|
|
|
|||||||
|
|||||||||
(Received for publication, June 21, 1994; and in revised form, October 24, 1994) From the
Previous studies have shown that formation of intermolecular DNA
triplexes at sequences that overlap protein binding sites inhibits DNA
binding by these proteins. We show that DNA cleavage by eukaryotic
topoisomerase II is blocked by triplex formation at sites overlapping
and adjacent to the triple binding site. To map precisely the
boundaries of triplex interference, we constructed a vector containing
enzyme binding sites of different lengths and flanked both 5` and 3` by
DNA triplexes. We call this method Triplex Interference Mapping by
Binding Element Replacement (TIMBER). Triplex regions within 3 bases 5`
or 7 bases 3` of cleavage sites blocked DNA cleavage; triplex formation
outside of this region had no effect upon cleavage activity. We
conclude that topoisomerase II binding requires unhindered access to
the major groove of a duplex DNA binding site in this 10-base region.
In addition, the inclusion of topoisomerase II inhibitors yielded the
same results for the triplex interference assays despite alterations in
DNA cleavage site selection. The statistical analyses of over 500
topoisomerase II cleavage sites (in the presence or absence of
inhibitors) suggest a model consistent with the region spanning
-3 to +7 (relative to the cleavage site) containing most of
the base-specific contacts for topoisomerase II. This triplex
interference assay may prove valuable in the characterization of DNA
binding sites for other proteins as well, particularly in conjunction
with deletion analysis.
Eukaryotic topoisomerase II is a nuclear enzyme that modulates
the topological states of DNA via transient double strand breaks in DNA
coupled with strand passage(1) . The enzyme is essential for
the segregation of daughter chromosomes in
mitosis(2, 3, 4) ; it is also the
intracellular target for a number of classes of clinically valuable
anticancer agents(5, 6, 7) . Most of these
agents, including doxorubicin, amsacrine, ellipticine,
epipodophyllotoxins, and quinolones, act through the trapping of the
covalent enzyme-DNA complex, subsequently inducing chromosome
breaks(7, 9, 10) . Recently, another class of
antitumor compounds was shown to stabilize a topoisomerase
II Topoisomerase II is a sequence-specific
DNA-binding protein, but it makes double strand breaks in DNA sites
that are not palindromic and are diverse in sequence content;
characterizations of cleavage sites in the absence of inhibitors have
revealed quite degenerate consensus
sequences(12, 13, 14) . The presence of
inhibitors alters the base preferences around the cleavage site,
leading to the derivations of different druginduced consensus sequences
(for example, see (14, 15, 16) . These
consensus studies have yielded results that are inconsistent in terms
of defining the size of the sequence-specific DNA recognition element
for topoisomerase II. To approach this problem in a different way, we
devised a mapping strategy based on our observation that the formation
of a DNA triple helix, in which a homopyrimidine oligonucleotide bound
to a homopurine/homopyrimidine target duplex, prevented topoisomerase
II cleavage at sites present on the DNA duplex alone(17) . We
call this method Triplex Interference Mapping by Binding Element Replacement (TIMBER). ( Intermolecular DNA
triplexes are formed under defined conditions when an appropriate
single strand oligonucleotide anneals to duplex DNA; these unusual DNA
structures offer new prospects for evaluating and modulating the
interaction between proteins and DNA(18, 19) . In this
DNA structure, pyrimidine oligonucleotides bind through the major
groove parallel to the Watson-Crick purine strand by Hoogsteen hydrogen
bonds. Thymine (T) in the third strand recognizes adenine-thymine (AT)
base pairs (T In this report, we show that intramolecular
DNA triplexes inhibit topoisomerase II induced breakage of DNA at
positions flanking the triplex regions. We evaluated the ability of the
DNA triplexes to interfere with topoisomerase II cleavage of DNA sites
located over a range of distances away from the two triplex blocks. We
found that triplex regions within 3 bases 5` or 7 bases 3` of a
cleavage site on either DNA strand define a minimum 10-bp core-binding
element required for topoisomerase II activity. This novel application
of DNA triplexes should prove useful for defining duplex DNA binding
minima for other proteins.
Figure 1:
Partial
sequence of the triplex vector pTX-0. The vector was constructed as
described under ``Materials and Methods,'' and a restriction
map is shown in the lowerfigure. Topoisomerase II
sites were cloned into the unique SmaI site defined at the
interface between the two triplex blocks. The upperdiagram shows the triplex blocks with their corresponding
pyrimidine oligos used to drive triplex formation (see Fig. 2).
