}
249: \begin{center}
250: \begin{minipage}{35mm}
251: \includegraphics[width=35mm,angle=0,clip]{bio-fig1a.eps}
252: \centerline{{\large\bf (a)}}
253: \end{minipage}
254: %%\hspace*{1mm}
255: \begin{minipage}{47mm}
256: \includegraphics[width=47mm,angle=0,clip]{bio-fig1b.eps} \\[6mm]
257: %%\vspace*{3mm} 
258: \centerline{{\large\bf (b)}}
259: \end{minipage}
260: \end{center}
261: %%\vspace{3mm}
262: \caption{(a) Schematic representation 
263: of a buckled chain.  The distance between the sites
264: near the bend becomes smaller  ($R_2<R_1$) lowering 
265: the coupling energy of a nonlinear 
266: mode that works against the increased bending energy and lead to 
267: a buckling instability. (b) Ground state spatial configurations 
268: of a semiflexible chain  ($\alpha=2$, $\theta_{\rm max}=\pi/3$,
269: and $\kappa=0.125$) at different values of $\chi$: 
270: 2.13~(i), 2.26~(ii), 2.44~(iii), 2.88~(iv), 3.00~(iv), 
271: 3.13~(iii), 3.18~(ii), and 3.25~(i). As the nonlinearity $\chi$ 
272: increases, the chain bending initially increases, reaches its 
273: maximum at $\chi\approx 2.96$, and then decreases.}
274: \label{fig:chain}
275: \end{figure}
276: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
277: 
278: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
279: %\vspace{2mm}
280: \begin{figure}
281: \centerline{\hbox{
282: \includegraphics[width=70mm,angle=0,clip]{bio-fig2.eps}}}
283: %%\vspace{3mm}
284: \caption{Phase diagram of the 
285: ground states of a semiflexible biopolymer chain calculated 
286: numerically (solid curves with squares) and analytically 
287: from Eqs.~(\protect\ref{kappa-cr})--(\ref{kin}) (dashed curve), at 
288: $\alpha=2$, $\theta_{\rm max}=\pi/3$, and $\sigma=\infty$.}
289: \label{fig:kappa-lambda}
290: \end{figure}
291: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
292: 
293: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
294: %%\subsection{Numerical Results}
295: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
296: 
297: We find the ground-state configurations of
298: the chain looking for the stationary solutions
299: of Eq. (\ref{eq-mot}) in the form
300: $\psi_n(t) = e^{i \Lambda t} \phi_n$ and 
301: $d \bbox{r}_n/dt = 0$. The corresponding 
302: system of nonlinear algebraic equations
303: has been solved by the Newton-Raphson iteration scheme.
304: The results of these calculations are presented 
305: in Fig. \ref{fig:kappa-lambda}. The most interesting feature of this
306: phase diagram is the existence of {\em three distinct 
307: types} of the ground states.
308: 
309: A stiff chain (with $\kappa>0.183$) is always straight 
310: and, regardless of the value of $\chi$,  it cannot 
311: be affected by the presence of a nonlinear excitation. 
312: However, for $\kappa<0.183$, there appears a finite domain 
313: of $\chi$, for which a straight chain becomes unstable 
314: in the presence of a nonlinear mode. Inside of this domain, 
315: the chain in its ground state {\em curves symmetrically} 
316: around the nonlinear excitation, with some finite overall buckling 
317: angle $\beta=\sum_n \theta_{n}$ [see Fig. \ref{fig:chain}(b)].
318: In Fig. \ref{fig:beta-lambda}, we plot the dependence of
319: the buckling angle $\beta$ vs. $\chi$ for 
320: different values of~$\kappa$. As is seen, for $0.124<\kappa<0.183$ the 
321: buckling angle $\beta$ is always finite and 
322: reaches its maximum at $\chi \simeq 3$.
323: 
324: When the bending rigidity of the chain decreases further,
325: i.e. for $\kappa<0.124$, a new type of the ground state,
326: {\em a collapsed chain}, emerges 
327: (see Figs. \ref{fig:kappa-lambda} and
328: \ref{fig:beta-lambda}). In this case, the distance
329: between the complementary particles of the chain 
330: [see Fig. \ref{fig:chain}(b)] decreases 
331: to an extent that the growing interaction  between the tails
332: of the nonlinear excitation turns buckling into collapse 
333: causing the coiling up of the chain. 
334: The conformation of the chain in the collapsed state 
335: (which is usually a three-dimensional compact globule, 
336: see Fig. \ref{fig:dyn}) 
337: is sensitive to the peculiarities of the potentials 
338: (\ref{poten-UB})--(\ref{poten-UR}) and initial conditions.
