The SV40 large T-antigen helicase can unwind four stranded DNA structures linked by G-quartets
The SV40 large T-antigen helicase can unwind four stranded DNA structures linked by G-quartets
Nava
Baran
,
Lev
Pucshansky
,
Yonit
Marco
,
Sima
Benjamin
and
Haim
Manor*
Department of Biology, Technion-Israel Institute of Technology,
Haifa
32,000,
Israel
Received October 22, 1996;
Accepted November 21, 1996
ABSTRACT
We describe a novel activity of the SV40 large T-ag helicase, the unwinding of four stranded DNA structures linked by stacked G-quartets, namely stacked groups of four guanine bases bound by
Hoogsteen hydrogen bonds. The structures unwound by the helicase were of two
types: (i) quadruplexes comprising four parallel strands that were generated by
annealing oligonucleotides including clustered G residues in a buffer containing Na
+
ions. Each parallel quadruplex consisted of four oligonucleotide molecules. (ii) Complexes comprising two parallel and
two antiparallel strands that were generated by annealing the above oligonucleotides in a buffer containing K
+
ions. Each antiparallel complex consisted of two folded oligonucleotide molecules. Unwinding of these unusual DNA structures by the T-ag was monitored by gel electrophoresis. The unwinding process required
ATP and at least one single stranded 3
'
-tail extending beyond the four stranded region. These data indicated that
the T-ag first binds the 3
'
-tail and moves in a 3
' ->
5
'
direction, using energy provided by ATP hydrolysis; then it unwinds the four
stranded DNA into single strands. This helicase activity may affect processes
such as recombination and telomere extension, in which four stranded DNA could
play a role.
INTRODUCTION
Short DNA sequence repeats, in which one strand consists of clusters of 3-8 guanine residues interspersed with other short sequences, are found in eukaryotic genomes at various chromosomal sites, including telomeres (
1
,
2
), immunoglobulin switch regions (
3
) and promoters (
4
,
5
). It has been shown that single stranded DNA sequences containing such repeats
can form unusual DNA structures stabilized by interactions between the guanine
bases (
6
-
8
). Of these structures, four stranded DNA complexes have been studied most extensively. In such complexes, groups of four guanine bases, one from each strand, are bound by Hoogsteen
hydrogen bonds, thereby forming stacked G-quartets that hold the chains together (
9
-
17
) (for a review see ref.
18
).
Four stranded DNA structures are formed
in vitro
in buffers containing physiological concentrations of Na
+
and K
+
ions at neutral pH. Therefore, it is conceivable that the same structures may
occur in living cells and affect essential cellular processes, such as
recombination and extension of telomeric sequences by the enzyme telomerase (
19
,
20
). In recombination, formation of four stranded DNA may account for initial
chromosome pairing, as previously suggested (
10
,
11
). At telomeres, formation of four stranded DNA structures could inhibit the
telomerase, as shown by
in vitro
assays of the enzyme (
21
). In both cases, the cells would be expected to possess a mechanism for
resolving the four stranded DNA into single strands. This task could be accomplished by enzymes which specifically recognize the unusual four stranded DNA and cause separation of the strands. However, recent studies have indicated
that helicases, which normally unwind DNA duplexes, can also unwind DNA triple
helices (
22
,
23
). These results have led us to examine the possibility that helicases might be
capable of unwinding four stranded DNA.
Here we report that the SV40 large tumour antigen (T-Ag), a
bona fide
helicase (
24
-
26
), can unwind four stranded DNA structures. The biochemical and biological
implications of this finding are further discussed below.