Figure 2:
Band shift experiment of pUC19 and pTX-0
with pyrimidine oligonucleotide. End-labeled oligonucleotides were
incubated with either pUC19 and pTX-0 under conditions optimal for
triplex formation as described under ``Materials and
Methods.'' Samples were loaded onto a 1% agarose gel and following
electrophoresis, the gel was stained with ethidium bromide and
photographed (B), dried, and exposed to X-ray film (A). The odd numbered lanes contained supercoiled (sc) DNA, and even lanes contained EcoRI-digested linear (lin) DNA; open circular DNA (oc) is marked in panelB. Pyrimidine oligonucleotides were added in lanes1, 2, 5, and 6.
Purine oligonucleotides were added in lanes3, 4, 7, and 8.
where H is the entropy (in bits), calculated as
follows:
in which p is the probability of base occurrence at
position x taken from the base proportions of the already
aligned sites, plus one strand of the current site. For example, as the
10th selected site is analyzed, the base proportions of the previously
aligned nine sites are used, with one orientation of the 10th site
added in; this information content is compared with that with the
opposite strand sequence added in. After a strand has been selected for
the last site of a data set, the process is repeated selecting the
sites in a different random order. After 10-20 such repetitions,
the alignment that yielded the greatest total information content is
used, and the base proportions of this set are determined. This method
was used for the analyses in Fig. 8, A-C, termed
the consensus single strand method; the information content, in bits,
at each position was then determined as in
Stormo(33, 35) . In Fig. 8, D and E, both the sequenced strand cleavage site and the
complementary strand site (located 4 bases 5`) were pooled for all
sites, with no strand selection; the information content at each
position was calculated directly from the frequencies of all pooled
sites. All computations were performed using the EDEN Genesys computer
software system (TEAM Associates, Westerville, OH).
Figure 8:
Information content analyses of
topoisomerase II cleavage site data sets. The information content (in
bits), is plotted against the sequence position relative to the
cleavage site for sets of topoisomerase II sites analyzed by consensus
single strands (A-C) or pooled double strands (D and E), as described in the text. The origins of the data
are indicated. The bracket marks the 10-base region from
-3 to +7, which was found by triplex interference to be the
essential duplex binding region; the identities of the bases at the
most highly conserved positions are indicated in the boxed region.
Incubation of an end-labeled
pyrimidine oligo probe (14-mer) with plasmid pTX-0 (see Fig. 1),
resulted in a clear band-shift of the probe to a position coincident
with the plasmid (Fig. 2A). As reported previously,
these results can be explained by formation of intermolecular DNA
triplexes(28) . Triplexes formed with closed circular
supercoiled (sc) and linear (lin) pTX-0 DNA (Fig. 2A, lanes5 and 6;
compare with the same gel stained with ethidium bromide and shown in Fig. 2B). Negative controls show that pUC19 DNA did not
band shift the pyrimidine oligo (Fig. 2A, lanes1-4) and purine oligos did not bind pTX-0 (Fig. 2A, lanes7 and 8);
thus, the pyrimidine oligo forms complexes specifically with pTX-0. The
reactions and gel electrophoresis in Fig. 2were carried out
with approximately the same molar concentrations of oligo and plasmid
(10 nM) in the pH 5.5 buffer. As triplexes block DNase I
access at the homopurine-homopyrimidine
sequences(17, 18, 22) , DNase I footprinting
was performed to localize sites of pyrimidine oligo binding in pTX-0 (Fig. 3). The 14-mer pyrimidine oligo specifically protected the
homopurine-homopyrimidine sequences (Fig. 3, lane6); however, footprints were not observed when a 14-mer
purine oligo was used (Fig. 3, lane7), and
the pUC 19 DNA fragments were not protected by any of the
oligonucleotides tested (Fig. 3, lanes1-3). These results, combined with the gel shift
data in Fig. 2A, show that pyrimidine oligonucleotides
formed stable intermolecular triplexes with pTX-0 at the sequences of
both target sites under our conditions. In Fig. 3, a DNase I
hypersensitive site can be seen at the 3` boundary of the triplex
region; a similar effect was observed at the 5` triplex boundary in
Chung et al.(17) .
Figure 3:
DNase I footprinting of triplexes. DNase I
digestions were performed as described under ``Materials and
Methods'' on an RsaI-PvuII fragment (labeled at
the RsaI site). Lanes1-4 contained
the pUC19 fragment as a control, and lanes5-8 contained the corresponding pTX-0 fragment, which differs in that
it has two 14-base pair homopurine-homopyrimidine regions (bracketed). Purine and pyrimidine oligonucleotides were
included as indicated above the gel lanes. Lanes4 and 8 show chemical sequencing markers. The filledbar marks the DNaseI protected
region.