339: 
340: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
341: %\vspace{2mm}
342: \begin{figure}
343: \centerline{\hbox{
344: \includegraphics[width=73mm,angle=0,clip]{bio-fig3.eps}}}
345: %%\vspace{3mm}
346: \caption{Buckling angle
347: $\beta$ vs. the nonlinearity $\chi$ for 
348: $\alpha=2$, $\theta_{max}=\pi/3$, $\sigma=\infty$, 
349: and varying $\kappa$.}
350: \label{fig:beta-lambda}
351: \end{figure}
352: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
353: 
354: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
355: %%\subsection{Variational Approach}
356: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
357: 
358: To gain a better insight into the physical 
359: mechanism of the buckling instability,
360: we have developed a simple variational approach. 
361: We assume that the chain bending occurs only at the center of the 
362: excitation ($n=0$). In this case, the chain is composed of two
363: semi-infinite lines with the overall buckling angle  $\beta \equiv
364: \theta_{0}$
365: between them. Obviously, such a geometry cannot describe the chain collapse 
366: because $\beta$ is restricted by  the maximum angle $\theta_{\rm
367: max}$,  but it should be valid as long as we are not interested in the
368: collapsed
369: phase but study only a threshold between the straight and buckled 
370: chain phases.
371: 
372: In such a simplified model, the ground state of the chain is determined
373: by the dependence of the system energy on $\beta$. 
374: We use a variational ansatz for the localized mode in the form
375: %%\begin{displaymath}
376: %%\label{var-phi}
377: $\phi_n=\sqrt{\tanh \gamma\,} \, \exp(-\gamma \, |n|)$, 
378: %%\end{displaymath}
379: which satisfies the constraint  $\sum_n \phi_n^2=1$, and we minimize $H$ 
380: with respect to $\beta$ and $\gamma$. 
381: Thus, we can show that the straight configuration of the chain 
382: ($\beta=0$) corresponds to the energy maximum (i.e., the straight 
383: chain is unstable) for $\kappa\,<\kappa_{\rm cr}$,
384: where the critical value of the bending stiffness
385: is given by the following analytical expression 
386: \begin{eqnarray}
387: \label{kappa-cr}
388: \kappa_{\rm cr}=\chi \, \alpha (e^{\alpha}-1)
389: \frac{3\,e^{\alpha+\gamma}-1}{3 \,
390: (e^{\alpha+\gamma}-1)^3} \, \tanh \gamma \; ,
391: \end{eqnarray}
392: and $\gamma$ is determined from the equations
393: \begin{eqnarray}
394: \label{kin}
395: \chi = \frac{8 \cosh^4\gamma}{(4 \cosh^2\gamma-3)}\,
396: \frac{\partial Q(\gamma)}{\partial \gamma} \; , 
397: \nonumber \\
398: Q(\gamma) = \frac{e^{\alpha} \, (e^{\gamma}-1)^2
399: (e^{\alpha+2\gamma} +1)}{(e^{2\gamma}+1) \,
400: (e^{\alpha+\gamma}-1)^2} \; .
401: \end{eqnarray}
402: In Fig. \ref{fig:kappa-lambda}, we compare the
403: threshold curve that separates the phases of the 
404: straight and buckled conformations 
405: predicted by the analytical formula 
406: (\ref{kappa-cr})--(\ref{kin}), 
407: with the same dependence calculated numerically. 
408: One can see that the variational method provides
409: reasonably fair analytical results for the phase diagram.
410: 
411: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
412: %%\section{Dynamical simulations}
413: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
414: 
415: An important question is how the buckling and
416: collapse instabilities manifest themselves in the 
417: dynamics of semiflexible biopolymer chains. To clarify this issue,
418: we consider a simple (but physically important) case of an 
419: initially straight chain of a finite  length whose left
420: end is free whereas the right end is fixed. 
421: First, we excite a single particle at the left end of the chain, and
422: then keep track of how this excitation propagates along the chain
423: and how it modifies the chain geometry. For small values 
424: of $\chi$, such a single-particle excitation disperses and  spreads out,
425: whereas for very large values of $\chi$ it does not propagate
426: and remains completely trapped at the end of the chain. 
427: However, for the intermediate values of $\chi$ (in our model, for 
428: $2.5 < \chi <3.5$), such a single-particle drive applied at the end of 
429: the chain generates {\em a moving localized mode}. 
430: It is remarkable that this interval of nonlinearity 
431: parameter $\chi$ matches closely the region in which 
432: the bending capability of the nonlinear mode 
433: reaches its maximum. 