MATERIALS AND METHODS
Preparation of parallel DNA quadruplexes
Oligonucleotides (purchased from Biotechnology General, Israel) were purified by polyacrylamide gel electrophoresis and were 5'-end labeled with
32
P, using the T4 polynucleotide kinase (
27
). One labeled oligonucleotide, or a mixture of two oligonucleotides, were dissolved in 10 [mu]l of a buffer containing 10 mM Tris-HCl pH 8.0, 1 mM EDTA (TE buffer) at a concentration of 50 nmol/ml. The solution was heated 5 min at 95oC and fast cooled to 4oC. At this stage, NaCl was added to a final concentration
of 0.50 M and the solution was annealed 40-72 h at 60oC, unless a different temperature was specified. The annealing was
performed in an 1.5 ml Eppendorf tube without covering the solution with
paraffin oil. This procedure was found to give a high yield of four stranded
structures. The mixture was then diluted 5-fold into TE buffer and stored at -20oC until further use. Formation of DNA quadruplexes was monitored by polyacrylamide gel electrophoresis, as
described in the section on helicase assays.
Preparation of four stranded DNA structures consisting of two folded chains
In this procedure,
32
P-labeled oligonucleotides were first heated 5 min at 95oC and fast cooled to 4oC. Then, the oligonucleotides were annealed, as described in the previous section, except that
the annealing solution contained 0.30 M Kcl, instead of 0.50 M NaCl.
Helicase assays
Helicase assays were carried out in 20 [mu]l of a solution containing 10 mM Tris-HCl pH 8.0, 10 mM NaCl, or 10 mM KCl, 10 mM MgCl
2
, 0.50 mM 1,4-dithiothreitol (DTT), 50 [mu]g/[mu]l bovine serum albumin, 2 mM ATP, 0.10 mM EDTA, 10% glycerol, 1.0 pmol/ml DNA
substrate and 60 pmol/ml hexameric recombinant SV40 large T-ag prepared in insect cells (
28
). The reactions were performed at 37oC for 45 min, unless different reaction times were specified, and were terminated, as described (
28
). The mixtures were electrophoresed in 12% polyacrylamide gels in a buffer containing 45 mM Trizma base, 1.2 mM Na
2
EDTA, 45 mM boric acid. The gels were dried on DE81 paper (Whatman) and autoradiographed.
The electrophoresis was carried out either at 23oC, or at 4oC, as specified in the figure legends.
Footprinting assays
32
P-labeled oligonucleotides were annealed in TE buffer containing 0.50 M NaCl, or 0.30 mM KCl, as described above. Other samples of the
same oligonucleotides were dissolved in 0.10 M NaOH, heated 5 min at 95oC and fast cooled to 4oC. These samples were neutralized by addition of HCl and Tris-HCl pH 8.0 to final concentrations of 0.10 M. Methylation by
dimethyl sulfate (DMS) was carried out 20 min at 22oC in a solution containing 10 mM Tris-HCl pH 8.0, 10 mM NaCl, or 10 mM KCl, 10 mM MgCl
2
, 0.50 mM 1,4-dithiothreitol (DTT), 2 mM ATP, 100 pmol/ml DNA and 0.10% DMS. The
reaction was terminated, as described (
23
), and the samples were electrophoresed in a 12% polyacrylamide gel. Bands
containing the single stranded oligonucleotides and the two types of four
stranded complexes were excised and incubated 16 h at 37oC in TE buffer containing 0.10 M NaCl to elute the radioactively labeled
DNA molecules. The DNA was next precipitated in ethanol, dried, dissolved in
1.0 M piperidine and heated 30 min at 90oC. The samples were electrophoresed in a 12% Long Ranger sequencing gel, as
previously described (
23
).
RESULTS
Formation of parallel DNA quadruplexes by association of oligonucleotides
containing clusters of G residues
It has been previously found that oligonucleotides containing clusters of G
residues, which are annealed in buffers containing Na
+
ions, form DNA complexes consisting of four parallel strands (
10
,
17
). Figure
1
a shows examples of such quadruplexes which are generated by self-association of identical molecules (X
4
, Y
4
and Z
4
), or by association of two different molecules (XY
3
). The single strands in these quadruplexes are held together by two groups of
four and five stacked G-quartets, which are indicated in these drawings by squares. A detailed
drawing of a G-quartet is illustrated in Figure
1
c; it consists of four guanine bases-one per DNA chain-which are linked by Hoogsteen hydrogen bonds.
Unwinding of parallel DNA quadruplexes by the SV40 large T-ag helicase
Figure
4
A shows a polyacrylamide gel analysis of A
4
DNA quadruplexes and of complexes having slower electrophoretic mobilities,
that were incubated at 37oC for various time periods in the presence of ATP and the SV40 large T-ag helicase. This incubation was performed in a buffer containing Na
+
ions. It can be seen that the T-ag helicase unwound the quadruplexes, as well as the slower migrating complexes, into monomeric oligonucleotide A molecules in a time dependent reaction. The reaction proceeded in a
linear fashion up to 15 min and then it slowed down. In the absence of the T-ag, these complexes remained stable. It should be noted that in these
assays, as in the other unwinding assays reported in this paper, the ratio T-ag/DNA substrates was 60:1. However, more recent studies have shown that
the T-ag was capable of efficiently unwinding quadruplexes even when the ratio T-ag/DNA substrate was 3:1 (Baran
et al
., unpublished results).
ATP and a 3
'
-tail are required for unwinding parallel quadruplexes by the SV40 large T-ag helicase
To find out whether unwinding of parallel quadruplexes was caused by T-ag helicase activity and not by a different mechanism, we sought to determine whether certain parameters of this reaction were
compatible with a helicase mechanism. Helicases, including the SV40 large T-ag helicase, require ATP, or other hydrolysable nucleoside triphosphates,
for unwinding duplex and triplex DNAs (
23
,
24
,
29
-
33
). Therefore, we tested the requirement for ATP as an energy source for unwinding quadruplexes. Figure
5
A shows an experiment in which the A
4
quadruplexes were incubated with the SV40 large T-ag helicase in the absence, or the presence of ATP, or in the presence of
the non-hydrolysable ATP analogue AMP-PNP. It can be seen that whereas unwinding of these complexes
proceeded very efficiently in the presence of ATP, no monomer oligonucleotide was released from the complexes in the absence of ATP. Furthermore, in the presence of AMP-PNP, the helicase unwinding reaction was very inefficient. These data
showed that ATP hydrolysis was required for unwinding the quadruplexes. Unless
otherwise specified, all the helicase assays reported below were carried out in
the presence of ATP.
Formation of antiparallel four stranded DNA structures generated by association
of two folded oligonucleotide molecules
It has been found that, in addition to parallel quadruplexes, oligonucleotides
containing G clusters can also form four stranded DNA structures consisting of two parallel and two antiparallel strands. Structures of this type are generated when the oligonucleotides are annealed in buffers containing K
+
ions, instead of Na
+
ions (
11
,
12
,
14
-
16
). Figure
1
b shows two examples of such complexes designated X
2
and X
2
'. Like the parallel quadruplexes shown in Figure
1
a, X
2
and X
2
' complexes are also held together through formation of G-quartets. It should be noted that X
2
and X
2
' represent just two of a larger number of fold-back antiparallel isomers (
16
,
18
) (see Discussion).
Figure
7
shows that self-annealing of the oligonucleotides A, D and F in a buffer containing K
+
ions generated complexes designated A
2
, D
2
and F
2
. Figure
3
b presents a DMS footprint of the D
2
complex. In this footprint, eight G residues in the two clusters of Gs were
fully protected against DMS methylation. The fifth G residue in the five G
cluster, designated as G17 and underlined in Table
1
, was not fully protected. In addition, the G residue designated as G23 and also
underlined in Table
1
, was unprotected. This pattern of protection differs from that of the parallel
quadruplex generated by self-annealing oligonucleotide D in a buffer containing Na
+
ions, which is shown in Figure
3
a; in that footprint, all nine G residues in the two clusters, as well as the
residue No. G23 were protected. The pattern of protection shown in Figure
3
b is consistent with the notion that pairs of oligonucleotide D molecules formed
complexes such as those illustrated in Figure
1
b, in which the encircled Gs correspond to the residues G17 and G23 described
above. The notion that D
2
complexes, as well as the complexes A
2
and F
2
, consisted of just two oligonucleotide molecules was supported by the
observation that these complexes
had faster electrophoretic mobilities than the corresponding parallel quadruplexes D
4
, A
4
and F
4
(compare
Figs
6
and
7
). However, these experiments could not establish which of the various possible
antiparallel four stranded isomers were obtained in the present studies.