The sequencing gel
in Fig. 4shows the results of topoisomerase II cleavage
reactions (in the presence or absence of the inhibitors m-AMSA
and VM-26) on either a pUC19 sequence (lanes1 and 2) or on the equivalent region of pTX-0 containing the two
triplex target sites (lanes4-9). Incubation of
the pUC19 fragment with the pyrimidine 14-mer used in Fig. 3did
not alter the specificity of topoisomerase II cleavage. In contrast, in
the reactions on pTX-0, all topoisomerase II sites (for no drug, m-AMSA or VM-26) within and immediately flanking the triplex
forming sequences were blocked by binding of the 14-mer (compare lanes5, 7, and 9 with 4, 6, and 8, respectively, in Fig. 4); external
to this region, triplex formation did not alter the cleavage patterns.
This experiment revealed that the inhibition of topoisomerase II
cleavage by DNA triplexes is independent of the presence of drugs, and
suggested that, in all cases, the enzyme requires contact with a
contiguous region of nontriplex (presumably B-form) DNA spanning a
minimum of 2 bases 5` and 6 bases 3` of a cleavage site. The
experiments described below were conducted in order to further define
this core duplex sequence required for topoisomerase II activity.
Figure 4:
Inhibition of topoisomerase II cleavages
intermolecular triplexes. Topoisomerase II cleavage reactions were
performed on the pUC-19 or pTX-0 fragments described in the legend to Fig. 3and in the presence (lanes2, 5, 7, and 9) or absence (lanes1, 4, 6, and 8) of
pyrimidine oligonucleotides. Lanes containing the pUC19 or pTX-0
fragments are indicated as are the constituents of each reaction
(inhibitor and oligo concentrations are specified under
``Materials and Methods''). Bracket shows the position of the
two 14-base pair homopurine-homopyrimidine
regions.
Figure 5:
Topoisomerase II cleavages on pTX-GT17. PanelA, reactions were carried out on a HpaII-EcoRI fragment (labeled at the HpaII
site) of pTX-GT17 (pTX-0 containing a 17-bp insert of GT). Lanes1-3 show DNase I footprinting protection, and lanes7-12 show topoisomerase II cleavage
reactions. Oligonucleotides and topoisomerase II inhibitors (VM is VM-26, AM is m-AMSA, and ND is no
drug) are indicated above the gel. Brackets to the left mark
the two homopurine-homopyrimidine regions and 17 base-pair CA repeat. PanelB, summary of topoisomerase II cleavages on
pTX-GT17 fragment. The 17-bp insert and partial flanking triplex
regions are shown. The x and y designators correspond
to sites marked in the sequencing gel in A. Sites marked by dottedlines correspond to cleavages that were
inhibited by triplex formation.
A
second plasmid was constructed to examine the cleavage site inhibition
when a shorter RY repeat was inserted. The plasmid, pTX-AC12 contains a
12-bp (dAdC)-(dGdT) repeat between the two triplex blocks. DNase I
footprinting shows the expected pattern of protection over the triplex
regions (Fig. 6A, lane3).
Topoisomerase II cleavages were also analyzed (Fig. 6A). The cleavage site located symmetrically in
the middle of the AC repeat (marked z) was not affected by
triplex formation; however, cleavage frequency of the flanking site y (2 bases 3` of triplex block) was significantly reduced by
triplex formation, as were sites x-1, x-2, and x-3 (see summary, Fig. 6B). Cleavage
protection by triplex formation on the bottom strand was identical to
that of top strand (Fig. 6B, data not shown). These
results, in conjunction with the protection pattern of pTX-GT17,
indicate that the minimal sequence requirement for topoisomerase II
cleavage is 10 base pairs of double strand DNA, spanning the sequence
from -3 to +7 (cleavage at -1/+1). It also
appears from these data that paired sites (defined as the top and
bottom strand cleavages staggered by 4 bp; see (31) ) are both
hindered by triplex formation, even when one of the paired sites is
clearly in a freely accessible duplex region. For example, considering
only the bottom strand in Fig. 6B, site y should be unhindered by triplex formation (it is between the two
paired z sites, which were unaffected by the triplex);
however, site y is blocked presumably because its paired site
(on the top strand) is hindered by the 5` top strand triplex insulator.
Figure 6:
Topoisomerase II cleavages on pTX-AC12. PanelA, reactions were carried out on a RsaI-PvuII fragment (labeled at the RsaI
site) of pTX-AC12 (containing a 12-bp insert of AC in pTX0). Lanes2-4 are DNase I footprinting, and lanes6-11 are topoisomerase II cleavage reactions.
