1: \documentclass{elsart}
2: \usepackage{amsmath,epsfig}
3: \begin{document}
4: \begin{frontmatter}
5: \title{Perturbed soliton excitations in DNA double helix }
6:
7: \author{M.~Daniel\corauthref{cor1}$^{a,b}$},
8: \ead{daniel@cnld.bdu.ac.in}
9: \author{V.~Vasumathi$^{a,b}$}
10:
11: \corauth[cor1]{Corresponding Author. Fax:+91-431-2407093}
12: \address{a. Centre for Nonlinear Dynamics, Department of Physics,
13: Bharathidasan University, Tiruchirappalli - 620 024, India.\\
14: b. The Abdus Salam International Centre for Theoretical Physics,
15: Strada Costiera-11,
16: 34014 Trieste, Italy}
17: \date{}
18:
19:
20: \begin{abstract}
21: We study nonlinear dynamics of inhomogeneous DNA double helical chain under
22: dynamic plane-base rotator model by considering angular rotation of bases
23: in a plane normal to the helical
24: axis. The DNA dynamics in this case is found to be governed
25: by a perturbed sine-Gordon equation while taking into account the interstrand hydrogen
26: bonding energy between bases and the intrastrand inhomogeneous stacking energy and by making an analogy
27: with the Heisenberg model of the Hamiltonian of an inhomogeneous anisotropic spin ladder with
28: ferromagnetic legs and antiferromagnetic rung coupling. In the homogeneous limit the dynamics is
29: governed by the kink-antikink soliton of the sine-Gordon equation which represents the
30: formation of open
31: state configuration in DNA double helix. The effect of inhomogeneity
32: in stacking energy in the form of
33: localized and periodic variations on the formation of open states in DNA
34: is studied under perturbation. The perturbed soliton is obtained using a multiple scale soliton perturbation theory by solving
35: the associated linear eigen value problem and by constructing the complete set of eigen functions.
36: The inhomogeneity in stacking energy is found to modulate the width and speed of the soliton
37: depending on the nature of inhomogeneity. Also it introduces fluctuations
38: in the form of train of
39: pulses or periodic oscillations in the open state
40: configuration.
41: \end{abstract}
42: \begin{keyword}
43:
44: Soliton \sep DNA \sep Multiple Scale Perturbation .\\
45: \\
46:
47:
48: \PACS
49:
50: 87.10.+e \sep
51: 87.15.He \sep
52: 66.90.+r \sep
53: 63.20.Ry
54: \end{keyword}
55:
56:
57: \end{frontmatter}
58:
59:
60: \section{Introduction}
61: A number of theoretical models have been proposed to describe nonlinear molecular
62: excitations in DNA double helix which plays an important role in the
63: conservation and transformation of genetic information in biological
64: systems \cite{ref1}.
65: These theoretical models are based on
66: longitudinal and transverse motions, as well as bending,
67: stretching and rotations \cite{ref2,ref3}.
68: Among the different motions,
69: the rotational motion of bases in DNA is found to contribute more towards the opening
70: of base pairs.
71: The first contribution towards nonlinear dynamics of DNA was made by Englander and his
72: co-workers \cite{ref4} who studied the dynamics of DNA open states by taking into account only the rotational
73: motion of nitrogenous bases, which made the main contribution towards the formation of open states. Yomosa \cite{ref6,ref7} developing this idea further proposed a dynamic plane base rotator model which
74: is a generalized version of the Frenkel-Kontrova \cite{ref8} model that was later improved by
75: Takeno and Homma \cite{ref9,ref10} in which attention was paid to the degree of freedom, characterizing base
76: rotations in the plane perpendicular to the helical axis around the backbone structure. In the above, the DNA dynamics was governed by the completely integrable sine-Gordon model admitting kink-type solitons. Then
77: Peyrard and Bishop \cite{ref11} studied the process of denaturation in which only the
78: transverse motion of
79: bases along the hydrogen bond was taken into account. There was one more
80: model studied by
81: Christiansen and his colleagues (see for e.g.\cite{ref12}) using Toda lattice model in which two types of internal motions namely, transverse motion
82: along the hydrogen bond direction and longitudinal motion along the backbone direction were found
83: to contribute to DNA denaturation process in terms of travelling solitary waves and standing waves.
84: These localized nonlinear excitations further explain conformation transition
85: \cite{ref14,ref15,ref16},
86: long range interaction of kink solitons in the double chain \cite{ref17,ref18}, regulation of transcription
87: \cite{ref17,ref19},
88: denaturation \cite{ref11} and
89: charge transport in terms of polarons and bubbles \cite{ref21}. Some of them have been successfully used for interpreting
90: experimental data related to
91: microwave
92: absorption \cite{ref22,ref23}. Further developments, in this approach for several years
93: was limited to small
94: improvements of the models involving only numerical methods of simulation of the internal
95: dynamics of DNA \cite{ref24,ref25,ref13}. Also, bubbles \cite{ref20} discrete breathers \cite{ref26,ref27,ref28} and non-breathing compacton-like modes were obtained by
96: solving the DNA-lattice model \cite{ref29}. Eventhough, the rotation of bases in DNA is mainly due to thermal forces the thermal fluctuation in DNA dynamics has been introduced through random forces in the recent past. For instance, Yakushevich et al \cite{ref13} has shown that topological solitons of the DNA chains are stable with respect to thermal oscillations. Since random thermal forces introduce only small fluctuations, it is not included in the present study. Thus, the study of nonlinear excitations in DNA molecular chain has become an important
97: task since it is related to its major functions.\\
98:
99: In all the above studies, DNA double helix with homogeneous stiff strands has been considered for the
100: analysis. However, in nature,
101: the presence of different sites along the strands such as promotor, coding, terminator etc. each
102: of which having a very specific sequence of bases and particular functions makes the strands
103: site-dependent or inhomogeneous(soft). Also, defects caused due to external molecules
104: in the sequence and the presence of
105: abasic site-like nonpolar mimic of thymine lead to inhomogeneity \cite{ref30,ref31}. When included, the DNA dynamics is governed by an inhomogeneous perturbed sine-Gordon equation and thus the problem boils down to solving the same and finding perturbed solitons. Also, in a different context, in the recent times, the study of wave propagation, especially solitons through inhomogeneous or disordered media assumed lot of interest \cite{ref5}. For instance, kink-impurity interaction and its scattering in the sine-Gordon model was studied in detail by Zhang Fei et al \cite{ref32}. With this in mind, in the present paper we study the nonlinear
106: dynamics of DNA double helix with inhomogeneous strands by considering a plane
107: base rotator model along the lines of Yomosa.
108: The paper is organized as follows.
109: In section 2 we introduce a dynamic plane base rotator model
110: for the site-dependent DNA double helix and derive the nonlinear dynamical equation.
111: In section 3, we study the effect of inhomogeneity in stacking energy on the open state of DNA in
112: terms of kink-antikink solitons by solving a perturbed sine-Gordon equation using a
113: multiple scale soliton perturbation theory. The results are concluded in section 4.
114: Detailed evaluation of few integrals using residue theorem is given in Appendix.
115:
116: \section{Plane-base rotator Model and Dynamical equation}
117:
118: In Fig.1a we have presented a schematic
119: structure of the B-form
120: DNA double helix. Here $S$ and $S'$ represent the two complementary strands in the DNA double helix.
121: Each arrow in the figure represents the direction of the base attached
122: to the strand and the dots between arrows represent the net hydrogen bonding effect between the
123: complement bases. While a horizontal
124: projection of the $n^{th}$ base pair in the XY-plane is represented in Fig.1b, in Fig.1c, we have given the
125: projection of the same in the XZ-plane.
126: The Z-axis is chosen along the helical axis of the DNA.
127: In Fig.1b, $Q_n$ and $Q'_n$ denote the tip of the $n^{th}$ bases belonging to the
128: complementary strands $S$ and $S'$. $P_n$ and $P'_{n}$ represent the points where the bases in the $n^{th}$ base
129: pair are attached to the strands $S$ and
130: $S'$ respectively. Let $(\theta_n,\phi_n)$ and $(\theta'_n,\phi'_n)$ represent the
131: angles of rotation of the bases in the $n^{th}$ base pair around the points $P_n$ and $P'_n$ in the XZ and
132: XY-planes respectively.\\
133: \begin{figure}
134: \begin{center}
135: \epsfig{file=fig1.ps,height =9cm, width=12cm}
136: \caption{(a) A schematic structure B-form DNA double helix. (b) A horizontal projection of the $n^{th}$ base
137: pair in the XY-plane. (c) A projection of the $n^{th}$ base pair in the XZ-plane.}
138: \end{center}
139: \end{figure}
140: The conformation and stability of DNA double helix is mainly determined by the stacking
141: energy between the intrastrand adjacent bases, the hydrogen bonding energy between the interstrand
142: complementary bases and other energies. From a heuristic argument it was assumed that the interstrand
143: base-base interaction or
144: hydrogen bonding energy of the given base pair depends on the distance between them. Thus, from Fig.1b we
145: can write down the square of the distance between the edges of the arrows $(Q_nQ'_n)^{2}$
146: as
147: \begin{eqnarray}
148: ( Q_nQ'_n)^{2}&=&2+4 r^{2}+(z_n-z'_n)^2+2 (z_n-z'_n)
149: \left(\cos\theta_n -\cos\theta'_n\right) \nonumber\\
150: &&-4r\left[\sin\theta_n\cos\phi_n
151: +\sin\theta'_n\cos\phi'_n\right]
152: +2\left[\sin\theta_n\sin\theta'_n
153: \right.\nonumber\\
154: &&\times\left(\cos\phi_n\cos\phi'_n+\sin\phi_n\sin\phi'_n\right)
155: \left.-\cos\theta_n\cos\theta'_n\right],\label{eq1}
156: \end{eqnarray}
157: where `$r$' is the radius of the circle depicted in Fig.1b.
158: The base-base interaction energy can be understood in a
159: more clear and transparent way by introducing quasi-spin operators ${\bf
160: S_n}=(S_n^{x}, S_n^{y}, S_n^{z})$ and $ {\bf S'_n}=(S_n^{'x}, S_n^{'y}, S_n^{'z})$ in the form \\
161: \begin{subequations}
162: \begin{eqnarray}
163: { S_n^{x}} = \sin\theta_n\cos\phi_n,\quad
164: { S_n^{y}} = \sin\theta_n\sin\phi_n,\quad
165: { S_n^{z}} = \cos\theta_n,\\
166: { S_n^{'x}} = \sin\theta'_n\cos\phi'_n,\quad
167: { S_n^{'y}} = \sin\theta'_n\sin\phi'_n,\quad
168: { S_n^{'z}} = \cos\theta'_n,
169: \end{eqnarray}
170: \end{subequations}
171: for the $n^{th}$ bases in the $S^{th}$ and $S^{'th}$ strands respectively. In view of
172: this, Eq.~(\ref{eq1}) can
173: be written in terms of the given spin operators as follows.
174: \begin{eqnarray}
175: (Q_n Q'_n)^2=2+4r^2+2\left[S_n^{x} S_n^{'x}+S_n^{y} S_n^{'y}
176: -S_n^{z} S_n^{'z}\right]
177: -4r\left[S_n^{x}+S_n^{'x}\right].\label{eq3}
178: \end{eqnarray}
179: While writing Eq.(\ref{eq3}) we have neglected the longitudinal compression along the direction of the
180: helical axis thereby choosing $z_n=z'_n$.
181: It may be noted that the form of $({Q_nQ'_n})^2$ given in Eq.(\ref{eq3}) is the same as the Hamiltonian
182: for a generalized form of the Heisenberg spin model. Therefore, the intrastrand base-base interaction
183: in DNA can be
184: written using the same consideration. It is known that stacking or base-base interaction is a dominant force that stabilizes the DNA double helix. It is much stronger than the hydrogen bonding force and in fact the stacking between adjacent bases often contributes more than half of the total free energy of the base pairs \cite{ref32a}. Several non-covalent forces including dipole-dipole interaction, van der Waals force etc stabilize stacking in DNA. It is also reasonable to think that if such a quasi-spin model
185: can be used in this problem, the double strand DNA and the rung-like base pairs can be
186: conceived as an anisotropic coupled spin chain model or spin ladder. This is schematically
187: presented in Fig.2. In the figure, $S$ and $S'$ which previously represented two strands of the DNA
188: here correspond to two ferromagnetic lattices of a spin ladder with
189: antiferromagnetic coupling among the rungs. In the case of spin chains each arrow represents
190: the magnetic moment corresponding to a group of atoms
191: at that lattice point. Due to antiferromagnetic type of rung coupling the arrows in the
192: lattice $S$ and $S'$ are marked anti-parallel to each other. Here Z-direction (i.e) the
193: direction of the helical axis is chosen as the easy axis of
194: magnetization in the spin chain. Further, the spin-spin exchange interaction is restricted. to
195: the nearest neighbours, (i.e) the $n^{th}$
196: spin is coupled to the spins at the $(n+1)^{th}$ and $(n-1)^{th}$ sites.
197: \begin{figure}
198: \begin{center}
199: \psfig{file=fig2.ps,height =3cm, width=9cm}
200: \caption{A schematic representation of DNA as
201: an anisotropic coupled spin chain model or spin ladder.}
202: \end{center}
203: \end{figure}
204: \\
205:
206: With this consideration we use the following Heisenberg model of the
207: Hamiltonian for an anisotropic coupled spin chain model or
208: spin ladder with site-dependent or inhomogeneous ferromagnetic-type
209: exchange interaction between nearest neighbouring spins in the same lattice (equivalent to coupling among bases in
210: the same strand i.e. intrastrand interaction) and antiferromagnetic rung-coupling (between
211: bases belonging to the complementary strands i.e. interstrand interaction).
212: \begin{eqnarray}
213: H&=&\sum_n \left[-J f_n\left({S^{x}_{n}} {S^{x}_{n+1}}+{S^{y}_{n}}
214: {S^{y}_{n+1}}\right)-K f_n\right.
215: {S^{z}_{n}} {S^{z}_{n+1}}-J' f'_n\left({S^{'x} _{n}} {S^{'x}_{n+1}}\right.\nonumber\\
216: &&\left.+{S^{'y} _{n}} {S^{'y}_{n+1}}\right)
217: -K' f'_n {S^{'z}_{n}} {S^{'z}_{n+1}}
218: +\eta\left(S_n^{x}
219: {S_n^{'x}} +{S^{y}_n} {S^{'y}_n}\right)+\mu {S_n^{z}} {S^{'z}_n}\nonumber\\
220: &&
221: \left.+ A (S_n^{z})^{2}+A'
222: ({S^{'z}_n})^{2}\right]. \label{eq4}
223: \end{eqnarray}
224: In the above Hamiltonian $J$ and $J^{'}$ correspond to the intrastrand interaction constant or the
225: stacking energy between the $n^{th}$ base and its nearest neighbours in the plane normal to the helical
226: axis in the strands $S$ and $ S'$ respectively.
227: When K and $K'$ are not equal to J and $J'$ respectively, they introduce anisotropy in the intrastrand
228: interaction. $\mu$ and $\eta $ represent a measure of
229: interstrand interaction or hydrogen bonding energy between the bases of similar sites in both the
230: strands
231: along the direction of the helical axis and in a plane
232: normal to it respectively. Here we have assumed that there exist on an average almost uniform
233: interactions
234: $A $ and $A^{'}$ that assume positive values which are the uniaxial anisotropy coefficients
235: leading to rotation of the bases in a plane normal to the helical axis. The quantities
236: $ f_n$ and $f'_n$ in Hamiltonian ~(\ref{eq4}) indicate that the intrastrand stacking energy
237: between bases in the $S^{th}$ and $S^{'th}$ strand varies in a specified site dependent
238: fashion which leads to inhomogeneity in the DNA
239: double helical
240: chain. In general $f_n$ and $f'_n$ may take different values for different base sequences.