434: 
435: 
436: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
437: %%\subsection{Bending of semiflexible polymer chains}
438: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
439: 
440: For a stiff biopolymer chain (i.e., for  $\kappa > 0.183$), 
441: the propagation of the generated nonlinear mode does not produce
442: any conformational change of the chain.  However, when $\kappa$ 
443: decreases, the nonlinear mode generates a local chain bending. 
444: In this case the nonlinear mode propagates
445: along the chain {\em being accompanied by a local  bending};  
446: the amplitude of such a bending decreases as the viscous damping $\nu$ 
447: increases. Thus, the buckling instability
448: we have described above 
449: appears to be physically visible, and it is responsible, in particular,
450: for large-amplitude  localized bending waves. 
451: 
452: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
453: %%\subsection{Collapse of flexible polymer chains}
454: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
455: 
456: Let us consider now a flexible biopolymer chain with even 
457: smaller $\kappa$, for which the nonlinear mode should produce a collapse
458: instability (see Fig. \ref{fig:kappa-lambda}). In this case, the 
459: evolution of the chain
460: clearly demonstrates a collapse dynamics (see Fig.
461: \ref{fig:dyn}), and the chain folds 
462: into a compact coil with 
463: several loops consisting of 7--10 particles each
464: \cite{note1}.
465: Importantly, even for large values of the damping parameter $\nu$ 
466: the collapse of the chain is only delayed, but not prevented. 
467: For instance, for the parameters used in 
468: Fig. \ref{fig:dyn} the collapse-on time $t_c$ grows almost 
469: linearly with $\nu$: $t_c \approx 87 + 1290 \nu + 283 \nu^2$ 
470: for $\nu \leq 1$. Besides, increasing of $\nu$ leads to 
471: decreasing of the distance between the place of the collapse 
472: nucleation and the left end of the chain. 
473: 
474: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
475: %%\section{Conclusions}
476: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
477: 
478: It should be mentioned that the instability 
479: we discuss in this Letter remains latent 
480: in a straight infinitely long chain, because the
481: bending of such a chain 
482: would require an infinite energy.  However, as we have 
483: demonstrated above, the instability
484: manifests itself as soon as we consider more realistic cases and
485: take into account a finite length of the chain.  
486: In this case, a nonlinear
487: localized mode, once generated at the end of the chain, 
488: propagates along the chain 
489: {\em accompanied by a local chain bending}, for 
490: intermediate values of the bending rigidity $\kappa$, 
491: and it {\em causes the chain folding}, 
492: for smaller $\kappa$. 
493: 
494: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
495: %\vspace{2mm}
496: \begin{figure}
497: \begin{center}
498: \includegraphics[width=78mm,angle=0,clip]{bio-fig4.eps}
499: \end{center}
500: %%\vspace{3mm}
501: \caption{Dynamics of a semiflexible biopolymer chain due to 
502: the collapse instability 
503: produced by a moving nonlinear mode generated from the 
504: single-particle excitation at the left end of the chain. 
505: The chain consists of 30 particles, with the right end fixed. 
506: Marked particles indicate location of the moving excitation 
507: (for these particles, $|\psi_n|^2>0.25$).
508: The parameters are: $\chi=3.2$, $\kappa=0.06$, $\nu=0.3$, 
509: $\sigma=10^3$,  $\alpha=2$, 
510: $\theta_{\rm max}=\pi/3$, $M=0.5$, $\delta=10^2$, and $d=0.6$.}
511: \label{fig:dyn}
512: \end{figure}
513: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
514: 
515: In conclusion, we have predicted analytically 
516: and demonstrated numerically a new role played by nonlinear 
517: excitations in the conformational 
518: dynamics of biopolymers. Taking into account
519: the coupling between the internal and mechanical degrees of freedom of
520: a semiflexible biopolymer chain, we have found that it may produce 
521: the buckling and collapse instabilities of an initially straight chain. 
522: In the case of collapse instability, the viscous damping of the aqueous 
523: environment only slows down the folding (coiling up) of the chain, 
524: but does not stop it even for relatively large damping ($\nu \simeq 1$). 
525: It should be stressed that these 
526: instabilities are most pronounced for {\em intermediate values} of the 
527: self-trapping nonlinearity, and they vanish completely 
528: in the linear limit (i.e., for $\chi=0$). 
529: We find that this effect is only weakly affected by 
530: the peculiarities of the interaction potentials \cite{note0,note2}, 
531: and thus it should be generic for different models of nonlinear 
532: semiflexible chains. 