Reaction conditions are shown above each gel lane. PanelB is a summary of topoisomerase II cleavages on the
pTX-AC12 fragment (the top strand sites are shown in A; the
bottom strand cleavages are from a different gel that is not shown).
The cleavages are marked as x, y, and z and
correspond to sites marked in A (see
text).
Figure 7:
Topoisomerase II cleavages on pTX-N10. PanelA, the reactions were performed on the RsaI-PvuII fragment (labeled at the RsaI
site) of pTX-N10. The lanes1-3 show the DNase
I footprinting data (sequencing ladders are in lanes4 and 5). Lanes6-9 show
topoisomerase II cleavage data; individual reactions are as marked above each lane. The singlearrow indicates the topoisomerase II cleavage within the 10-bp insert. PanelB, summary of topoisomerase II cleavage on pTX-
N10 fragment.
A total of 519 topoisomerase
II cleavage sites were analyzed, from four different sets of sites.
These data sets were 134 VM26-induced chicken enzyme sites, (
The TIMBER analysis strongly suggests a 10-bp binding core
for topoisomerase II. The question persists as to whether a duplex
element smaller than 10 bp (say 9 bp) might be sufficient for cleavage
activity. It would be difficult to prove that 9 bp is necessary and
sufficient for cleavage since failure to detect cleavage in the
presence of triplexes might result if the 9-bp test sequence simply is
not a topoisomerase II cleavage site. Parenthetically, the
topoisomerase II consensus sequences defined to date are larger than 9
bp; thus, only trial and error could be used to design an appropriate
insert less than 10 bp to test. Since there are 4
Our data clearly show that cleavage inhibition is symmetrical (i.e. impeded at both top and bottom strand sites when either
site is within 2 bases of a triplex). Thus, top site 1 is eliminated by
analogy with top strand site x-1 in Fig. 5B.
Since the bottom strand site x-1 (Fig. 5B) is
also blocked by triplex, we can also eliminate bottom strand site 1 in
the 9-mer. Site 2 cleavages in the 9-mer above are eliminated by the
equivalent experiment summarized in Fig. 6B site y (topstrand) and its corresponding paired bottom
strand site. By the same reasoning, one can see that neither bottom
site 3 nor 4 would be cleavable by topoisomerase II (since both are
<3 bases from the triplex). While we cannot categorically prove that
a sequence exists less than 10 bp that might be cleavable in the
triplex, the statistical analysis of cleavage sites indicates that a
10-bp window from -3 to +7 is conserved in topoisomerase II
cleavage sites (see below). Moreover, as noted above, nuclease
footprinting data (40) and minimal site binding experiments (41) suggest that the core is equal to or greater than 10 bp as
opposed to smaller than 10 bp. The conformation of triplex/duplex
junctions may also be considered in interpreting our results. For
example, Hartman et al.(42) identified distortions at
these junctions, and we observed DNaseI hypersensitivity at the 3`
triplex/duplex boundary. However, Huang et al.(43) reported no inhibition of protein binding by triplex
formation immediately 5` of the minimal DNA binding element.
Furthermore, the topoisomerase II cleavage sites characterized in Fig. 8show strong base preferences immediately at the triplex
boundaries, suggesting that the junction does not impede enzyme
activity. It is possible that topoisomerase II contacts a larger core
binding site than indicated by these TIMBER experiments due to
``breathing'' of the triplex strand binding or to
displacement of adjacent triplex base pairs by enzyme binding; however,
rigorous tests of these dynamic interactions would be difficult. An
important question concerning the mechanism for the stabilization of
topoisomerase II-DNA covalent complexes by these anti-cancer agents is
how they alter the interaction of the protein with its substrates.
These agents are known to modify enzyme cleavage
patterns(39, 44, 45, 46, 47) .
The triplex interference results, however, revealed the same protection
results regardless of whether drugs were included in the reactions.
This suggests that the topoisomerase II inhibitors interact with the
protein-DNA intermediate only at sequences proximal to the cleavage
sites and do not alter the size of the recognition element.
Volume 270,
Number 11,
Issue of March 17, 1995 pp. 5932-5943
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
ATPDNA complex in a conformation incompatible with DNA
cleavage (11) .
)AT triplet) and protonated cytosine (CH+)
recognizes guanine-cytosine (GC) base pairs (CH+GC
triplet)(20, 21, 22, 23, 24, 25, 26, 27) .
Due to the requirement for protonation of cytosine, triplexes involving
C-rich sequences are more stable under acidic conditions (pH
5-6)(26, 28) . The pH stability of
intermolecular triplex formation thus depends on the sequence
composition of oligonucleotides employed; however, intermolecular
triplexes have been shown to exist at neutral
pH(21, 24, 25, 26, 29) .
Previous findings have additionally revealed that pyrimidine
oligonucleotides can inhibit restriction endonuclease and methylase
activities at sites located in or near purpyr
tracts(18, 19) . Furthermore, triplex-forming
oligonucleotides blocked binding of the eukaryotic transcription factor
Sp1 to its consensus sequence(18) . These results clearly show
that DNA-binding proteins can be inhibited by oligonucleotide-directed
triplex formation. An interesting and useful extension of this idea
relates to recent reports using triplexes to limit the systemic
specificity of restriction enzyme in high complexity
genomes(30) .
Enzyme and Materials
Topoisomerase II was
purified from whole chicken blood according to procedures previously
described(15, 31) . Restriction enzymes and T4
nucleotide kinase were purchased from Life Technologies, Inc. and U. S.
Biochemical Corp., respectively. 
- and
-[
P]ATP were from ICN. Oligonucleotides
were synthesized on an Applied Biosystem DNA synthesizer (at the Ohio
State University Biochemical Instrumentation Center). The topoisomerase
II inhibitors 4`-(9-acridinylamino)methanesulfon-m-anisidide (m-AMSA)and VM-26 (teniposide) were provided by the National
Cancer Institute divisions of synthetic and natural products,
respectively. DNA manipulation, fragment end labeling, and DNA
sequencing were performed as described elsewhere(32) .Plasmid Constructions
The triplex plasmid pTX-0
was constructed as follows: pUC19 DNA was digested with HincII, and 14-bp homopurine-homopyrimidine fragments were
inserted (see Fig. 1) such that a new SmaI site was
generated in the middle of the 14-base elements (the original SmaI site of pUC19 was removed by previous manipulation).
Constructs pTX-GT17, pTX-AC12, and pTX-N10 were prepared by inserting
appropriate fragments into SmaI site of pTX-0.
Band Shift Analysis
Formation of triplex
structures was assayed using a modification of the band shift technique
described previously(28) . End-labeled oligonucleotides (10
nM, 10,000 cpm) were mixed with 10 nM supercoiled or
linearized DNA (20 µl, total volume) in triplex buffer (25 mM Tris acetate, pH 5.5, 70 mM NaCl, 10 mM MgCl
, 10 mM 2-mercaptoethanol, 0.4 mM spermidine, 3 mM ATP, and 100 µg/ml bovine serum
albumin). Reactions were incubated for 30 min at 25 °C, and, after
addition of 2 µl of tracking dye (0.25% bromphenol blue, 40%
sucrose), they were loaded onto a 1% agarose gel (running buffer, 40
mM Tris acetate, pH 5.5, 100 mM sodium acetate with
circulation). After electrophoresis, the gel was stained with ethidium
bromide, destained, photographed, dried, and exposed to film at
-70 °C.DNase I Footprintng and Topoisomerase II Cleavage
Reactions
Reactions (20 µl) were carried out in triplex
buffer (see above) and contained 5` end-labeled fragment (2
nM, 20,000 cpm) and the specified oligonucleotide at 3
µM. Reactions were preincubated 30 min at 25 °C before
the addition of enzymes. For DNase I footprinting experiments,
digestions were initiated by the addition of 1 µl of 40 mM CaCl and 1 µl of 50 µg of DNaseI/ml. Reactions
were incubated 1 min at 25 °C and terminated by the addition of 2
µl of 0.5 M EDTA. Topoisomerase II cleavage reactions were
initiated by the addition of 8 units (5.6 nM) of purified
chicken topoisomerase II in the presence or absence of drugs (m-AMSA at 50 µg/ml and VM26 at 500 µg/ml) as
indicated, incubated 30 min at 30 °C, and then terminated by the
addition of SDS to 1% followed by proteinase K digestion (50 µg/ml,
56 °C for 30 min). Samples were ethanol-precipitated and analyzed
on an 8% sequencing gel.Statistical Analysis of Topoisomerase II Cleavage Site
Data Sets
Previous analyses of topoisomerase II cleavage sites
suggested that despite the double-stranded nature of these events, the
sequence information common to all cleavage sites is conserved
primarily on only one of the two strands(15) . The
determination of cleavage site locations from sequencing gels does not
reveal which of the two strands contains the information to specify the
cleavage. To select the conserved strand from each cleavage pair, we
used a computer algorithm that maximizes the information content of the
data set (i.e. to yield the most nonrandom base proportions).
Sites were taken in random order and tested to see which strand
contributed to a greater information content from the set (33, 34) using .
The Vector pTX-0 Forms DNA Triplexes at Two Adjacent
Sequences
We previously observed that DNA triplex formation over
a duplex DNA sequence containing topoisomerase II cleavage sites
inhibited enzymatic cleavage at those sites; furthermore, this triplex
interference was propagated over proximal cleavage sites flanking the
triplex region(17) . Having shown that DNA triplex formation
blocks topoisomerase II activity at sites both within and external to a
triplex region(17) , the next experiments were designed to map,
at base pair resolution, the exact 5` and 3` boundaries at which
triplex formation interfered with topoisomerase II cleavage of DNA. Our
previous experiments (17) were conducted in a neutral pH buffer
(6.8). The following experiments were conducted at pH 5.5 (as described
under ``Materials and Methods'') since triplexes with
cytosine containing oligos are stabilized at lower pH(28) ;
moreover, topoisomerase II displays very good DNA cleavage activity at
this pH (36) . ()In terms of DNA relaxation
activity, topoisomerase II has high catalytic activity pH 5.5 (90% of
the activity seen at pH 6.8, data not shown). The plasmid pTX-0, shown
in Fig. 1, contains two 14-base pair polypurine-polypyrimidine
sequences (identical and in opposite orientation) with a central SmaI to allow insertion of additional DNA sequences between
these two triplex-forming sites.
Inhibition of Topoisomerase II Cleavage of pTX-0
Sequences by DNA Triplexes
For several reasons, topoisomerase II
is particularly amenable to site mapping by triplex interference. The
enzyme cleaves DNA at specific sites, making it possible to determine
directly the binding site locations without resorting to band shifts or
footprinting. In addition, while sequence-specific in its binding,
topoisomerase II sites are found rather frequently (on average, one
site every 25.9 bases in random DNA sequences, (15) ).
Furthermore, a large number of binding sites have been characterized,
and high affinity sites have been
defined(13, 15, 37) .
Triplex Interference: Analysis of Flanking Sequences
Required by Topoisomerase II
To evaluate the minimal duplex DNA
sequence required for topoisomerase II activity and to determine the
exact boundaries of triplex interference, pTX-0 (triplex vector) was
modified by inserting topoisomerase cleavage sites between the two
14-bp triplex blocks, such that sites would be at varying distances
(both 5` and 3`) from the triplex boundaries. We took advantage of our
previous results showing that topoisomerase II cleaves DNA at every
second base in sequences of alternating purine-pyrimidine (RY), and
that these sites are conserved when topoisomerase inhibitors are
included(37) . The construct pTX-GT17 has a 17 bp (dGdT)-(dAdC)
repeat cloned between the two triplex blocks in the SmaI site
(see Fig. 1). DNase I footprinting revealed that both triplex
regions were protected from digestion in the presence of the pyrimidine
oligonucleotide (Fig. 5A, lane2); in
addition, a novel hypersensitive site appeared at the 3` triplex
boundary. Topoisomerase II cleavage reactions were also carried out on
pTX-GT17 with or without oligonucleotides (lanes7-12). Topoisomerase II cleavages at triplex and
flanking regions were inhibited by the pyrimidine oligonucleotides
(compare lane7 with 8, 9 with 10, and 11 with 12). As reported
previously(37) , topoisomerase II cleaves within the AC repeat
at every second base (Fig. 5A, lanes7-10; in reactions lacking the cleavage-stimulating
inhibitors, much longer exposures of the autoradiogram were required to
visualize the cleavages, lanes11 and 12).
The cleavage sites located centrally within the insert (through the AC
repeat) were not significantly affected by triplex formation. Cleavage
data on the pTX-GT17 target DNA are summarized in Fig. 5B. The sites marked x-1 and x-2 on the top strand were clearly inhibited by triplex formation;
however, the enzyme could access the internal sites (e.g.y-1 to y-4). The paired sites on the bottom
strand (4 bases 5`) reflect the results observed on the top strand, as
marked in Fig. 5B (sequencing data not shown). From
these data, it appears that 3 bases of duplex DNA 5` of a cleavage site
are necessary for enzyme activity. For example, considering the top
strand of a given topoisomerase II site, this corresponds to nucleotide
positions -3 to -1 (cleavage at -1/+1).
Topoisomerase II Cleavages on pTX-N10 Containing One
Cleavage Site between Two Triplex Blocks
In order to confirm
that a 10-base pair duplex binding site is sufficient for topoisomerase
II-mediated cleavage, we constructed the triplex plasmid pTX-N10, in
which a single 10-bp topoisomerase II cleavage site was inserted into
the SmaI site of pTX-0 (see Fig. 7B). Triplex
formation is shown by the DNase I protection pattern in lane2 of Fig. 7A. Topoisomerase II cleavages
on this fragment are shown in lanes6-9 of Fig. 7A. Again, the topoisomerase II cleavages were
blocked directly over triplex regions and at flanking sites when
triplexes were formed by the pyrimidine oligonucleotides. However, the
cleavage site within the 10-base pair insert (Fig. 7A,
the site indicated by arrow) was not inhibited by triplex
formation (compare lanes6 and 7, 8 and 9). Furthermore, it is clear that while the addition
of a topoisomerase II inhibitor (in this case, m-AMSA) may
alter the sequence specificity of the enzyme, it does so within this
10-base minimal binding site, because the triplex interference results
appear independent of drug inclusion.
Statistical Analyses of Topoisomerase II Cleavage
Sites
Our results suggest that most of the base-specific
contacts of topoisomerase II with DNA lie in the 10-bp region spanning
-3 to +7 relative to the cleavage site. We wanted to compare
our triplex interference mapping results with results obtained from
statistical analyses of large numbers of topoisomerase II cleavage
sites to test whether the bulk of conserved base-specific contacts
identified also lay within this 10-bp region. We used the measure of
information content to reveal which positions (relative to the cleavage
sites) contained conserved base occurrences within sets of
topoisomerase II cleavage sites. This analysis identifies positions
with nonrandom base frequencies, and, in contrast with 
-square
analysis, the information content at a base position is independent of
sample size; the maximum value is two bits of information for a
position at which one particular base occurs in all sites examined, and
the minimum value is zero bits for random base
frequencies(33, 38) .
)111 VM26-induced mouse topoisomerase II
sites(39) , 197 m-AMSA-induced mouse enzyme
sites(39) , and 77 Drosophila topoisomerase II sites
observed in the absence of inhibitors(12) . The
data were treated for either a single strand (consensus sequence type)
analysis or a pooled double strand analysis method (see
``Materials and Methods''). The information content plots in Fig. 8, A and B, show that for both VM26 data
sets, the bulk of the total conserved base information does lie in the
10-bp region between -3 and +7; in fact, the only strongly
conserved position for each is at -1, which in each case
represents a high frequency of cytosines. The analysis of the m-AMSA sites in Fig. 8C also reveals that the
bases conserved among these sites are located within this 10-bp region;
the identities of these preferred bases are indicated. Fig. 8D shows the analysis of these same 197 m-AMSA sites analyzed by the pooled double strand method; the
plot appears quite similar to that of Fig. 8C. Finally,
the information content plot in Fig. 8E, representing
77 ``no drug'' Drosophila topoisomerase II sites,
also reveals that the bases conserved among these sites are found in
the 10-bp region spanning -3 to +7 relative to the cleavage
site. Thus, for these topoisomerase II cleavage data, the base-specific
information is found in the region identified as the core binding
element by our TIMBER results and is independent of the origin of the
enzyme or whether inhibitors were included in the reactions.
DNA Triplex Inhibition of Topoisomerase II
Cleavages
Repetitive topoisomerase II cleavage sites were cloned
into the SmaI site of the pTX-0 vector (located between the
two triplex-forming regions) in order to reveal the 5` and 3`
boundaries of triplex interference. On all of the DNA targets tested,
topoisomerase II cleavage sites were not affected if located at least 3
bases 3` and 7 bases 5` of triplex regions; sites located more
proximally to the triplexes were inhibited. Therefore, we conclude that
topoisomerase II has a minimal requirement of 10 base pairs of duplex
DNA at a binding site. Topoisomerase II is a homodimeric enzyme; in its
double-stranded cleavages, both subunits cleave the DNA, with a 4-base
5` overhang between the two paired cuts on opposite strands.
Considering the double-stranded nature of topoisomerase II cleavages,
the triplex interference data suggest a symmetrical 10-base pair
recognition element for the enzyme (see Fig. 7B). These
data support a model in which the cleavage step is carried out by each
subunit of the homodimer acting on single strands, and both sites must
be available in order for the enzyme to initiate cleavage activity. The
data suggest that it needs access to the major groove over a span of 10
bp. It is likely that the enzyme makes contacts further 3` and 5` of
the 10-bp site based upon nuclease footprinting
experiments(40) , minimal site binding analyses(41) ,
and consensus sequence derivations (12, 15) . Our
results suggest, however, that these external interactions are either
not major groove contacts or are not essential for topoisomerase II
activity.
possible
sequence combinations in a 9-mer experiment, proving that cleavage never occurs is not practical; however, we consider that a
9-mer would not be a cleavage substrate in a TIMBER experiment for the
following reason. Placement of a cleavage site anywhere within the 9 bp
will violate the 5` or 3` positional requirements for cleavage as
defined in Fig. 5Fig. 6Fig. 7. For example,
consider the following 9-mer insert experiment with potential cleavages
at sites 1 to 4 (N is any base and top site 1 is paired with bottom
site 1).
Analyses of Base Information Content in Sets of
Topoisomerase II Cleavage Sites
We sought to test the validity
of our triplex interference model by comparing its results with those
of statistical analyses to determine which base positions were
conserved among topoisomerase II sites. We determined the base
conservation of topoisomerase II binding sites (by position relative to
the cleavage site) using the measure of information
content(15, 38) , which indicates the deviation from
random base frequencies at each position. Four sets of topoisomerase II
cleavage sites were selected to cover a range of origin of the enzyme
and the presence or absence of different inhibitors. Both sets of
VM26-induced topoisomerase II cleavage sites (134 chicken enzyme sites
in Fig. 8A and 111 mouse enzyme sites in Fig. 8B) exhibited the majority of base conservation
within the 10-bp region spanning -3 to +7, with each
possessing only a single strongly conserved base, a cytosine at
position -1 relative to the cleavage site. The information
content plot of 197 m-AMSA-induced mouse topoisomerase II
sites analyzed using the consensus single strands revealed conserved
bases at five positions (-3, -2, +1, +4,
+7), all of which are within the 10-bp TIMBER model core element. Fig. 8D shows a plot of the same data, but with the 197
pairs of sites (from each strand of the cleavage sites) pooled and
treated as 394 individual sites, all aligned by the position of DNA
cleavage. This method obviates the need for consensus single strand
selections and reveals base information content that is conserved
symmetrically in topoisomerase II sites; the bases common to these
cleavage sites are located at positions -3, +1, +4, and
+7 relative to the cleavage sites (i.e. within the
-3 to +7 core). Finally, the information content of 77 Drosophila topoisomerase II cleavage sites (observed in the
absence of inhibitors), was analyzed symmetrically as 154 sites (Fig. 8E); conserved base occurrences are found
throughout the -3 to +7 region, but little information is
found outside this 10-bp region. The statistical analyses of
topoisomerase II sites reported by others (39, 45, 46) is in close agreement with the
-3/+7 conserved region derived by the TIMBER model.General Considerations for Triplex Interference Mapping
of DNA Binding Sites
An important parameter in characterizing a
DNA binding protein is determining the size of a DNA binding site and
in particular, the region in which base-specific contacts occur.
Generally this is addressed using a battery of biochemical and chemical
nucleases to search for protected versus accessible sites.
Triplex interference analysis complements these existing methods by
allowing the investigator to combine the power of triplex
selectivity(29, 30) with a systematic reduction in
access to the major groove of the DNA site. Recent studies of DNA
triplexes have shown that a variety of techniques permit their
formation under physiological conditions so that our triplex
interference assay may be easily extended for use in other systems. For
example, Duval-Valentin et al.(1992) (48) showed that
triplexes formed at pH 7.8 and inhibited in vitro transcription, and that chemical modifications to the
polypyrimidine oligonucleotide further stabilized its binding to duplex
DNA. In addition, Roy (49) demonstrated an inhibition of gene
transcription by formation of triplexes using a purine-rich
oligonucleotide. Dervan and co-workers (8) have published a
number of reports characterizing conditions for stabilizing DNA
triplexes. The triplex interference assay may prove quite useful in
conjunction with the traditional deletion analysis of a protein's
DNA binding sites.
)))
We thank Dr. Joseph Spitzner (TEAM Associates,
Westerville, Ohio 43081, U.S.A.) for providing the EDEN-GENESYS
computer software used to analyze the cleavage data sets. We thank D.
Luo for characterizing pH effects on topoisomerase II activity.
Critical reading of the paper by C. Sommer-Furbee and D. Subramanian is
gratefully acknowledged.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Duca, P. Vekhoff, K. Oussedik, L. Halby, and P. B. Arimondo The triple helix: 50 years later, the outcome Nucleic Acids Res., September 1, 2008; 36(16): 5123 - 5138. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |