Abstracts: Tenth Conversation
Note that web at present does not support
special characters such as greek etc. and as such, be careful in reading
certain abstracts in which the greek is replaced with english alphabet.
For example, alpha = a, beta = b and so on.
The Use of Suicide Substrates
to Probe the Mechanism of E. coli Topoisomerase I
Camille J. Roche, Chang-Xi Zhu and Yuk-Ching Tse-Dinh
New York Medical College,
Valhalla, NY 10595
Fax 914-993-4058; phone 914-993-4061; Yuk-Ching_Tse-dinh@nymc.edu
E. coli topoisomerase I cleaves
single stranded DNA, passes another strand through the break and reseals
the bond to alter the linking number of DNA by one. Oligonucleotides that
are too short to be religated were used as suicide substrates to analyze
the binding characteristics and the kinetics of E. coli topoisomerase
I. Sequences designed to mimic longer stretches of DNA, used in earlier
studies1,2 were shown to be cleaved with some selectivity. Substrates 8
and 9 bases long were used to determine the minimum requirement on the 3'
and 5' sides of the cleavage site for efficient binding and cutting by the
enzyme. It was found that the E. coli enzyme requires a minimum of
4 bases 3' to the cleavage site to effectively anchor the substrate. An
unfavorable interaction on the 5' side of the cleavage site will necessitate
increasing that requirement to 5 or more bases. In addition, the kinetic
mechanism of the enzyme, in the presence of an oligonucleotide substrate,
was determined to consist of two sequential first order processes. The relative
rates of these two steps were used to demonstrate sequence selectivity by
the enzyme. The study also included the effects of magnesium on the cleavage
rate. Two of the substrates were used to photoaffinity label the enzyme
to help identify interactions at the active site. These covalent adducts
were visualized on an SDS Page gel, and isolated by hplc in preparation
for sequencing.
References and Footnotes
1. Tse-Dinh, Y-C., McCarron, B. G. H., Arentzen, R., &
Chowdhry, V., Nucleic Acids Res. 11, 8691-8701 (1983).
2. Kirkegaard, K. & Wang, J. C., J. Mol. Biol., 185, 625-637
(1982).
Architectural Role of p53
in the Spatial Organization of the DNA Response Element. Computer Modeling
and Comparison with Experiment
V. Zhurkin (1), A. Gorin (2), E. Appella (3), R. Harrington
(4), R. Jernigan (1) and S. Durell (1)
(1) Laboratory of Experimental and Computational
Biology,
National Cancer Institute, NIH,
Bg.12B, Rm. B116, 12 South Drive,
Bethesda, MD 20892-5677
(2) Cellular Biochemistry and Biophysics Program,
Memorial Sloan-Kettering Cancer Center,
New York, NY 10021
(3) Laboratory of Cell Biology,
National Cancer Institute, NIH,
Bethesda, MD 20892
(4) Department of Biochemistry,
School of Medicine,
University of Nevada,
Reno, NV 89557
p53, the ubiquitous tumor suppressor protein, cooperatively
binds to DNA as a tetramer [1]. The solution data indicate that binding
of the p53 tetramer to the cognate response element causes the DNA to bend
by a significant amount, ~50-60° [1, 2]. This is in contrast to the
straight conformation of DNA observed in the complex with three p53 DNA
binding domains (DBD) determined by X-ray crystallography [3]. However,
considering that only one p53 DBD in the co-crystal appears to be bound
in a native manner [3], it is likely that the conformation of the DNA does
not accurately reflect the tetramer-bound state in solution. To investigate
the conformational restraints imposed by the protein tetramer, we have constructed
stereochemically feasible molecular models of the (p53 DBD)--(Waf1 DNA)
complex based on the specific p53-DNA contacts observed in the crystal [3],
high resolution footprinting of the peptide-DNA interactions in solution
[2] and sequence-dependent flexibility of DNA.
The p53 binding sites contain four pentamers in the following
orientation: (RRRCW-WGYYY)--(RRRCW-WGYYY), where R is purine, Y is pyrimidine
and W is A or T. The spacer between the two decamers in parentheses is variable,
but in the functionally important case of the Waf1 site it has zero length.
The rules of DNA mechanics predict that this 20-mer is only moderately intrinsically
bent (curved) when isolated in solution. Our analysis shows that when four
p53 monomers bind the response element in its equilibrium or "straight"
B-DNA configuration, numerous steric clashes occur among the different p53
subunits. Importantly, one set of clashes involves highly conserved amino
acid residues which are commonly mutated in tumors (i.e., at the Zn-binding
helix H1 and loop L3 [3]). Another set of clashes involves residues of beta-sheet
domains which are conserved among different animal species and human p53.
We propose that to relieve these clashes, DNA has to be bent further and/or
unwound in the two consensus tetramers C(A/T)(T/A)G separated by 10 bp.
This bending is directed toward the major groove, in agreement with the
intrinsic conformational anisotropy of the CA:TG and AG:CT dimers established
by energy calculations and numerous X-ray data. This can explain the conservation
of the two AT pairs in the central part of the decamers, which do not directly
interact with the p53 protein in the co-crystal structure [3].
Based on systematic energy calculations of the p53-p53
and p53-DNA interactions, we found a close coupling between the DNA bending
and twisting: to relieve the clashes between p53 subunits DNA has to be
either bent toward the major groove, or untwisted. Changing both parameters
simultaneously leads to smaller distortions in the DNA structure. Thus,
the actual changes in conformation would appear to depend on the intrinsic
conformational propensities of the specific DNA sequences. In our "basic"
model for the Waf1 response element (without substantial untwisting), the
DNA is bent by ~50° (see Figure 1), which is consistent with the experimental
measurement of the 50-60° bend [1, 2]. Recently observed binding of
p53 to the four-way junctions [4] is also in accord with the notion of significant
DNA distortion (bending and twisting) proposed here. This overall lateral
arrangement of the four p53 subunits with respect to DNA comprises a novel
type of nucleoprotein assembly that has not been reported previously in
other complexes. The directionality of DNA bending predicted by our model
is currently being tested in new phasing and cyclization experiments.
Finally, the stereochemical analysis of the p53-DNA nucleo-protein
superstructure, both theoretical and experimental, is important for elucidating
the general architecture of p53 binding to DNA packed in chromatin, and
in principle opens perspectives for understanding the molecular mechanisms
of subsequent trans-activation of the DNA repair apparatus.
Figure 1: Putative side-by-side
arrangement of four p53 DNA binding domains on the Waf1 response element.
DNA bending relieves the steric hindrances between the a-helices H1 (cylinders
on the left) and leads to formation of the salt bridges Glu 180 -- Arg 181
(not shown). The recognition helices H2 interacting with the DNA major groove
are on the right. Notice the spatial complementarity between the two p53
sub-domains projecting toward the viewer (in the center); in the case of
straight DNA they are overlapping. In this model, the N-terminal ends of
the p53 subunits, which attach to the trans-activation domains of the intact
p53, are exposed on the outside of the DNA loop and hence are fully accessible.
The C-terminal ends (linked to the oligomerization p53 domains) are located
at the inside of the bend, thus facilitating tetramerization of the wild
type protein.
References and Footnotes
1. Balagurumoorthy, P., Sakamoto, H, Lewis, M., Zambrano,
N., Clore, G.M., Gronenborn, A.M., Appella, E. & Harrington, R.E.
Proc. Natl. Acad. Sci. USA 92, 8591-8595 (1995).
2. Nagaich, A.K., Balagurumoorthy, P., Sakamoto, H, Zhurkin, V.B., Clore,
G.M., Gronenborn, A.M., Appella, E. & Harrington, R.E. Abstract in this
volume.
3. Cho, Y.J., Gorina, S. & Pavletich, N. Science 265, 346-355
(1994).
4. Lee, S., Cavallo, L. & Griffith, J. J. Biol. Chem. 272,
in press (1997).
Protein-DNA Recognition:
A Role for CH...O Interactions in the Major
Groove
Y. Mandel-Gutfreund (1), H. Margalit (1), R.L. Jernigan
(2) and V.B. Zhurkin (2)
(1) Department of Molecular Genetics and Biotechnology,
The Hebrew University-Hadassah Medical School,
Jerusalem 91120, Israel
(2) Laboratory of Experimental and Computational Biology,
National Cancer Institute,
National Institutes of Health, Bg.12B, Rm. B116,
12 South Drive,
Bethesda, MD 20892-5677, USA
General stereochemical principles for specific protein-DNA
recognition, applicable to all protein-DNA complexes, have not yet been
determined. However, many details are now understood based on a wealth of
biochemical and structural information. To make a further step in elucidating
the stereochemical origins of protein-DNA recognition, we analyzed all direct
(side chain) - (base edge) interactions in the major groove, based on 43
crystallographically solved protein-DNA complexes. In addition to the "conventional"
hydrogen bonds, interactions with carbons were also included (exemplified
in Figure 1).
We found that close CH...O contacts involving the methyl
group of thymidine and position C5 of cytosine are rather frequent. Surprisingly,
the number of these CH...O interactions with a C--O distance less then 3.5Å
is comparable to the number of protein-DNA contacts involving nitrogen and
oxygen atoms as donors and acceptors. In the case of cytosine the C5--O
contacts can be interpreted either as primary (attractive), or as secondary,
caused indirectly by the attraction of the negatively charged oxygens to
the amino group N4. The latter explanation, however, does not hold for thymidine
with its electro-negative carbonyl group O4 in the equivalent position,
where the protein carbonyls still interact abundantly with the methyl group.
Thus, the interactions between protein oxygens and the C5(Cyt)/C5-Met(Thy)
positions per se seem to be attractive.
Moreover, analysis of the distributions of the (C5)H--O
distances (C5--O and H--O) and the angles C5-H-O and H-O-C' (where C' is
a protein atom) indicates that these CH--O interactions are remarkably similar
to the weak hydrogen bonds CH...O observed in the high resolution crystals
of small molecules [1]. This similarity with hydrogen bonds is understandable
due to the relatively high polarizability of aromatic carbons leading to
modest positive charges on the protons, especially in the case of C5 of
cytosine.
Based on this analysis we suggest that C5(Cyt) and C5-Met(Thy)
form weak CH...O hydrogen bonds with Asp, Asn, Glu, Gln, Ser and Thr, that
contribute to the specificity of recognition. For example, considering attractive
interactions between C5-Met(Thy) and side chain oxygens of Ser and Thr can
explain the experimentally observed preference of these amino acids for
thymidine [2], that otherwise remains unaccounted for. Including these interactions
in addition to the classical protein-DNA hydrogen bonds, first put forward
by Seeman, Rosenberg and Rich [3], enables extraction of simple structural
guidelines for amino acid-base recognition consistent with electrostatic
considerations.
Figure 1: Stereo view of the
interaction between Gln29 and 5'-TG-3' dimer in the 434 Cro/OR1 complex
[4]. Broken lines are hydrogen bonds: NH2(Gln)...O6(Gua) suggested in the
original paper [4], and CH3(Thy)...O(Gln) proposed here; the N--O and C--O
distances are 3.00 and 2.94Å, respectively. The heavy atom coordinates
are from the Protein Data Bank (entry 3cro); the protons are built by us
with ideal sp3 and sp2 geometries for CH3 and NH2 groups. Notice that the
TG dimer is rolled toward that major groove. We propose that the glutamine
bridge between Thy and Gua, including NH...O and CH...O hydrogen bonds,
may be one of the factors stabilizing the DNA bend.
References and Footnotes
1. Taylor, R. and Kennard, O. J. Am. Chem. Soc.
104, 5063-5070 (1982)
2. Choo, Y. and Klug, A. Proc. Nat. Acad. Sci. USA 91, 11168-11172
(1994)
3. Seeman, N.C., Rosenberg, J.M. and Rich, A. Proc. Natl. Acad. Sci.
USA 73, 804-808 (1976)
4. Mondragon, A. and Harrison, S.C. J. Mol. Biol. 219, 321-334
(1991)
Hydration, Electrostatics
and Dynamics of the HIV Tar-Tar* RNA Kissing Loops
Chang-Shung Tung and Angel E. García
Theoretical Biology and Biophysics Group,
T10, MS K710, Los Alamos National Laboratory,
Los Alamos, New Mexico 87545, USA
The structural hydration, ionic environment and structural
fluctuations of the HIV TAR-TAR* RNA kissing loops motif is studied by a
3.0 nanoseconds (ns) molecular dynamics (MD) simulation in aqueous solution
and with excess salt. In this RNA motif the complementary loop regions of
two hairpins are base-paired to form a helix. Each RNA hairpin has 16 bases;
five base pairs in the stem and six bases in the loop. The secondary structure
for this motif is shown in Fig. 1. The tertiary structure for the motif
has been modeled by C.S. Tung at al. and is shown in Fig. 2. The system
consists of 2920 water molecules, 60 Na+ ions and 30 Cl- ions (corresponding
to a 0.5 M NaCl excess salt and a 17 mM RNA complex concentration). The
total number of atoms in the system is 9878. The system is simulated with
periodic boundary conditions, using the particle mesh Ewald (PME) algorithm
for long ranged electrostatics and at constant temperature (300 K) and pressure
(1 atm).
Figure 1: Primary and secondary structure of the HIV
TAR-TAR*
Figure 2: Tertiary structure of the HIV TAR-TAR* RNA kissing loops.
Structural fluctuations sampled by the system during the
MD simulation are studied by means of the Molecule Optimal Dynamic Coordinate
(MODC) method described by García (García, A.E., Phys.
Rev. Lett. 68:2696-2699 (1992)). This analysis shows that the
system samples a hierarchy of energy minima (Fig. 3.). The largest root
mean square (rms) distance between minima is around 1.0 A. The mean time
between large structural charges is around 0.5 ns. Principal modes describing
these fluctuations show relative motions of the three helical regions while
conserving the secondary structure of the complex.
Figure 3: Projection of the 3 nanoseconds molecular
dynamics trajectory of the HIV TAR-TAR* RNA Kissing loops on a plane spanned
by two collective coordinates that best describe the system fluctuations.
In this trajectory the protein clearly samples five basins with inter-basin
transitions occurring in the 0.5 nanoseconds timescale. Arrows indicate
the region of this plane sampled along the trajectory at 0.5 ns time intervals.
The water and ionic distributions around the surface of
the system are studied by means of site-water and site-ions radial distribution
functions for all the RNA molecules atoms. The local density of water and
ions are also calculated on a grid around the RNA complex.. Of particular
interest is the localization of water by A11 C2-H, U22 C5-H and A28 C2-H
atoms. This water localization suggests a hydrogen bond between C-H atoms
in RNA bases and water in agreement with calculations in the tRNA (ASP)
anticodon hairpin loops by Auffinger and Westhof (Aufinger, P. and Westhof,
E., Biophys. J. 71:940-954 (1996)). We find that A11 and A28
C2 atoms are exposed to bulk solvent and water survival time in the C2 atom
hydration shell is short (ps). However, U22 does not form Watson-Crick base
pairings and is partially buried inside a cavity formed by other bases.
Water molecules solvating U22 C5 atom are long lived (0.5 ns). However,
these water molecules diffuse within this cavity without making any preferential
contacts. The solvation of U22 C5 appears to be opportunistic, and therefore
entropic, and not due any preferential hydrogen bonding interactions.
Aggregation of Guanine Dinucleotides:
An Analytical Ultracentrifugation Study
Consuelo Walss, Borries Demeler*, Jeffrey C. Hansen*
and Judith A. Walmsley
Division of Earth & Physical Sciences,
University of Texas at San Antonio,
San Antonio, TX 78249
*Department of Biochemistry,
University of Texas Health Science Center at San Antonio,
San Antonio, TX 78284.
Guanine nucleotides are known to self-associate and form
ordered aggregates via the stacking of guanine tetramers (G-tetrads). Alkali
metal cations, especially K+, promote the formation of extremely large species.
In the present study, the size and shape of guanine dinucleotide aggregates
was investigated by analytical ultracentrifugation. Aqueous solutions of
Nad(GpG) (sodium 2'-deoxyguanylyl-(3'-5')-2'-deoxyguanosine) and Et3NH(GpG)
(triethylammonium guanylyl-(3'-5')- guanosine were prepared in the presence
of KCl. Sedimentation velocity experiments were performed at 40,000 rpm
and 25 oC and demonstrated that at nucleotide concentrations of 1 to 8 mM
the aggregates formed were large enough to sediment in the presence of added
K+. Some measurements were also made at 5 oC. Aggregate formation is cooperative.
The sedimentation coefficient (s) distribution plots indicated that
there was considerable heterogeneity in the size of the aggregates, especially
at the lower nucleotide concentrations, and that an increase in the K+ concentration
greatly increased their size. Furthermore, a marked concentration dependence
of s was observed, which increased with increasing nucleotide and
salt concentrations, indicating that the aggregates formed are long and
rod-like in shape and that interaction between rods is likely. Various computer
models were used to estimate the molecular weights of the d(GpG) and GpG
aggregates; the most reliable estimates were obtained with the Sedvfin model
which takes intermolecular association into account. Values of 200,000-400,000
Da, depending on the nucleotide and K+ concentrations, were obtained for
solutions which exhibited the least non-ideality. The aggregates formed
by Nad(GpG) are much larger than those formed by Et3NH(GpG).
Figure: 2 mM Nad(GpG) + 4 mM
KCl in D2O, 25 oC. (A1) Sedimentation velocity data (A2) Van Holde-Weischet
plot
Are There Universal Statistical
Differences Between Coding and Noncoding DNA?
Ivo Grosse(1)*, Hanspeter Herzel(2), Sergey V. Buldyrev(1)
and H. Eugene Stanley(1)
(1)Center for Polymer Studies and the Department of Physics,
Boston University,
Boston, MA, 02215, USA
(2)Innovationskolleg Theoretical Biology,
Humboldt University,
Invalidenstr. 42,
D-10115 Berlin, Germany
*E-mail: ivo@bu.edu
The identification of genes in raw DNA sequences by statistical
means is important, as genome projects turn from mapping to large scale
sequencing. Many computer algorithms have been developed that carry out
this identification with a high level of accuracy. However, most of them
have the disadvantage that, in order to recognize unknown genes from a particular
organism, they need to be trained on sets of known genes from that organism
[1]. The scientific question we are asking is: "Are there statistical
features in coding and noncoding DNA that are species independent?"
Inspired by recent studies of long-range correlations in
DNA sequences, we analyze the Mutual Information Function Function (MIF)
of coding and noncoding DNA sequences [2] and find that the Average Mutual
Information (AMI) is signicantly higher in coding DNA than in noncoding
DNA sequences of all taxonomic classes [3]. Moreover, we find that the AMI
distributions for coding DNA as well as for noncoding DNA is quantitatively
the same throughout the entire phylogenetic tree. This universality implies
that the AMI may be used to identify genes in all species, without the requirement
of prior training on any DNA dataset.
References and Footnotes
[1] J. W. Fickett, Comput. Chem. 20 , 103{118 (1996).
[2] H. Herzel and I. Grosse, Physica A 216, 518{542 (1995); I. Grosse,
S. V. Buldyrev, H. Herzel, and H. E. Stanley, J. Biomol. Struct. &
Dynamics 12 (6) a078 (1995).
[3] I. Grosse, H. Herzel, S. V. Buldyrev, and H. E. Stanley, "Mutual
information of coding and noncoding DNA" (preprint).
Data Mining of Large Gene
Datasets Using the Mutual Information Function
Ivo Grosse(1)*, Kenneth A. Marx(2), Sergey V. Buldyrev(1),
Georges Grinstein(2), Hanspeter Herzel(3), Pat Hoffman(2), Anzhi Li(2),
Claudio Meneses(2) and H. Eugene Stanley(1)
(1) Center for Polymer Studies and the Department of Physics,
Boston University,
Boston, MA, 02215, USA
(2) Center for Intelligent Biomaterials and the Institute for Visualization
and Perception Research,
Departments of Chemistry and Computer Science,
University of Massachusetts,
Lowell, MA, 01854, USA
(3) Innovationskolleg Theoretical Biology,
Humboldt University,
Invalidenstr. 42,
D-10115 Berlin, Germany
*E-mail: ivo@bu.edu
The identification of unknown coding sequences in raw DNA
sequence is an important problem. There are public domain software methods
that have been developed to carry out this identification with a high level
of accuracy. However, these approaches have the disadvantage that they rely
upon training the software methods to recognize unknown genes from a particular
organism, using training sets of known genes from that organism. Recently,
a non-linear function called the Mutual Information Function (MIF) has been
used to construct a statistical measure, the Average Mutual Information
(AMI), which recognizes coding sequences irrespective of the origin from
which they originate [1].
The AMI has been found capable of of discriminating coding
from noncoding DNA with an accuracy as good as or better than the alternate
methods for classes of organisms spanning the vertebrates, invertebrates,
and plants. Moreover, the AMI has the advantage that it does not need to
be trained on representative coding sequences from any of these organisms.
Going beyond the simple identification of protein coding
regions, we are investigating whether the MIF and the AMI are capable of
revealing additional structural or functional information about the coding
sequences identified. As a starting point we are using coding and noncoding
sequence databases compiled by Fickett et al. [2]. We are developing integrated
data mining and visualization software to explore multi-dimensional datasets
derived from the MIF and AMI applied to the Fickett databases. We will report
on our initial attempts at uncovering patterns in the data using the Data
Mining Package Clementine.
We acknowledge funding from NIH, Pfizer Central Research
Corp., and DFG.
References and Footnotes
[1] Gro e, I., Herzel, H., Buldyrev, S. V. & Stanley,
H. E., \Mutual information of coding and noncoding DNA" (preprint).
[2] Fickett, J. W. & Tung C.-S. Nucl. Acids Res. 20, 6441{6450 (1992).
Asymmetric Cleavage of Nucleosomal
DNA by Calicheamicing1I
Prasad N. Kuduvalli, Craig A. Townsend and Thomas D.
Tullius
Department of Chemistry,
The Johns Hopkins University,
3400 N. Charles St.,
Baltimore, MD 21218
Calicheamicin g1I (CLM) cleavage of a uniquely positioned
nucleosome on the 5S rRNA gene of Xenopus borealis (this sequence
is denoted XB) showed a "hotspot" for cleavage at the end of a
polypyrimidine tract located 1 turn away from the dyad axis of symmetry
of the nucleosome. There was no cleavage at a site located 1 turn away from
the axis in the opposite direction, which seemed to indicate an asymmetric
cleavage pattern.1
We have conducted CLM cleavage studies on a series of nucleosome
positioning sequences and mapped CLM cleavage on the nucleosomes to examine
whether asymmetric cleavage was peculiar to the XB sequence or if it could
be applied to mononucleosomes in general. A 134 bp sequence of the yeast
DED1 promoter region is known to uniquely position a nucleosome in vitro.
This sequence (denoted HSAT) is of particular interest since it also contains
a polypyrimidine tract like the XB sequence2. The cleavage of nucleosomal
DNA for the HSAT sequence is very similar to that of the XB sequence. A
cleavage hotspot is observed at a position one helical turn away from the
dyad axis of symmetry in one direction only. This hotspot is also located
within the polypyrimidine tract.
A series of sequences, resulting from an experiment to
select out mouse genomic DNA that bound histone octamers3, were also subject
to cleavage studies. These sequences positioned the octamer in such a manner
as to yield two possible translational positions and hence two possible
dyad axes of symmetry which were mapped by hydroxyl radical footprinting4.
CLM cleavage of nucleosomal DNA once again yielded asymmetric cleavage patterns
in the dyad regions.
Current literature on chromatin structure focusses on the
asymmetric binding of the linker histone H1 to nucleosomal DNA about 65
bp away from the dyad in one direction, thus rendering the whole particle
asymmetric5,6. It has been suggested that the sequence within the nucleosome
is predisposed to favor the asymmetric binding of linker histone H1. Travers
et al. have also pointed out that the asymmetric cleavage we observed
with the XB sequence was probably reflective of asymmetry in the dyad region
of the nucleosome7. We are currently of the opinion that the drug is able
to recognize the asymmetric structure within the dyad region of the nucleosome,
leading to the asymmetric hotspot we observe in each case.
References and Footnotes
1. Kuduvalli, P. N., Townsend, C. A. & Tullius, T.
D., Biochemistry 34, 3899-3906 (1995).
2. Schieferstein, U. & Thoma, F., Biochemistry 35, 7705-7714
(1996).
3. Widlund, H., Kubista, M., Crothers, D. M., Submitted for Publication
(1997).
4. Hayes, J. J., Tullius, T. D. & Wolffe, A. P., Proc. Natl. Acad.
Sci. USA 87, 7405-7409 (1990).
5. Pruss, D., Bartholomew, B., Persinger, J., Hayes, J., Arents, G., Moudrianakis,
E. N. & Wolffe, A. P., Science 274, 614-617 (1996).
6. Hayes, J. J., Biochemistry 35, 11931-11937 (1996).
7. Travers, A. & Muyldermans, S. V., J. Mol. Biol. 257, 486-491
(1996).
Relationships between Crystal
Packing and Conformation in Crystals of DNA Oligonucleotides
Helen M. Berman, Samuel Engel, Michael Mehnert and John
Westbrook
Rutgers University,
Department of Chemistry,
Piscataway NJ 08855-0939
The DNA double helices contained in the Nucleic Acid Database
(NDB) were classified as to their crystal environments. All isomorphous
structures were given common identifiers; each crystal type was further
categorized as to its packing motif. Analyses of possible relationships
among packing motif, conformation, and sequence were performed and it was
found that certain steps are particularly sensitive to their environment.
The implications of these findings with respect to protein-DNA recognition
will be discussed.
The Nucleic Acid Database is funded by the NSF and the
DOE.
Effects of Histone Acetylation
on Nucleosome Structure
E. M. Bradbury(1,2), F. Dong(1), J. Gatewood(2), S.
Usachenko(1) and P. Yau(1)
(1)Dept. Biological Chemistry,
School of Medicine,
U.C. Davis,
Davis, CA 95616
(2)Life Sciences Division,
Los Alamos National Laboratory,
Los Alamos, NM 98545
Currently there is considerable interest in the nucleosome
structure/function relationships of reversible histone acelytations. Hyperacelytations
of histones have been strictly correlated with all aspects of DNA processing;
transcription, replication and also with spermiogenesis. Recently, families
of genes have been identified for both histone acetyltransferases and histone
deacetylases [1]. Of note are the findings that the histone acetyltransferases
so far identified are proteins already known to be associated with transcription
factors in the control of gene expression.
We have studied the effects histone hyperacetylation on
nucleosome structure and on DNA supercoiling. All states of acetylated histones
have been isolated and purified. Histone octamers [(H2A, H2B)2 (H32 H42)]
of known states of acetylations have been assembled with 195 bp and 207
bp DNA which contain a strong nucleosome positioning sequence. Oligonucleosomes
have been assembled on head to tail tandem repeats of the 207 bp, i.e.,
207-18 DNA (kindly provided by Dr. R. T. Simpson) and also on the circular
form of 207-18. Previously we have shown that DNA supercoiling was affected
by the full acetylation of the histone octamer. Whereas the DNA linking
number change, DLk, of the control nucleosome was -1.04±0.08 that
of the fully acetylated nuclesome was found to be -0.82±0.05 (2).
We have extended these studies to the effects of fully acetylated H3 and
H4 [DLk= -0.81±0.05] (3) and recently to fully acelyted H3 alone
[Dlk= -1.05±0.01] and fully acetylated H4 alone [DLk= -0.81±0.015](4).
It would appear that the acetylation induced change nucleosome in DLk is
caused largely by the acetylation of histone H4. Initial studies using atomic
force microscopy of the effects of histone acetylation on oligonucleosomes
assembled on linear 207-18 indicate that the length of DNA constrained by
the histone octamer is reduced in hyperacetylated nucleosomes. X-ray and
neutron scatter studies of 195 bp nucleosome particles assembled with control
and fully acetylated histone H4 support a model for the acetylated particle
in which DNA is released from the ends of particle.
To identify the effects of histone acetylation on histone:
DNA contacts in native nucleosomes we have used chemically induced zero
length histone: DNA crosslinking in intact nuclei. 146 bp nucleosome core
particles have been isolated from native hyperacetylated and hypoacetylated
chromatin domains which also contained different stoichiometeries of histone
H1 subtypes. Strong histone-DNA contacts observed for the hypoacetylated
core particles around sites 0, ±1, ±4, and ±5 of the
core particle DNA were not found in the hyperacetylated core particles.
These results suggest that in addition to the losening of the DNA ends of
the particle structural chanes take place throughout the nucleosome core
particle.
Acknowledgements
This work was supported by a grant from the U.S. Department
of Energy (DEFG03-88ER-60673) to E.M.B.
References and Footnotes
(1) Pennisi, E., Research News Opening the Way to Gene
Activity, Science 275, 155-157 and references (1997).
(2) Norton, V. G., Imai, B. S., Yau, P., Bradbury, E. M., Cell 57,
449-457 (1989)..
(3) Norton, V. G., Marvin, K. W., Yau, P., and Bradbury, E. M., J. Biol.
Chem. 265, 19848-19852 (1990).
(4) Cao, Y.-C., Yau, P., and Bradbury; E. M., in preparation.
Structure/Function Relationship
of the Insulin-Linked Polymorphic Region: Implications in the Length Polymorphism
and Gene Regulation
Paolo Catasti(1,2), X. Chen(1,2), E. Morton Bradbury(2,3)
and Goutam Gupta(1)
(1) Theoretical Biology and Biophysics,
T-10, MS-K710
(2) Life Science Division,
LS-8. MS-M888,
Los Alamos National Laboratory,
Los Alamos, NM 87545.
(3) Department of Biological Chemistry,
School of Medicine,
University of California at Davis,
Davis, CA 95616
The insulin-linked polymorphic region (ILPR), a 14 base-pairs-long
tandem repeat: 5'-(ACAG4TGTG4)n-3', is located 363 bases upstream of the
human insulin gene. A locus for insulin-dependent diabetes mellitus (IDDM),
an autoimmune disease, has been mapped to the ILPR. We show by 2D NMR analyses
that both strands of the ILPR form multiply folded single-stranded structures.
The G-rich strand of the ILPR adopts a telomere-like hairpin G-quartet structure
with melting temperature of 85oC (Catasti et al., 1997a) as proven by a
combination of 1D/2D NMR, CD, P1 digestion and gel electrophoresis (Catasti
et al., 1996), whereas the C-rich strand forms intercalated hairpin structures
with hemi-protonated parallel C+%C base pairs (i-motif) already at neutral
pH (Catasti et al., 1997b). 1D NMR studies show that the ILPR forms hairpin
structures even in presence of the complementary strand, with the equilibrium
shifted toward the Watson-Crick duplex for neutral solutions (pH>6) and
low temperatures.
References and Footnotes
1. Catasti, P., Chen, X., Moyzis, R.K., Bradbury, E.M.,
& Gupta, G., Structure-function correlations of the insulin-linked polymorphic
region, J. Mol. Biol. 264 (3), 534-545 (1996).
2. Catasti, P., Chen, X., Bradbury, E.M., & Gupta, G.,The insulin minisatellite
forms a hairpin G-quartet structure in solution, Biophys. J., in
press (1997a).
3. Catasti, P., Chen, X., Moyzis, R.K., Bradbury, E.M., & Gupta, G.,Cytosine-rich
strands of the insulin minisatellite adopt hairpins with intercalated cytosine+%cytosine
pairs, J. Mol. Biol., manuscript submitted (1997b).
4. Kennedy, G.C., German, M.S., & Rutter, W.J., The minisatellite in
the diabetes susceptibility locus IDDM2 regulates insulin transcription,
Nature Genetics, 9 (3), 293-298 (1995).
Theoretical and Experimental
Investigations of the Conformational Transitions of Complexes of DNA with
Ligands Having Several Types of Binding Sites of Different Forms of DNA
Armen T. Karapetian(1), Genady A. Terzikian, Ara P.
Antonian, Pogos O. Vardevanian(2) and Maxim D. Frank-Kamentskii(3)
(1) Yerevan Institute of Architecture and Construction,
375009 Yerevan, Armenia
(2) Yerevan State University,
375009 Yerevan, Armenia
(3) Center for Advanced Biotechnology,
Departement of Biochemical Engineering,
Boston University,
36 Cummington Street,
Boston, MA 02215, USA
The presentation will focus on two aspects of DNA-ligand
interation. The first will be the theoretical traitement of conformational
transitions of the DNA-ligand complexes allowing for the existence of different
binding parameters of the ligand (multimodal ligands) to different DNA conformations.
Obtained formula expresses the dependence of the experimentally estimated
values of the changes of transition point and transition width on the ligand
concentration. These expressions make possible to obtain the thermodynamic
parameters of conformational transitions (helix-coil, B-A, B-Z etc.) of
"naked" DNA, binding constants and other thermodynamic parameters
of the interaction of multimodal ligands with DNA by comparing theory and
experiment. The second portion of the talk will focus on the experimental
data obtaind for Ethidium and Actinomycin D and comparision of these data
with the theory.
The Topological Complementarity
of the Histone Octamer and DNA in the Nucleosome
Khrapunov S.N., Dragan A.I. and Sivolob A.I.
T.Shevchenko University of Kiev,
Vladimirskaya str, 64,
Kiev 252617, Ukraine
Fluorescence spectroscopy was used to investigate ionic
strength-dependent changes in the histone octamer complexed with high-molecular-weight
DNA. The changes in the position of the spectral maximum and anisotropy
of histone tyrosine fluorescence reflect structural transitions in the nucleosome
in the ranges of 0,5-3 and 20-30 mM NaCl. Under conditions of the interaction
of polynucleosomic fibrils (100-600 mM NaCl) more substantial conformational
changes in the histone octamer are observed. These changes are associated
with the complete destruction of specific contacts between the (H2A-H2B)
dimer and (H3-H4)2 tetramer. The results obtained indicate that several
states of the histone octamer can exist within chromatin, including states
differing substantially from the structure of the octamer in 2 M NaCl.
We have suggested the concept of topological complementarity
of histone octamer and DNA in the nucleosome. We determine by this definition
such mutual location of positive charged groups of histones and negative
charged groups of DNA, which promotes to their most effective interactions.
The basis of such complementarity between histone octamer and DNA in the
nucleosome is the histone octamer spatial structure determining the exact
distribution of cationic groups on the surface of the histone octamer.
Insignificant structural changes in the histone core of
nucleosome due to amino acids modifications, binding of nonhistone proteins
or other regulators as well as the conformational changes of DNA (for example
local B-A or B-Z transition) could disturb the topological complementarity
of interacting groups and induce a transition of given chromatin region
from one into another structural state.
One can see the consequences for gene activation from this
assumption.
Structural Characterization
of Modified Nucleic Acids
Samuel S. Woodley#, Linda Eckel*, Charles W. Bailey*,
Christopher Lawrence# and Thomas R. Krugh*
*Department of Chemistry
#Department of Biochemistry and Biophysics
University of Rochester
Rochester, NY 14627
The structures of DNA oligomers containing mutagens are
being characterized by NMR spectroscopic techniques. The mutagens include
guanine-C8 adducts with both N-2-(acetylamino)fluorene (AAF) adduct, and
N-2-(amino)-fluorene (AF), as well as the pyrimidine (6-4) pyrimidone photoproduct,
which is referred to as a TT(6-4) photoadduct. Two DNA oligomers containing
a TT(6-4) photoadduct were studied. In the first DNA oligomer, a guanine
is positioned opposite the 3'-modified thymine of the TT(6-4) moiety, while
an adenine occupies this position in the second oligomer.
The two TT(6-4) oligomers have similar structures. Unmodified
DNA bases away from the TT(6-4) lesion form normal base pairs, as expected.
However, the TT(6-4) lesion itself does not hydrogen bond with either adenine
or guanine. The TT(6-4) lesion rotates towards the minor grove of the DNA
double helix. The adenine and guanine bases opposite the lesion are stacked
within the helix, but do not hydrogen bond with the photoadduct. One G·C
base pair, located next to TT(6-4) on its 5'- side, is disrupted.
The aromatic amide N-2-acetylaminofluorene (AAF) has been
shown to induce -1 and -2 frameshift deletions. We have used one- and two-dimensional
NMR to characterize the structures of AAF-modified DNA oligomer duplexes
that serve as a model of frameshift deletion. The NMR data show the presence
of more than one conformation. The fluorene moiety is stacked with adjacent
base pairs while the AAF-G6 base is displaced into the groove.
Experimental and Computational
Strategies for NMR Structure Determination of Troublesome Proteins
F. Y. Luh(1), S. J. Archer(2), B. O. Smith(1), A. R.
C. Raine(1), W. Boucher(1), S. L. Brenner(2), P. J. Domaille(2) and E. D.
Laue(1)
(1) Department of Biochemistry,
University of Cambridge,
Tennis Court Rd, Cambridge
CB21QW, United Kingdom.
(2) DuPont Merck Pharmaceutical Company,
Wilmington, DE 19880-0328, USA.
We have been using NMR to study several proteins which
regulate progression through the eukaryotic cell cycle. Not surprisingly,
since many of these proteins up- or down-regulate function by interaction
with other proteins in large complexes they are susceptible to aggregation
and precipitation when isolated as single species. Sample concentrations
are frequently limited to < 1mM. We will describe our recent attempts
to simplify spectra both by uniform deuterium labeling (in addition to 13C,
15N) and selective protonation of specific residue types. A variety of 3D
and 4D heteronuclear NOE data have been used to determine the global fold
using these partial data. Additionally we will describe the implementation
of a restrained molecular dynamics, simulated annealing computational strategy
(Nilges; 1995, 1996) which retains the crosspeak ambiguity inherent in heavily
overlapped spectra to refine the structures.
tRNAAsp-Aspartyl-Trna
Synthetase: Two Base-Pairs Make the Difference
D. Moras
Laboratoire de Biologie Structurale
IGBMC
1 rue Laurent Fries - BP 163
67404 Illkirch Cedex - France
It is accepted that tRNA recognition by aminoacyl-tRNA
synthetases involves specific interactions between the protein and few bases
of the nucleic acid called the identity determinants. Other contacts have
been observed with the sugar-phosphate backbone but could not be related
to the specificity of the recognition process.
The crystal structure of the non productive heterologous
complex between yeast tRNAAsp and E. coli AspRS, together with those
of the two corresponding homologous complexes, provide the first experimental
evidence for an indirect readout of the first two base pairs of the acceptor
stem through phosphate backbone-protein contacts. The structural correlation
is confirmed by mutagenesis experiments at the tRNA level. A mechanism for
tRNA induced AspRS activation can thus be proposed which accounts for all
experimental observations and explains some otherwise unexplained sequence
conservations.
Modes of Motions and Mechanisms
of Regulation in the Regulatory Domains of pp60 c-Src
Linda Nicholson
Section of Biochemistry, Molecular and Cell Biology
Cornell University,
Ithaca, NY 14853
E-mail: lkn2@cornell.edu; Fax: (607) 255-2428; Phone: (607) 255-7208
http://www.bio.cornell.edu/biochem/nicholson/nicholson.html
******* abstract to follow very soon **********
Conformational Analysis
of a Fourteen Amino Acid Mussel Adhesive Protein Peptide
Marion P. Olivieri(1,2), Robert M. Wollman(3), Emily
A. Mysliwiec-Levandusky(1) and James L. Alderfer(3)
(1)D'Youville College,
320 Porter Avenue,
Buffalo, New York 14201 USA
(2)NSF Industry University Cooperative for Biosurfaces,
State University of New York at Buffalo,
Buffalo, NY 14214 USA.