533: 
534: We believe that the nonlinearity-induced buckling 
535: and collapse instabilities may be of a considerable 
536: importance in the conformational dynamics of the 
537: globular macromolecules 
538: with a multitude of quasi-isoenergetic conformations 
539: \cite{Taylor:2001-apr:RPP}, 
540: where nonlinear modes excited even far from the chain ends 
541: should cause {\em conformational transitions} between 
542: different types of the globular states of the same or similar topology.
543: Presumably, another related problem is the energy 
544: transduction in molecular motors, for which it was recently 
545: suggested that the free energy 
546: from ATP hydrolysis is initially converted into the 
547: hinge-bending motion of the motor sub-unit 
548: \cite{Wang:1998:Nature}. 
549: 
550: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
551: 
552: The authors are indebted to M.~Peyrard, 
553: K.~Rasmussen, S.~Takeno, 
554: J.~Tuszynski, and A.~Zolotaryuk for useful and 
555: encouraging comments. 
556: S.~M. and Yu.~G. acknowledge support from the European 
557: Commission RTN project LOCNET (HPRN-CT-1999-00163) and the 
558: MIDIT center at the Technical University of Denmark, S.~M. 
559: acknowledges supports from the the Performance and Planning 
560: Fund of the Institute for Advanced Studies, the Australian 
561: National University, and Yu.~G. from the DLR Project No. 
562: UKR-002-99 and the SRC ``Vidhuk'' (Kiev).
563: 
564: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
565: \begin{thebibliography}{99}
566: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
567: 
568: \bibitem{myoglobin}
569: A. Xie, L. Meer, W. Hoff, and R. H. Austin, 
570: Phys. Rev. Lett. {\bf 84}, 5435 (2000); 
571: J. Edlers, P. Hamm, and A. C. Scott, 
572: {\em Femtosecond study of self-trapped vibrational excitons
573: in crystalline acetanilide}, submitted for publication.
574: 
575: \bibitem{peyrard}
576: {\em Nonlinear Excitations in Biomolecules}, 
577: Ed.: M. Peyrard (Springer, Berlin, 1995).
578: 
579: \bibitem{davydov}
580: A. S. Davydov, {\em Solitons in Molecular Systems} 
581: (Reidel, Dordrecht, 1985); A.C. Scott, 
582: Phys. Rep. {\bf 217}, 1 (1992).
583: 
584: \bibitem{dna}
585: L. V. Yakushevich, {\em Nonlinear Physics of DNA}
586: (Wiley, New York, 1998), and references therein. 
587: 
588: \bibitem{scott}
589: A. Scott, {\em Nonlinear Science: Emergence and Dynamics of 
590: Coherent Structures} 
591: (Oxford University Press, New York, 1999), and references therein. 
592: 
593: \bibitem{cruzeiro}
594: L. Cruzeiro-Hansson, Phys. Rev. Lett. {\bf 73}, 2927 (1994).
595: 
596: \bibitem{Taylor:2001-apr:RPP}
597: See, e.g., 
598: W. R. Taylor, A. C. W. May, N. P. Brown, and A. Aszodi, 
599: Rep. Prog. Phys. {\bf 64}, 517 (2001).
600: 
601: \bibitem{geom}
602: See, e.g., 
603: A. L. Kholodenko and T. A. Vilgis,  Phys. Rep. {\bf 298}, 254 (1998), 
604: and references therein. 
605: 
606: \bibitem{polymers}
607: A. J. Heeger {\em et al.}, Rev. Mod. Phys. {\bf 60}, 781 (1988).
608: 
609: \bibitem{curv}
610: Yu. B. Gaididei, S. F. Mingaleev, and P. L. Christiansen,
611: Phys. Rev. E {\bf 62}, R53 (2000);
612: P. L. Christiansen, Yu. B. Gaididei, and S. F. Mingaleev,
613: J. Phys.: Cond. Matter {\bf 13}, 1181 (2001).
614: 
615: \bibitem{note0}
616: Discussed effects are tolerant to 
617: a choice of $J_{nm}$, provided the interaction extends 
618: beyond the approximation of the nearest neighbors. 
619: We have verified the validity of our predictions for 
620: other types of $J_{nm}$, in particular, for the 
621: dipole-dipole interaction.
622: 
623: \bibitem{note2}
624: We have checked that the use of other types of the 
625: bending potential leads to qualitatively same results. 
626: 
627: \bibitem{note1}
628: We use $\theta_{\rm max}=\pi/3$, so
629: that 7 is a minimum number of particles in such 
630: a loop.
631: 
632: \bibitem{Wang:1998:Nature}
633: H. Wang and G. Oster, Nature {\bf 396}, 279 (1998).
634: 
635: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
636: \end{thebibliography}
637: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
638: 
639: \end{multicols}
640: \end{document}
641: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
642: 
643: