J. theor. Biol.

(1977) 66, 379-387

Solitons and Energy Transfer Along Protein Molecules

A. S. DAVYDOV

The Institute of Theoretical Physics, Academy of Sciences of the

Ukrainian S.S.R., Kiev, U.S.S.R.

(Received 2 July 1976, and in revisedform 10 August 1976)

It is shown that the energy released in the hydrolysis of ATP molecules
can be transferred in the form of vibration solitons along a-helical protein
molecules. The vibration solitons are collective excitations travelling
along a chain of successively arranged peptide groups and corresponding
to amide I vibrations. The exceptional stability of solitons in one-dimen-
sional structures can account for the small probability of their energy
transforming into that of disordered heat motion.

1. htroduction
Many biological phenomena are known to be associated with spatial energy
transfer along protein molecules. To elucidate the mechanism for this
energy transfer at the molecular level is one of the most urgent problems
in bioenergetics.
It is now established that the energy (about 0.43 eV or 10 kcal/mol)
released in the hydrolysis of adenosine triphosphate molecules (ATP) is
the universal energy unit used in all biological phenomena. This energy is
only 20 times greater than the average energy of heat motion at physiologic
temperatures. It is insufficient to excite the electronic states of molecules.
Therefore the importance of vibration excitations (with an energy less than
0.4 eV) of separate groups of atoms in protein molecules has been repeatedly
indicated.
The opponents of this viewpoint have frequently mentioned the fact that
the lifetime (with respect to transformation into heat) of such excitations
is very small (of the order of 10-‘2-10-‘3 s). At the same time it is usually
forgotten that when the molecular system is organized in a certain manner
the lifetime of excitations of some of its degrees of freedom can increase
by many orders. For example, in condensed luminescent molecular systems
(in particular, in crystalline anthracene) the electronic excitations of a system
of conjugated rc-electrons give off almost all their energy by radiation during
the time (1O-*-1O-g s) exceeding considerably the time of thermal relaxation.
7.8. 379 25
380 A. S. DAVYDOV

In addition to the fact that the energy released in the hydrolysis of ATP
molecules is insufficient to excite electronic states, the excitation of the
latter is also disadvantageous because their lifetime relative to radiation
is considerably smaller than the corresponding lifetime of vibration states.
For example, the lifetime (relative to radiation) of valence vibrations C=O
(amide I with an energy of O-21 eV) in peptide groups of proteins is equal
to about 0.01 s. It exceeds the lifetime of electronic excitations of the same
group (N 4.4 eV of energy) by a factor of (4.4/O*21)4 = I.9 x IO’.
In alpha-helical protein molecules the peptide groups (H-N-C=O) are
situated along the moleculeinthreechains. Ineach of thesechains theneighbour-
ing peptide groups are held together by hydrogen bonding between atomic H
and 0. Just these bonds provide for a relatively rigid structure of the a-helix.
Amide I vibrations are quite characteristic, i.e. they are weakly connected
with other normal vibrations of protein molecules. This fact is confirmed
by a very small difference in the amide I vibration frequencies in proteins
of different kinds.
The collective excitations of a chain of successively arranged peptide
groups which correspond to amide I vibrations are described by “solitary
waves”-the vibration solitons. They have a particle-like character, propagate
rapidly along the chain and have big lifespan, because they interact with
other degrees of freedom of the molecule much more weakly than separate
amide I vibrations.
The solitons were first observed as solitary, rapidly moving water loops
in narrow channels filled with water. The “tsunami” covering huge distances
in the oceans is also identified with solitons by some people. Numerous
investigations (Hasegawa & Tappert, 1973 ; Scott, Chu & McLaughlin, 1973 ;
Toda, 1975) showed that a remarkable property of the solitons described
by the same non-linear equation as vibration solitons is their exceptional
stability for outer disturbances.
The formation of stable excitations such as vibration solitons in the
cr-helical portions of protein molecules guarantees the high efficiency of the
transfer of energy, released in the hydrolysis of ATP molecules, by
protein. The exceptional stability of solitons in one-dimensional structures
can account for the small probability of their energy transforming into that
of disordered heat motion.