Unwinding of antiparallel four stranded DNA structures by the SV40 large T-ag helicase
Figure
7
shows that the antiparallel four stranded complexes A
2
, D
2
and F
2
were efficiently unwound by the T-ag helicase. It should be noted that these experiments and the subsequent
experiments on unwinding of antiparallel complexes were all performed in a
buffer containing K
+
ions. Figure
8
A shows that ATP hydrolysis was required for unwinding the A
2
complexes. Figure
8
B shows that the antiparallel complexes C
2
and B
2
, which do not contain 3'-tails, were not unwound by the T-ag helicase under conditions in which the complexes A
2
, D
2
and F
2
were unwound by the enzyme. Thus, ATP and a 3'-tail were required for unwinding the antiparallel, as well as the
parallel four stranded structures.
DISCUSSION
The experiments reported in this article have shown that the SV40 large T-ag helicase can unwind parallel DNA quadruplexes and four stranded
complexes containing two parallel and two antiparallel strands. The unwinding process requires ATP hydrolysis and at least one single stranded 3'-tail. Based on these data, we suggest that the unwinding proceeds
as follows: first, the T-ag binds a single stranded 3'-tail. Next, the enzyme migrates in a 3' -> 5' direction along the single stranded DNA,
using the energy provided by ATP hydrolysis, until it encounters the G-quartets; then, it causes disruption of these quartets.
The parallel quadruplexes of the types shown in Figure
1
a were generated from oligonucleotides including clustered G residues in a
buffer containing Na
+
ions, and were unwound by the T-ag helicase in buffers containing Na
+
ions (Figs
4
A,
5
and
6
). However, as shown in Figure
4
B, quadruplexes formed in a buffer containing Na
+
ions and then transferred into a buffer containing K
+
ions could not be efficiently unwound by the helicase. This result was not due
to a major structural transition induced by the K
+
ions, since the DMS footprint of the quadruplexes was not altered by the ion
replacement, as shown in Figure
3
a. It could be caused by the selective stability imparted on these structures by
K
+
ions binding to the O6 atoms in the cavities formed between pairs of stacked G-quartets (
11
,
14
).
The antiparallel four stranded structures (Fig.
1
b), were generated by annealing the oligonucleotides containing the clustered G residues in a
buffer containing K
+
ions. As Figures
7
and
8
show, these structures were efficiently unwound by the T-ag in a buffer containing K
+
ions. Apparently, these complexes were not stabilized by the K
+
ions to the same extent as the parallel quadruplexes. It should be noted that
the isomer X
2
' shown in Figure
1
b was found by X ray analysis to occur in crystals formed in the presence of K
+
ions (
15
). However, a different isomer was found by NMR spectroscopy to occur in
solution (
16
). As already pointed out in the Results, these isomers represent just two of a
larger number of possible isomers. The footprints shown in Figure
3
b could not reveal which isomer was generated by our annealing procedure. Thus,
the question whether the T-ag can unwind all possible fold-back antiparallel four stranded isomers formed in the presence of K
+
ions, or just a subset of these isomers, remains open. It should also be noted that the intracellular concentration of K
+
ions is considerably higher than that of Na
+
ions (
34
). Hence, the unwinding of the structures generated in buffers containing K
+
ions might be physiologically more relevant.
In a recent study we have found that the SV40 large T-ag helicase was capable of unwinding the third polypyrimidine strand of
polypyrimidine[middot]polypurine[middot]polypyrimidine DNA triple helices, which is linked to the
polypurine strand by Hoogsteen hydrogen bonds (
23
). However, the data obtained in that study have not completely ruled out the
possibility that the third strand and the polypyrimidine strand in the duplex
were transiently exchanged, such that D-loop containing molecules were formed at a very low concentration and
served as intermediates in the unwinding reaction. In this case, the helicase
could have released the third strand by unwinding the duplex in these
hypothetical D-loop intermediates. Unlike the triplexes, the four stranded DNA
structures, in which the strands are solely linked by Hoogsteen hydrogen bonds,
could not possibly undergo a transition into normal duplex DNA. Hence, the data
shown in the present article constitute a more conclusive demonstration that a helicase is capable of dissociating unusual DNA structures through unwinding DNA strands which are linked by hydrogen bonds other than those of the
Watson-Crick type.
One way in which the T-ag helicase might unwind four stranded DNA structures is through a
`passive' creeping mechanism which is analogous to a mechanism previously
proposed for unwinding DNA duplexes and triplexes by helicases (
23
,
31
,
33
). In this process, the Hoogsteen hydrogen bonds linking one of the G residues
in the 3'-terminal quartet to the other Gs might be transiently opened. Then,
the released G residue could be trapped by the helicase. The hydrogen bonds
between the remaining Gs of the quartet could then be destabilized and
spontaneously opened. Subsequently, the helicase might translocate, using the energy derived from ATP hydrolysis, and similarly disrupt the next G-quartet. Other `active' mechanisms of unwinding four stranded DNA would
require ATP hydrolysis not just for translocation along the single stranded
DNA, but also for the unwinding process itself (
31
). Such mechanisms would also require binding of the enzyme to the four stranded
DNA, for which there is no evidence. Clearly, more detailed studies are needed to unravel
the mechanism by which the T-ag unwinds four stranded DNA.
Although biological roles for four stranded DNA structures have not been
conclusively demonstrated, several studies appear compatible with functions
suggested for these structures, such as regulation of telomere synthesis (
20
,
21
) and pairing of chromosomes in recombination (
10
,
35
). First, proteins which selectively bind four stranded DNA structures have been
reported (
36
,
37
). Secondly, a subunit of a protein which specifically binds single strand
extensions in telomeres was found to stimulate formation of four stranded DNA (
38
). Thirdly, the enzyme telomerase was found to be inhibited
in vitro
by formation of four stranded DNA structures (
20
,
21
). Fourthly, a yeast endonuclease that specifically cuts single stranded DNA next to four stranded structures has been reported (
39
). Furthermore, genetic evidence has been provided indicating that this
endonuclease may affect telomere metabolism (
40
,
41
). As already pointed out, four stranded DNA structures are readily generated
under physiological conditions. However, once formed, these structures are very
stable and do not spontaneously dissociate (
18
). Nevertheless, such dissociation would presumably be required if four stranded
DNA were formed in the course of telomere synthesis and during recombination
processes. Our finding that the SV40 large T-ag helicase, a
bona fide
helicase, can unwind four stranded DNA structures, suggests that helicases may provide an enzymatic mechanism for dissociating these structures. It should be noted that helicases were found to be involved
in both recombination (see, for example, ref.
42
,
43
) and telomere synthesis (
44
), at stages which apparently require unwinding of duplex DNA, or DNA-RNA hybrids. Further studies would be needed to find out whether
helicases are also involved in these processes through unwinding of four
stranded DNA structures.
ACKNOWLEDGEMENTS
This research was supported by grant 93-00156 from the United States-Israel Binational Science Foundation and by a project grant from
the Israel Cancer Research Fund.
26 Fanning,E. and Knippers,R. (1992) Annu. Rev. BioChem.61, 55-85.MEDLINE Abstract
27 Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
32 Matson,S.W., Bean,D.W. and George,J.W. (1994) BioEssays, 16, 13-22.MEDLINE Abstract
33 Borowiec,J. (1996) In DePamphilis,M.L. (ed.), DNA Replication in Eukaryotic Cells. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
34 Alberts,B., Bray,D., Lewis,J., Raff,M., Roberts,K. and Watson,J.D. (1994) Molecular Biology of the Cell. Garland Publishing, Inc. New York & London.