241: The inhomogeneity in DNA double
242: helix may arise due to any one of the reasons mentioned in the previous section. The effect of
243: inhomogeneity in DNA nonlinear dynamics has been studied in the past by several
244: authors in various contexts such as variational mass density of bases
245: \cite{ref33} and defect \cite{ref26}. \\
246:
247: To proceed further we rewrite Hamiltonian~(\ref{eq4})
248: in terms of the variables $(\theta_n,\phi_n)$ and $ (\theta'_n,\phi'_n)$ using
249: Eqs.(2) and obtain
250: \begin{eqnarray}
251: H&=&\sum_n\left[-J f_n \sin\theta_n\sin\theta_{n+1} \cos
252: (\phi_{n+1}-\phi_n)\right.
253: -K f_n \cos\theta_n\cos\theta_{n+1}\nonumber\\
254: &&-J' f'_n\sin\theta'_n\sin\theta'_{n+1}
255: \cos (\phi'_{n+1}-\phi'_n)-K' f'_n \cos\theta'_n\cos\theta'_{n+1}\nonumber\\
256: &&+\eta \sin\theta_n\sin\theta'_n
257: \cos (\phi_n-\phi'_n)+\mu\cos\theta_n\cos\theta'_n+A\cos\theta_n^{2}\nonumber\\
258: &&\left.+A'
259: {\cos\theta'_n}^{2}\right].\label{eq5}
260: \end{eqnarray}
261: The quasi-spin model thus introduced implies that the dynamics of bases in DNA can be described by the
262: following equations of motion.
263: \begin{eqnarray}
264: \dot\theta_n=\frac{1}{\sin\theta_n} \frac{\partial{H} } {\partial {\phi_n}},~
265: \dot\phi_n=\frac{-1}{\sin\theta_n}\frac{\partial{H}} {\partial {\theta_n}} ,~
266: \dot\theta'_n=\frac{1}{\sin\theta'_n} \frac{\partial{H} } {\partial {\phi'_n}},~
267: \dot\phi'_n=\frac{-1}{\sin\theta'_n}\frac{\partial{H}} {\partial {\theta'_n}} .\label{eq6}
268: \end{eqnarray}
269: In Eq.(\ref{eq6}) the overdot represents time derivative.
270: When the anisotropy energies $A$ and $A'$ are much larger than the other
271: interactions, (i.e) when $A,A'>>J,J',K,K',\eta,\mu$, then the equations of
272: motion ~(\ref{eq6}) on substituting the Hamiltonian ~(\ref{eq5}) become
273: \begin{eqnarray}
274: \dot \phi_n= 2 A \cos\theta_n,~ \dot\phi'_n= 2 A'\cos\theta'_n. \label{eq7}
275: \end{eqnarray}
276: The other two equations in (\ref{eq6}) satisfy identically.
277: Using Eq.~(\ref{eq7}) in the Hamiltonian ~(\ref{eq5}) we obtain
278: \begin{eqnarray}
279: H&=&\sum_{n}\left[\frac{I}{2} {\dot\phi_n}^{2}+\frac{I'}{2}{\dot\phi_n}^{'2}
280: -J f_n \sin\theta_n\sin\theta_{n+1}\right.
281: \cos(\phi_{n+1}-\phi_n)
282: -K f_n \cos\theta_n\nonumber\\
283: &&\times\cos\theta_{n+1}-J' f'_n\sin\theta'_n\sin\theta'_{n+1}
284: \cos (\phi'_{n+1}-\phi'_n)
285: -K' f'_n \cos\theta'_n\cos\theta'_{n+1}\nonumber\\
286: &&+\eta \sin\theta_n\sin\theta'_n
287: \left.\cos (\phi_n-\phi'_n)+\mu\cos\theta_n\cos\theta'_n\right],\label{eq8}
288: \end{eqnarray}
289: where $I=\frac{1}{2A}$ and $ I'=\frac{1}{2A'}$.
290: The above Hamiltonian can be rewritten in the limits of plane-base rotator
291: $(\theta_n=\theta'_n=\pi/2)$ and absolute minima of
292: potential as
293: \begin{eqnarray}
294: H&=&\sum_n\left[ \frac{I}{2} {\dot\phi_n}^{2}+\frac{I'}{2} {\dot\phi_n}^{'2} +J f_n
295: \left[1-\cos (\phi_{n+1}-\phi_n)\right]\right.\nonumber\\
296: &&+J' f'_n \left[1-\cos (\phi'_{n+1}-\phi'_n)\right]
297: \left. -\eta\left[1-\cos (\phi_n-\phi'_n)\right]\right]. \label{eq9}
298: \end{eqnarray}
299: It may be noted that the above limit corresponds to the XY-spin model of two coupled inhomogeneous
300: ferromagnetic spin system.
301: In Hamiltonian (\ref{eq9}) the first two terms represent the kinetic energies of the rotational motion
302: of the $n^{th}$ nucleotide bases accompanied by the potential energy associated with the $n^{th}$ nucleotide sugar
303: and phosphate and its complementary unit around the axes at $P_n$ and $P'_n$ (see Fig.1b).
304: $I$ and $I'$ are the moments of inertia of the
305: nucleotides around the axes at $P_n$ and $P'_n$ respectively. It may be further noted that in the new Hamiltonian the term proportional
306: to $\mu$ vanishes.
307: Now, using Hamiltonian (\ref{eq9}), the Hamilton's equations of motion can be immediately written as
308: \begin{subequations}
309: \begin{eqnarray}
310: I\ddot\phi_n&=&J \left[f_n \sin (\phi_{n+1}-\phi_n)-f_{n-1}\sin
311: (\phi_n-\phi_{n-1})\right]
312: +\eta \sin (\phi_n-\phi'_n), \label{eq10a}\\
313: I'\ddot\phi'_n&=&J' \left[f'_n \sin (\phi'_{n+1}-\phi'_n)-f'_{n-1}\sin
314: (\phi'_n-\phi'_{n-1})\right]
315: +\eta \sin (\phi'_n-\phi_n).\label{eq10b}
316: \end{eqnarray}
317: \end{subequations}
318: Eqs.(\ref{eq10a}) and (\ref{eq10b}) describe the dynamics of DNA in a plane-base rotator model at the discrete level
319: while considering the dominant angular rotation of the bases in a plane
320: normal to the helical axis and ignoring all other small motions of the bases.\\
321:
322: It is expected that in the B-form of DNA double helix the difference in the angular rotation of bases with respect to
323: neighbouring bases along the two strands is small \cite{ref9,ref10}. Hence we assume that
324: $\sin(\phi_{n\pm 1}-\phi_n)\approx(\phi_{n\pm 1}-\phi_n)$ and
325: $ \sin(\phi'_{n\pm 1}-\phi'_n)\approx(\phi'_{n\pm 1}-\phi'_n)$
326: in Eqs.(\ref{eq10a},~b). Also, as the length of the DNA chain is very large involving several thousands of base pairs compared to the distance between the
327: neighbouring bases along the strands we make a continuum approximation as done by several authors in the past \cite{ref2,ref3,ref4,ref6,ref7,ref8,ref9,ref10,ref11} which is valid in the long wavelength, low temperature limit
328: by introducing two fields of rotational angles,
329: $\phi_n(t)\rightarrow\phi(z,t)$, $\phi'_n(t)\rightarrow\phi'(z,t)$ and
330: two inhomogeneous stacking fields, $f_{n}\rightarrow f(z)$ and $f'_{n}\rightarrow f'(z)$ along with
331: the following expansions.
332: \begin{eqnarray}
333: \phi_{n\pm1}=\phi(z,t)\pm a\frac{ \partial{\phi}}{\partial{z}}+\frac{
334: a^{2}}{2!}\frac{\partial^{2}{\phi}}{{\partial{z}}^{2}}\pm...,
335: f_{n\pm 1}=f(z)\pm a \frac{\partial{f}}{\partial{z}}+\frac{ a^{2}}{2!}
336: \frac{\partial^{2}{f}}{{\partial{z}}^{2}}\pm...,\label{eq 11}
337: \end{eqnarray}
338: where `$a$' is the lattice parameter along both
339: the strands $S$ and $S'$. In a similar way we write down expansions for $\phi'_{n\pm 1}$ and $f'_{n\pm 1}$.
340: As the inhomogeneity here is site-dependent and associated with the bases themselves, we have chosen same
341: lattice parameter `$a$' for both $\phi_{n\pm 1}$ and $f_{n\pm 1}$.
342: Under this continuum approximation the equations of motion (\ref{eq10a},~b) upto O($a^{2}$) is written as
343: \begin{subequations}
344: \begin{eqnarray}
345: I\phi_{tt}&=& Ja^{2} \left[ f(z) \phi_{zz}+ f_z \phi_z\right]+ \eta\sin
346: (\phi-\phi'),\\
347: I' \phi'_{tt}&=& J'a^{2} \left[ f'(z) \phi'_{zz}+ f'_z \phi'_z\right] +\eta\sin
348: (\phi'-\phi).
349: \end{eqnarray}
350: \end{subequations}
351: In Eqs.~(12) the suffices $t$ and $z$ represent partial derivatives with respect
352: to time $t$
353: and the spatial variable $z$ respectively.\\
354:
355: In a DNA chain, the two strands are expected to exhibit similar type of macroscopic physical
356: behaviour and
357: hence we assume that $I=I',~ J=J'$ and $ f=f'$.
358: In view of this, Eqs.~(12)
359: after rescaling the time variable as
360: $\hat{t}=\sqrt{\frac{Ja^{2}}{I}} t
361: $ and choosing $\eta=-\frac{Ja^{2}}{2}$ for future convenience can be written as
362: \begin{subequations}
363: \begin{eqnarray}
364: \phi_{\hat t\hat t}&=&\left[f(z)\phi_{zz}+f_z \phi_{z} \right]
365: -\frac{1}{2} \sin (\phi-\phi'),\label{eq13a}\\
366: \phi'_{\hat t\hat t}&=&\left[f(z)\phi'_{zz}+f_z
367: \phi'_{z}\right]-\frac{1}{2} \sin (\phi'-\phi).\label{eq13b}
368: \end{eqnarray}
369: \end{subequations}
370: It is more convenient to describe the transverse motion of bases in DNA strands in terms of the centre of
371: mass co-ordinates. For this, we rewrite Eqs.~(\ref{eq13a}) and (\ref{eq13b})
372: by subtracting and adding them respectively.
373: \begin{subequations}
374: \begin{eqnarray}
375: (\phi-\phi')_{\hat t\hat t}&=&f(z)(\phi-\phi')_{zz}+f_z(\phi-\phi')_{z}
376: -\sin (\phi-\phi'),
377: \label{eq14a}\\
378: (\phi+\phi')_{\hat t\hat t}&=&f(z)(\phi+\phi')_{zz}+f_z(\phi+\phi')_{z}.
379: \label{eq14b}
380: \end{eqnarray}
381: \end{subequations}
382: In a different context while studying the magnetoelastic effect induced by interaction between two
383: ferromagnetically coupled XY-spin chains in the static limit Dandoloff and
384: Saxena \cite{ref36}
385: obtained similar equations. To commence the open state configuration in DNA, the two complementary bases are expected to
386: rotate in opposite directions so
387: that $\phi=-\phi'$, and in this case Eq.~(\ref{eq14b}) satisfies identically and
388: Eq.~(\ref{eq14a}) becomes
389: \begin{eqnarray}
390: \Psi_{\hat t\hat t}=f \Psi_{zz}+f_z\Psi_z-\sin\Psi , \label{eq15}
391: \end{eqnarray}
392: where $\Psi=2\phi$. Assuming small inhomogeneity along the strands by
393: choosing $f(z)=1+\epsilon g(z)$ where $\epsilon$ is a small parameter
394: and $g(z)$ a measure of the inhomogeneity,
395: Eq.~(\ref{eq15}) can be written as
396: \begin{eqnarray}
397: \Psi_{\hat t\hat t}-\Psi_{zz}+\sin\Psi=\epsilon \left[g(z)\Psi_{z}\right]_{z}.\label{eq16}
398: \end{eqnarray}
399: Eq.~(\ref{eq16}) describes the dynamics of bases under rotation in a plane-base rotator model of an inhomogeneous DNA double helical
400: chain. When $\epsilon=0 $,
401: Eq.~(\ref{eq16}) reduces to the completely
402: integrable sine-Gordon equation which admits kink and antikink-type of soliton solutions and
403: hence we call Eq.(\ref{eq16}) as a perturbed sine-Gordon equation.
404: The integrable sine-Gordon equation $(\epsilon=0$ case) was originally solved for N-soliton
405: solutions using the most celebrated Inverse Scattering Transform (IST) method by Ablowitz and his
406: co-workers \cite{ref37}. The kink and antikink soliton solutions of the integrable sine-Gordon equation
407: (Eq.(\ref{eq16}) when $\epsilon=0$) are depicted in Figs.3a and 3b.
408: The kink-antikink solitons of the sine-Gordon equation describe an open
409: state configuration in the DNA double helix. The formation of open state configuration in terms of
410: kink-antikink pair in DNA double helical chain is schematically represented in Fig.3c. In this figure the base
411: pairs are found to open locally in the form of kink-antikink
412: structure in each strand and propagate along the direction of the helical
413: axis.
414: \begin{figure}
415: \begin{center}
416: \epsfig{file=fig3.eps,height =10cm, width=12cm}
417: \caption{(a) Kink and (b) antikink soliton solutions of the sine-Gordon equation (Eq. (\ref{eq16}) when
418: $\epsilon=0$ ).
419: (c) A sketch of the formation of open state configuration in terms of
420: kink-antikink solitons in DNA double helix.}
421: \end{center}
422: \end{figure}
423:
424: \section{ Effect of stacking energy Inhomogeneity on the Open State }
425: \subsection{ A Perturbation approach}
426: When inhomogeneity in stacking is present (i.e) when terms proportional to
427: $\epsilon$ are present in Eq.(\ref{eq16}), the
428: inhomogeneity is expected to perturb the kink and antikink solitons corresponding
429: to the open state of DNA. One of the most powerful techniques in dealing with perturbed soliton
430: equations is the soliton perturbation theory which is based on the IST method
431: \cite{ref38,ref39}.
432: However, as the method is
433: very sophisticated it is very difficult to use the same in
434: several cases. In view of this, a direct method to study the soliton perturbation was first
435: introduced by Gorshkov and Ostrovskii \cite{ref40} and later many authors
436: used different types of direct methods to study soliton perturbation (see for e.g. refs.
437: \cite{ref41,ref42,ref43,ref44,ref45,ref46}).
438: The characteristic feature of this method is that the perturbed
439: nonlinear equation is linearized by expanding its solution about the
440: unperturbed solution and the eigen functions for the
441: operator associated with the linearized equation are found out. The complete solution is then written
442: in terms of these eigen functions \cite{ref47,ref48}. In
443: these methods the basic fact that the presence of perturbation not
444: only modifies the shape of the soliton by a correction of linear dispersion
445: but also undergoes a slow time change of the soliton parameters have been acknowledged.
446: In this paper, we use one such direct perturbation method which is also dealt in reference
447: \cite{ref49} in a different context
448: to solve the perturbed sine-Gordon equation (\ref{eq16}) and to understand
449: the effect of inhomogeneity in stacking
450: energy on the open state configuration of DNA.
451: The procedure we adapt here is based on the derivative
452: expansion method to linearize the perturbed sine-Gordon equation in the
453: co-ordinate frame attached to the moving frame. The parameters of the
454: kink-antikink soliton are assumed to depend on a slow time scale in order to
455: eliminate the secular terms. The linearized equations will be solved using
456: the method of separation of variables which ultimately will be related to an eigen
457: value problem, the eigen functions of which form the bases of the
458: perturbed solution. The eigen functions contain information about
459: the time dependence of the soliton parameters and help to calculate the
460: perturbed soliton. In the following we use the above approach to find
461: the perturbed soliton solution of Eq.(\ref{eq16}).
