Transcript Holliday Poster - The HeliX group

Understanding the DNA Holliday Junction
Benjamin C. Gale, James H. Thorpe, Nick H. Hopcroft and Christine J. Cardin. The School of Chemistry, University of Reading.
Crystallographic Studies
Homologous recombination (HR) is key for life, acting to
create genetic diversity and to repair double strand breaks
in DNA. However the key role now appears to be in the
repair and resetting of DNA replication forks that have
stalled or collapsed at sites of DNA damage1. Either way
HR is characterised by the
formation of branched DNA
molecules called Holliday
Junctions (HJ’s)2.
The structures of the DNA
HJ’s solved in this group
have
provided
strong
evidence of features now
associated with this motif.
In
particular
the
requirement of the central
d(ACC) core and quite
specific stabilising contacts
have been shown to be
key3,4,5,6.
Until
recently
however7,8 the role of
metal ions in the crystal
structure of the HJ has
been limited despite their
dramatic effect in solution9.
2+
d(TCGGTACCGA)
2+
This
work
has
investigated
two
sequences,
d(TCGGTACCGA) and d(CCGGTACCGG) in the presence of
Ca2+ and Sr2+. In the presence of Sr2+, d(TCGGTACCGA)
forms the half-X structure with a spine of five Sr2+ sites
that spiral down each B-DNA arm. By contrast in the
presence of Ca2+, only two
ion sites have been refined
in the short arms, removed
from the terminal bases.
2+
d(CCGGTACCGG)
Sr
C2 64.3 25.0 36.8,  =110.0
Ca
C2 66.6 23.6 37.2,  = 111.2
Sr
C2 65.8 24.0 77.3,  = 114.7
5 sites form a spine in each B-DNA
chain.
2 sites in short arms removed from
terminus.
1 in short arm at terminus
1 in long arm removed.
Extensive hydration coordinated by
ion sites
Crossover angle = 43.3º
Inter Phosphate separation
Minor groove = 6.98 Å
Major groove = 6.80 Å
Pronounced holes either side of the
junction
Crossover angle = 43.2º
Inter Phosphate separation
Minor groove = 6.91 Å
Major groove = 7.37 Å
Not refined sufficiently
Not refined sufficiently
Ca2+
C2 66.4 23.8 37.1,  = 110.1
2 sites in long arms removed from
terminus
2 in short arms at terminus.
Hydration roduces less pronounced
holes
Crossover angle = 37.6º.
Inter Phosphate separation
Minor groove = 6.21 Å
Major groove = 6.91 Å
For the second sequence,
four Ca2+ ion sites can be
located. Two are in the long
arms, found removed from
the terminal bases whereby
in the short arms the two
sites are found at the
terminus of the sequence
causing the junction to be
closed
up.
Preliminary
results for Sr2+ indicate two
ion sites, one in the short
arm at the terminus and
one in the long arm
removed from the terminal
bases.
Competitive Dialysis
Competitive dialysis or competition dialysis was initially
described by Ren and Chaires1 and has since then been
adapted in this laboratory. The aim of such an experiment
is to determine which sequence a
drug preferentially binds to in
solution and hence provides great
insight into the drugs that should be
crystallised with specific sequences.
The method works by equilibrium
dialysis. A macromolecule is paced
inside a semi-permeable membrane,
called a dispodialyser, that have pore
sizes to prevent escape of the
macromolecule
but
allow
the
surrounding drug solution to enter.
The dialysers are then left to
equilibrate in a beaker containing a
stirred solution of drug, which can
cross
the
membrane
of
the
dispodialysers and bind to the most
preferred DNA sequence. More
technically the drug in the beaker will first enter the
dispodialysers by simple osmosis, to equilibrate the
pressure difference between both sides of the membranes
of the
dialysers. The drug will then
intercalate into some of the oligonucleotide sequences
and further displace the osmosis equilibrium so that
more
drug will pass from the beaker to the
dispodialysers to compensate for the
binding to DNA. Osmotic pressure
and intercalation will therefore start
to compete up to a point when
equilibrium is reached. The amount
of drug up taken by each sequence
is
compared
by
UV-visible
spectroscopy
after
equilibrating
overnight.
This work has initially focussed on
adopting the protocol described by
Ren and Chaires to suit our
requirements. Though there has
been only limited success to date it
is expected that this technique will
provide
the
rationalisation
for
crystallisation studies within this
research area. A range of drugs are
available for use with this study including XR5944, which
has been shown to have sub-nanomolar IC50 values in
tumour cell lines and entered Phase 1 clinical trials in the
UK in July 2003.