(3)Roswell Park Cancer Institute,
Elm & Carlton Streets,
Buffalo, New York 14263 USA
Mussel Adhesive Protein (MAP) is derived from the Common
Blue Sea Mussel which produces MAP as a bioadhesive that is capable of attachment
to various underwater surfaces (1). In addition, when bound to a surface,
MAP is capable of non-specific cellular attachment (2). Therefore, this
protein serves as an excellent model to study the general aspects of bioadhesion.
Our prior studies characterized the surface properties of films created
using the intact MAP as well as the peptide (Ala-Lys-Pro-Ser-Tyr-Hydroxyproline-Hydroxyproline-Thr-L-DOPA-Lys)(2).
It has been reported that MAP is composed of primarily 70 to 80 of these
repeating decameric units (1). Continuing studies include the structural
characterization of several peptide fragments of MAP.
This current study examines a MAP fourteen amino acid sequence
(Pro-Ser-Tyr-Hydroxyproline-Hydroxyproline-Thr-Tyr-Lys-Ala-Lys-Pro-Ser-Tyr-Hydroxy=
proline {SynPep, Inc., Dublin, CA}) using nuclear magnetic resonance (NMR)
spectroscopy coupled with molecular modeling techniques (Sybyl, Tripos,
St. Louis, MO). NMR was performed on a Bruker AMX 600 MHz instrument. One
dimensional, 2D COSY, TOCSY, and ROESY experiments were acquired for a 3.6
mM sample in two solvents (D2O and 90% H2O/10% D2O) at 303 K.
Resonance assignments were made for the peptide. Two dimensional
nOe experiments were used to get inter-proton distance constraints. NMR
spectroscopy and molecular modeling techniques have elucidated a family
of possible conformations of this peptide unit. This information in conjunction
with earlier work (3,4) is used to illustrate interactions that may take
place between the peptide and a surface. Since adsorbed films of this protein
are used as cellular attachment agents, these models describe the portions
of the molecules that offer themselves to surface as well as cellular attachment.
A model of the peptide will be presented and compared a model of the decameric
repeat unit obtained using similar methods. This structural comparison is
a first-step approach toward the understanding of the general aspects related
to the phenomena of bioadhesion.
References and Footnotes
1. J. H. Waite, T. J. Housley and M. L. Tanzer, Biochemistry
24:5014 (1985).
2. M. P. Olivieri, R. E. Baier and R. E. Loomis, Biomaterials 13
No. 14 pp. 1000-1008 (1992).
3. M. P. Olivieri, R. M. Wollman, D. M. Dembik, E. A. Mysliwiec and J. L.
Alderfer, Abstracts from the Ninth Conversation Issue of the Journal
of Biomolecular Structure and Dynamics, 12, No. 6 Abstract 178 (1995).
4. M. P. Olivieri, R. M. Wollman, and J. L. Alderfer, NMR Spectroscopy of
Mussel Adhesive Protein Repeating Peptide Segment, Submitted to the Journal
of Peptide Research (1997).
This work was supported by a NSF MCB-951-3390, a NSF EEC-98416826
RUI grant and a NIH CA-16056 which supports in part the NMR Facility at
Roswell Park Cancer Institute.
Structure of the A-Site of
E. Coli 16s rRNA Complexed with an Aminoglycoside Antibiotic
Dominique Fourmy and Joseph D. Puglisi
Department of Chemistry and Biochemistry
and Center for Molecular Biology of RNA,
University of California,
Santa Cruz, CA 95064 USA
The A site of E. coli 16S ribosomal RNA is the site
of the tRNA anticodon stem loop-mRNA codon interaction. It is also the binding
site of aminoglycoside antibiotics, which affect protein synthesis by inducing
codon misreading and inhibiting translocation. Features of the translation
mechanism may be revealed by structure determination of an aminoglycoside-A-site
RNA complex. The solution structure of an oligonucleotide encompassing the
highly conserved A-site 16S RNA sequence (C1404-C1412 and G1488-G1497) of
E. coli free and complexed with paromomycin has been determined using
nuclear magnetic resonance spectroscopy. The RNA, free and in the complex
forms a continuous A form helix closed by non-canonical base pairs, with
a single bulged adenine. The antibiotic binds in the major groove of the
A-site within a pocket created by an A-A base pair and a the single bulged
adenine. Specific interactions are observed between the aminoglycoside chemical
groups important for antibiotic activity and conserved nucleotides in the
RNA. Binding of different neomycin-class aminoglycosides to the model oligonucleotide
has also been studied by NMR. The results explains binding of diverse aminoglycosides
to the ribosome, their specific activity against prokaryotic organisms and
various resistance mechanism. Comparison of the free and bound forms of
the RNA revealed that two universally conserved nucleotides are displaced
towards the minor groove and suggests a possible mechanism for translational
inhibition.
A Link Between Conformational
Change and Ribozyme Specificity in Group II Introns
Anna Marie Pyle
Department of Biochemistry and
Molecular Biophysics,
Columbia University,
630 W. 168th Street,
New York, NY 10032
Phone: 212-305-5430; Fax: 212-305-7932
Group II introns are large, structurally complex catalytic
RNA molecules that catalyze self-splicing, together with a variety of hydrolysis
and transesterification reactions. Group II introns are unusual in that
they contain few conserved nucleotides and lack Watson-Crick covariations
in base-pairing between functional domains of the molecule. Instead, folded
structures within the intron are stabilized by networks of unusual RNA tertiary
interactions, making the group II intron an important model system for understanding
RNA folding and conformational stability. An additional reason for intense
interest in group II introns is the recent finding that they are mobile
genetic elements, capable of targeting and reverse-splicing themselves into
DNA or RNA. In the ai5g intron from yeast mitochondria, the targets are
recognized by an extensive set of thirteen base-pairing interactions between
Domain 1 of the intron and sequences within the target strand. Thirteen
nucleotides of base-pairing is sufficient for recognition of a single site
on a particular mRNA within a human genome. Thus, if the ribozyme is found
to be highly specific for target site choice, group II introns may potentially
important agents for application in gene therapy.
Using a multi-piece ribozyme derivative of this intron,
it has been possible to explore the basis for group II intron target specificity
and to relate specificity to conformational states of the intron. A quantitative
estimate for substrate specificity is obtained by comparing the kcat/Km
values for cleavage of the correct target to the kcat/Km values for cleavage
of a mismatched target (the specificity index). To obtain estimates of this
index in the ai5g group II intron, the sequence of a substrate target RNA
was changed at six different positions and different types of mismatches
were incorporated at several places within the ribozyme-substrate recognition
helices. In contrast to other ribozymes studied thus far, large effects
on relative kcat/Km were observed, indicating that depending on sequence,
mismatches result in 100-10,000X effects on the ability of the ribozyme
to discriminate the proper target. Thus, group II intron ribozymes are highly
specific for recognition of specific sequences on RNA and DNA.
This finding was surprising in light of the fact that the
extensive base-pairing between substrate and ribozyme should result in a
Kd in the picomolar range. With such tight binding, the relative effect
of a single mismatch within the recognition helices would be expected to
be small. However, extensive studies of substrate/ribozyme association using
direct binding assays, fluorescence determinations of association and dissociation
rate constants, together with estimates of Km from activity assays have
all revealed that the ribozyme-substrate Kd is ~6 nM, approximately six
kcal/mol weaker than that expected from the extensive base-pairing interactions.
This is consistent with the fact that the ribozyme is highly specific for
correct target site choice and implies that there is an energetic folding
penalty for substrate binding. Fluorescence and chemical modification folding
studies on the intron reveal that the substrate binding induces conformational
changes within the intron, and these may exact an energetic cost. Considering
the specificity and energetic results together, it is intriguing to consider
that the intron may have evolved an energetic strategy for retaining extensive
molecular recognition capability (through numerous base-pairs) while reducing
total binding energy, thereby resulting in a highly specific ribozyme capable
of recognizing a very long substrate recognition site. The structural, thermodynamic
and conformational basis for the energetic penalty is now being investigated.
Crystal Structure of an
IHF-DNA Complex: A Protein-Induced DNA U-Turn
Phoebe A. Rice (1), Shu-wei Yang (2), Kiyoshi Mizuuchi
(1), and Howard Nash (2)
(1)Laboratory of Molecular Biology, NIDDK
(2)Laboratory of Molecular Biology, NIMH,
NIH, Bethesda MD 20892
Integration host factor (IHF) is a small heterodimeric
protein that specifically binds to DNA and functions as an architectural
factor in many cellular processes in prokaryotes. We have solved crystal
structure of IHF complexed with 35 base pairs of DNA.
The DNA executes a U-turn as it wraps around the protein.
If only the 35bp found in one asymmetric unit are considered, the bend angle
is ~160degrees. The DNA fragments in the crystal are packed end - to - end
to form a pseudo-continuous helix, and if contacts to these symmetry related
pieces of DNA are taken into account the overall bend may actually exceed
180 degrees. Most of the bending occurs at two large kinks, spaced 9 bp
apart, where a proline at the tip of each arm is intercalated between base
pairs.
The DNA lies largely in a single plane, making a dihedral
angle of only ~10-15 degrees, and although there are significant local variations
in the helical twist, the average over the entire structure is 33.3 degrees.
IHF contacts the DNA exclusively via the phosphodiester
backbone and the minor groove, and relies heavily on indirect readout to
recognize its binding sequence. One such readout involves a six base A tract
whose structure is very similar to that of A-tracts seen in crystal structures
of naked DNA dodecamers.
DNA Methylation, X Chromosome
Inactivation, and Epigenetic Mechanisms
Arthur D. Riggs, Zhenggang Xiong, Lijun Wang and Jeanne
M. LeBon
Biology Department,
Beckman Research Institute of the City of Hope,
Duarte, CA 91010 USA
In addition to a review of DNA methylation, X chromosome
inactivation, I will present our recent data on methylation dynamics, including
rates of methylation and demethylation at a particular CpG site. The implications
of these data for epigenetic heritability will be discussed, as will the
relevance of these data to the Fragile X syndrome, a triplet repeat disease
with unusual heredity. If time permits, I will also give a progress report
on our studies of chromatin differences between the active and inactive
X chromosome. Since many of those in attendance may not be familiar with
DNA methylation and epigenetics, a brief introduction to these subjects
follows below.
The DNA of vertebrates and plants contains 5 bases, not
four, and this fact should be taken into consideration for all studies of
DNA structure, chromatin structure, and gene function. Approximately 4%
of DNA cytosine in mammals is enzymatically converted, at the DNA level,
to 5-methylcytosine (m5C) by an enzyme called DNA (cytosine-5)-methyltransferase
(MTase). m5C is found mainly in CpG dinucleotides, and 60-80% of CpG sites
are symmetrically methylated, i.e., C is converted to m5C is in both strands.
Hemimethylated CpG sites, having m5C in only one strand, are not normally
detectable. It is now well established that the mammalian genome has cell-type-specific
patterns of methylated and unmethylated CpG sites, and the methylation pattern
is somatically heritable (1). Enzymatic DNA methylation significantly changes
the structural and chemical properties of DNA (2). In particular, the interaction
of proteins with DNA is affected, and thus, chromatin structure and gene
function is affected.
Epigenetic changes are defined as mitotically (or meiotically)
heritable changes in gene function that are not due to changes in base sequence.
Mammalian development requires stable, somatically heritable epigenetic
switches which are dependent on DNA methylation and/or chromatin structure.
As evidenced by differential gene function depending on whether a gene comes
from the mother or father (parental imprinting), some epigenetic modifications
are differentially engraved in male and female gametes, and these modifications
affect the next generation. Gene knock-out experiments have established
that DNA methylation is necessary for mammalian development, normal parental
imprinting, and X chromosome inactivation. X chromosome inactivation (which
results in the silencing of most genes on one of the two X chromosomes in
all female mammalian cells) displays numerous epigenetic features, including
parental imprinting. Some of the changes leading to metastatic cancer and
other diseases are epigenetic in nature and involve DNA methylation (3).
Of particular relevance to this symposium is the involvement of X inactivation
and DNA methylation in the manifestation of fragile X syndrome, one of the
best examples of triplet expansion diseases and one of the leading causes
of mental retardation.
References and Footnotes
1. Russo, E., Martienssen, R. and Riggs, A.D. eds., Epigenetics
Mechanisms of Gene Regulation, Cold Spring Harbor Press (1996).
2. Zacharias, W., Methylation of cytosine influences DNA structure. in DNA
Methylation: Molecular Biology and Biological Significance, eds. Jost, J.P.,
and Saluz, H.P. Birkhauser Verlag, Boston, pp 27-38 (1993).
3. Laird, P. W. and Jaenisch, R., The Role of DNA Methylation in Cancer
Genetics and Epigenetics, Annu.Rev.Genet., 30, 441-464 (1996).
Interaction of an Anti-HIV-1
Hammerhead Ribozyme With a 17-mer DNA Analog
Substrate of HIV-1 gag RNA: A 750 MHz NMR and Computer Experiments*
Sarma, Ramaswamy H., Dhingra, M. M. and Sarma, Mukti H.
Institute of Biomolecular Stereodynamics,
Chemistry, The University,
Albany NY 12222
email: rhs07@cnsvax.albany.edu; fax: 518-452-4955; phone: 518-456-9362
Turner, Christopher J.
National Magnet Lab,
MIT,
Cambridge, MA 02139
Setlik, R. F., Shibata, M., Rein, R., Kazim, A. L. and Cairo, A.
Department of Biophysics and Biopolymer Facility,
Roswell Park Cancer Institute,
Buffalo NY 14263
Ornstein, R. L.
Molecular Science Research Center,
Pacific Northwest Laboratory,
Richland, WA 99352
Hammerhead ribozymes have become potential therapeutic
agents against AIDS. A 30-mer anti-HIV-1 DNA/RNA chimeric hammerhead ribozyme
(1, 2) and its 17-merDNA substrate analog were synthesized. By following
the imino proton signals in water a 1:1 complex between the two were formed.
Structure of this anti-HIV-1 ribozyme-DNA abortive substrate complex was
investigated by PE COSY, NOESY and TOCSY at 750 MHz. Assignments of the
imino resonances of stems I and III were made from appropriate NOESY walk.
In addition several of the base protons, and H1' protons were assigned.
The derived NMR data, along with available crystallographic data were used
to arrive the structural model of the complex. Energy minimization was then
performed on the final model with the MAXIMIN module of SYBYL using AMBER
all-atom parameters of Kollman and coworkers. The model was then allowed
to relax for 200 steps of deepest descent, minimization followed by conjugate
gradient until a convergence criterion of a maximum energy change of 0.05
Kcal/mole was achieved. The final model had sugar puckers resemling the
A form of DNA, and we are proposing a catalytic mechanism for ribozyme action
based on the derived structure. It is believed that the present studies
will lead to efficient ribozyme engineering to combat HIV.
*Supported by NFCR, NASA, NSF and DOE
References and Footnotes
1. Sarma, R. H., Sarma, M. H., Rein, R., Shibata, M., Setlik,
R. S., Ornstein, R. L., Kazim, A. L., Cairo, A., and Tomasi, T. B., Secondary
Structure in Solution of Two Anti-HIV-1 Hammerhead Ribozymes as Investigated
by Two-Dimensional 1H 500 MHz NMR Spectroscopy in Water, FEBS LETTERS
357, 317-323 (1995).
2. Setlick, R. F., Shibata, M., Sarma, R. H., Sarma, M. H., Kazim, A. L.,
Ornstein, R. L., Tomasi, T. B., and Rein, R., Modeling of a Possible Conformational
Change Associated with the Catalytic Mechanism in the Hammerhead Ribozyme,
J. Biomolecular Structure and Dynamics 13, 515-522 (1995).
Common Structural Elements
in Eukaryotic Promoters
Tanja Schätz and Jörg Langowski
German Cancer Research Center (DKFZ),
INF 280 (0830),
69120 Heidelberg, Germany
The TATA binding protein (TBP) is a DNA-binding protein
obligatory required for transcription initiation in eukaryotes. In pol II
promoters it has been shown that TBP recognizes and binds a conserved promoter
element, the classical TATA box (TATAAA) and that the formation of this
complex is already sufficient to initiate basal transcription. During the
last years it could be shown that the classical TATA box is not the only
element recognized by TBP and promoting transcription. A variety of sequence
elements can functionally replace TATAAA, stimulating transcription with
equal or increased efficiency. Our work deals with the question whether
there are common structural features between non-classical and classical
TBP binding sites. One structural element, DNA curvature, is known to play
an important role in gene expression and has been found in various promoter
contexts. To calculate the distribution of curvature along an arbitrary
DNA sequence we developed the program Curvature. The program can be used
with any bending model that offers values of the parameters twist, roll
and tilt which describe the helix axis deflection at each dinucleotide step.
Calculating the curvature of 504 eukaryotic promoters predicted by the models
of Bolshoy et al. (1), Cacchione et al. (2), Calladine et al. (3) and Satchwell
et al. (4) we found in each case a correlation between TBP binding sites
and DNA curvature. Characterizing the TBP binding sites revealed that in
addition to the classical TATA box there are five more elements occurring
significantly often in the promoters, nearly all of them being one point
mutations of the classical TATA box element. The strong curvature that could
be detected for the binding sites is about 26-47% higher than the overall
curvature of the promoters. These results support the proposition that intrinsic
curvature is an important criterion for the recognition of different DNA
elements by TBP.
References and Footnotes
1. Bolshoy, A., McNamara, P., Harrington, R. E. and Trifonov,
E. N., Proc. Natl. Acad. Sci. USA, 88 (6), 2312-2316 (1991).
2. Cacchione, S., de Santis, P., Foti, D., Leoni, L., Palleschi, A., Risuleo,
G. and Savino, M., Biochemistry, 28, 8706-8713 (1989).
3. Calladine, C. R., Drew, H. R. and McCall, M. J., J. Mol. Biol., 201,
127-137 (1988).
4. Satchwell, S. C., Drew, H. R. and Travers, A. A., J. Mol. Biol., 191,
659-675 (1986).
A Global Overview of Protein
Folds
Dikeos Mario Soumpasis and Martin Christian Strahm
Biocomputation Group,
Max-Planck-Institute for Biophysical Chemistry,
37070 Goettingen, FRG
We have devised an exact, quantitative and superfast technique
to charac- terize the spatial folding of proteins (or any other polymers)
based on metric curvature (kappa) and torsion (tau) concepts. We have used
it to analyze all 4000 structures in the PDB and will discuss the picture
presently emerging. The kappa-tau approach opens entirely new ways to harness
the structural information explosion facing us and understand the geometric
nature of the folds. It can be used efficiently in evolutionary and taxonomic
studies of the structures, sequence-structure correlations and the structural
homoloy based de novo design of proteins.
An RNA Bridge in the l N
Antitermination Complex Requires Tryptophan-Purine Stacking and Base Flipping
Leila Su, James T. Radek, Patrick Hermanto, Klaas Hallenga,
Huifen Chen, Ging Chan, Laua A. Labeots and Michael A. Weiss
Departments of Biochemistry & Molecular Biology,
Chemistry and Center for Molecular Oncology,
The University of Chicago,
Chicago, IL 60637-5419
Binding of the phage l N arginine-rich motif to an RNA
enhancer element provides a model of RNA-mediated signaling in transcriptional
antitermination. The complex -- remarkable for costabilization of novel
peptide and RNA structures -- provides an "RNA bridge" between
N and RNA polymerase. Bridging RNA surfaces each contain an exposed purine
required for efficient antitermination. The adenine at the N-binding surface
interacts with tryptophan by direct stacking. The 19 possible amino-acid
substitutions each confer reduced RNA affinity in vitro in accord
with genetic analysis of N-mediated antitermination in vivo. RNA
mutagenesis further demonstrates that stacking of tryptophan is purine-specific
and coupled to induction of a specific internal RNA architecture and pattern
of base flipping. The interrelation of induced fit and peptide affinity
rationalizes genetic analysis of the l N protein and RNA nut site.
We propose that the indole ring of tryptophan functions as a "pseudo-base"
to extend an otherwise incomplete RNA motif. N-dependent antitermination
in phage l thus requires extended aromatic stacking and base flipping to
build an RNA bridge.
Evidence That the DNA A-Tract
is Bent and Rigid in Solution Obtained by Time-Resolved Fluorescence Resonance
Energy Transfer
Lukasz Milos (undergraduate
at the College of The University of Chicago),
Sherri Oslick and Michael A. Weiss
Departments of Biochemistry & Molecular Biology and Chemistry,
The University of Chicago,
Chicago IL 60637-5419.
Short runs of adenines (A-tracts) with proper phasing are
associated by anomalous electrophoretic mobility with curvature of the DNA
double helix. Two models have been proposed to explain the observed curvature;
the wedge model ascribes bending to axial deflections (wedges) between successive
AA.TT dinucleotides. The net bend is proposed to be smooth along the length
of the A-tract: the overall result of addition of the small wedges. The
second model -- supported by a series of crystallographic studies -- posits
that A-tracts are straight. This junction model proposes that the overall
curvature arises at the junction of the A-tract and its flanking sequences:
respective helical axes are oriented at different angles. In the present
study a 15-mer A-tract duplex fluorescently tagged with fluorescein and
rhodamine; F-ATCACAAAAAAGCGT : R-ACGCTTTTTTGTGAT, was examined using time-resolved
fluorescence resonance spectroscopy (FRET). Structural and dynamic information
about the distance between the fluorescent probes (i.e., end-to-end distances
in the heteropolymer) was recovered; data analysis assumes a gaussian distance
distribution between donor and acceptor. Comparison between A-tract and
non-A-tract oligonucleotides yielded dramatic differences. In the A-tract
the dispersion of end-to-end distances is half that of the corresponding
distribution in a non-A tract control duplex. The results demonstrate that
the A-tract DNA is more rigid than the non-A tract. Moreover, the r-6-averaged
mean end-to-end distance is shorter in the A-tract than the non-A tract,
indicating a small net curvature in the former. These experiments explicitly
demonstrate sequence-dependent bending and bendability of DNA and suggest
a general approach for future studies of the dynamics of protein-induced
DNA bending.
Exploring the Energy Landscape
of Proteins: The Temperature Dependence of Dynamics of Ribonuclease H
Arthur G. Palmer, III
Department of Biochemistry and Molecular Biophysics
Columbia University
630 West 168th Street,
New York, NY 10032
E-mail: agp6@columbia.edu; ftp:cuagpa.bioc.columbia.edu; Fax: (212) 305-7932
phone: (212) 305-8675; WWW: http://convex.hhmi.columbia.edu/palmer
The dynamical properties of Escherichia coli ribonuclease
H have been studied at temperatures of 285 K, 300 K and 310 K by using NMR
spectroscopy to measure spin-lattice relaxation rate constants, spin-spin
relaxation rate constants, and steady state nuclear Overhauser enhancements
of the backbone and side chain nitrogen-15 nuclear spins. Relaxation data
were analyzed by using the modelfree formalism to determine order parameters,
effective internal correlation times, and conformational exchange rate constants.
The temperature dependence of order parameters defines a characteristic
temperature for librational motions that in turn parameterizes the energy
surfac for backbone motions. The apparent conformational exchange rate constants
indicate that the rates of motions on microsecond time scales are strongly
temperature dependent with approximate Arrhenius behavior. The activation
barriers for microsecond time scale dynamics are ~20-50 kJ/mol.
Back to Table of Contents: Abstracts
Back to Table of Contents: Tenth
Conversation
Go to Abstracts Section VI