462: \subsection{Linearization of the perturbed sine-Gordon equation}
463: When the perturbation is absent (i.e) when $\epsilon =0$ in
464: Eq.~(\ref{eq16}) the unperturbed integrable sine-Gordon equation
465: provides N-soliton solutions and the one soliton solution (see also Figs.3a and 3b is written as
466: \begin{eqnarray}
467: \Psi(z,\hat t) =4 arc\tan{exp[\pm m(z-v\hat t)]} ,\quad m=\frac{1}{\sqrt{1-v^2}}. \label{eq17}
468: \end{eqnarray}
469: In Eq.(\ref{eq17}) while the upper sign corresponds to kink soliton, the lower sign represents the antikink
470: soliton.
471: Here $ v$ and $m^{-1}$ are real parameters that determine the velocity and width of the soliton
472: respectively. In order to study the effect of perturbation the time variable $\hat t$ is transformed
473: into several variables as $t_n=\epsilon^{n} \hat t$ where n=0,1,2,... and $\epsilon$ is a very small parameter.
474: In view of this, the
475: time derivative in Eq.~(\ref{eq16}) is replaced by the expansion
476: \begin{eqnarray}
477: \frac{\partial }{\partial \hat t}=\frac{\partial }{\partial t_0}+\epsilon~
478: \frac{\partial }{\partial t_1}+\epsilon^2 \frac{\partial }{\partial
479: t_2}+....\label{eq18}
480: \end{eqnarray}
481: Simultaneously $\Psi$ is expanded in an asymptotic series as
482: \begin{eqnarray}
483: \Psi=\Psi^{(0)}+\epsilon \Psi^{(1)}+\epsilon^2 \Psi^{(2)}+....\label{eq19}
484: \end{eqnarray}
485: Using the above expansions for $\hat{t}$ and $\Psi$ in Eq.~(\ref{eq16}) and equating
486: the coefficients of different powers of $\epsilon$, we obtain the following
487: equations.
488: \begin{subequations}
489: \begin{eqnarray}
490: \epsilon^{(0)}:\qquad \Psi^{(0)}_{t_0t_0}-\Psi^{(0)}_{zz}+\sin\Psi^{(0)}=0,
491: \qquad\qquad\qquad\qquad \qquad\quad\label
492: {eq20a}\\
493: \epsilon^{(1)}:\qquad
494: \Psi^{(1)}_{t_0t_0}-\Psi^{(1)}_{zz}+\cos\Psi^{(0)}\Psi^{(1)}=g
495: \Psi^{(0)}_{zz}
496: +g_z \Psi^{(0)}_z-2 \Psi^{(0)}_{t_0 t_1}, \label {eq20b}
497: \end{eqnarray}
498: \end{subequations}
499: \quad etc.\\
500: The initial conditions for perturbation in the case of the single soliton given in
501: Eq.~(\ref{eq17}) is written as
502: \begin{subequations}
503: \begin{eqnarray}
504: \Psi^{(0)} (z,0)=4arc\tan\exp{mz},
505: \Psi^{(n)} (z,0)=\Psi^{(n)}_{t_{0}}(z,0)=0,~ n=0,1,....\quad\label {eq21b}
506: \end{eqnarray}
507: \end{subequations}
508: The equation obtained at order of $\epsilon^{(0)}$, (i.e) Eq.(\ref{eq20a}) is just the integrable sine-Gordon
509: equation for $\psi^{(0)}$, the single soliton solution of which can be written from Eq.~(\ref{eq17})
510: immediately as
511: \begin{eqnarray}
512: \Psi^{(0)}(z,t_0)= 4 arc\tan\exp\zeta,~\zeta= \pm m_{0}(z-\xi),~\xi_{t_0}=
513: v_{0},\label{eq22}
514: \end{eqnarray}
515: where $v_{0}$ is the velocity of soliton in the $t_{0}$-time scale.
516: Due to perturbation, the soliton parameters namely $m$ and $\xi$ are now treated
517: as functions of the slow time variables $t_0,t_1,t_2, .... $ However m is treated independent
518: of $t_0$. In view of the above, Eq.~(\ref{eq20b}) becomes
519: \begin{subequations}
520: \begin{eqnarray}
521: {\Psi^{(1)}_{t_0t_0}}-{\Psi_{zz}^{(1)}}+(1-2{sech^{2}} \zeta)\Psi^{(1)}=F^{(1)}
522: (z) ,\label{eq23a}
523: \end{eqnarray}
524: where
525: \begin{eqnarray}
526: F^{(1)} (z)= 2\left[g(z) sechz \right]_{z} + 4v_{0} sechz
527: \left[m_{t_{1}}+(m^{2}\xi_{t_{1}}-zm_{t_{1}})\tanh z
528: \right].\quad \label{eq23b}
529: \end{eqnarray}
530: \label{eq23}
531: \end{subequations}
532: While writing the above equation we have used the result $\cos\Psi^{(0)}=1-2sech^{2}
533: \zeta$ obtained from the solution of the unperturbed equation
534: (\ref{eq20a}).
535: In order to represent everything in a co-ordinate system moving with the soliton, we
536: transform $z=\frac{\zeta}{m}+v t_{0}$ and $t_0=t_0+\zeta$, so that Eqs.~(\ref{eq23})
537: and the initial conditions given in Eq.
538: (\ref{eq21b}) can be written as
539: \begin{subequations}
540: \begin{eqnarray}
541: {\Psi^{(1)}_{t_0t_0}}-2mv_{0}{\Psi_{t_0
542: \zeta}^{(1)}}-{\Psi_{\zeta\zeta}^{(1)}}
543: +(1-2{sech^{2}}\zeta)\Psi^{(1)}=F^{(1)}
544: (\zeta,t_{0}), \label{eq24a}
545: \end{eqnarray}
546: where
547: \begin{eqnarray}
548: F^{(1)}(\zeta,t_{0})=2\left[g(\zeta) sech\zeta\right]_{\zeta}+4v_{0} sech\zeta
549: \left[m_{t_{1}}+(m^{2}\xi_{t_{1}}-\zeta
550: m_{t_{1}})\tanh\zeta
551: \right],\label{eq24b}
552: \end{eqnarray}
553: and
554: \begin{eqnarray}
555: \Psi^{(1)}(\zeta,0)=0, \quad \Psi_{t_{0}}^{(1)} (\zeta,0)=0. \label{eq24c}
556: \end{eqnarray}
557: \end{subequations}
558: We now introduce one more transformation
559: $\tau=\frac{t_0}{2m}-\frac{(1+v_{0})\zeta}{2}$ on the independent variable to
560: eliminate the first term in the left hand side of Eq.~(\ref{eq24a}). Thus on
561: using the above transformation Eq.~(\ref{eq24a}) becomes
562: \begin{eqnarray}
563: \Psi_{\tau\zeta}^{(1)}-{\Psi_{\zeta\zeta}^{(1)}}+(1-{sech^{2}}\zeta)\Psi^{(1)}=F^{(1)}
564: (\zeta,\tau),\label{eq25}
565: \end{eqnarray}
566: where $ F^{(1)}(\zeta,\tau)$ equals exactly the right hand side of Eq. (\ref{eq24b}).
567: The solution of Eq.~(\ref{eq25}) is searched by assuming
568: \begin{eqnarray}
569: \Psi^{(1)} (\zeta,\tau)=X(\zeta)T(\tau),\quad
570: F^{(1)} (\zeta,\tau)= X_{\zeta}(\zeta)H(\tau).\label{eq26}
571: \end{eqnarray}
572: Substituting Eq.~(\ref{eq26}) in Eq.~(\ref{eq25}) we obtain
573: \begin{eqnarray}
574: \frac{1}{X_{\zeta}}\left[X_{\zeta\zeta}+(2
575: sech^{2}\zeta-1)X\right]=\frac{1}{T}\left[T_{\tau}-H(\tau)\right].\label{eq27}
576: \end{eqnarray}
577: In Eq.~(\ref{eq27}) the left hand side is independent of $\tau$ and the right hand side
578: is independent of the variable $\zeta$. Hence we can equate the left and right hand
579: sides of Eq.~(\ref{eq27}) to a constant say $\lambda_{0}$ and write
580: \begin{subequations}
581: \begin{eqnarray}
582: X_{\zeta\zeta}+(2 sech^{2}\zeta-1)X=\lambda_{0} X_{\zeta}, \label{eq28a}\\
583: T_{\tau}-\lambda_{0} T= H(\tau).\label{eq28b}
584: \end{eqnarray}
585: \label{eq28}
586: \end{subequations}
587: Thus, the problem of constructing the
588: perturbed soliton at this moment turns out to be solving
589: Eqs.~(\ref{eq28a}) and (\ref{eq28b}) by constructing the eigen functions and
590: finding the eigen values. It may
591: be noted that Eq.~(\ref{eq28a}) is a generalized eigen value problem which is
592: not a self-adjoint eigen value problem and differing from the normal
593: eigen value problem with $X_{\zeta}$ in the right hand side instead of $X$ and
594: Eq.~(\ref{eq28b}) is a first order linear inhomogeneous differential
595: equation which can be solved using known procedure.
596: \subsection{Solving the eigen value problem }
597: Before actually solving the eigen value equation (\ref{eq28a})
598: we first consider it
599: in a more general form given by
600: \begin{eqnarray}
601: \qquad L_{1} X=\lambda \tilde{X} ,\qquad L_{1}=\partial_{\zeta\zeta}+ 2
602: sech^{2}\zeta -1, \label{eq29}
603: \end{eqnarray}
604: where $\lambda$ is the eigen value.
605: In order to find the adjoint eigen function to $X$, we consider another eigen value problem
606: \begin{eqnarray}
607: \qquad L_{2} \tilde{X}=\lambda X, \label{eq30}
608: \end{eqnarray}
609: where the operator $L_{2}$ is still to be determined.
610: The eigen value equations (\ref{eq29}) and (\ref{eq30}) can be combined to give
611: \begin{eqnarray}
612: L_{2}L_{1} X=\lambda^{2} X, \quad
613: L_{1}L_{2} \tilde{X}=\lambda^{2} \tilde{X} .\label{eq31b}
614: \end{eqnarray}
615: Since we already know the form of $L_1$ as given in Eq.(\ref{eq29}), if
616: we choose the operator $L_{2}$ as $ L_{2}=\partial_{\zeta\zeta}+ 6sech^{2}\zeta -1 $,
617: it can be verified that $L_{1}L_{2}$ is the adjoint of $L_{2}L_{1}$. Thus X and
618: $\tilde{X}$ are expected to be adjoint eigen functions.\\
619:
620: Now for solving the eigen value equations ~(\ref{eq29}) and (\ref{eq30}), we
621: choose the eigen functions $X$ and $\tilde{X}$ to be in the form
622: \begin{eqnarray}
623: X(\zeta,k)=p(\zeta,k) e^{ik\zeta},\quad
624: \tilde {X}(\zeta,k)=q(\zeta,k) e^{ik\zeta}, \label{eq32}
625: \end{eqnarray}
626: where $p(\zeta,k)$ and $q(\zeta,k)$ are assumed to have the asymptotic behaviour
627: $p(\zeta,k)\rightarrow$ a constant and $q(\zeta,k)\rightarrow$ a constant as
628: $\zeta\rightarrow \pm\infty$ and $k$ is the propagation constant. On using
629: these asymptotic forms for $p(\zeta,k)$ and $q(\zeta,k)$ in
630: Eq.~(\ref{eq32}) and then substituting the resultant
631: $X(\zeta,k)$ and $\tilde {X}(\zeta,k)$
632: in Eqs.~(\ref{eq29}) and (\ref{eq30}) we obtain the
633: eigen value as
634: \begin{eqnarray}
635: \lambda =-(k^{2}+1). \label{eq33}
636: \end{eqnarray}
637: On substituting the exact forms of $X$ and $\tilde{X}$ from Eq.~(\ref{eq32}) in Eqs.~(\ref{eq29}) and (\ref{eq30}) we get the
638: following set of ordinary differential equations for $p(\zeta,k)$ and
639: $q(\zeta,k)$ .
640: \begin{subequations}
641: \begin{eqnarray}
642: (L_{1}-k^{2})p+2ikp_{\zeta}+(1+k^{2})q=0, \label{eq34a}\\
643: (L_{2}-k^{2})q+ 2ikq_{\zeta}+(1+k^{2})p=0. \label{eq34b}
644: \end{eqnarray}
645: \end{subequations}
646: For solving the above equations we expand $p(\zeta,k)$ and $q(\zeta,k)$ in
647: the following series \cite{ref49}.
648: \begin{subequations}
649: \begin{eqnarray}
650: p (\zeta,k)&=&p_0+p_1 \frac{\sinh\zeta}{\cosh\zeta}+p_2 \frac{1}{\cosh^{2}\zeta}+p_3
651: \frac{\sinh\zeta}{\cosh^{3}\zeta}
652: +p_4\frac{1}{\cosh^{4}\zeta}+...,\label{eq35a}\\
653: q (\zeta,k)&=&q_0+q_1\frac{\sinh\zeta}{\cosh\zeta}+q_2 \frac{1}{\cosh^{2}\zeta}+q_3
654: \frac{\sinh\zeta}{\cosh^{3}\zeta}
655: +q_4\frac{1}{\cosh^{4}\zeta}...,\qquad\label{eq35b}
656: \end{eqnarray}
657: \end{subequations}
658: where the coefficients $p_j$ and $q_{j}$, j=0,1,2,... are functions of $k$ which are to be
659: determined. We substitute the series expansions given in
660: Eqs.~(\ref{eq35a}) and (\ref{eq35b})
661: in Eqs.(34) and collect the coefficients of $ 1,
662: \frac{\sinh\zeta}{\cosh\zeta}, \frac{1}{\cosh^{2}\zeta}$,... and obtain the following algebraic equations.
663: \begin{subequations}
664: \begin{eqnarray}
665: p_{0}=q_{0}, \quad p_{1}=q_{1}, \qquad\qquad\qquad \label{eq36a}\\
666: 2p_{0}+2ik p_{1}+(3-k^{2})p_{2}-4ik p_{3}+(1+k^{2})q_{2}=0,\label{eq36b}\\
667: 6q_{0}+2ikq_{1}+(3-k^{2})q_{2}-4ik q_{3}+(1+k^{2})p_{2}=0,\label{eq36c}\\
668: -4ik p_{2}+(3-k^{2})p_{3}+(1+k^{2})q_{3}=0,\label{eq36d}\\
669: 4q_{1}-4ik q_{2}+(3-k^{2})q_{3}+(1+k^{2})p_{3} =0.\label{eq36e}\\
670: etc.\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\nonumber
671: \end{eqnarray}
672: \end{subequations}
673: By assuming $p_{j}=q_{j}=0$ for $j\geq 3$, and substituting these values in
674: Eqs.(\ref{eq36d}) and (\ref{eq36e}), we obtain
675: \begin{eqnarray}
676: p_{2}=0,\quad q_{2}=-i\frac{q_{1}}{k}. \label{eq37}
677: \end{eqnarray}
678: On substituting the results given in Eq.(\ref{eq37}) in Eqs.
679: (\ref{eq36b}) and (\ref{eq36c})
680: we get
681: \begin{eqnarray}
682: p_{1}=q_{1}=-\frac{2ik p_{0}}{(1-k^{2})}, \quad q_{2}=-\frac{2p_{0}}{(1-k^{2})}.\label{eq38}
683: \end{eqnarray}
684: For our convenience we choose $p_{0}=q_{0}=c (1-k^{2})$ and on substituting this
685: in the above equations, we obtain the other coefficients as follows.