2. Solitons in One-Dimensional Chains

The dipole-dipole interaction between the adjacent peptide groups
situated along a hydrogen-bonded chain in the a-helical protein molecules
is an order of magnitude higher than a similar interaction between the
 SOLITONS AND ENERGY TRANSFER 381
nearest groups of neighbouring chains. Therefore the excitations of the three
chains may thus be considered independently. For the sake of simplicity
we shall further investigate the excited states of only one of these chains.
Consider an infinite chain of weakly bound peptide group with mass M
situated along axis x at distance R. Assume that the internal excitations of
peptide groups are characterized by an energy E (equal to 0.21 eV) and an
electric dipole moment d directed along the chain. To investigate the col-
lective excitations of this chain we shall use the results obtained by the
author and Kislukha (1976).
Let x,, be a co-ordinate characterizing the position of the nth group and
B+ the creation operator for an amide I vibration excitation.
The excited state of such a chain is defined by the wave function
w> = c %OP,+ lo> (1)
”
with the normalized condition
<w>lw> = F IU”@)12= 1. (2)

Here 10) is the function of the ground state. The function (1) must satisfy
the Schrodinger equation

Since the effective mass ii4 of the peptide group is large, we treat their
motion classically, and the amide I vibration excitations terms of quantum
mechanics.
The Hamiltonian of the chain (provided the interaction between adjacent
groups only is taken into account) can then be written as
H = T+U + c [(E-DD~B,+B,-J(B,++~B,+B,+~B,+)] (3)
n
where the function -D, characterizes the change in the static interaction
(the energy of the hydrogen bond) of the nth group with neighbouring
groups when it is excited. Making allowance for the interaction with nearest
neighbours we have

D” = %(lx,+1- xnl)+%(lxn-I-X$= (1 +;P>,

where
in = R-IA-xn-11
is the decrease in equilibrium distances between peptide groups; D = 2.9(R)
and /I are the positive parameters of a theory. J = 2d2/R3 is the operator
382 A. S. DAVYDOV

of the resonance interaction determining the transition of an excitation

from one group to another;

are the kinetic and the potential energies of the displacement of peptide
groups; w is the elasticity constant of the chain.
The set of functions a,(t) and p,(t) determine the distribution of excita-
tions of peptide groups in the chain and the relative changes in the equilibrium
distances between them. They obey the set of equations

ih%= [e+T+U-(l+~p,)D]a,-J(a,+,-a,-,). (4)

ah
2 (2P,-P,,, -Ll) + ~(2lu.l’- 1%+,12- )~“-112). (5)
at2 - - M
Equations (4) and (5) define the collective excitations and: deformations of
the chain.
We are interested only in excitations distributed over some region of
the chain. Therefore we can apply the continual model. At this approxi-
mation we can replace discrete functions u,(t) and p,(t) by the continuous
functions a((, t) and ~(5, t) of the dimensionless co-ordinate 5 = x/R.
Using the expression

we can transform equations (4) and (5) to

a
A - PD p(< t) + J-E a(( t)= 0
1
ih - -
[ K ’ 1 >
a<2 ’ ’ (6)

where
[ii 2
1
- v,”$ P(5,t)+ g $ I4rso12= 0, (7)

A=e+T+U-D-25,

vz E w/M, therefore V, = v,R is the longitudinal sound velocity in the

chain.
We shall look for solution of the set of equations (6) and (7) in the form
of excitations propagating along the chain with velocity V = vR. For this
 SOLITONS AND ENERGY TRANSFER 383

we set
PCS, 0 = p(5-4, a(<, t) = @(t-N) exp [ie(& t)].
We then obtain from equation (7)
p(t3r)=MR(v,2
B4G-v’)’
w (9)

Substituting this value in equations (6) and (8) we transform them to

(, & - I + J $ + G&(5, r)l*} a(<, t) = 0,

where
2 2
G, = GO G,=$ 2’ (12)
1 -(V/V,)”
The exact normalized solution of the non-linear Schrodinger equa-
tion (10) is caIled soliton. It has the form

(13)
cash CM - to - ~11 ’
where
PY= WJ; (14)
co is the arbitrary constant due to the translational invariance of equations (6)
and (7). Each value V = vR corresponds to the soliton energy
h2Vz G;(z2+$z4-$z6)
E,d&=ht-&,+~+
24J(l- z2)3 ’
(15)
where z z V/V, and
hQ, = s-D-2J-G;/48J (16)
is the internal energy of the soliton.
It follows from the expression (15) that when the soliton velocity approaches
the longitudinal sound velocity V, the soliton energy tends to infinity. At
all finite values of the soliton energy its velocity is therefore always less
than V,. The soliton energy thus cannot transform into the energy of
longitudinal sound waves. This is one of the reasons of the great stability
of the solitons.
384 A. S. DAVYDOV

At small velocities of the soliton when V < V, is valid, the soliton energy
(15) transforms to a simple expression
E, = hf&,++n*V2, (171
where

is effective mass of the sol&on. In the same approximation the parameter

is not dependent upon the soliton velocity.

Substituting equation (13) into equation (9) we find the distribution of
the change in relative distances between the peptide groups. At small
velocities of the soliton we shall get

p(& t) = e$ sech’ I$,,(< - &, - vt)].

According to equation (19), the maximal relative changes in distances

between peptide groups at 5 = lo+ vt moves along the chain with the
velocity V = vR. The deformation moving along the chain is accompanied
by the internal excitation of the peptide groups. The region of the chain
embraced by the excitation and deformation is defined by
nR 4rcJwR=
RAt (20)
-=pzDz*
P

3. Excitons in One-Dimensional Chains

In an absolutely rigid chain w = co, so that the non-linearity parameter
(12) is equal to zero and equation (10) reduces to linear Schrodinger equation
which has solutions in the form of plane waves. In this case an excitation is
uniformly distributed along the whole chain. The excitations of this type is
called excitons. The internal energy of the exciton is
E,, = ~-D-225. w
The effective mass of the exciton is
mz = h2/2R2J . (22)
At the final value w the non-linear equation (10) has also solution in the form
 SOLITONS AND ENERGY TRANSFER 385
of plane wave (exciton)
a(<, t) = A exp (i[k<R-w(k)t]} (23)
which characterizes the uniform distribution of an exciton along the whole
chain. To each wave (23) there corresponds the energy
ho(k) = E, + h2k2/2m,*.
The internal energy exciton E, and an effective mass rn: are determined by
expressions (21) and (22). Therefore the internal energy of the exciton
exceeds the internal energy (16) of the soliton by the value Gi/48J. Thus
the excitations (23) in one-dimensional molecular systems are unstable.
In order to satisfy the normalization condition (2), we must set up in a
chain of infinite length a wave packet with values of k in some narrow
interval k,-Ak < k < k, +Ak, using the solution (23). The excitation
described by such a wave packet propagates with the group velocity
V, = hk,/mE. In this state the energy and the momentum of the exciton
have not well-defined values. The region of the chain embraced by the
excitation increases, because these wave packets are known to “spread”.
If the excitation was initially localized on segment (A& = 2nlAk this
segment increases eventually, so that in the time t
(Ax), N” J(Ax)$ +h2t2/[m~(Ax),J2.

4. The Stability of Soiitons

A remarkable property of the vibration solitons at small velocities is that
the distribution of excitation and deformation characterized by the function
(19) remains completely unchanged in motion. The value of the maximum,
the shape and the width (20) are independent of the velocity (energy) of a
soliton and are defined only by the properties of the chain, i.e. the quantities
J, G,, R. When the soliton slows down, its velocity changes, leaving the shape
and the size of the excitation region unchanged. The stability property of
solitons enables us to regard them as particle-like formations.
The stability of solitons is connected with the competition of two processes :
the spreading of excitation region, determined by the resonance interaction J,
and the tendency towards excitation localization which is due to the non-
linear terms of equation (10) proportional to the parameter (12). In protein
molecules the formation of solitons corresponding to the vibration amide I
is favoured by the one-dimensional arrangement of peptide groups, the
relative softness of hydrogen bonds between peptide groups (the small w)
and the orientation of electric dipole moments of vibrations along the chain.
Only if the electric dipole moments in the excitation region are arranged
386 A. S. DAVYDOV

in this manner the distances between peptide group decrease, thus promoting
excitation localization. In very rigid chains the ratio Go/J is very small and
the formation of solitons is impossible.
Solitons can be formed only under a strong coupling of the internal
vibration with a local deformation of the chain. This coupling is proportional
to the non-linearity parameter Go = j?2D2/~R2. This parameter is big
enough in comparatively soft chains of the peptide groups in protein
molecules.
The condition of the formation solitons in a-helical protein molecules
resembles the condition of the super conductivity formation in a metal. The
superconductivity state due to the interaction of the electrons of a metal
with phonons (the shifts of ions). Good conductors (silver, copper, gold)
do not turn into a superconductive state. Superconductive states arise only
in the metals which have a strong enough electron-phonon interaction. The
strong electron-phonon interaction leading to a large resistivity in the normal
state of metal makes the formation a superconducting state without any
resistivity.
The above-considered solitons transfer with a very great efficiency the
energy of excitation between different portions of the protein molecule.
The energy transfer is accompanied by the deformation moving along the
protein molecule. These solitons have a longitudinal polarization, since
the electric dipole moment is directed along the chain.
The formation of solitons can occur only under the action of local
excitations. These types of excitations also include the chemical energy
released in the hydrolysis of ATP molecules which often attach themselves
to certain portions of protein molecules (for example, to the heads of myosin
molecules in the muscles). The vibration solitons cannot be excited by the
infrared radiation of suitable frequency (N 1660 cm-‘), since the wave length
of this radiation is very large (- 6 x IO4 A), and the photon localization is
impossible with an accuracy higher than the wave length of radiation.
The concept of solitons which occur in myosin molecules was used by
the present author? (Davydov, 1973, 1974,1975) to elucidate the contraction
mechanism of transversely striated muscles at the molecular level. Only
the kinetic energy of solitons is actively used in muscle contraction, since
the sliding of thin filaments relative to thick ones in the sarcomeres is due
to the movement of solitons in myosin molecules contained in thick
filaments.

-1In these papers it was assumed that the distortion of the chain follows in an inert
manner the movement of excitation. The notion of inert motion of peptide groups is
valid only if the soliton velocity is much less than the longitudinal sound velocity in chain
of the peptide groups.
 SOLITONS AND ENERGY TRANSFER 387

REFERENCES
DAVYDOV, A. S. (1973). J. theor. Biol. 38, 559.
DAVYDOV, A. S. (1974). Studia Biophysicu 47, 221.
DAWDOV, A. S. (1975). Ukr. fiz. Zh. 20, 179.
DAVYDOV, A. S. & KJSLUKHA, N. I. (1976). Phys. stat. sol. 75, 735.
HASEGAWA, A. & TAPPERT, F. (1973). Appl. Phys. Lett. 23, 142.
&m-r, A., CHU, E. Y. & MCLAUGHLIN, D. W. (1973). Proc. IEEE 61, 1443.
TODA, M. (1975). Phys. Reports (Sec. C Phys. Lett.) 18, 1.

T.B. 26