INTERMOLECULAR FORCES, RECOGNITION, AND DYNAMICS
Donald
C. Rau, PhD, Head, Section on
Molecular Recognition and Assembly Nina
Sidorova, PhD, Staff Fellow Brian
Todd, PhD, Postdoctoral Fellow Shakir
Muradymov, Summer Intern |
|
Our
laboratory focuses on elucidating the coupling of the forces, structure, and
dynamics of biologically important macromolecules. The next challenge in
structural biology is to understand the physics of interactions between
molecules in aqueous solution. The ability to take advantage of the
increasing number of protein and nucleic acid structures determined by X-ray
crystallography and solution NMR will depend critically on knowledge of
structural biology so that we can understand the strength and specificity of
interactions among biologically important macromolecules that control
cellular function and then rationally design agents that can effectively
compete with those specific interactions associated with disease. Our earlier
results have shown that experimentally measured forces differ markedly from
those predicted by current, conventionally accepted theories of
intermolecular interactions. We have interpreted the observed forces as
indicating the dominant contribution of water-structuring energetics. Our
research program uses osmotic stress and X-ray scattering to measure directly
forces between biological macromolecules in macroscopic condensed arrays. To
investigate the role of water in the interaction of individual molecules, we
measure and correlate changes in binding energies as well as in the hydration
accompanying specific recognition reactions of biologically important
macromolecules in dilute solution, particularly of sequence-specific
DNA-protein complexes. 1.
Direct Force Measurements The
ability to measure directly forces between biopolymers in macroscopic
condensed arrays has greatly changed our understanding of how molecules interact
at close spacings, i.e., at 10 to 15Å separation. The universality of the
force characteristics observed for a wide variety of macromolecules,
including DNA, proteins, lipid bilayers, and carbohydrates, has led us to
conclude that the energy associated with structuring water between close
surfaces dominates intermolecular forces. We are currently focusing on
understanding the connection between hydration force magnitudes and the
chemical natures of the interacting surfaces. Exclusion
of solutes from macromolecular surfaces Rau Using
direct force measurement, we have further characterized the exclusion of
alcohols from DNA. Forces between DNA helices in macroscopic condensed arrays
can be measured by monitoring the distance between helices, using X-ray
diffraction as a function of the osmotic pressure of a polymer solution in
equilibrium with the DNA phase. The sensitivity of DNA forces to the
concentration of alcohols or other solutes permits extraction of the spatial
distribution of alcohols or other solutes around DNA. We have previously seen
that repulsive hydration forces underlie the preferential hydration of DNA in
the presence of alcohols. We have observed little difference in the spatial
dependence associated with the exclusion of methyl pentanediol (MPD) from
spermidine (Spd3+)-DNA, Co3+-DNA, or NaBr-DNA arrays.
The hydration of phosphate groups on the DNA backbone likely dominates
forces. Any additional contribution from electrostatics and a lowered
solution dielectric constant is minimal. To
analyze the chemical features of alcohols that determine exclusion, we are
examining the interaction of 17 alcohols with Spd3+-DNA arrays. To
a surprisingly good first-order approximation, the amplitudes of the observed
repulsive hydration forces simply scale with the number of alkyl carbons in
the alcohol. In contrast, carbons with bound hydroxyl groups are nearly
invisible to DNA. Methanol and ethylene glycol are very slightly excluded
from Spd3+-DNA arrays; glycerol is neither included nor excluded;
threitol and sorbitol are slightly included, as is the six-carbon polyol
sorbitol, more so than threitol with four hydroxylcarbons. Our current
hypothesis posits that these polyols can replace strings of water bound to
DNA with a subsequent gain in entropy. Hultgren A, Rau DC. Exclusion of alcohols from spermidine-DNA
assemblies: probing the physical basis of preferential hydration. Biochemistry
2004;43:8272-8280. Spermidine-DNA
assemblies Rau; in collaboration
with Yang We
have completed our investigation into the unusual spermidine-mediated
transition between two attractive states of DNA, a hexagonal phase seen at
lower spermidine (Spd) concentrations and a cholesteric phase with a larger
interhelical spacing that appears at higher Spd concentrations. The transition
has been widely interpreted as overcharging of DNA, i.e., the binding of more
Spd than is necessary to neutralize the DNA charge. However, careful
examination of the dependence of the critical Spd concentration at the
transition point on NaCl concentration shows that overcharging does not
occur. The data strongly suggest that, at higher Spd concentrations, the bulk
solution contains a mixture of Spd3+ and SpdCl2+ ions.
The data are best fit with a Cl– binding constant of about
100 mM. The structural transition is most likely attributable to a
replacement of trivalent Spd with a more weakly attractive or perhaps even
repulsive divalent Spd-Cl complex. Single-molecule
force measurements Todd, Rau; in
collaboration with Parsegian We
have finished constructing a magnetic tweezers apparatus for single-molecule
force measurements. Our initial experiments will focus on measuring the
interhelical attractive energies for DNA condensed with various divalent,
trivalent, and higher-valence cation ions, including naturally occurring
protamines used for tightly packaging sperm DNA in several organisms. The
goal is to connect the dependence of attractive energies on equilibrium
interhelical spacing to hydration forces. 2.
Hydration Changes Linked to Sequence-Specific DNA-Protein Recognition
Reactions Our
ultimate goal is to apply the lessons from direct force measurements to the
recognition reactions that control cellular processes. We have started by
measuring differences in water sequestered by complexes of four
sequence-specific DNA binding proteins with varying DNA sequences, with
particular emphasis on correlating the incorporated binding energy and water
as well as on the energy necessary to remove hydrating water from complexes. The
energetics of removing water from noncognate EcoRI-DNA complexes Sidorova, Rau We
have continued investigating the role of water in the sequence-specific
recognition reaction of the restriction nuclease EcoRI. The exceptionally
stringent specificity of restriction endonucleases in general and of the
EcoRI in particular is a paradigm for DNA-protein recognition reactions. We
measure the number of waters coupled to the DNA-protein binding reaction from
the dependence of the binding free energy on water chemical potential or, equivalently,
on osmotic pressure. We previously found that the nonspecific complex of the
EcoRI sequesters about 110 more waters than the complex with the specific
recognition sequence and that the water is likely located at the protein-DNA
interface. In contrast to this extremely strong sensitivity to the osmotic
stress (or water activity), the relative specific-nonspecific binding
constant has a particularly weak salt dependence and no pH dependence. We
have further shown that the dissociation rate of the specific complex is
highly sensitive to osmotic pressure requiring the net binding of 120 to 150
waters for the dissociation reaction. During
the past year, we completed our project on the dependence of the number of
waters sequestered on DNA sequence. Sequences that differ by even a single
base pair (“star” sequences) from the EcoRI-specific recognition
sequence (GAATTC) bind to protein only slightly better than completely
nonspecific sequences. We demonstrated that this abrupt decrease in binding
energy with even a single base-pair change is accompanied by an abrupt
increase in the water sequestered by the EcoRI-DNA complex. We developed a
new quantitative technique allowing us to measure reliably the dissociation
rate of the EcoRI from noncognate DNA sequences. At low stresses, we observed
that the osmotic sensitivities of both the competitive equilibrium constants
and dissociation rates of complexes with two “star” sequences
that differ by a single base pair out of six from the specific recognition
sequence and with nonspecific DNA are practically indistinguishable. At low
osmotic stresses, the complexes with noncognate sequences sequestered about
the same 110 waters as nonspecific complexes. At least some of the water,
however, can be removed from the noncognate complexes by applying
sufficiently high osmotic pressures. Given that equilibration times at high
stresses are too long for convenient measurement of binding constants, we
used dissociation rates to determine loss of sequestered water. At higher pressures,
the behavior of “star” sequence complexes differed strikingly
from that of the nonspecific complex. We observed a clear correlation between
the osmotic dependence of the dissociation rate and DNA sequence. The kinetic
measurements showed that the TAATTC “star” sequence complex
(strongest “star” site for the EcoRI) loses about 90 of its
initially sequestered waters at an osmotic pressure of 150 atmospheres. The
osmotic work required to remove these waters was about 3 Kcal/mole. Over the
same range of pressures, the CAATTC complex, the weaker “star”
site, lost only about 20 to 30 of its waters; we estimated that about 6
Kcal/mole would be necessary to remove substantially all the water from this
complex. There was no apparent loss of water from the nonspecific complex.
The amount of water we were able to remove from a noncognate complex
correlated with the ability of the enzyme to cleave the sequence. In
principle, any sequestered water can be removed by applying sufficiently high
osmotic stress, but the work necessary to dehydrate complexes naturally
depends on resulting unfavorable DNA-protein contacts. Sidorova NY, Rau DC. Differences between EcoRI nonspecific and
“star” sequence complexes revealed by osmotic stress. Biophys
J 2004;87:2564-2576. Differences
in sequestered water between specific and nonspecific complexes of BamHI:
comparison with X-ray structures Muradymov, Sidorova,
Rau We
previously observed that the osmotic sensitivity of the EcoRI
specific-nonspecific equilibrium binding constant was only very slightly
dependent on the nature of the solute used to set water activity. That
observation led us to conclude that the extra water retained by the
nonspecific complex is sequestered in a space sterically inaccessible to
solutes, most probably at the DNA-protein interface. As no structure of the
nonspecific complex of EcoRI is available, we could not directly confirm our
conclusion. To validate the connection between osmotic stress measurements
and structure, we are now investigating the difference in water release and
DNA binding strength for a second restriction endonuclease, Bam HI. The
significant advantage of this system is that X-ray structures are available
for both specific and nonspecific BamHI-DNA complexes. A comparison of the
two structures showed that the nonspecific complex loses the direct contacts
of the protein with DNA bases in the major groove that is characteristic of
the specific sequence complex. While extensive contacts still exist between
the sugar-phosphate backbone and the protein, a large cavity separates the
major groove of the DNA and the recognition protein surface. We estimated the
volume of the cavity as equivalent to about 150 waters. We performed a series
of kinetic and equilibrium competition experiments, initially with a
commercially available protein and subsequently with a highly purified sample
of the enzyme. For both protein samples, we observed very strong osmotic
dependence of the specific dissociation rate constant, with half-life time
changing from two minutes with no osmolyte present to about 40 minutes in the
presence of 0.8 osmolal betaine or methyl glucoside. Initial equilibrium
competition measurements indicated that the nonspecific BamHI-DNA complex
sequesters about 130 waters more than the specific complex of the enzyme,
correlating well with the structural data. 3.
Hydration Changes Associated with Specific Lambda Cro-DNA Binding We
are continuing to investigate the correlation between water and binding
strength of DNA complexes with lambda Cro repressor. This protein is
responsible for the induction of the lytic phase of lambda phage. The
recognition stringency for the repressor is more typical of sequence-specific
DNA-protein binding systems. Single base-pair changes from the optimal
recognition sequence reduce binding energies by only a few kT. We have
constructed a series of DNA fragments with consensus and altered Cro
recognition sequences that span a range of about 1,000 in binding constant.
Results for the osmotic pressure dependence of dissociation rates and of the
competitive equilibrium binding constants for the various complexes indicated
that the number of sequestered waters increases as the binding energy
decreases. For about every 10-fold decrease in the binding constant, the
complex incorporates an extra 10 to 15 water molecules. The insensitivity of the
number of extra water molecules sequestered by these complexes to solute
identity indicated that the waters are also likely at the DNA-protein
interface, mediating suboptimal contacts between the two surfaces. Sensitivity
of Thermus aquaticus MutS-DNA–specific binding to solution
conditions Sidorova The
mismatch repair system, which is highly conserved in both prokaryotes and
eukaryotes, plays a crucial role in maintaining genetic stability. Among a
large group of proteins required for mismatch repair, only MutS recognizes
and binds to heteroduplex DNA containing mispaired or unpaired bases. MutS
binding then triggers a cascade of events ultimately leading to repair of the
mismatched sequence. All members of MutS family possess ATPase activity, and
both mismatch recognition and the ATPase activity of MutS are critical for
the overall performance of the mismatch repair system. However, it is still
unclear how MutS recognizes a broad range of mismatches (one to four
consecutive unpaired bases or mispaired bases such as GT, AG, and so forth).
Broad substrate specificity is typical for the proteins of the MutS family.
The binding preferences and specificities of MutS proteins differ markedly
from the sequence-specific DNA-binding proteins such as restriction endonucleases
or lambda Cro repressor that we have previously studied. The
structure of Taq MutS in complex with DNA was recently solved. Both
biochemical studies and the crystal structure indicated that the MutS dimer
is responsible for specific mismatch recognition. Dimerization of the
protein, which buries about 2,400 Å2 of molecular surface,
extensive interactions between kinked heteroduplex DNA, and four MutS domains
plus the presence of an extra channel, whose function in the MutS-DNA complex
structure is still unknown, suggest that MutS-DNA binding could be
osmotically sensitive. We are now investigating the effect of solution
thermodynamic parameters (salt activity, water activity, and pH) for mismatch
recognition by MutS. We
employed gel mobility shift assays to study DNA binding of Taq MutS
protein to two DNA heteroduplexes containing one unpaired base each. Binding
curves of MutS to a 60 bp DNA heteroduplex containing unpaired C have a
distinct sigmoidal shape, typical of positively cooperative binding.
Scatchard plots confirmed that MutS binding to this DNA substrate is highly
cooperative. We suggest that observed positive cooperativity results from
monomer-dimer equilibrium coupled with binding of the dimeric form of Taq MutS
to DNA. A theoretical model that accounts for both processes allows us to
quantitate the apparent binding constant and estimate its sensitivity to
solution conditions. In parallel with gel mobility shift experiments, we are
also employing fluorescence anisotropy to measure MutS-DNA binding.
Dissociation kinetic experiments performed with no osmolyte and in the
presence of one osmolal betaine glycine or triethylene glycol showed that the
stability of the MutS-DNA specific complex increases significantly in the
presence of solutes with the half-life time increasing about 9-fold between
zero and one osmolal betaine or triethylene glycol. COLLABORATORS Per Lyngs Hansen, PhD, Peggy Hsieh, PhD, Genetics and Biochemistry
Branch, NIDDK, Sergey Leikin, PhD, Section on Molecular
Forces and Assembly, OD, NICHD, Shimon Mizrahi, PhD, Technion-Israel
Institute of Technology, Galina Obmolova, PhD, Genetics and
Biochemistry Branch, NIDDK, V. Adrian Parsegian, PhD, Laboratory of
Physical and Structural Biology, NICHD, Rudi Podgornik, PhD, Laboratory of Physical
and Structural Biology, NICHD, Jie Yang, PhD,
|