686: \begin{eqnarray}
687: p_{1}=q_{1}=-2ik,~p_{2}=0,~q_{2}=-2c.\label{eq39a}
688: \end{eqnarray}
689: Here `$c$' is an arbitrary constant which will be determined. Using the above values in Eqs.~(\ref{eq35a})
690: and (\ref{eq35b}), and then in Eq.~(\ref{eq32}), we finally obtain
691: \begin{subequations}
692: \begin{eqnarray}
693: X(\zeta,k)&=&c (1-k^{2}-2ik\tanh\zeta) e^{ik\zeta}, \label {eq40a}\\
694: \tilde{X} (\zeta,k)&=&c (1-k^{2}-2ik\tanh\zeta-2sech^{2}\zeta) e^{ik\zeta}. \label {eq40b}
695: \end{eqnarray}
696: \end{subequations}
697: On comparing Eqs.~(\ref{eq40a}) and (\ref{eq40b}) we can write
698: \begin{eqnarray}
699: \tilde{X} (\zeta,k)=\frac{{X_{\zeta}(\zeta,k)}}{ik}. \label{eq41}
700: \end{eqnarray}
701: Using Eq.(\ref{eq41}) in the right hand side of Eq.(\ref{eq29}) and comparing
702: the resultant equation with (\ref{eq28a}), we can write down the eigen value
703: $\lambda_{0}$ as
704: \begin{eqnarray}
705: \lambda_{0}=i\frac{(1+k^{2})}{k}.\label{eq42}
706: \end{eqnarray}
707: \\
708:
709: Now to determine the constant `$c$' we use the orthonormality relation between
710: $X(\zeta,k)$ and $\tilde {X}(\zeta,k)$ given by
711: \begin{eqnarray}
712: \int_{-\infty}^{\infty} X(\zeta,k) {\tilde{X}} ^{\ast} (\zeta,k') d\zeta=\delta
713: (k-k'). \label{eq43}
714: \end{eqnarray}
715: On substituting the eigen functions $X(\zeta,k)$ and $\tilde {X}(\zeta,k)$ given
716: in Eqs.~(\ref{eq40a}) and ~(\ref{eq40b}) respectively in Eq.(\ref{eq43})
717: and after evaluating the integral we obtain
718: \begin{eqnarray}
719: 2\pi c^{2} \left[ (1-k^{2}) (1-k'^{2})+4 kk'\right] \delta (k-k')=\delta
720: (k-k'), \label{eq44}
721: \end{eqnarray}
722: which gives $c^{2}=\frac{1}{2\pi}(1+k^{2})^{-2}$. The correct form of the eigen
723: functions $X(\zeta,k)$ and $\tilde {X}(\zeta,k)$ is written down after using
724: the above value of `$c$' in
725: Eqs.~(\ref{eq40a}) and (\ref{eq40b}).
726: \begin{subequations}
727: \begin{eqnarray}
728: X(\zeta,k)&=&\frac{(1-k^{2}-2ik\tanh\zeta)}{\sqrt{2\pi}(1+k^{2})}e^{ik\zeta} ,
729: \label{eq45a}\\
730: \tilde{X} (\zeta,k)&=&
731: \frac{(1-k^{2}-2ik\tanh\zeta-2sech^{2}\zeta)}{\sqrt{2\pi}(1+k^{2})} e^{ik\zeta} .
732: \label{eq45b}
733: \end{eqnarray}
734: \end{subequations}
735: It can be verified that the operator $L_{2}$ also has the following two
736: discrete eigen functions.
737: \begin{eqnarray}
738: \tilde{X}_{0} (\zeta)=(1-\zeta\tanh\zeta)sech\zeta ,\quad
739: \tilde{X}_{1} (\zeta)=sech\zeta \tanh\zeta. \quad\label{eq46}
740: \end{eqnarray}
741: It may be further noted that the eigen function $\tilde {X}_{1}$ corresponds
742: to the discrete eigen value $\lambda=0$. That is
743: \begin{eqnarray}
744: L_{2} \tilde{X}_{1}(\zeta)=0. \label{eq47}
745: \end{eqnarray}
746: \subsection{Complete set of orthonormal basis}
747: Having found the eigen functions $X(\zeta,k)$ and $\tilde {X}(\zeta,k)$, we now
748: check the completeness of them
749: by writing
750: \begin{eqnarray}
751: \int_{-\infty} ^{\infty} X(\zeta,k) {\tilde{X}}^{\ast} (\zeta',k) dk+f(\zeta,\zeta')=
752: \delta (\zeta-\zeta'), \label{eq48}
753: \end{eqnarray}
754: where $f(\zeta,\zeta')$ is an arbitrary function to be determined.
755: First we evaluate the integral in the left hand side of Eq.~(\ref{eq48}) after
756: substituting the values of the eigen
757: functions $X(\zeta,k)$ and $ \tilde{X} (\zeta,k)$ given in Eqs.~(\ref{eq45a})
758: and (\ref{eq45b}). Thus we have the following integrals to evaluate.
759: \begin{eqnarray}
760: \int_{-\infty} ^{\infty} X(\zeta,k) {\tilde{X}}^{\ast} (\zeta',k) dk&=&\delta
761: (\zeta-\zeta')
762: -\frac{1}{\pi}\left[ \int_{-\infty}^{\infty}
763: \frac{dk}{(1+k^{2})}e^{ik(\zeta-\zeta')} \right.\nonumber\\
764: && \times\{2-sech^{2}\zeta'-ik(\tanh\zeta -\tanh\zeta')
765: \}\nonumber\\
766: &&-2\int_{-\infty}^{\infty}\frac{dk}{(1+k^{2})^{2}}
767: e^{ik(\zeta-\zeta')}\tanh\zeta'\nonumber\\
768: &&\times(1-ik\tanh\zeta)
769: \left.(ik+\tanh\zeta')\right].\label{eq49}
770: \end{eqnarray}
771: The integrals in the right hand side of Eq.(\ref{eq49}) are evaluated
772: with the aid of the residue theorem. It
773: may be noted that the integrands in these two integrals as functions of the
774: complex variable $k$ are analytic everywhere in the complex $k$-plane except at the two
775: poles $k=\pm i$ of first and second order respectively. Let
776: $R_{1}$ and $R_{2}$ be the residues corresponding to the functions
777: $\frac{1}{(1+k^{2})}e^{ik(\zeta-\zeta')}
778: \{2-ik(\tanh\zeta -\tanh\zeta')
779: -sech^{2}\zeta'\}$ and $\frac{1}{(1+k^{2})^{2}}
780: e^{ik(\zeta-\zeta')}(1-ik\tanh\zeta)
781: (ik+\tanh\zeta')\tanh\zeta'$ in Eq.~(\ref{eq49}) at the pole $k=i$ of first and second order
782: respectively. On calculating the residues $R_1$ and $R_2$ using standard procedure we obtain
783: \begin{subequations}
784: \begin{eqnarray}
785: R_1&=&\frac{-i}{2}
786: \left[2+\tanh\zeta-\tanh\zeta'-sech^{2}\zeta\right]
787: e^{(\zeta-\zeta')},\label{eq50a}\\
788: R_2&=&\frac{-i}{4}
789: \left[(1-\zeta+\zeta')(\tanh\zeta-\tanh\zeta')-(\zeta-\zeta')\right.\nonumber\\
790: &&\left.\times(1-\tanh\zeta\tanh\zeta')\right] e^{(\zeta-\zeta')}.
791: \label{eq50b}
792: \end{eqnarray}
793: \end{subequations}
794: On summing up the above results Eq.~(\ref{eq49}) becomes
795: \begin{eqnarray}
796: \int_{-\infty}^{\infty} X(\zeta,k){\tilde{X}}^{\ast} (\zeta',k) dk &=& \delta
797: (\zeta-\zeta')-\left[sech\zeta sech\zeta'(1-\zeta'\tanh\zeta')\right.\nonumber\\
798: &&\left.+\zeta sech\zeta sech\zeta'\tanh\zeta'\right],\label{eq51}
799: \end{eqnarray}
800: which can be identified as
801: \begin{eqnarray}
802: \int_{-\infty}^{\infty} X(\zeta,k){\tilde{X}}^{\ast} (\zeta',k) dk=\delta
803: (\zeta-\zeta')
804: -\left[X_{0} (\zeta){\tilde{X}}_{0} (\zeta')+X_{1} (\zeta) {\tilde{X}}_{1}
805: (\zeta') \right],\label{eq52}
806: \end{eqnarray}
807: upon using the values of $\tilde {X}_{0}(\zeta)$ and $\tilde{X}_{1}(\zeta)$
808: as given in Eq.(\ref{eq46}). By
809: substituting Eq.(\ref{eq52}) in Eq.(\ref{eq48}), we can evaluate the
810: following value of $f(\zeta,\zeta')$ in terms of the discrete
811: orthogonal states.
812: \begin{eqnarray}
813: f(\zeta,\zeta')=\sum_{j=0,1} X_{j}(\zeta) {\tilde{X}}_{j} (\zeta). \label{eq53}
814: \end{eqnarray}
815: Thus, by comparing Eqs.(\ref{eq51}) and (\ref{eq52}), we can write two additional
816: orthogonal discrete states given by
817: \begin{eqnarray}
818: X_{0} (\zeta)= sech\zeta ,\quad
819: X_{1} (\zeta)= \zeta sech\zeta . \label{eq54}
820: \end{eqnarray}
821: It may be checked that the following relations exist between
822: the discrete states $X_{0}(\zeta),\tilde {X}_{0}(\zeta),X_{1}(\zeta)$ and
823: $\tilde {X}_{1}(\zeta)$.
824: \begin{eqnarray}
825: L_{1} X_{0}=0,~L_{1}X_{1}(\zeta)= -2\tilde{X}_{1}(\zeta),~ L_{2} \tilde{X}_{0}(\zeta)=2
826: X_{0}(\zeta), \label{eq55}
827: \end{eqnarray}
828: in addition to the relation given in Eq.(\ref{eq47}).
829: In conclusion, we have a set of two complete orthonormal
830: bases $\{X\}$ and $\{\tilde{X}\}$ given
831: by $\{X(\zeta,k),X_{0}(\zeta),X_{1}(\zeta)\}$ and
832: $\{\tilde{X}(\zeta,k),\tilde{X}_{0}(\zeta),\tilde{X}_{1}(\zeta)\}$ respectively.
833: These set of orthonormal basis functions
834: will be used to construct the perturbed soliton solution.
835: \subsection{Evaluation of $T(\tau)$}
836: Having solved Eq.(\ref{eq28a}) by finding $X(\zeta)$, in order to construct
837: $\psi^{(1)}(\zeta,\tau)$,
838: we now find $T(\tau)$ by solving
839: Eq.~(\ref{eq28b}). For this first we rewrite Eq.~(\ref{eq28b}) by replacing
840: the function $H(\tau)$ in the right hand side
841: using Eq.~(\ref{eq26}) and
842: (\ref{eq41}).
843: \begin{eqnarray}
844: T_{\tau}-\lambda_{0} T=\frac{F^{(1)}(\zeta,\tau)}{i k \tilde{X}(\zeta,k)}. \label{eq56}
845: \end{eqnarray}
846: On multiplying and dividing the right hand side of Eq.~(\ref{eq56})
847: by ${X}^{\ast}(\zeta,k)$ and on integrating with respect to $\zeta$
848: between the limits $-\infty$ to $+\infty$ we obtain due to orthonormality
849: of the functions $X$ and $\tilde{X}$ (see Eq.(\ref{eq43})) the following equation.
850: \begin{eqnarray}
851: T_{\tau}-\lambda_{0} T=\frac{1}{ik} \int_{-\infty}^{\infty} F^{(1)}(\zeta,\tau)
852: {X}^{\ast}(\zeta,k) d\zeta, \label{eq57}
853: \end{eqnarray}
854: which can be explicitly written after substituting the values of
855: $F^{(1)}(\zeta,\tau)$ and $X^{\ast}(\zeta,k)$ from Eqs.~(\ref{eq24b}) and (\ref{eq45a})
856: respectively as
857: \begin{eqnarray}
858: T_{\tau}-\lambda_{0} T&=&\frac{2}{i\sqrt{2\pi} k (1+k^2)} \int_{-\infty}^{\infty}d\zeta
859: \left(\left[g(\zeta) sech\zeta\right]_{\zeta} \right.
860: +2v_0~sech\zeta\left[m_{t_{1}}\right.\nonumber\\
861: &&\left.\left.+(m^{2}\xi_{t_{1}}-\zeta
862: m_{t_{1}})\tanh\zeta
863: \right]\right) (1-k^{2}+2ik\tanh\zeta)e^{-ik\zeta}.\label{eq58}
864: \end{eqnarray}
865: On solving the above equation using standard procedure, we obtain
866: \begin{eqnarray}
867: T(\tau,k)=C(k) e^{\lambda_{0}\tau}-\frac{2}{i \sqrt{2\pi}\lambda_{0}k(1+k^{2})}
868: \int_{-\infty}^{\infty} d\zeta
869: \left(\left[g(\zeta) sech\zeta\right]_{\zeta}+2v_0~sech\zeta\right.\nonumber\\
870: \left.\times\left[m_{t_{1}}+(m^{2}\xi_{t_{1}}-\zeta
871: m_{t_{1}}) \tanh\zeta
872: \right]\right)(1-k^{2}+2ik\tanh\zeta)e^{-ik\zeta},\qquad
873: \label{eq59}
874: \end{eqnarray}
875: where C(k) is a constant which can be found by using the initial
876: condition $T(\tau,k)=0$ when $\tau\rightarrow
877: -\frac{(1+v)\zeta}{2}$. Thus we obtain
878: \begin{eqnarray}
879: C(k)&=&\frac{2}{i \sqrt{2\pi}\lambda_{0}k(1+k^{2})}\int_{-\infty}^{\infty} d\zeta
880: \left(\left[g(\zeta) sech\zeta\right]_{\zeta}\right.
881: +2v_0~sech\zeta\left[m_{t_{1}}\right.
882: \nonumber\\
883: &&\left.\left.+(m^{2}\xi_{t_{1}}-\zeta
884: m_{t_{1}})\tanh\zeta
885: \right]\right) (1-k^{2}+2ik\tanh\zeta)e^{\lambda_{0}(1+v_{0})\frac{\zeta}{2}-ik\zeta},\qquad
886: \label{eq60}
887: \end{eqnarray}
888: and hence
889: \begin{eqnarray}
890: T(\tau,k)&=&\frac{2}{\sqrt{2\pi}i\lambda_{0} k(1+k^{2})}\int_{-\infty}^{\infty}
891: d\zeta' \left(\left[g(\zeta') sech\zeta' \right]_{\zeta'}\right.
892: +2v_0~sech\zeta'\left[m_{t_{1}}\right.\nonumber\\
893: &&\left.\left.+(m^{2}\xi_{t_{1}}-\zeta'
894: m_{t_{1}})\tanh\zeta' \right] \right)
895: (1-k^{2}+2ik\tanh\zeta') e^{-ik\zeta'}\nonumber\\
896: && \times\left(e^{\lambda_{0}
897: [\tau+\frac{(1+v_{0})}{2}\zeta']}-1\right).
898: \label{eq61}
899: \end{eqnarray}
900: The allowed discrete values of $T$ will be determined in the next section while
901: constructing the perturbed part of the soliton
902: $\psi^{(1)}(\zeta,\tau)$ using the values of $X(\zeta)$ and $T(\tau)$.
