| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Biophys J, September 2001, p. 1588-1599, Vol. 81, No. 3
Institute of General, Inorganic, and Theoretical Chemistry, University of Innsbruck, A-6020 Innsbruck, Austria
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
METHODS
RESULTS
SUMMARY AND CONCLUSION
REFERENCES
|
|---|
Minor groove binding ligands are of great interest due to their extraordinary importance as transcription controlling drugs. We performed three molecular dynamics simulations of the unbound d(CGCGAATTCGCG)2 dodecamer and its complexes with Hoechst33258 and Netropsin. The structural behavior of the piperazine tail of Hoechst33258, which has already been shown to be a contributor in sequence-specific recognition, was analyzed. The simulations also reveal that the tails of the ligands are able to influence the width of the minor groove. The groove width is even sensitive for conformational transitions of these tails, indicating a high adaptability of the minor groove. Furthermore, the ligands also exert an influence on the BI/BII backbone conformational substate behavior. All together these results are important for the understanding of the binding process of sequence-specific ligands.
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
SUMMARY AND CONCLUSION
REFERENCES
|
|---|
Sequence-specific minor groove binding ligands
(Ren and Chaires, 1999
; Fishleigh et al., 2000; Geierstanger and
Wemmer, 1995; Park and Breslauer, 1992; Goodsell et al., 1995; Sponar
and Votavova, 1996; Wemmer and Dervan, 1997; Steinmetzer and Reinert,
1998) are able to influence the expression of specific genes
(Gottesfeld et al., 1997; Dickinson et al., 1998; Wittung-Stafshede,
1998; Ho et al., 1994). Thus, small ligands such as Netropsin (Chen et
al., 1996; Singh and Kollman, 1999; Zakrzewska et al., 1983; Zimmer et al., 1982; Duong and Zakrzewska, 1997; Perez and Portugal, 1990; Patel, 1982; Coll et al., 1989; Tarbernero et al., 1993; Kopka et
al., 1985; Rentzeperis et al., 1995; Nunn et al., 1997; Lah and
Vesnaver, 2000; Sriram et al., 1992b), Hoechst33258 (Sriram et
al., 1992a; Spink et al., 1994; Squire et al., 2000; Teng et al., 1988;
Carrondo et al., 1989; Vega et al., 1994), or small polyamides (Herman
et al., 1999a,b; Kielkopf et al., 1998a) are of interest as anti-tumor,
anti-viral, and anti-microbial agents. A great variety of such ligands
have been synthesized and investigated with different experimental and
theoretical methods. Now it is possible to distinguish between all four
possible base pair steps (A-T, T-A, G-C, C-G) (Kielkopf et al., 1998b;
Ellervik et al., 2000) in the minor groove. To recognize one specific
DNA sequence out of the human genome (3 × 109 base pairs), the ligands have to interact at
least with 17 base pairs (Thuong and Hélène, 1993). Thus,
to improve the selectivity, the lengths of the ligand molecules are
extended or hybrids of ligands are used (Ketterle et al., 1996;
Perree-Fauvet and Gresh, 1994; Becker and Norden, 1999).
Netropsin and Hoechst 33258 have affinity for A+T-rich regions, and
several studies of the drugs complexed with DNA are reported. The
N-methyl piperazine (Pip), the two benzimidazole (Bz1 and Bz2), and the phenol (Phe) group are the four planar structural segments of the Hoechst 33258 ligand (Fig.
1). The torsion angles between these
groups are named according to Quintana et al. (1991) 
1, 2, and
3. Free rotation is possible around the connecting bonds. The
Netropsin ligand consists of a guanidinium (Gua), two pyrrole (Py), and
a propylamidinium (PrAm) part. Both molecules adopt a convex
conformation (arc-like conformation) in DNA complexes, thus fitting
exactly in the concave shape of the minor groove.
|
Footprinting experiments (Murray and Martin, 1988; Harshman and Dervan,
1985) performed on Hoechst33258 underline the A-T preference of the
ligand. The exocyclic amino group of guanine prevents the binding of
Hoechst33258 in G+C-containing sequences although it has a tolerance
for G-C base pairs at the end of the binding site. Recent molecular
dynamics simulations of Netropsin (Wellenzohn et al., 2000b)
and polyamide DNA (Wellenzohn et al., 2001) complexes and experimental
investigations also proposed the importance of the ligand tails for
recognition processes. Cheryl et al. (2000) suggest that the tails are
responsible for the binding orientation of small ligands, and Becker
and Norden (1999, 2000) attribute sequence specificity to the
interaction of a cationic piperazinium tail with the minor groove (Ren
et al., 1999; Wilson et al., 1985). Thus, an exact structural knowledge of the interaction behavior between the ligand tails and the DNA is of
extraordinary interest in the design of new ligands.
Behind the direct readout arranged through ligand-DNA contacts the
indirect readout also contributes to sequence specificity and
selectivity (Neidle, 1997; Dickerson, 1998; Giese et al., 1997; Chen
and Prohofsky, 1995; von Hippel, 1994; Strauss et al., 1996;
Bareket-Samish et al., 1998; Wenz et al., 1996; Steitz, 1993; W. Flader, B. Wellenzohn, R. H. Winger, A. Hallbrucker, E. Mayer,
and K. R. Liedl, submitted; Flader et al., 1995; Gehring et al., 1994;
Bewley et al., 1998). Indirect readout can be mediated by means of
changes in structural parameters such as bending, unwinding, and the
groove width. Two different models are used to explain the
heterogeneity in the minor groove. One model explains the groove width
by the repulsion of the negative phosphate groups, thus proposing an
influence of positive charges on the minor groove width (Hamelberg et
al., 2000; Shui et al., 1998; Tereshko et al., 1999; Feig and Petitt,
1999; Hud and Feigon, 1997; Young and Beveridge, 1998). The second
model makes the short-range interaction of DNA bases responsible for
the size of the minor groove (Wing et al., 1980; Drew and Dickerson,
1981; Chiu et al., 1999).
The BI/BII conformational
substates are defined by the 
and angles of the B-DNA backbone
or by the angle difference ( 
). In the
BI state the corresponding and angles are
between 120° and 210° (trans) and 235°-295°
(gauche
), respectively; for
BII, the angle lies between 210° and 300° (gauche), between 150° and 210°
(trans) (Schneider et al., 1997; Berman, 1997; Hartmann and
Lavery, 1996). The angle difference ( ) is close to
90° for BI and +90° for
BII phosphates (Fratini et al., 1982). Molecular
dynamics simulations compared with experimental results have shown that
force fields are able to describe the BI/BII substate pattern in
a correct way (Winger et al., 1998; Rüdisser et al., 1997;
Pichler et al., 1999, 2000a,b). The complexation of the minor
groove with ligands influences the
BI/BII behavior of the DNA.
It has been proposed that these conformational substates are able to
contribute to sequence recognition (van Dam and Levitt, 2000; Song et
al., 1997; Wellenzohn et al., 2000b, B. Wellenzohn, W. Flader, R. H. Winger, A. Hallbrucker, E. Mayer, and K. R. Liedl, submitted; W. Flader, B. Wellenzohn, R. H. Winger, A. Hallbrucker, E. Mayer, and K. R. Liedl, submitted; Pichler et al., 2000a).
We performed two 5-ns molecular dynamics simulations of complexes of
the Drew Dickerson dodecamer (d(CGCGAATTCGCG)2).
In the first simulation the dodecamer is complexed with Netropsin and in the second simulation with Hoechst33258. As reference we use a 10-ns
simulation of the unbound dodecamer. The comparison of all three
simulations allows us to investigate on the one hand the common effects
induced by minor groove binding ligands on DNA, and on the other hand
we are able to detect differences between Hoechst33258 and Netropsin.
Both Hoechst33258 and Netropsin undergo structural transitions at the
end of the molecule. These structural changes of the tails of the
ligand are not separable from those of the DNA because they induce
changes in the minor groove width. Thus, the results indicate that the
minor groove width exhibits great flexibility and changes the structure
to fit the ligand exactly in the groove. In such a case a rigid ligand
should lead to an entropic penalty due to stiffening of the DNA by
complexation, or if the DNA keeps its pliability the direct interaction
is weakened. This structural and dynamic knowledge of the complexation
is of importance in the ligand design because, for example, the
interaction of a piperazine part with the minor groove is able to
introduce sequence specificity to intercalating ligands (Becker and
Norden, 1999). The two minor groove binders are also able to change the pattern of the BI/BII
substates, supporting recent suggestions of the influence of these
substates in sequence recognition (van Dam and Levitt, 2000; Song et
al., 1997; Wellenzohn et al., 2000b; W. Flader, B. Wellenzohn, R. H. Winger, A. Hallbrucker, E. Mayer, and K. R. Liedl, submitted; Pichler
et al., 2000a,b).
| |
METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
SUMMARY AND CONCLUSION
REFERENCES
|
|---|
Molecular dynamics simulations of DNA and DNA complexes are able to provide complementary information to experimental evidence. Thus, molecular dynamics simulations are an essential tool in the field of biomolecular research. The inclusion of the long-range interactions via the Ewald summation in the form of the particle mesh Ewald method leads to stable B-form DNA trajectories. We performed two simulations of DNA complexes (simulations A and B) and one reference simulation of the unbound DNA (simulation C) (Fig. 2).
|
Simulation A
As a starting point the crystal structure of the
Nt/d(CGCGAATTCGCG)2 complex was used (Nt
represents Netropsin). The structure has the protein data bank (PDB)
code 1D86. Each strand of the DNA has 11 PO
) package.
Subsequently, solvation of the DNA with TIP3P Monte Carlo water boxes
requiring a 12-Å solvent shell in all directions resulted in a system
with the dimension 67.1 × 50.6 × 48.7 Å3 containing 4642 water molecules. The
corresponding 
-value (water/nucleotide) is 193.4. The simulation was
carried out using the AMBER5 (Case et al., 1995) package with the
all-atom force field of Cheatham et al. (1999). The procedure of the
parameter development for the ligand has already been described
(Wellenzohn et al., 2000b). Standard protocols (Young et al., 1997a,b;
de Souza and Ornstein, 1997a,b; Winger et al., 1998) were adapted for
our needs. At the beginning, minimizations were carried out with
harmonic restraints on DNA and counterion positions. The restraints
were stepwise relaxed, and at the end, a 500-step minimization without
restraints was performed. For equilibration the system was heated from
50 K to 300 K during 10 ps under constant volume conditions and
harmonic restraints. Subsequently, the restraints were once again
relaxed, and finally an unrestrained 5-ps equilibration was carried
out. After this procedure the system was switched to constant
temperature and pressure and simulated for 5 ns.
Simulation B
The procedure described for simulation A was also used for
simulation B. As a starting point for the
Hoe/d(CGCGAATTCGCG)2 complex (Hoe represents
Hoechst33258) the x-ray structure with the NDB-code gdl012 (Quintana et
al., 1991) was used.
Simulation C
For the simulation of the unbound DNA (used as reference
simulation) a similar protocol as described for simulations A and B was
used that is described elsewhere (W. Flader, B. Wellenzohn, R. H. Winger, A. Hallbrucker, E. Mayer, and K. R. Liedl, submitted). All
simulations produced the B-form of DNA. This is consistent with recent
infrared spectroscopic studies of the Drew Dickerson dodecamer that
have shown that it persists in the B-form even at low water activity
(Pichler et al., 2000a,b).
| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
SUMMARY AND CONCLUSION
REFERENCES
|
|---|
The energy of the systems was stable during the simulations and
the root mean square values with respect to the starting structures were in the range of ~2-3 Å and showed no drift. An analysis of the
ligands during the simulation indicates that the tails of the ligands
undergo structural transitions. Structural transitions found in the
Netropsin complex were recently published (Wellenzohn et al., 2000b),
so we concentrate on the changes in the Hoechst33258 ligand. The
torsion angles 1, 2, and 3 of the Hoechst33258 molecule are
shown in Fig. 3, pointing out transitions
only in 3. Thus, the piperazine ring rotates while the rest of the
molecule stays in the starting x-ray (Quintana et al., 1991)
conformation during the simulation.
|
The piperazine end of the molecule extends partially into the GC region of the DNA. The minor groove of this GC region is wider than that in the A-tract, which may be the explanation of the enhanced flexibility of this side of the Hoechst33258 ligand. Fig. 3 indicates that three distinct substates occur, and representative snapshots are shown in Fig. 4. The detailed analysis of such snapshots leads to the conclusion that a nitrogen inversion is the origin of the structural transitions. The nitrogen transition is responsible for only two substates, but as seen above three different substates occur. The third substate arises from an additional torsion that occurs only in one of the two nitrogen inversion states.
|
The atomic type for the inverting nitrogen was chosen to be N3 (see
Figs. 1 and 4), which represents an
sp3-hybridized nitrogen (MD et al., 1995;
Cheatham et al., 1999). This is in contrast to the types normally used
for a nitrogen bound to an aromatic system
(sp2-hybridization) taking into account ab initio
calculations (Sponar et al., 1996). The calculations of Sponar
suggest the nonplanarity of amino groups that are bound to aromatic
systems such as in aniline or in the nucleic acid bases. The nonbonded
interaction energy between the piperazine and the DNA does not alter
significantly during the structural changes, suggesting entropy as the
driving force for the transitions.
An analysis of the Hoechst33258 ligands in the x-ray structures of DNA
ligand complexes indicates that such nitrogen inversions as above
described are also experimentally observed. Fig.
5 shows the piperazinium part of three
different experimental structures taken from the PDB, and at both
piperazine nitrogens inversions occur. The nitrogen inversions torsions
about 3 are also found in the different experimental structures.
Thus, our simulations indicate that the piperazinium tail of
Hoechst33258 must be considered as a highly flexible part, explaining
the structural variability in the crystallographic structures.
|
The contribution to sequence specificity of such piperazine tail
interactions with the minor groove of DNA have been reported recently
(Becker and Norden, 1999, 2000; Ren et al., 1999), which underlines the
importance of this exact structural and dynamical understanding.
The consequence of minor groove binding on the groove width is shown in Fig. 6, and it indicates that the two ligands affect the minor groove in different ways. In the case of Netropsin the complexation leads to a reduction of the minor groove width with the exception of the distances P21-P8 (number 4) and P20-P9 (number 5). In this region Netropsin is bound to the DNA with the two pyrrole parts preventing a too small minor groove by steric hindrance. The minor groove widening of this part of the DNA with respect to the unbound case is not significant, indicating that the pyrroles fit very well in the groove. In all other cases, complexation with Netropsin reduces the groove width. This effect is extended over the whole DNA and not restricted to the binding region, having also implications for additional binding sites.
|
The Hoechst33258 ligand exerts a different influence on the
groove width. In contrast to the Netropsin case, the binding of the
ligand induces a widening in the minor groove. This widening is
introduced only on this end of the DNA on which the piperazine part is
bound. As described above the piperazine undergoes structural changes
during the simulation, and a comparison of these changes with the time
dependence of the groove width shows (Fig.
7) that the structural changes also
affect the groove width. The correlation coefficient calculated between
the structural changes indicated by 3 and the groove width of
P19-P10 (number 6) is ~0.5. The structural transitions after 500 ps
lead to a widening of the minor groove of ~1.2 Å.
|
The correlation between the structural transitions in the ligand and the minor groove width could also be observed in the case of complexation with Netropsin (not shown). The interaction energy between the piperazine and the DNA does not alter significantly during the structural changes, suggesting entropy as the driving force for the transitions. Thus, a rigid tail of the ligand should lead to an entropic penalty due to stiffening of the DNA by complexation, or if the DNA keeps its pliability the direct interactions should be weakened.
The BI/BII conformational
substate behavior of the three simulations also differs in several
points, which underlines the ability of DNA to react on distortions
such as binding of ligands. In the uncomplexed DNA (Fig.
8), the A-tract (numbers 6-9 and 18-21
in Figs. 8 and
9) in
which the ligands bind exhibits almost no BII in
agreement with known results (Winger et al., 1998). The simulation also
indicates that two successive base pairs are never in the
BII substate at the same time.
|
|
|
In contrast to the unbound state, in the case of the DNA-ligand
complexes, the A-tracts contain phosphates in the
BII substate (Fig. 9). This is an unusual
behavior of such base pair steps and therefore assigned to the binding
of the ligand. Thus, we conclude that minor groove binding ligands are
able to influence the
BI/BII substate pattern.
The differences in the
BI/BII behavior between
Netropsin and Hoechst33258 show that the ligands differ in the way they
influence the backbone conformations. In a recent simulation
(Wellenzohn et al., 2001) it was shown that the binding of two
polyamides bound to the same minor groove position freeze out the
DNA-backbone flexibility. In contrast to this, the binding of Netropsin
and Hoechst33258 rather leads to an enhanced
BI/BII substate transition
flexibility. The freezing-in of the phosphates was explained as a
result of an optimization of the nonbonded contacts, which are
the main contributor to the binding of minor groove ligands. Together
with our new results we believe that the freezing-in in the
polyamide-DNA complexes is due to steric hindrance. The two ligands
Netropsin and Hoechst33258 are bound as monomers and are thus
sterically much less demanding than complexation with two polyamides
and therefore do not reduce the backbone conformational flexibility.
As described above in uncomplexed DNA no successive base pairs are at
the same time in BII. In the complexes this
condition is fulfilled only with some exceptions. In the Netropsin
simulation the successive phosphates 4 and 5 are in the
BII state over a long period (Fig. 9,
top) in the simulation. A detailed analysis of both angles indicated that P5 is in BII during the
whole simulation and that P4 is neither in a stable
BII nor in BI. The mean
value of this angle is at 220°, which is between the angles
of the two substates (Fig. 10, top). For
comparison the angle of P16 (shown in Fig. 10, bottom)
is approximately half the time in BI and half the
time in BII. Thus, its mean value is about the
same (220°) as that of P4 (Fig. 10, top). Recently made
x-ray studies also found DNA-phosphates that do not belong to either
the BI or BII
conformational group (Schuerman and Van Meervelt, 2000).
It is worth pointing out that the
BI/BII substate pattern
reported as a function of time in our first study of uncomplexed Drew
Dickerson dodecamer over 3 ns shows about the same behavior as shown
here in Fig. 8. The most pronounced differences being the
BII substate population at P11, although the
force field was changed from the Cornell et al. (1995) force field to
the improved one of Cheatham et al. (1999). The changes in the
BI/BII substate behavior on
complexation with Netropsin or Hoechst33258 shown in Figs. 8 and 9 for
the same force field are much more pronounced. This is strong support
that these changes are caused by interaction with the ligand and not by
the force field applied in the simulations. All together, complexation
influences these BI and BII
backbone conformational substates, and thus these substates may be able to contribute to sequence-specific binding of a protein
| |
SUMMARY AND CONCLUSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
SUMMARY AND CONCLUSION
REFERENCES
|
|---|
We performed three simulations of the Drew Dickerson dodecamer alone and complexed with Hoechst33258 and Netropsin. The tails of the ligands undergo a variety of structural transitions during the simulation in agreement with the structural variability of these tails in the crystallographic studies. The conformational changes of the tails are correlated with the time dependence of the groove width. Thus, the results indicate that the minor groove exhibits a great flexibility, fitting the ligand exactly in the groove. A more rigid ligand tail would lead either to entropic cost due to stiffening of the DNA by complexation or, if the DNA keeps it pliability, weakening of the direct contacts. Furthermore, the binding of the ligand influences the BI/BII conformational substate behavior, having possible implications for protein recognition processes. All these results are of importance for the understanding of the binding process valuable in the design of new sequence-specific minor groove binding ligands.
| |
ACKNOWLEDGMENTS |
|---|
This work was supported by a grant of the Austrian Science Fund (grant P13845-TPH).
| |
FOOTNOTES |
|---|
Received for publication 20 February 2001 and in final form 24 May 2001.
Address reprint requests to Dr. Klaus R. Liedl, Department of Theoretical Chemistry, Institute of General, Inorganic, and Theoretical Chemistry, University of Innsbruck, Innrain 52a, A-6020 Innsbruck, Austria. Tel.: 43-512-507-5164; Fax: 43-512-507-5144; E-mail: klaus.leidl{at}uibk.ac.at.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
SUMMARY AND CONCLUSION
REFERENCES
|
|---|
BII conformer substate dynamics is coupled with water migration.
J. Phys. Chem. B.
102:8934-8940
Biophys J, September 2001, p. 1588-1599, Vol. 81, No. 3
© 2001 by the Biophysical Society 0006-3495/01/09/1588/12 $2.00
This article has been cited by other articles:
 |
|
 |
|
  J. Dolenc, R. Baron, C. Oostenbrink, J. Koller, and W. F. van Gunsteren Configurational Entropy Change of Netropsin and Distamycin upon DNA Minor-Groove Binding Biophys. J., August 15, 2006; 91(4): 1460 - 1470. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Dolenc, C. Oostenbrink, J. Koller, and W. F. van Gunsteren Molecular dynamics simulations and free energy calculations of netropsin and distamycin binding to an AAAAA DNA binding site Nucleic Acids Res., February 1, 2005; 33(2): 725 - 733. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. L. Olsen, E. A. Louie, G. P. Drobny, and S. Th. Sigurdsson Determination of DNA minor groove width in distamycin-DNA complexes by solid-state NMR Nucleic Acids Res., September 1, 2003; 31(17): 5084 - 5089. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Wellenzohn, W. Flader, R. H. Winger, A. Hallbrucker, E. Mayer, and K. R. Liedl Exocyclic groups in the minor groove influence the backbone conformation of DNA Nucleic Acids Res., December 15, 2001; 29(24): 5036 - 5043. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
