Vol.
178,
No.
July
15, 1991
1, 1991
BIOCHEMICAL
BIOPHYSICAL
AND
RESEARCH
COMMUNICATIONS Pages
Conformational
Space A Molecular
of cr l-4
glvcosidic
Dynamics
Study
192-l
97
linkage:
M.Prabhakaran Department
of Medicinal
Chemistry
University
of Illinois
P. Box 6998, Received
May 10,
Molecular
Dynamics
The conformational
using
connecting
the simulations.
The
is also brought
nature
the different
preferred
structural
by both theoretical was investigated
Hence,
in a unique
glucose
and has been studied of the molecule
angles
(MD)
as a model
pressure
and
step of 0.005 saved during
in representing
ps.
two virtual
were found
the
in both
the glycosidic
Press, Inc.
important
feature
in determining
The nature
simulation,
dependence
of linkage
is
out in both vacua
ring in terms
of Brady
(3) have
of the pucker
angle of
the polysaccharide on the ring
seven ~~(1-4) linked complexes
for enzyme
carried
(2) and the studies
between
consisting
structure
means (1). In this study, the polysaccharide
dynamics
it can form inclusion
linkage
of polysaccharides.
on a-cyclodextrin
much
a time
by calculating
0 1991 Academic
space of the glucose
a cyclic molecule structure,
was studied
is the most
the linkage
at constant
with
units for the 2000 structures
arrangements
conformational we followed
of solvent
dihedral
linkage
studies
out for 100 ps on crystal
for cr l-4 glycosidic
and experimental
way (4) without
-cyclodextrin, its annular
linkage
regions
by molecular
Our earlier
the large
60680
algorithm,
out from these studies.
studied
shown
inclusion MD
use of these virtual
linkage
glucose.
explicit
the GROMOS
of the glycosidic
tertiary
and in water.
IL
have been carried
the successive
Three
The
781)
at Chicago,
Chicago,
space of the glycosidic
two simulations. linkage
simulations
in vacua and with
temperature
dihedrals
(M/C
1991
of P-cyclodextrin constant
and Pha.rma,cognosy
activity
was solved at a very high resolution
with
molecules,
conformation. glucose
and hydrogen by Saenger
We choose p
residues.
a wide variety
Because
of
of guest molecules
bonding. group
defined
The structure
(5).
Method The /?-cyclodextrin model
(5).
0006-291X/91 Copyright All rights
The MD
coordinates simulations
refined
by neutron
were carried
$1.50
0 1991 by Academic Press, Inc. of reproduction in any form reserved.
192
out using
diffraction
are taken as the starting
the molecular
dynamics
package
Vol. 178, No. 1, 1991
E&UZL
BIOCHEMICAL
Dihedral8 4: 04 ... Cl 11: Cl - 0; [ - denotes
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
Definition: - or, - c; - c; ... 0; bonded atoms; ... denotes nonbonded
atoms]
GROMOS (Groningen Molecular Simulation), using the parameters developed by us and published elsewhere (6). A 100 picosecond simulation was carried out on the system both in vacua and in aqueous environment with a time step of 0.005 ps. 2000 structures were saved during the simulation.
To scan adequate conformational space, the system was
assignednew velocities at every 5 pe. After every 20 pe, the molecuhr was minimized with conjugate gradient minimization and a fresh simulation was started. To simulate aqueous environment, the minimized molecule was placed in a box of TIPSP water molcculee with the dimcneione 22.1 A, 23.1 A and 25.1 A. Th c water moleculce within 2.3 A distance from the cyclodcxtrin atoms were removed. Further minimization wa8 carried on the whole system, constraining the solute molecule in the starting configuration and adjusting the configuration of 407 water molecules. After adequate heating and equilibration, the simulation was carried out with suitable boundary conditions and at constant temperature and at constant pressure. Both the velocities of the solute and solvent molcculce are scaled separately to maintain constant temperature. The lengths of the boxes were adjusted during the simulation. After every 20 ps, the simulation wa.s restarted with a new box of molecules after adequate minimization. 2000 structures, saved from 100 ps run, were used for data analysis. The polysaccharide linkage as defined by Ramachandran’s(4) group is shown in Fig 1. The virtual dihedral connecting 04, Cl of the first residue with 04 and C4 of the second residue is defined as 4 and the virtual dihedral connecting Cl of the first residue 04 and C4 of the second and 04 of the third residue is dcfincd as $. As only one atom from the ring is chosenfor these dihedrrds, these dihedral8 would be the representative of the tertiary linkage between the polysaccharidel without much dependence on the local donforimation. Each structure containr consecutive 7 4 and 7 $ dihedra+langlee‘ Our data base coneists of two set8 of 14000 values of 4 and $ derived from in vacua and in aqueous simulatione. Results and
Di8CU88iOn
The stereo view of cyclodextrin molecule during the eimulation ir rhown in Fig 2. The molecule placed in a box of water molecules ir shown in Fig 3. The o&l&one continued smoothly throughout the oecillations. To rhow the dynamic conformational apace of the molecuka the continuous 50 snap ehotr 8aved during the trajectory ir depicted in Fii 4, 193
Vol.
178,
No.
BIOCHEMICAL
1, 1991
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
03 Figure
A stereo view of P-cyclodextrin
E&UJCL P-cyclodextrin molecules is 407.
showing
the structure around
(7). The glucose to -180” (Fig
picture
is the same range is measured
by the dihedrals
by 360 to have a continuous
5) shows
the distribution
noticeable: distribution
3) the conformational
Figure
three
space is limited
linkage in aqueous dihedral
derived
from
from sca.t-
the simulation.
from both
points
are shifted
from crystallographic
during
the simulation
4, The dynamic trajectory of glucose unit during snap shots were saved at every 0.2 ps. 194
Three
a (8) a.nd p
regions.
is not crystallographically
10
in water
ps period.
solution The
occupancy
and $=170
is
values
0” to 360”.
derived
of
dihedral
specific
1) the high density near 4=240
scale from
movement
the maintenance
4 and 4. The
structures
10 ps period.
the dynamic
in glycosidic
of 4 and + angles
the values of crystallography
2)The
to show
that
for a-cyclodextrin
is added
The plots display
within
steps to investigate
as observed
The circles denote
tures.
unit
The fluctuation
(5) cyclodextrins. immediately
glucose
in order
out necessary
to the equilibrium.
linkage
22.1 A, 23.1 A, 25.1 A. The number of water
of a single
in this
We have carried closer
10” which
ter plots
movement
is attempted
of the molecule.
-1”
in a box of dimensions
the conformational
No superposition
during the simulation.
features
are struc-
observable.
due to damping
These 56 continuous
as
Vol.
178, No. 1, 1991
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
b
Scatter plot of q5and +: a) In vacua. b) In water during 100 ps simulation. 7000 points are shown in each plot. The angles between -1” and -180” are added by 300”. Circles denote crystal structure vahes.
observed in similar studies (9). In Homans’ (10) caiculations on @l-4 linkage, the minimum cncrgy conformations were aIso observed other that found in crystal structure. We have also cslculated the linkage as defined by Tvaroshat et al (11) and find three similar regions of occupation as shown in these scatter plots. The dampening of torsional
oscil-
lations, when water molcculcs were included, could be due to the hydrogen bonds around the solute, as indicated by Edge et al (9). Wc h ave also plotted the frequency of occurrence of 4 and $ during both the simulations (Fig 6). The maximum occurrence of 4 occurs near 165 in both the oscillations. As shown is limited
in water
in contrast
to in vacua
from the sca.tter plot,
simulation.
It is interesting
the distribution to note that
the
conformational spaceis limited between 120 and 240 in both the simulations, whereas the middle region is heavily populated in water simulation. The distribution of II, shows a slight bimodal distribution, having peaks at 145 and at 225. In water, the distribution is limited to 225, whereas the distribution in vacua extends beyond 280. A central minimum at 200 is noticeable at both the simulations. Pullman’s group from their quantum mechanical calculations (12) have observed the global minimum for 4 90, asd 180 and $ between near 110 and 240~ Allowed conformations with higher energy arc ahro found for 4 between 180 to 240 and $ between 180 and 240. Similar conformational spacc was observed earlier by Rao et al (4). By using this virtual dihedral, the anomalous linkage found in a glycosidic linkage of a-cyclodextrin was brought out by Manor and Sacngcr (13). Thus our dynamics simulation exhibits the dynamic space of the ghrcosc a 1-4 Iinkage and the use of this virtuai dihcdrsis to represent the tcrtiuy linkage of the polysaccharidea The conformationrl space of further linkage could be scanned by this method& 195
Vol.
178,
No.
1, 1991
160.0
180.0
BIOCHEMICAL
200.02200240.0 PHI
AND
260.02mo
BIOPHYSICAL
RESEARCH
hi,nqlrln 1I
8
300.0
loo.0
COMMUNICATIONS
1400
lo.0 1800
2oo.o PHI
2200240.0
260.0
280.0
3ac.o
d
II
IO 0.0 1200
fieur~
6.
diagram
of the 4 and
a) $ in vacua,
Bar
b) q5 in water,
+ distribution
14o.c
during
c) 4 in vacua,
I-lh,7
160.0
160.0
200.02200240.0 PSI
,
260.0280.0
3000
the 100 ps simulation.
d) 1c, in water.
Acknowledgments I am thankful Michael
to Profs Stephen
E. Johnson
Association
C. Harvey
for their support.
of Chicago
Senior
of University
This
Fellowship
of Alabama
work was supported
at Birmingham
in part by American
and Heart
and grant.
Reference8 1) Bender,
M. L and Komiyama,
Chemistry.
Springer
2) Prabhakaran, 3) Brady,
5) Betzel
S. C. (1987)
J. A m. Chem. Sot.
4) Rao, V. S. R., Sundararajan, (1967)
Cyclodextrin
Verlag.
M. and Harvey,
J. W (1986)
M. (1978)
Bioploymers
108,8153-8160.
P. R., Ramakrishnan,
Conformation
of Biopolymers
Vol II Academic
Ch.,
W., Hingerty,
B. E and Brown,
Saenger,
26, 1087-96. C. and
Ramachandran,
G. N
Press NY p 721-737. G. M. (1984)
J. Amer.
Chem.
G. C., Roden,
L. and
sot. 106,7545-7557. 6) Ramakrishnan, Harvey,
N., Bo-young,
S. C. (1990)
J. Biol.
C., Prabhakaran, Chem 265,
M., Ekbory,
18256-62.
Vol.
178,
No..
7) Koehler,
1, 1991
BIOCHEMICAL
J. E. H., Saenger,
AND
BIOPHYSICAL
W. and van Gunsteren,
RESEARCH
W. F. (1988)
COMMUNICATIONS
J. Mol.
Viol.
203,
241-250. 8) Lindner, 9) Edge,
K. and Saenger, C. J., Singh,
T. W. (1990) 10) Homans, 11) Tvaroska, 12) Giacomini, 13) Manor,
W. (1981)
U. C., Bazza,
I., Perez,
Biochemistry
S., Noble
M., Pullman, P. C and Saenger,
Cry&a
R., Taylor,
B38,
203-210.
G. L., Dwek,
R. A., and Rademacher,
29,1971-74.
Biochemistry
S. W. (1990)
Acta
29, 9910-9118.
0. and Taravel,
B. and Maigret, W (1974)
F. (1987)
B. (1970)
J. Am.
197
Biopolymers
Theoret.
Chim.
26, 1499-1508. Acta 19,347-364.
Chem. Sot. 96, 3630-3639.
178,
No.
July
15, 1991
1, 1991
BIOCHEMICAL
BIOPHYSICAL
AND
RESEARCH
COMMUNICATIONS Pages
Conformational
Space A Molecular
of cr l-4
glvcosidic
Dynamics
Study
192-l
97
linkage:
M.Prabhakaran Department
of Medicinal
Chemistry
University
of Illinois
P. Box 6998, Received
May 10,
Molecular
Dynamics
The conformational
using
connecting
the simulations.
The
is also brought
nature
the different
preferred
structural
by both theoretical was investigated
Hence,
in a unique
glucose
and has been studied of the molecule
angles
(MD)
as a model
pressure
and
step of 0.005 saved during
in representing
ps.
two virtual
were found
the
in both
the glycosidic
Press, Inc.
important
feature
in determining
The nature
simulation,
dependence
of linkage
is
out in both vacua
ring in terms
of Brady
(3) have
of the pucker
angle of
the polysaccharide on the ring
seven ~~(1-4) linked complexes
for enzyme
carried
(2) and the studies
between
consisting
structure
means (1). In this study, the polysaccharide
dynamics
it can form inclusion
linkage
of polysaccharides.
on a-cyclodextrin
much
a time
by calculating
0 1991 Academic
space of the glucose
a cyclic molecule structure,
was studied
is the most
the linkage
at constant
with
units for the 2000 structures
arrangements
conformational we followed
of solvent
dihedral
linkage
studies
out for 100 ps on crystal
for cr l-4 glycosidic
and experimental
way (4) without
-cyclodextrin, its annular
linkage
regions
by molecular
Our earlier
the large
60680
algorithm,
out from these studies.
studied
shown
inclusion MD
use of these virtual
linkage
glucose.
explicit
the GROMOS
of the glycosidic
tertiary
and in water.
IL
have been carried
the successive
Three
The
781)
at Chicago,
Chicago,
space of the glycosidic
two simulations. linkage
simulations
in vacua and with
temperature
dihedrals
(M/C
1991
of P-cyclodextrin constant
and Pha.rma,cognosy
activity
was solved at a very high resolution
with
molecules,
conformation. glucose
and hydrogen by Saenger
We choose p
residues.
a wide variety
Because
of
of guest molecules
bonding. group
defined
The structure
(5).
Method The /?-cyclodextrin model
(5).
0006-291X/91 Copyright All rights
The MD
coordinates simulations
refined
by neutron
were carried
$1.50
0 1991 by Academic Press, Inc. of reproduction in any form reserved.
192
out using
diffraction
are taken as the starting
the molecular
dynamics
package
Vol. 178, No. 1, 1991
E&UZL
BIOCHEMICAL
Dihedral8 4: 04 ... Cl 11: Cl - 0; [ - denotes
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
Definition: - or, - c; - c; ... 0; bonded atoms; ... denotes nonbonded
atoms]
GROMOS (Groningen Molecular Simulation), using the parameters developed by us and published elsewhere (6). A 100 picosecond simulation was carried out on the system both in vacua and in aqueous environment with a time step of 0.005 ps. 2000 structures were saved during the simulation.
To scan adequate conformational space, the system was
assignednew velocities at every 5 pe. After every 20 pe, the molecuhr was minimized with conjugate gradient minimization and a fresh simulation was started. To simulate aqueous environment, the minimized molecule was placed in a box of TIPSP water molcculee with the dimcneione 22.1 A, 23.1 A and 25.1 A. Th c water moleculce within 2.3 A distance from the cyclodcxtrin atoms were removed. Further minimization wa8 carried on the whole system, constraining the solute molecule in the starting configuration and adjusting the configuration of 407 water molecules. After adequate heating and equilibration, the simulation was carried out with suitable boundary conditions and at constant temperature and at constant pressure. Both the velocities of the solute and solvent molcculce are scaled separately to maintain constant temperature. The lengths of the boxes were adjusted during the simulation. After every 20 ps, the simulation wa.s restarted with a new box of molecules after adequate minimization. 2000 structures, saved from 100 ps run, were used for data analysis. The polysaccharide linkage as defined by Ramachandran’s(4) group is shown in Fig 1. The virtual dihedral connecting 04, Cl of the first residue with 04 and C4 of the second residue is defined as 4 and the virtual dihedral connecting Cl of the first residue 04 and C4 of the second and 04 of the third residue is dcfincd as $. As only one atom from the ring is chosenfor these dihedrrds, these dihedral8 would be the representative of the tertiary linkage between the polysaccharidel without much dependence on the local donforimation. Each structure containr consecutive 7 4 and 7 $ dihedra+langlee‘ Our data base coneists of two set8 of 14000 values of 4 and $ derived from in vacua and in aqueous simulatione. Results and
Di8CU88iOn
The stereo view of cyclodextrin molecule during the eimulation ir rhown in Fig 2. The molecule placed in a box of water molecules ir shown in Fig 3. The o&l&one continued smoothly throughout the oecillations. To rhow the dynamic conformational apace of the molecuka the continuous 50 snap ehotr 8aved during the trajectory ir depicted in Fii 4, 193
Vol.
178,
No.
BIOCHEMICAL
1, 1991
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
03 Figure
A stereo view of P-cyclodextrin
E&UJCL P-cyclodextrin molecules is 407.
showing
the structure around
(7). The glucose to -180” (Fig
picture
is the same range is measured
by the dihedrals
by 360 to have a continuous
5) shows
the distribution
noticeable: distribution
3) the conformational
Figure
three
space is limited
linkage in aqueous dihedral
derived
from
from sca.t-
the simulation.
from both
points
are shifted
from crystallographic
during
the simulation
4, The dynamic trajectory of glucose unit during snap shots were saved at every 0.2 ps. 194
Three
a (8) a.nd p
regions.
is not crystallographically
10
in water
ps period.
solution The
occupancy
and $=170
is
values
0” to 360”.
derived
of
dihedral
specific
1) the high density near 4=240
scale from
movement
the maintenance
4 and 4. The
structures
10 ps period.
the dynamic
in glycosidic
of 4 and + angles
the values of crystallography
2)The
to show
that
for a-cyclodextrin
is added
The plots display
within
steps to investigate
as observed
The circles denote
tures.
unit
The fluctuation
(5) cyclodextrins. immediately
glucose
in order
out necessary
to the equilibrium.
linkage
22.1 A, 23.1 A, 25.1 A. The number of water
of a single
in this
We have carried closer
10” which
ter plots
movement
is attempted
of the molecule.
-1”
in a box of dimensions
the conformational
No superposition
during the simulation.
features
are struc-
observable.
due to damping
These 56 continuous
as
Vol.
178, No. 1, 1991
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
b
Scatter plot of q5and +: a) In vacua. b) In water during 100 ps simulation. 7000 points are shown in each plot. The angles between -1” and -180” are added by 300”. Circles denote crystal structure vahes.
observed in similar studies (9). In Homans’ (10) caiculations on @l-4 linkage, the minimum cncrgy conformations were aIso observed other that found in crystal structure. We have also cslculated the linkage as defined by Tvaroshat et al (11) and find three similar regions of occupation as shown in these scatter plots. The dampening of torsional
oscil-
lations, when water molcculcs were included, could be due to the hydrogen bonds around the solute, as indicated by Edge et al (9). Wc h ave also plotted the frequency of occurrence of 4 and $ during both the simulations (Fig 6). The maximum occurrence of 4 occurs near 165 in both the oscillations. As shown is limited
in water
in contrast
to in vacua
from the sca.tter plot,
simulation.
It is interesting
the distribution to note that
the
conformational spaceis limited between 120 and 240 in both the simulations, whereas the middle region is heavily populated in water simulation. The distribution of II, shows a slight bimodal distribution, having peaks at 145 and at 225. In water, the distribution is limited to 225, whereas the distribution in vacua extends beyond 280. A central minimum at 200 is noticeable at both the simulations. Pullman’s group from their quantum mechanical calculations (12) have observed the global minimum for 4 90, asd 180 and $ between near 110 and 240~ Allowed conformations with higher energy arc ahro found for 4 between 180 to 240 and $ between 180 and 240. Similar conformational spacc was observed earlier by Rao et al (4). By using this virtual dihedral, the anomalous linkage found in a glycosidic linkage of a-cyclodextrin was brought out by Manor and Sacngcr (13). Thus our dynamics simulation exhibits the dynamic space of the ghrcosc a 1-4 Iinkage and the use of this virtuai dihcdrsis to represent the tcrtiuy linkage of the polysaccharidea The conformationrl space of further linkage could be scanned by this method& 195
Vol.
178,
No.
1, 1991
160.0
180.0
BIOCHEMICAL
200.02200240.0 PHI
AND
260.02mo
BIOPHYSICAL
RESEARCH
hi,nqlrln 1I
8
300.0
loo.0
COMMUNICATIONS
1400
lo.0 1800
2oo.o PHI
2200240.0
260.0
280.0
3ac.o
d
II
IO 0.0 1200
fieur~
6.
diagram
of the 4 and
a) $ in vacua,
Bar
b) q5 in water,
+ distribution
14o.c
during
c) 4 in vacua,
I-lh,7
160.0
160.0
200.02200240.0 PSI
,
260.0280.0
3000
the 100 ps simulation.
d) 1c, in water.
Acknowledgments I am thankful Michael
to Profs Stephen
E. Johnson
Association
C. Harvey
for their support.
of Chicago
Senior
of University
This
Fellowship
of Alabama
work was supported
at Birmingham
in part by American
and Heart
and grant.
Reference8 1) Bender,
M. L and Komiyama,
Chemistry.
Springer
2) Prabhakaran, 3) Brady,
5) Betzel
S. C. (1987)
J. A m. Chem. Sot.
4) Rao, V. S. R., Sundararajan, (1967)
Cyclodextrin
Verlag.
M. and Harvey,
J. W (1986)
M. (1978)
Bioploymers
108,8153-8160.
P. R., Ramakrishnan,
Conformation
of Biopolymers
Vol II Academic
Ch.,
W., Hingerty,
B. E and Brown,
Saenger,
26, 1087-96. C. and
Ramachandran,
G. N
Press NY p 721-737. G. M. (1984)
J. Amer.
Chem.
G. C., Roden,
L. and
sot. 106,7545-7557. 6) Ramakrishnan, Harvey,
N., Bo-young,
S. C. (1990)
J. Biol.
C., Prabhakaran, Chem 265,
M., Ekbory,
18256-62.
Vol.
178,
No..
7) Koehler,
1, 1991
BIOCHEMICAL
J. E. H., Saenger,
AND
BIOPHYSICAL
W. and van Gunsteren,
RESEARCH
W. F. (1988)
COMMUNICATIONS
J. Mol.
Viol.
203,
241-250. 8) Lindner, 9) Edge,
K. and Saenger, C. J., Singh,
T. W. (1990) 10) Homans, 11) Tvaroska, 12) Giacomini, 13) Manor,
W. (1981)
U. C., Bazza,
I., Perez,
Biochemistry
S., Noble
M., Pullman, P. C and Saenger,
Cry&a
R., Taylor,
B38,
203-210.
G. L., Dwek,
R. A., and Rademacher,
29,1971-74.
Biochemistry
S. W. (1990)
Acta
29, 9910-9118.
0. and Taravel,
B. and Maigret, W (1974)
F. (1987)
B. (1970)
J. Am.
197
Biopolymers
Theoret.
Chim.
26, 1499-1508. Acta 19,347-364.
Chem. Sot. 96, 3630-3639.