903: \subsection{Perturbation of soliton}
904: As mentioned, we now write down the first order perturbation correction
905: $\Psi^{(1)}(\zeta,\tau)$ in terms of the basis functions
906: $\{X\}\equiv \{X(\zeta,k),X_{0}(\zeta),X_{1}(\zeta)\}$ and
907: $ \{T\}\equiv\{T(\tau,k),T_{0}(\tau),T_{1}(\tau)\}$. As per Eq.(\ref{eq26}),
908: in terms of the basis functions the
909: perturbed part of the soliton can be written as
910: \begin{eqnarray}
911: \Psi^{(1)}(\zeta,\tau)=\int_{-\infty}^{\infty} X(\zeta,k) T(\tau,k)
912: dk+\sum_{j=0,1}X_{j} (\zeta) T_{j}(\tau). \label{eq62}
913: \end{eqnarray}
914: However, it should be noted that in the basis $\{T\}, T_{0}(\tau)$ and
915: $T_{1}(\tau)$ are yet to be determined. Therefore before
916: evaluating the values of $\Psi^{(1)}(\zeta,k)$,
917: we determine $T_{0}(\tau)$ and $T_{1}(\tau)$. This is carried out
918: by substituting Eq.(\ref{eq62}) in
919: Eq.(\ref{eq25}) which finally becomes
920: \begin{eqnarray}
921: \int_{-\infty}^{\infty} ik\left[ T_{\tau} (\tau,k)-\lambda_{0} T(\tau,k)\right]\tilde{X}(\zeta,k)
922: dk+{T_{1}}_{\tau} (\tau)\tilde{X}_{0} (\zeta)\nonumber\\
923: -\left[ {T_{0}}_{\tau} (\tau)-2 T_{1} (\tau)\right]
924: \tilde{X}_{1}(\zeta)=F^{(1)}(\zeta,\tau), \label{eq63}
925: \end{eqnarray}
926: upon using Eqs.(\ref{eq29}),(\ref{eq41}),(\ref{eq47}) and (\ref{eq55}). Now multiplying
927: Eq.(\ref{eq63}) by $X^{\ast}(\zeta,k), X_{0}(\zeta)$ and $X_{1}(\zeta)$
928: separately and using the orthonormal relations such as
929: $ \int_{-\infty}^{\infty}\tilde{X}(\zeta,k)
930: X^{\ast}(\zeta,k) dk=1,\int_{-\infty}^{\infty} \tilde{X}(\zeta,k) X_{j}(\zeta)
931: d\zeta\equiv\int_{-\infty}^{\infty}X(\zeta,k)\tilde{X}_{j}(\zeta)d\zeta=0,
932: \int_{-\infty}^{\infty}X_{j}(\zeta)\tilde{X}_{l}(\zeta)
933: d\zeta=\delta_{jl}, j,l=0,1$, we get
934:
935: \begin{subequations}
936: \begin{eqnarray}
937: T_{\tau}(\tau,k)-\lambda_{0} T(\tau,k)&=&\frac{1}{ik} \int_{-\infty}^{\infty} F^{(1)}
938: (\zeta,\tau)X^{\ast} (\zeta,k) d\zeta, \label{eq64a}\\
939: { T_{1}}_{\tau} (\tau)&=&\int_{-\infty}^{\infty} F^{(1)} (\zeta,\tau) X_{0} (\zeta)
940: d\zeta , \label{eq64b}\\
941: {T_{0}}_{\tau} (\tau)-2T_{1} (\tau)&=&-\int_{-\infty}^{\infty} F^{(1)}
942: (\zeta,\tau)
943: X_{1} (\zeta) d\zeta. \qquad\label{eq64c}
944: \end{eqnarray}
945: \end{subequations}
946: As $F^{(1)}(\zeta,\tau)$ given in Eq.~(\ref{eq24b}) does not contain time
947: `$\tau$'
948: explicitly, the right
949: hand side of Eqs.~(\ref{eq64a}-c) also should be independent of time. Then it is
950: easy to verify that for the values of $F^{(1)}(\zeta,\tau)$,
951: $X^{\ast}(\zeta,k), T(\tau,k)$ and $ \lambda_{0}$ given respectively in Eqs.
952: (\ref{eq24b}), (\ref{eq45a}), (\ref{eq61}) and (\ref{eq42}), Eq.~(\ref{eq64a})
953: satisfies identically. As the right hand sides of Eqs.~(\ref{eq64b}) and
954: ~(\ref{eq64c}) are also independent of time, they give rise to secularities and
955: hence the nonsecular conditions can be written as
956: \begin{subequations}
957: \begin{eqnarray}
958: \int_{-\infty}^{\infty}F^{(1)} (\zeta,\tau) X_{0} (\zeta) d\zeta=0, \label{eq65a}\\
959: \int_{-\infty}^{\infty}F^{(1)} (\zeta,\tau) X_{1} (\zeta) d\zeta=0. \label{eq65b}
960: \end{eqnarray}
961: \end{subequations}
962: Using Eqs.~(\ref{eq65a}) and (\ref{eq65b}) back in Eqs.~(\ref{eq64b}) and ~(\ref{eq64c}),
963: we obtain
964: $T_{1}(\tau)=0$ and $\quad T_{0} (\tau)=C_{1}$ which has to be determined. For
965: this, we substitute $T_{1}(\tau)=0$ in Eq.~(\ref{eq64c}) and
966: integrate with respect to $\tau$ to obtain
967: \begin{eqnarray}
968: T_{0}(\tau)\equiv C_{1}&=&(1+v)\int_{-\infty}^{\infty}d\zeta\left(\left[g(\zeta) sech\zeta \right]_{\zeta}+2v_0~sech\zeta\right.\nonumber\\
969: &&\times\left.\left[m_{t_{1}}+(m^{2}\xi_{t_{1}}-\zeta
970: m_{t_{1}})\tanh\zeta \right]\right)\zeta^{2} sech\zeta.\quad
971: \label{eq66}
972: \end{eqnarray}
973: \subsection{Variation of soliton parameters}
974: We now estimate the nonsecularity conditions (\ref{eq65a}) and (\ref{eq65b})
975: by evaluating the
976: integrals after substituting the
977: values of $F^{(1)}(\zeta,\tau), X_{0}(\zeta)$ and $X_{1}(\zeta)$ respectively
978: from Eqs.(\ref{eq24b}) and (\ref{eq54}). The results give
979: \begin{subequations}
980: \begin{eqnarray}
981: m_{t_{1}}&=&-\frac{1}{2v_0} \int_{-\infty}^{\infty} \left[ {g(\zeta) sech\zeta}\right]
982: _{\zeta} sech\zeta d\zeta, \label{eq67a}\\
983: \xi_{t_{1}}&=&-\frac{1}{2m^{2} v_0}\int_{-\infty}^{\infty} \left[{g(\zeta) sech\zeta}\right]
984: _{\zeta} \zeta sech\zeta d\zeta . \label{eq67b}
985: \end{eqnarray}
986: \end{subequations}
987: Eq.~(\ref{eq67a}) describes the time evolution of the inverse of the
988: width of the soliton and
989: Eq.~(\ref{eq67b}) gives the velocity of the soliton. $g(\zeta)$ that appears in
990: the above nonsecularity relation
991: is related to the inhomogeneity in stacking energy of DNA. In order to evaluate
992: the integrals in Eqs.~(\ref{eq67a}) and (\ref{eq67b}) explicitly
993: we have to substitute specific value for $g(\zeta)$.
994: We consider $g(\zeta)$ in the form of
995: localized and
996: periodic functions. A localized $g(\zeta)$ corresponds to the intercalation of a
997: compound between neighbouring base pairs without disturbing the base pairs
998: and their sequence in the
999: DNA double helical chain. The periodic nature of $g(\zeta)$ represents a
1000: periodic
1001: repetition of similar base pairs along the helical chain.
1002: We consider the localized form of $g(\zeta)$ as (i)
1003: $g(\zeta)=sech\zeta$ and the periodic form
1004: of $g(\zeta)$ as (ii) $g(\zeta)=\cos\zeta$. We substitute the above values of $g(\zeta)$
1005: one by one in Eqs.~(\ref{eq67a}) and
1006: (\ref{eq67b}) and evaluate the integrals in the right hand side to understand the time evolution of the width
1007: of the soliton and its velocity. At this point it is worth mentioning that Dandoloff and Saxena \cite{ref36}
1008: realized that in the case of XY-spin chains the model of which identifies with our plane-base rotator model,
1009: the ansatz $g(\zeta)=sech\zeta$ energetically favours the deformation of spin chains.\\
1010:
1011: When we substitute $g(\zeta)=sech\zeta$ in Eqs.~(\ref{eq67a}) and (\ref{eq67b})
1012: and on evaluating the integrals, we obtain
1013: \begin{eqnarray}
1014: m_{t_{1}}=0, \quad \xi_{t_{1}}=\frac{\pi}{6m^{2}v_{0}}. \label{eq68}
1015: \end{eqnarray}
1016: The above equations can be rewritten in terms of the original time variable $ \hat t $
1017: by using the transformation
1018: $\frac{\partial }{\partial \hat t}=\frac{\partial }{\partial t_0}+\epsilon
1019: \frac{\partial }{\partial t_1}$ or in otherwords
1020: $m_{\hat t}=m_{t_{0}}+\epsilon m_{t_{1}}$ and $ \quad \xi_{\hat t}=\xi_{t_{0}}+\epsilon
1021: \xi_{t_{1}}.$ As m is independent of $t_{0} (m_{t_{0}}=0)$ and $\xi_{t_{0}}=v_{0}$, we can write
1022: \begin{eqnarray}
1023: m=m_{0}, \quad \xi_{\hat{t}}\equiv v=v_{0}+\frac{\epsilon\pi}{6m_{0}^{2}v_{0}}, \label{eq69}
1024: \end{eqnarray}
1025: where $1/m_{0}$ is the initial width of the soliton. The first of Eq.(\ref{eq69})
1026: says that when $g(\zeta)=sech\zeta$, the width $(m^{-1})$
1027: of the soliton remains constant. However from the second of Eq.(\ref{eq69}), we
1028: find that the velocity of the
1029: soliton gets a correction. As the correction term is a definite positive quantity the velocity of the soliton
1030: increases in this case. Interestingly, the amount of increment in velocity depends on the initial width and
1031: initial velocity
1032: of the soliton. Wider the soliton, greater the increment in velocity due to inhomogeneity. Also slowly moving solitons gain more
1033: speed. The increase in speed helps to overcome the barrier of the local inhomogeneity (which may
1034: be due to the presence of abasic site or intercalation of a molecule) and the
1035: solitons representing the open state will propagate easily along the chain without formation of a
1036: bound state. In a similar study on resonant kink-impurity interaction and kink scattering in the sine-Gordon model
1037: Zhang Fei et al. \cite{ref32} observed
1038: that if the initial velocity of the kink is smaller than a critical velocity it will be either trapped or reflected
1039: by the impurity. In fact they have showed that for most of the initial velocities, the kink is trapped except in the case of some special initial velocities the kink may be totally reflected by the impurity. It was further found that when the kink velocity is greater than the critical velocity, it will pass through the impurity. It is interesting to note that our results on the velocity of soliton found in Eq. (\ref{eq69}) is similar to the last case of Zhang Fei et al \cite{ref32} where the soliton will pass through by overcoming the barrier of the local inhomogeneity.\\
1040:
1041: Next, we substitute
1042: the periodic function $g(\zeta)=\cos\zeta$ in
1043: Eqs.(\ref{eq67a}) and (\ref{eq67b})
1044: and evaluate the integrals to obtain
1045: \begin{eqnarray}
1046: m_{t_{1}}=0, \quad
1047: \xi_{t_{1}}=\frac{\pi^{2}}{16 m^{2}v_{0}}.\label{eq72}
1048: \end{eqnarray}
1049: From the above the parameters m and $\xi$ can be written in terms of the original
1050: variable $\hat t$ after solving Eq.(\ref{eq72}) as
1051: \begin{eqnarray}
1052: m=m_{0}, \quad
1053: \xi_{\hat {t}}\equiv v=v_{0}+\frac{\epsilon\pi^{2}}{16 m^{2} v_{0}}.\label{eq73}
1054: \end{eqnarray}
1055: From
1056: Eq.~(\ref{eq73}) we
1057: find that the width of the soliton in this case remains
1058: constant and the soliton velocity increases similar to the case
1059: $g(\zeta)=sech\zeta$. On comparing Eqs.~(\ref{eq69}) and~(\ref{eq73}),
1060: we observe that the increase in velocity
1061: of the soliton is less in this case. This is because in this case, the
1062: inhomogeneity occurs periodically in the entire DNA chain in terms of sequence.
1063: \subsection{First order perturbed soliton}
1064: Now, we explicitly construct the first order perturbation correction to the one soliton
1065: for the different cases of $g(\zeta)$
1066: by substituting the values of the basis functions $\{X\}
1067: \equiv\{X(\zeta,k), X_{0}(\zeta), X_{1}(\zeta)\}$ and
1068: $\{T\}\equiv\{T(\tau,k), T_{0}(\tau), T_{1}(\tau)\}$ after using the values
1069: of $F^{(1)}(\zeta,\tau), m_{t_{1}}$ and $\xi_{t_{1}}$ for the
1070: respective $g(\zeta)$ values in Eq.~(\ref{eq62}).
1071: Thus in the case of $g(\zeta)=sech\zeta$, we substitute the values of
1072: $X(\zeta,k),
1073: X_{0}(\zeta)$ and $ X_{1}(\zeta)$ from Eqs.~(\ref{eq45a}) and
1074: (\ref{eq54}) and $T(\tau,k),~T_{0}(\tau)$ from Eqs.~(\ref{eq61}) and
1075: (\ref{eq66}) and $F^{(1)}(\zeta,\tau)$ from Eq.~(\ref{eq24b}) and use the
1076: values of $m_{t_{1}}$ and $\xi_{t_{1}}$ from Eqs.~(\ref{eq68}) in Eq.~(\ref{eq62}) to obtain
1077: \begin{eqnarray}
1078: \Psi^{(1)} (\zeta,t_{0})&=&-\frac{1}{3\pi}\int_{-\infty}^{\infty}
1079: \frac{dk}{(1+k^{2})^{3}} (1-k^{2}-2ik\tanh\zeta)
1080: e^{ik\zeta}\int_{-\infty}^{\infty}d\zeta\nonumber\\
1081: &&\times(1-k^{2}+2ik\tanh\zeta)
1082: (\pi-6sech\zeta) sech\zeta\tanh\zeta \nonumber\\
1083: &&\times\left[e^{-ik\zeta} -e^{i\frac{(1+k^{2})}{k}\alpha}e^{i\beta\zeta}\right],
1084: \label{eq74}
1085: \end{eqnarray}
1086: where $\alpha=\frac{t_{0}-m(1+v_{0})\zeta}{2m}$
1087: and $\beta={\frac{(1+v_{0})(1+k^{2})}{2k}}-k$.
1088: While writing the above, we have also used $T_{1}(\tau)=0, \tau=\frac{1}{2m}[t_{0}-m(1+v_{0})\zeta]$ and the eigen value
1089: $\lambda_{0}=\frac{i(1+k^{2})}{k}$.
1090: The integrals in the right hand side of Eq.(\ref{eq74}) can be evaluated using
1091: standard residue
1092: theorem . The details are given in Appendix-A.
1093: The final form of $\Psi^{(1)}(\zeta,t_{0})$ after evaluating the integrals
1094: becomes
1095: \begin{eqnarray}
1096: \Psi^{(1)}(\zeta,t_{0})&\approx&\frac{80}{27\sqrt{2}}\sqrt{sech\zeta}~e^{-\frac{3\zeta}{2}}
1097: +\frac{64}{27\sqrt[3]{2}}\tanh\zeta
1098: \sqrt[3]{sech\zeta}~e^{-\frac{5}{3}\zeta}\nonumber\\
1099: && +\frac{\pi}{6v^{2}}
1100: \left[2 v(1+v)\zeta
1101: +v^{2}+4\alpha v-1\right]sech\zeta. \label{eq75}
1102: \end{eqnarray}
1103: While constructing the above perturbed solution, we have used the
1104: values of m's and $\xi$'s and of course the
1105: corresponding $F^{(1)}(\zeta,t_0)$ values.
1106: Finally, the perturbed one soliton solution that is
1107: $\Psi(z,t_{0})=\Psi^{(0)}(z,t_{0})+
1108: \Psi^{(1)}(z,t_{0})$ (choosing $\epsilon=1$) is written
1109: using Eqs.~(\ref{eq22}) and (\ref{eq75}) as
1110: \begin{figure}
1111: \begin{center}
1112: \epsfig{file=fig5.ps,height =10cm, width=12cm}
1113: \caption{(a) The perturbed kink-soliton and (b) the perturbed antikink-soliton for the inhomogeneity
1114: $g( z)=sech z$ and $v_0=0.4$. (c) A sketch of the base pair opening in DNA double helix with fluctuation.}
1115: \end{center}
1116: \end{figure}
1117: \begin{eqnarray}
1118: \Psi(z,t_{0})&\approx& 4
1119: arc\tan\exp[\pm m_{0}(z-v_{0}t_{0})]+\frac{80}{27\sqrt{2}}
1120: \sqrt{sech[\pm m(z- v t_{0})]}\nonumber\\
1121: &\times& e^{\mp\frac{3(m(z-v t_{0})}{2}}
1122: +\frac{64}{27\sqrt[3]{2}}\tanh[\pm m(z-v t_{0})]
1123: \sqrt[3]{sech[\pm m(z-v t_{0})]} \nonumber\\
1124: &\times& e^{\mp\frac{5}{3}m(z-v t_{0})}+\frac{\pi}{6 m v^{2}}
1125: \left[m (v^{2}-1)+2 t_{0} v\right]sech[\pm m(z-v t_{0})]. \label{eq76}
1126: \end{eqnarray}
1127: In Eq.(\ref{eq76}) while the upper sign corresponds to perturbed kink-soliton the lower sign represents
1128: the perturbed antikink-soliton.
1129: The rotation of bases denoted by $\phi(z,t_{0})$ can be immediately written down by using the
1130: relation $\phi=\frac{\Psi}{2}$.
1131: In Figs. 4a and 4b, we plot $\phi(z,t_{0})$ (rotation of bases under perturbation) for the
1132: parametric choice
1133: $v_{0}=0.4$. From the figure, we observe that there appears fluctuation in the form of
1134: a train of pulses closely resembling the shape of the inhomogeneity profile in the width of the soliton as
1135: time progresses. Also, as time passes, the amplitude of the pulses generating this fluctuation
1136: increases. However, in the asymptotic region of the soliton there is no change in the topological character
1137: and no fluctuations appear in that region. It shows that
1138: the localized inhomogeneity in stacking energy in DNA in the form of a pulse ($g(z)=sech z$) does not
1139: affect very much the opening of bases except
1140: fluctuations in the form of a train of pulses in the localized region of the kink and antikink-soliton.
1141: We have schematically represented this in Fig.4c.\\
1142:
1143: We then repeat the procedure for constructing the perturbed one-soliton solution in the case of $g(\zeta)=\cos\zeta$ which is written as
1144: \begin{eqnarray}
1145: \Psi^{(1)} (\zeta,t_{0})&\approx&\frac{1}{\pi}\int_{-\infty}^{\infty}
1146: \frac{dk}{(1+k^{2})^{3}} (1-k^{2}-2ik\tanh\zeta) e^{ik\zeta} \int_{-\infty}^{\infty}
1147: d\zeta(1-k^{2}\nonumber\\
1148: && +2ik\tanh\zeta)[\sin\zeta+(\cos\zeta -\frac{\pi^{2}}{8})\tanh\zeta]sech\zeta
1149: \nonumber\\
1150: &&\times\{e^{i\frac{(1+k^{2})}{k}\alpha}e^{i\beta\zeta}-e^{-ik\zeta}\}
1151: +(1+v)sech\zeta\nonumber\\
1152: && \times\int_{-\infty}^{\infty}
1153: d\zeta ~\zeta^{2} [\sin\zeta+(\cos\zeta-\frac{\pi^{2}}{8})\tanh\zeta] sech^{2}\zeta .\label{eq79}
1154: \end{eqnarray}
1155: The details of values of the integrals in the above equation are given in
1156: Appendix-B.
1157: The perturbed kink (upper sign)-antikink (lower sign) one soliton solution in this case is finally
1158: obtained as
1159: \begin{eqnarray}
1160: \Psi(z,t_{0})&\approx&4~
1161: arc\tan\exp[\pm m_{0}(z-v_{0} t_{0})]
1162: +\frac{\pi^{2}}{16 m v^{2}}\left[m(v^{2}-1)+2 v t_{0}\right]\nonumber\\
1163: && \times sech[ m(z-v t_{0})].\label{eq80}
1164: \end{eqnarray}
1165: After finding $\phi(z,t_{0})$ from the relation $\phi=\frac{\Psi}{2}
1166: $ we plot it in Figs.5a and 5b for the same value of the parameter as before. From the figures we observe
1167: that periodic oscillations appear in the width of the soliton without any change
1168: asymptotically.
1169: \begin{figure}
1170: \begin{center}
1171: \epsfig{file=fig7.ps,height =10cm, width=12cm}
1172: \caption{ The perturbed (a) kink-soliton and (b) antikink-soliton for the inhomogeneity
1173: $g(z)=\cos z $ with $v_{0}=0.4$. (c) A sketch of the open state configuration in DNA
1174: with small fluctuations.}
1175: \end{center}
1176: \end{figure}
1177: \section{Conclusions }
1178: In this paper, we
1179: studied the nonlinear dynamics of DNA double helix with stacking inhomogeneity
1180: by considering the
1181: dynamic plane-base rotator
1182: model. The dynamical equation which finally appeared in the form of a perturbed
1183: sine-Gordon equation was derived from a suitable Hamiltonian in analogy with Heisenberg model of
1184: an inhomogeneous anisotropic coupled spin chain or spin ladder with ferromagnetic legs and antiferromagnetic
1185: rung coupling in the continuum limit. In the unperturbed limit
1186: which is also the homogeneous limit, the dynamics is governed by the kink-antikink soliton of the integrable sine-Gordon
1187: equation which represents the open state configuration of base pairs in DNA double helix.
1188: Even though DNA double helix is a large molecular chain involving very large number of base
1189: pairs, base pair opening is limited to a very few number of base pairs forming
1190: localized coherent structure in the form of kink-antikink solitons which is obtained as a balance
1191: between dispersion and nonlinearity traveling with constant speed and amplitude without
1192: loosing its energy along the
1193: helical chain. To understand the
1194: effect of stacking energy inhomogeneity on the open state configuration of base pairs we carried out a
1195: multiple scale soliton perturbation analysis. For implementing this we linearized
1196: the perturbed sine-Gordon equation using multiple-scale expansion
1197: to obtain linearized equations in the form of eigen value problems. The perturbed kink-antikink
1198: soliton solutions were constructed for different forms of inhomogeneities by solving the associated eigen value
1199: problem. By using the complete set of eigen functions thus obtained as the basis functions the perturbed
1200: solutions were constructed.\\
1201:
1202: The perturbation
1203: not only modifies the shape of the soliton but also undergoes a slow time
1204: change of the soliton parameters namely the width and velocity for the different forms
1205: of the inhomogeneity chosen. We chose the inhomogeneity in the form of localized and
1206: periodic functions. The results show that when the inhomogeneity is either in the form
1207: $g(z)=sech z$ or $g(z)=\cos z$, the width of the soliton remains constant. Thus, the number
1208: of base pairs participating in the opening do not change due to the above pattern of
1209: inhomogeneities. However, in this case the
1210: speed of the soliton increases with a correction that is proportional to the square of the
1211: initial width of the soliton and inversely proportional to its initial velocity. Thus the inhomogeneity increases the speed with which the base pairs
1212: are opening and closing or winding and unwinding.
1213: From the perturbed solutions corresponding to different inhomogeneities (see Figs. 4,5) we observe
1214: that the perturbation due to
1215: inhomogeneity in stacking energy along the strands introduces fluctuation only in the width of the
1216: solitons. The nature of the fluctuation varies depending on the type of inhomogeneity.
1217: In particular, when the inhomogeneity is in the form of $g(z)=sech z$, we find that fluctuation in the form of
1218: pulse trains resembling the shape of the inhomogeneity is generated in the width of the soliton
1219: representing open state configuration. It is noted that in the long time limit, eventhough
1220: neighbouring pulses overlap the train of pulse-like fluctuations maintain their character without
1221: affecting the soliton. In a similar way, when the inhomogeneity is of the form
1222: $g(z)=\cos z$, the fluctuation appears in the form of periodic oscillations in the width of the
1223: soliton. In all these cases, asymptotically the kink-antikink soliton shape is preserved.
1224: The results indicate that inhomogeneity in stacking energy in DNA double helix can (i) introduce small
1225: fluctuations which may merge asymptotically during the process of opening and closing of base pairs (ii) increase or decrease the number of base pairs
1226: participating in the open state configuration and (iii) change the speed with which the open state
1227: configuration can travel along the double helical chain. Thus in conclusion the inhomogeneity in
1228: stacking does not affect the general pattern of base pair opening. Even though the size of the base pair
1229: is big and the solvent effect on pure rotation of base pairs is negligible, the soliton,
1230: a coherent structure which is formed by involving few base pairs move along the helical chain
1231: without dissipation or any other form of deformation. Similar conclusion was also arrived in
1232: the case of propagation of soliton representing base pair opening in a discrete site-dependent
1233: DNA \cite{ref24} and propagation of bubble in a heterogeneous DNA chain \cite{ref20}. As nature selects generally inhomogeneous DNA,
1234: the functions such as replication and transcription can be explained more viably through formation of open
1235: states through our inhomogeneous model rather than the homogeneous case. This is also because,
1236: it is known that transcription and replication are sequence dependent. This is further similar to
1237: what was observed in the case of proteins where inhomogeneity of the sequence leads to
1238: inhomogeneous fluctuations, enhanced by the nonlinear effect \cite{ref50}. Eventhough the
1239: relevance of these effects in the DNA to biological processes are not yet clearly, established
1240: a recent study suggests that thermally induced base pair opening agrees with
1241: experimental observation on DNA base pair opening detected by potassium permanganate foot
1242: printing \cite{ref51}. We will make numerical analysis of the discrete dynamical equation separately to understand the discreteness effect which will be published elsewhere. It is also equally important to analyse the nonlinear
1243: dynamics of DNA double helix when the hydrogen bonding energy depends on the distance between the bases
1244: (inhomogeneity in hydrogen bonds) and the study is under progress. Also, the nonlinear dynamics
1245: study of open state configuration facilitated by enzymes (protein) is under progress.
1246: \section{Acknowledgements}
1247: The authors thank the anonymous referees for useful comments.
1248: The major portion of the work is done within the framework of the Associateship Scheme of the Abdus Salam
1249: International Centre for Theoretical Physics, Trieste, Italy and the financial support is acknowledged. The work of M. D also forms part of a major
1250: DST project. V. V also thanks SBI for financial support.
1251: \appendix
1252: \section{ Evaluation of integrals in Eq.(\ref{eq74}) using residue theorem}
1253: In this appendix we evaluate the integrals found in Eq.(\ref{eq74})
1254: using standard residue theorem. Eq.(\ref{eq74}) is of the form
1255:
1256: \begin{eqnarray}
1257: \Psi^{(1)} (\zeta,t_{0})&=&-\frac{1}{3\pi}\int_{-\infty}^{\infty}
1258: \frac{dk}{(1+k^{2})^{3}} (1-k^{2}-2ik\tanh\zeta)
1259: e^{ik\zeta}\int_{-\infty}^{\infty}d\zeta(1-k^{2}\nonumber\\
1260: &&+2ik\tanh\zeta)
1261: (\pi-6sech\zeta) sech\zeta\tanh\zeta\nonumber\\
1262: &&\times \left[e^{-ik\zeta} -e^{i\frac{(1+k^{2})}{k}\alpha}e^{i\beta\zeta}\right].
1263: \label{eqa1}
1264: \end{eqnarray}
1265: The evaluation of various integrals in Eq.~(\ref{eqa1}) can be
1266: facilitated by first finding the values of the
1267: following two integrals \cite{ref52}.
1268: \begin{eqnarray}
1269: I_{1}=\int_{-\infty}^{\infty} sech\zeta ~e^{i\chi\zeta}d\zeta, \label{eqa2}
1270: \end{eqnarray}
1271: \begin{eqnarray}
1272: I_{2}=\int_{-\infty}^{\infty} \tanh\zeta ~e^{i\chi\zeta}d\zeta, \label{eqa3}
1273: \end{eqnarray}
1274: where $\chi$ can take values $-k$ and $\beta$.
1275: The integrand in $I_{1}$ is found to be analytic
1276: everywhere except at the pole $\zeta\rightarrow
1277: i(2n+1)\frac{\pi}{2},~ n=0,1,2,....$ The residue of the function
1278: $sech\zeta ~ e^{i\chi\zeta}$ can be written as
1279: \begin{eqnarray}
1280: Res(\zeta=i(2n+1)\frac{\pi}{2})= \overset{\infty}{\underset{n=0}{\sum}}
1281: ~\underset{\zeta\rightarrow i(2n+1)\frac{\pi}{2}}{lim}
1282: \frac{e^{i\chi\zeta}}{\frac{d}{d\zeta}(\cosh\zeta)} ,\label{eqa4}
1283: \end{eqnarray}
1284: which can be simplified to give
1285: \begin{eqnarray}
1286: Res(\zeta=i(2n+1)\frac{\pi}{2})=\frac{1}{2i\cosh{\frac{\pi}{2}\chi}} .
1287: \qquad\qquad\label{eqa5}
1288: \end{eqnarray}
1289: Thus the integral value of $I_{1}$ in Eq.~(\ref{eqa2}) is written as
1290: \begin{eqnarray}
1291: I_{1}\equiv\int_{-\infty}^{\infty} sech\zeta e^{i\chi\zeta}d\zeta
1292: =\frac{\pi}{\cosh{(\frac{\pi}{2}\chi)}}. \label{eqa6}
1293: \end{eqnarray}
1294: Similarly, using the same procedure we evaluate the integral $I_{2}$
1295: in which the integrand $\tanh\zeta ~ e^{i\chi \zeta}$ contains poles at $\zeta\rightarrow
1296: i(2n+1)\frac{\pi}{2},~ n=0,1,2,...$ and obtain
1297: \begin{eqnarray}
1298: I_{2}\equiv\int_{-\infty}^{\infty}
1299: \tanh\zeta ~e^{i\chi\zeta} d\zeta=\frac{i\pi}{\sinh{(\frac{\pi}{2}\chi)}} .\label{eqa7}
1300: \end{eqnarray}
1301: Now using the value of the integral $I_{1}$ as given in Eq.~(\ref{eqa6}), we
1302: evaluate the following integrals found in Eq.~(\ref{eqa1}) by
1303: integrating them by parts successively to obtain
1304: \begin{subequations}
1305: \begin{eqnarray}
1306: \int_{-\infty}^{\infty} sech\zeta\tanh\zeta~ e^{i\chi\zeta}
1307: d\zeta=\frac{i\pi \chi} {\cosh{(\frac{\pi}{2}\chi)}} ,\quad\quad\label{eqa8a}\\
1308: \int_{-\infty}^{\infty}sech\zeta\tanh^{2}\zeta ~e^{i\chi\zeta}
1309: d\zeta=\frac{\pi(1-\chi^{2})}{2\cosh{(\frac{\pi}{2}\chi)}}.\qquad\label{eqa8b}
1310: \end{eqnarray}
1311: \end{subequations}
1312: Similarly, using Eq.~(\ref{eqa7}) we evaluate the following
1313: integrals found in Eq.~(\ref{eqa1}) by making successive
1314: integrations by parts and obtain
1315: \begin{subequations}
1316: \begin{eqnarray}
1317: \int_{-\infty}^{\infty} sech^{2}\zeta\tanh\zeta ~e^{i\chi\zeta}
1318: d\zeta=\frac{i\pi \chi^{2}}{2\sinh{(\frac{\pi}{2}\chi)}} ,\qquad\label{eqa9a}\\
1319: \int_{-\infty}^{\infty} sech^{2}\zeta\tanh^{2}\zeta~ e^{i\chi\zeta}
1320: d\zeta=\frac{\pi \chi(2-\chi^{2})}{6\sinh{(\frac{\pi}{2}\chi)}}.\quad\quad\label{eqa9b}
1321: \end{eqnarray}
1322: \end{subequations}
1323: Here also $\chi$ can take values $-k$ and $\beta$. Substituting the values of the
1324: integrals found in Eqs.~(\ref{eqa8a}), (\ref{eqa8b}) (\ref{eqa9a}) and
1325: (\ref{eqa9b}) in Eq.~(\ref{eqa1}), we obtain
1326: \begin{eqnarray}
1327: \Psi^{(1)} (\zeta,t_{0})&=&-\frac{i}{12}\left[\int_{-\infty}^{\infty}dk
1328: \frac{(1-k^{2}-2ik\tanh\zeta)}
1329: {k^{2}(1+k^{2})^{2}}\left( \pi (1-v^{2}) k(1+k^{2})\right.\right.\nonumber\\
1330: && \times sech{\frac{\pi}{2}\beta}
1331: -\{(2-v)(1+v)^{2}(1+k^{2})^{2}-4 k^{2}(1+v+k^{2})\} \nonumber\\
1332: &&\left. \times cosech{\frac{\pi}{2}\beta}\right)e^{i\frac{(1+k^{2})}{k}\alpha+ik\zeta}\nonumber\\
1333: &&\left.+4\int_{-\infty}^{\infty}
1334: dk \frac{(k^{2}-k^{4}-2ik^{3}\tanh\zeta)}{(1+k^{2})^{2}\sinh{\frac{\pi}{2}k}}e^{ik\zeta}
1335: \right].\quad\label{eqa10}
1336: \end{eqnarray}
1337: \\
1338:
1339: Now, before writing down the final form of $\psi^{(1)}(\zeta,t_0)$ we evaluate the integrals
1340: with respect to $k$ in the right hand side of the above
1341: equation. For this first we rearrange
1342: the integrand in
1343: Eq.(\ref{eqa10}) by simple multiplication and
1344: call them $I_{3} , I_{4}, I_{5}$ and $I_{6}$ as given below.