1 McGlynn, P. et al, PNAS, 2001. 98: p. 8227-8234. 2 Holliday, R., Genet. Res., 1964. 5: p. 282-304. 3 Eichman, B. F. et al, PNAS, 2000. 97: p. 3971-3976. 4 Ho, P.S. et al, Curr. Opin. Struct. Biol. 2001. 11: p. 302-308. 5 Ortiz-Lombardia, M. et al, Nat. Struct. Biol. 1999 6: p.
913-917. 6 Thorpe, J.H. et al Acta Cryst, 2002. D58: p. 567-569. 7 Thorpe, J.H. et al, J. Mol. Biol, 2003. 327: p. 97-109. 8 Vargason, J.M. et al, J. Biol. Chem, 2002. 277: p. 21041-21049. 9 Lilley, D.M., DNA-Protein: Structural Interactions, OUP
1 Ren, J. and Chaires, J.B. Methods in Enzymology, 2001. 340: p. 99 - 108.
p53 Consensus Sequence
Fundamental Points
The pivotal role of the tumour suppressor protein p53 in
cancer is well documented with over 50% of all cancers
having associated with them, mutations in p531,2. Previous
work has identified the p53 consensus sequence, which
defines the DNA sequence elements with which p53
interacts3.X-ray
crystallography
has revealed how
one core domain,
the central half of
p53 binds to a
quarter site4 and
further a model
has
been
proposed in which
four
identical
domains
can
occupy all four
quarter
isites
in the full consensus sequence though this has never been
seen in an X-ray crystallography experiment4. Further to
this it has been demonstrated through NMR that human
p53 can bind to the Holliday junction5. In this work it was
demonstrated that 80-96% of p53 were specifically
located at the junction with only 4% at the ends. For an
~565 base pair sequence, this represents a high level of
specificity.
Crystallographic Studies
Competitive Dialysis
This study will look to characterise the half site 5'PuPuPuC(A/T)(T/A)GPyPyPy-3’ of the p53 consensus
sequence, which
has
a
striking
similarity to the
HJ
forming
sequences to see
if
it
is
predisposed to being
a HJ or if the
binding of p53 or
a range of anticancer
drugs
effects
the
conformation
adopted.
 Homologous recombination is characterised by the
formation of branched DNA molecules called Holliday
Junctions.
 This experiment will seek to determine which sequence
Crystallisation studies are currently underway for the
sequence d(GGGCTAGCCC), which has been initially
screened with the Hampton Research Nucleic Acid screen
to yield small microcrystals. Optimised conditions are now
being utilised and it is hoped that suitable crystals will be
available for data collection in the later half of this year.
1 Levine, A.J. et al, Nature, 1991. 351. 2 Hollstein, M. et al, Science, 1991. 253: p. 49-53. 3 el-Deiry, W. S. et al, Nat Genet, 1992. 1: p. 45-49. 4 Cho, Y. et al, Science, 1994, 15: p. 346-355. 5 Suman, L. et al, J Biol Chem, 1997. 272: p. 7532 - 7539.
a drug preferentially binds to in solution.
 The role of metal ions in solution is key to the HJ.
 This will allow crystallographic studies to be focussed
on the most relevant sequences and drugs.
 The crystal structures reveal the binding of metal ions
to the HJ.
 The method is based on simple equilibrium dialysis.
 Small differences gives rise to quite dramatic effects
which could be important to junction recognition.
 A range of drugs are available for study including those
that have sub-nanomolar IC50 values in tumour cell lines
and have entered Phase 1 clinical trials.
p53 Consensus Sequence
Acknowledgements
 The tumour suppressor protein p53 is mutated in over
 For funding I thank the EPSRC, the BBSRC and the
50% of all human cancers.
Association for International Cancer Research.
 The p53 consensus sequence defines the DNA sequence
elements with which p53 interacts.
 Thanks also to Xenova, and in particular to Peter
Charlton for supplying many of the drugs being used.
 The half site of this sequence has a striking similarity to
the HJ forming sequences.
 For useful correspondence on the competition dialysis
method I would like to thank Jonathan Chaires.
 The sequence d(GGGCTAGCCC) will be characterised to
see if it adopts the HJ conformation.
 Finally I would like to thank the many colleagues who
have helped me including staff at DESY in Hamburg.