1345: \begin{subequations}
1346: \begin{eqnarray}
1347: I_{3}&=&\int_{-\infty}^{\infty}
1348: dk \frac{(k^{2}-k^{4}-2 i k^{3}\tanh\zeta)}{{(1+k^{2})^{2}\sinh{\frac{\pi}{2}k}}}e^{ik\zeta},\label{eqa11a}\\
1349: I_{4}&=&\int_{-\infty}^{\infty}dk
1350: \frac{(1-k^{2}-2ik\tanh\zeta) }
1351: {(1+k^{2})^{2}\sinh{\frac{\pi}{2}\beta}}(1+v+k^{2})
1352: e^{i\frac{(1+k^{2})}{k}\alpha+ik\zeta},\label{eqa11b}\\
1353: I_{5}&=&\int_{-\infty}^{\infty}dk
1354: \frac{(1-k^{2}-2ik\tanh\zeta)}{k^{2}\sinh{\frac{\pi}{2}\beta}}
1355: e^{i\frac{(1+k^{2})}{k}\alpha+ik\zeta},\qquad\label{eqa11c}\\
1356: I_{6}&=&\int_{-\infty}^{\infty}dk
1357: \frac{(1-k^{2}-2ik\tanh\zeta)}
1358: {k(1+k^{2})\cosh{\frac{\pi}{2}}\beta}
1359: e^{i\frac{(1+k^{2})}{k}\alpha+ik\zeta}.\qquad\label{eqa11d}
1360: \end{eqnarray}
1361: \end{subequations}
1362: The integral $I_{3}$ can be evaluated by finding the residue of the integrand
1363: $ \frac{(k^{2}-k^{4}-2ik^{3}\tanh\zeta)}{(1+k^{2})^{2}\sinh{k\frac{\pi}{2}}}e^{ik\zeta} $ at the poles
1364: $k=i$ of order two
1365: and at the simple pole $k=2in$. The results are given by
1366: \begin{subequations}
1367: \begin{eqnarray}
1368: Res(k=i)=\frac{1}{2}(\zeta-2) sech\zeta,\qquad\qquad\qquad\qquad\qquad\qquad \qquad\quad\label{eqa12a}\\
1369: Res(k=2in)=\frac{40}{9\sqrt{2}~\pi} \sqrt{sech\zeta}e^{-\frac{3\zeta}{2}}
1370: +\frac{32}{9\sqrt[3]{2}~\pi} \tanh\zeta\sqrt[3]{sech\zeta}e^{-\frac{5\zeta}{3}}. \label{eqa12b}
1371: \end{eqnarray}
1372: \end{subequations}
1373: While writing the first term in Eq.(\ref{eqa12b}), we have approximated the results obtained
1374: as a series in terms of suitable functions.
1375: Adding the above two residue values we obtain the value of the integral $I_{3}$ as
1376: \begin{eqnarray}
1377: I_{3}&=&2\pi i\left[\frac{1}{2}(\zeta-2) sech\zeta+\frac{40}{9\pi\sqrt{2}}
1378: \sqrt{sech\zeta}e^{-\frac{3\zeta}{2}}+\frac{32}{9\sqrt[3]{2}~\pi} \tanh\zeta\right.\nonumber\\
1379: &&\left.\times \sqrt[3]{sech\zeta}~ e^{-\frac{5\zeta}{3}}\right].
1380: \label{eqa13}
1381: \end{eqnarray}
1382: In the case of the integral $I_4$, the integrand possesses a second order pole at $k=i$ and there is
1383: one more simple pole at $k=\frac{(2n+1)}{(1-v)}(-i\pm\sqrt{\frac{1-v^{2}}{(2n+1)^{2}}-1})$
1384: which is however out of the contour. Thus the residue for the integrand function
1385: $\frac{(1-k^{2}-2ik\tanh\zeta)}
1386: {(1+k^{2})^{2}\sinh{\beta\frac{\pi}{2}}}(1+v+k^{2})e^{i\frac{(1+k^{2})}{k}\alpha+ik\zeta} $ is obtained as
1387: \begin{eqnarray}
1388: Res(k=i)=\frac{1}{2}(2+v\zeta+2v\alpha)sech\zeta,\label{eqa14}
1389: \end{eqnarray}
1390: and hence the right hand side of Eq.~(\ref{eqa14}) represents the value of the integral $I_{4}$.\\
1391:
1392: The integrand of the integral $I_{5}$ namely
1393: $\frac{(1-k^{2}-2ik\tanh\zeta)}
1394: {k^{2}\sinh{\beta\frac{\pi}{2}}}e^{i\frac{(1+k^{2})}{k}\alpha+ik\zeta} $ admits an essential singularity at
1395: $k=0$, in addition to a simple pole at $k=\frac{(2n+1)}{(1-v)}(-i\pm\sqrt{\frac{1-v^{2}}{(2n+1)^{2}}-1})$
1396: which is also out of the contour. The residue corresponding to the essential singularity is found by
1397: collecting the coefficient of $\frac{1}{k}$ after expanding the above function. Thus we obtain
1398: \begin{eqnarray}
1399: Res(k=0)&=&\overset{\infty}{\underset{n,n'=0}{\sum}}
1400: \frac{(-1)^{n+n'}(\frac{\pi}{2})^{2n-1} b^{n-1}(b-1)^{n}\alpha^{n'}}{(n'!)}
1401: (\zeta+\alpha)^{n'}\left[\frac{B_{n}}{(n!)^{2}(n'!)}\right.\nonumber\\
1402: && \times(2^{2n-1}-1)+\frac{(\frac{\pi}{2})^{2}b^{2}B_{n+1}}{(n+1)!^{2}(n'!)}
1403: (2^{2n+1}-1)\left[1-2\tanh\zeta\right.\nonumber\\
1404: &&\times\frac{(\zeta+\alpha)}{(n'+1)}
1405: \left.+\frac{(\zeta+\alpha)^{2}}{(n'+1)(n'+2)}\right]
1406: +\frac{(\frac{\pi}{2})^{4}b^{4}B_{n+2}}{(n+2)!(n+3)!}\nonumber\\
1407: &&\times\frac{(n+1)(2^{2n+3}-1)}{(n'+2)!}\left[1-2\tanh\zeta\frac{(\zeta+\alpha)}{(n'+3)}\right.\nonumber\\
1408: &&\left.\left.+\frac{(\zeta+\alpha)^{2}}{(n'+3)(n'+4)}\right]+...\right],
1409: \label{eqa15}
1410: \end{eqnarray}
1411: where $B_{n}'$s are Bernoulli numbers and $b=\frac{(1+v)}{2}$.
1412: Thus the right hand side of Eq.~(\ref{eqa15}) gives the value of the integral $I_{5}$. In the case of
1413: the integral $I_{6}$, the integrand
1414: $\frac{(1-k^{2}-2ik\tanh\zeta)}
1415: {k(1+k^{2})\cosh{\frac{\pi}{2}}\beta}
1416: e^{i\frac{(1+k^{2})}{k}\alpha+ik\zeta} $ possesses a
1417: second order pole at $k=i$ and an essential singularity at $k=0$.
1418: In addition, we have a simple pole at
1419: $k=\frac{(2n+1)}{(1-v)}(-i\pm\sqrt{\frac{1-v^{2}}{(2n+1)^{2}}-1})$
1420: which is again out of the contour. The residues at the pole $k=i$ and at the
1421: essential singularity $k=0$
1422: are respectively found to be
1423: \begin{subequations}
1424: \begin{eqnarray}
1425: Res(k=i)&=&\frac{-1} {\pi v^{2}}(1-2 v\zeta-4v\alpha)
1426: sech\zeta,\label{eqa16a}\\
1427: Res(k=0)&=&\overset{\infty}{\underset{m,n,n'=0}{\sum}}\overset{m}{\underset{j=0}{\sum}}
1428: \frac{(-1)^{n+n'}
1429: (b-1)^{n}E_{n+m}(\frac{\pi}{2})^{2n+2m}b^{n+2m}\alpha^{n'} }{(n!)(n+2m)!(n')!(n'+2j)!}\nonumber\\
1430: &&\times (\zeta+\alpha)^{n'+2j}\left[1-\frac{\frac{\pi}{2}^{2}b^{2}}{(n+1+2m)(n+2+2m)}\right.\nonumber\\
1431: &&\left.-\frac{2(\frac{\pi}{2})^{2} b^{2}(\zeta+\alpha)\tanh\zeta}
1432: {(n+1+2m)(n+2+2m)(n'+1+2j)}\right],\quad\label{eqa16b}
1433: \end{eqnarray}
1434: \end{subequations}
1435: where $E_{n}'$s are Euler numbers. We evaluated the value of the residue given in (\ref{eqa16a}) using
1436: Mathematica.
1437: The value of the integral $I_{6}$ is the sum of the right hand sides of Eqs.~(\ref{eqa16a}) and
1438: ~(\ref{eqa16b}).\\
1439:
1440: Now the value of $\psi^{(1)}(\zeta,t_{0})$ is found by combining
1441: Eqs.~(\ref{eqa13}),~(\ref{eqa14}),~(\ref{eqa15}),
1442: (\ref{eqa16a}) and ~(\ref{eqa16b}) and the final form is written as
1443: \begin{eqnarray}
1444: \Psi^{(1)}(\zeta,t_{0})&\approx&\frac{80}{27\sqrt{2}}\sqrt{sech\zeta}~e^{-\frac{3\zeta}{2}}
1445: +\frac{64}{27\sqrt[3]{2}}\tanh\zeta \sqrt[3]{sech\zeta}~e^{-\frac{5}{3}\zeta}\nonumber\\
1446: && +\frac{\pi}{6v^{2}}\left[2 v(1+v)\zeta+v^{2}+4\alpha v-1\right]sech\zeta.
1447: \label{eqa17}
1448: \end{eqnarray}
1449: While writing the above, we have
1450: dropped few higher order terms due to smallness in values
1451: that appeared in the residue of the essential singularity.
1452: \section{ Evaluation of the integrals in Eq. (\ref{eq79}) using residue theorem}
1453: In this appendix, we evaluate the integrals found in Eq.~(\ref{eq79}) using standard
1454: residue theorem as done in the previous two cases. Eq.(\ref{eq79}) is written as
1455: \begin{eqnarray}
1456: \Psi^{(1)} (\zeta,t_{0})&\approx&\frac{1}{\pi}\int_{-\infty}^{\infty}
1457: \frac{dk}{(1+k^{2})^{3}} (1-k^{2}-2ik\tanh\zeta) e^{ik\zeta} \int_{-\infty}^{\infty}
1458: d\zeta(1-k^{2}\nonumber\\
1459: && +2ik\tanh\zeta)[\sin\zeta+(\cos\zeta -\frac{\pi^{2}}{8})\tanh\zeta]sech\zeta
1460: \nonumber\\
1461: &&\times\{e^{i\frac{(1+k^{2})}{k}\alpha}e^{i\beta\zeta}-e^{-ik\zeta}\}
1462: +(1+v)sech\zeta\left[\int_{-\infty}^{\infty}
1463: d\zeta ~\zeta^{2}\right.\nonumber\\
1464: && \times\left. [\sin\zeta+\cos\zeta]sech^{2}\zeta-\frac{\pi^{2}}{8}\int_{-\infty}^{\infty}
1465: d\zeta ~\zeta^{2}sech^{2}\zeta\tanh\zeta \right] .\label{eqc1}
1466: \end{eqnarray}
1467: It may be verified that the last integral $\int_{-\infty}^{\infty}d\zeta~\zeta^{2} sech^{2}\zeta\tanh\zeta$
1468: in the right hand side of Eq.(\ref{eqc1}) on evaluation vanishes.
1469: For evaluating some of the integrals found in Eq.~(\ref{eqc1}) we use the values of the
1470: integrals given in
1471: Eqs.~(\ref{eqa6}), (\ref{eqa8a}) and (\ref{eqa8b}) and also the value of the following integral.
1472: \begin{eqnarray}
1473: I_{8}=\int_{-\infty}^{\infty} \zeta^{2} \tanh\zeta e^{i\zeta} d\zeta.\label{eqc2}
1474: \end{eqnarray}
1475: The integrand in Eq.(\ref{eqc2}) is found to be analytic everywhere except at the pole
1476: $\zeta=i(2n+1)\frac{\pi}{2}$, n=0,1,2,... The residue of the function
1477: $ \zeta^{2} \tanh\zeta e^{i\zeta}$
1478: is then found to be $
1479: -\frac{\pi^{2}}{8\sinh(\frac{\pi}{2})}
1480: \left[1+2\frac{e^{-\frac{\pi}{2}}}{\sinh(\frac{\pi}{2})}
1481: +\frac{e^{\frac{\pi}{2}}(1-e^{-2\pi\chi})}{2\sinh^{3}(\frac{\pi}{2})}\right]$ and hence the value of the
1482: integral $I_8$ is written as
1483: \begin{eqnarray}
1484: I_{8}\equiv\int_{-\infty}^{\infty} \zeta^{2} \tanh\zeta e^{i\zeta}
1485: d\zeta=-\frac{i\pi^{3}}{4\sinh(\frac{\pi}{2})}
1486: \left[1+2\frac{e^{-\frac{\pi}{2}}}{\sinh(\frac{\pi}{2})}
1487: +\frac{e^{\frac{\pi}{2}}(1-e^{-2\pi})}{2\sinh^{3}(\frac{\pi}{2})}\right].\label{eqc3}
1488: \end{eqnarray}
1489: Now, using the value of the integral given in Eq.~(\ref{eqc3}), we evaluate the following integrals found in
1490: Eq.~(\ref{eqc1}) by integrating them by parts successively.
1491: \begin{subequations}
1492: \begin{eqnarray}
1493: \int_{-\infty}^{\infty}d\zeta\zeta^{2}sech^{2}\zeta e^{\pm
1494: i\zeta}&=&\frac{\pi^{2}}{\sinh\frac{\pi}{2}}\left[\pm(1\mp\frac{\pi}{4})
1495: +\frac{(1\mp\frac{\pi}{2})e^{\mp\frac{\pi}{2}}}{\sinh\frac{\pi}{2}}\right.\nonumber\\
1496: &&\left.\pm\frac{\pi e^{\pm\frac{\pi}{2}}(1-e^{\mp2\pi})}
1497: {\sinh^{3}\frac{\pi}{2}}\right],\label{eqc4a}\\
1498: \int_{-\infty}^{\infty}d\zeta\zeta^{2}sech^{2}\zeta\tanh\zeta e^{ \pm i\zeta}&=&\frac{i\pi}{\sinh\frac{\pi}{2}}
1499: \left[(\pi\mp1\mp\frac{\pi^{2}}{8})\pm
1500: \frac{\pi(2\mp\frac{\pi}{2})
1501: e^{\mp\frac{\pi}{2}}}{2\sinh\frac{\pi}{2}}\right.\nonumber\\
1502: &&\left.-\frac{i\pi^{2}e^{\pm\frac{\pi}{2}}(1-e^{\mp 2\pi})}{16\sinh^{3}\frac{\pi}{2}}\right].\quad\label{eqc4b}
1503: \end{eqnarray}
1504: \end{subequations}
1505: On substituting Eqs.~(\ref{eqc4a}) and (\ref{eqc4b}) in Eq.~(\ref{eqc1}) we obtain
1506: \begin{eqnarray}
1507: \Psi^{(1)}(\zeta,t_{0})&=&\frac{(1+v)i\pi^{3}}{16\sinh\frac{\pi}{2} }
1508: \left[\frac{\cosh\frac{3\pi}{2}}{\sinh^{3}\frac{\pi}{2}}
1509: -\cosh\frac{\pi}{2}\left(\frac{4}{\sinh\frac{\pi}{2}}+1\right )
1510: \right] sech\zeta\nonumber\\
1511: &&+\frac{i}{2}\int_{-\infty}^{\infty}dk
1512: \frac{(1-k^{2}-2ik\tanh\zeta)}{k(1+k^{2})^{3}}
1513: e^{\frac{i(1+k^{2})\alpha}{k}+ik\zeta}\left[ \{k^{2}+b(1-b)\right.\nonumber\\
1514: &&\times(1+k^{2})^{2}\}
1515: \left(sech{\frac{\pi}{2}(\beta+1)}
1516: +sech{\frac{\pi}{2}(\beta-1)}\right)
1517: -\frac{\pi^{2}}{4}b(1-b)\nonumber\\
1518: &&\left.\times(1+k^{2})^{2}sech{\frac{\pi}{2}\beta}\right]
1519: -\frac{i}{2}\int_{-\infty}^{\infty}
1520: dk \frac{ (k-k^{3}-2ik\tanh\zeta)}{(1+k^{2})^{3}}\nonumber\\
1521: &&\times e^{ik\zeta}\left(sech{\frac{\pi}{2}(1-k)}+sech{\frac{\pi}{2}(1+k)}\right).
1522: \label{eqc5}
1523: \end{eqnarray}
1524: Before evaluating the integrals in Eq.(\ref{eqc5}) we rewrite the same appropriately.
1525: We then evaluate the integrals in Eq.~(\ref{eqc5}) one by one by finding the values of the residues
1526: at the pole $(k=i)$ at different orders.
1527: The residue for the integrand function $\left(sech{\frac{\pi}{2}(\beta+1)}
1528: +sech{\frac{\pi}{2}(\beta-1)}\right)\frac{k(1-k^{2}-2ik\tanh\zeta)}
1529: {(1+k^{2})^{3}}e^{\frac{i(1+k^{2})\alpha}{k}+ik\zeta}$ at the pole $ k=i$ of order three is found to be
1530: \begin{eqnarray}
1531: Res(k=i)=-\frac{\pi}{16} [2(2b-1)\zeta+4(2b-1)\alpha+1)]
1532: sech\zeta.\label{eqc6}
1533: \end{eqnarray}
1534: Next, we find the residue for the function
1535: $e^{\frac{i(1+k^{2})\alpha}{k}+ik\zeta}\frac{(1-k^{2}-2ik\tanh\zeta)}
1536: {k(1+k^{2})\cosh\beta\frac{\pi}{2}}$ at the simple poles
1537: $k=i$ and $k=\frac{(2n+1)}{(1-v)}(-i\pm\sqrt{\frac{1-v^{2}}{(2n+1)^{2}}-1})$ ( which is out of
1538: contour). Further, at $ k=0$ there is an essential singularity, the residue of which is not shown here due its
1539: unwieldy form. Thus
1540: the residue for the above function at the first order pole $ k=i$ is given by
1541: \begin{eqnarray}
1542: Res(k=i)=\frac{1}{\pi (1-2b)^{2}} [2(2b-1)\zeta
1543: +4(2b-1)\alpha-1] sech\zeta.\label{eqc7}
1544: \end{eqnarray}
1545: The residues for the function $\left(sech{\frac{\pi}{2}(k-1)}
1546: +sech{\frac{\pi}{2}(k+1)}\right)\frac{(k-k^{3}-2ik^{2}\tanh\zeta)}{(1+k^{2})^{3}}e^{ik\zeta}
1547: $ at the
1548: pole $k=i$ of order three and at the simple poles $k=i(2n+1)+1$ and $k=i(2n+1)-1$
1549: are written as
1550: \begin{subequations}
1551: \begin{eqnarray}
1552: Res(k=i)&=&\frac{\pi}{16}(2\zeta-1)sech\zeta, \label{eqc8a}\\
1553: Res[k=i(2n+1)+1]&=&\frac{-2 i}
1554: {\pi}e^{(i-1)\zeta}\sum_{n=0}
1555: \frac{(-1)^{-n}[1+i(2n+1)]e^{-2n\zeta}}{[2-(2n+1)^{2}+2i(2n+1)]^{3}}\nonumber\\
1556: &&\times\{[(2n+1)^{2}-2 i(2n+1)]\nonumber\\
1557: &&-2i\tanh\zeta [1+i(2n+1)]\},\label{eqc8b}\\
1558: Res[k=i(2n+1)-1]&=&\frac{2 i}
1559: {\pi}e^{-(1+i)\zeta}\sum_{n=0}
1560: \frac{(-1)^{-n}[-1+i(2n+1)]e^{-2n\zeta}}{[(2n+1)^{2}-2+2i(2n+1)]^{3}}\nonumber\\
1561: &&\times\{[(2n+1)^{2}+2 i(2n+1)]\nonumber\\
1562: &&-2i\tanh\zeta [-1+i(2n+1)]\}.\label{eqc8c}
1563: \end{eqnarray}
1564: \end{subequations}
1565: It can be verified that the residue of the function $\left(sech{\frac{\pi}{2}(\beta+1)}
1566: +sech{\frac{\pi}{2}(\beta-1)}\right)$ $\frac{(1-k^{2}-2ik\tanh\zeta)}{k(1+k^{2})}
1567: e^{\frac{i(1+k^{2})\alpha}{k}+ik\zeta}
1568: $ at $k=i$ vanishes. On summing up the values of the residues given in Eqs. (\ref{eqc6}), (\ref{eqc7})
1569: and (B8) (dropping higher order terms in Eqs. (\ref{eqc8b}) and (\ref{eqc8c}) while
1570: adding) we obtain the following expression for $\psi ^{(1)}(\zeta,t_{0})$.
1571: \begin{eqnarray}
1572: \Psi^{(1)}(\zeta,t_{0})&\approx&\frac{\pi^{2}}{4(1-2b)^{2}}sech\zeta
1573: \left[b(2b-1)\zeta
1574: +\{b(2\alpha+b-1)-\alpha\}+\frac{4}{122825}\right.\nonumber\\
1575: &&\times\{ (\cosh 2\zeta-\sinh 2\zeta)
1576: \left[25\cosh 2\zeta(523\cos\zeta-1512\sin\zeta)+289\right.\nonumber\\
1577: &&(19\cos\zeta+8\sin\zeta)+15(766cos\zeta
1578: \left.\left. -2393\sin\zeta)\sinh 2\zeta\right]\}\right] .\label{eqc9}
1579: \end{eqnarray}
1580: As residues due to the singularities in the plane except along the
1581: imaginary axis, lead to secular terms in the solutions, we take into
1582: account only singularities along the imaginary axis that is at $k=i$, dropping
1583: residues due to the other two singularities at $k=i(2n+1)+1$ and
1584: $k=i(2n+1)-1$. Thus the first order correction $\psi^{(1)}(\zeta,t_{0})$ given in Eq.(\ref{eqc9}) finally
1585: becomes
1586: \begin{eqnarray}
1587: \Psi^{(1)}(\zeta,t_{0})&\approx&\frac{\pi^{2}}{4(1-2b)^{2}}
1588: \left[b(2b-1)\zeta
1589: +\{b(2\alpha+b-1)-\alpha\}\right]sech\zeta.
1590: \end{eqnarray}
1591: \begin{thebibliography}{03}
1592: \bibitem{ref1}
1593: L. Stryer, Biochemistry. $4^{th}$ ed (W. H. Freeman and Company, New York, 1995).
1594: \bibitem{ref2}
1595: L. V. Yakushevich, Physica D 79 (1994) 77.
1596: \bibitem{ref3}
1597: L. V. Yakushevich, Nonlinear Physics of DNA ( Wiley-VCH, Berlin, 2004).
1598: \bibitem{ref4}
1599: S. W. Englander, N. R. Kallenbanch, A. J. Heeger, J. A. Krumhansl and S. Litwin, Proc. Natl. Acad. Sci. U.S.A 77 (1980) 7222 .
1600: \bibitem{ref6}
1601: S. Yomosa, Phys. Rev. A 27 (1983) 2120.
1602: \bibitem{ref7}
1603: S. Yomosa, Phys. Rev. A 30 (1984) 474.
1604: \bibitem{ref8}
1605: J. Frenkel and T. Kontrova, J. Phys. (USSR) 1 (1939) 137.
1606: \bibitem{ref9}
1607: S. Takeno and S. Homma, Prog. Theor. Phys. 70 (1983) 308.
1608: \bibitem{ref10}
1609: S. Takeno and S. Homma, Prog. Theor. Phys. 72 (1984) 679.
1610: \bibitem{ref11}
1611: M. Peyrard and A. R. Bishop, Phys. Rev. Lett. 62 (1989) 2755.
1612: \bibitem{ref12}
1613: P. L. Christiansen, P. C. Lomdahl and V. Muto, {\it Nonlinearity} 4 (1991) 477.
1614: \bibitem{ref14}
1615: P. Jensen, M. V. Jaric and K. H. Bannenmann, Phys. Letts. A 95 (1983) 204.
1616: \bibitem{ref15}
1617: A. Khan, D. Bhaumik and B. Dutta-Roy, Bull. Math. Biol. 47 (1985) 783.
1618: \bibitem{ref16}
1619: V. K. Fedyanin and V. Lisy, Studia Biophys. 116 (1986) 65.
1620: \bibitem{ref17}
1621: R. V. Polozov and L. V. Yakushevich, J. Theor. Biol. 130 (1988) 423.
1622: \bibitem{ref18}
1623: J. A. Gonzalez and M. M. Landrove, Phys. Letts. A 292 (2002) 256.
1624: \bibitem{ref19}
1625: L. V. Yakushevich, Nanobiology 1 (1992) 343.
1626: \bibitem{ref21}
1627: G. Kalosakas, K. Q. Rasmussen and A. R. Bishop, Synthetic Metals, 141 (2004) 93.
1628: \bibitem{ref22}
1629: V. Muto, J. Halding, P. L. Christiansen and A. C. Scott, J. Biomol. Struc. Dyn.
1630: 5 (1988) 873.
1631: \bibitem{ref23}
1632: C. T. Zhang, Phys. Rev. A 40 (1989) 2148.
1633: \bibitem{ref24}
1634: M. Salerno, Phys. Rev. A 44 (1991) 5292.
1635: \bibitem{ref25}
1636: M. Salerno, Phys. Letts. A 167 (1992) 49.
1637: \bibitem{ref13}
1638: L. V. Yakushevich, A. V. Savin and L. I. Manevitch, Phys. Rev . E 66 (2002) 016614.
1639: \bibitem{ref20}
1640: A. Campa, Phys. Rev. E 63 (2001) 021901.
1641: \bibitem{ref26}
1642: K. Forinash, M. Peyrard and B. Malomed, Phys. Rev. E 49 (1994) 3400.
1643: \bibitem{ref27}
1644: J. A. D. Wattis, S. A. Harris, C.R. Grindon and C. A. Laughton, Phys. Rev. E 63 (2001) 061903.
1645: \bibitem{ref28}
1646: J. Cuevas, F. Palmero, J. F. R. Archilla and F. R. Romero, Phys. Letts. A 299 (2002) 221.
1647: \bibitem{ref29}
1648: S. Takeno, Phys. Letts. A 339 (2005) 352.
1649: \bibitem{ref30}
1650: J. Ladik and J. Cizek, Int. J. Quantum Chem. 26 (1984) 955.
1651: \bibitem{ref31}
1652: E. Cubero, E. C. Sherer, F. J. Luque, M. Orozco and C. A. Laughton, J. Am. Chem. Soc. 121 (1999) 8653.
1653: \bibitem{ref5}
1654: O. M. Braun and Y. S. Kivshar, Phys. Rev. B 43 (1991) 1060.
1655: \bibitem{ref32}
1656: F. Zhang, Y. S. Kivshar and L. Vazquez, Phys. Rev. A 45 (1992) 6019.
1657: \bibitem{ref32a}
1658: E. T. Kool, Annu. Rev. Biophys. Biomol. Struct. 30 (2001) 1.
1659: \bibitem{ref33}
1660: M. Hisakado and M. Wadati, J. Phys. Soc. Jpn. 64 (1995) 1098.
1661: \bibitem{ref36}
1662: R. Dandoloff and A. Saxena, J. Phys.: Condens. Matter 9 (1997) L667.
1663: \bibitem{ref37}
1664: M. J. Ablowitz, D. J. Kaup, A. C. Newell, and H. Segur, Stud. Appl. Math. 53 (1974) 249.
1665: \bibitem{ref38}
1666: D. J. Kaup, SIAM J. Appl. Math. 31 (1976) 12.
1667: \bibitem{ref39}
1668: G. L. Lamb, Elements of Soliton Theory (Wiley, New York, 1980).
1669: \bibitem{ref40}
1670: K. A. Gorshkov and L. A. Ostrovskii, Physica D 3 (1981) 428.
1671: \bibitem{ref41}
1672: D. J. Kaup, Phys. Rev. B 27 (1983) 6787.
1673: \bibitem{ref42}
1674: D. J. Kaup, Phys. Rev. B 29 (1984) 1072.
1675: \bibitem{ref43}
1676: J. P. Keener and D. W. McLaughlin, J. Math. Phys. 18 (1977) 2008.
1677: \bibitem{ref44}
1678: J. P. Keener and D. W. McLaughlin, Phys. Rev. A 16 (1977) 777.
1679: \bibitem{ref45}
1680: R. L. Herman, J. Phys. A 23 (1990) 1063.
1681: \bibitem{ref46}
1682: R. L. Herman, J. Phys. A 23 (1990) 2327.
1683: \bibitem{ref47}
1684: J. Yan and Y. Tang, Phys. Rev. E 54 (1996) 6816.
1685: \bibitem{ref48}
1686: Y. Tang and W. Wang, Phys. Rev. E 62 (2000) 8842.
1687: \bibitem{ref49}
1688: J. Yan, Y. Tang, G. Zhou and Z. Chen, Phys. Rev. E 58 (1998) 1064.
1689: \bibitem{ref50}
1690: H. Frauenfelder, Int. J. Quantum Chem. 35 (1989) 711.
1691: \bibitem{ref51}
1692: G. Kalosakas, K. O. Rasmussen, A. R. Bishop, C. H. Choi and A. Usheva,
1693: Europhys. Lett. 68 (1) (2004) 127.
1694: \bibitem{ref52}
1695: E. Kreyszig, Advanced Engineering Mathematics (John-Wiley, New York, 2002).
1696: \end{thebibliography}
1697: \end{document}
1698:
1699: