DNA triple helix formation at oligopurine sites containing multiple contiguous pyrimidines
DNA triple helix formation at oligopurine sites containing multiple contiguous pyrimidinesDarren M. Gowers and Keith R. Fox*
Division of Biochemistry and Molecular Biology, School of Biological Sciences, University of Southampton, Bassett Crescent East, Southampton SO16 7PX, UK
Received July 14, 1997;Revised and Accepted August 19, 1997
ABSTRACT
We have used DNase I footprinting to assess the formation of triple helices at 15mer oligopurine target sites which are interrupted by several (up to four) adjacent central pyrimidine residues. Third strand oligonucleotides were designed to generate complexes containing central (X·TA)n or (X·CG)n triplets (X = each base in turn) surrounded by C+·GC and T·AT triplets. It has previously been shown that G·TA and T·CG are the most stable triplets for recognition of single TA and CG interruptions. We show that these triplets are the most useful for recognizing consecutive pyrimidine interruptions and find that addition of each pyrimidine residue leads to a 30-fold decrease in third strand affinity. The addition of 10 [mu]M naphthylquinoline triplex-binding ligand stabilizes each complex so that all the oligonucleotides produce footprints at similar concentrations (0.3 [mu]M). Targets containing two pyrimidines are only bound by oligonucleotides generating (G·TA)2 and (T·CG)2 with a further 30-fold decrease in affinity. (G·TA)2 is slightly more stable than (T·CG)2. In the presence of the triplex-binding ligand the order of stability is (G·TA)2 > (C·TA)2 > (T·TA)2 > (A·TA)2 and (T·CG)2 > (C·CG)2 > (G·CG)2 = (A·CG)2. No oligonucleotide footprints are generated at target sites containing three consecutive pyrimidines, though addition of 10 [mu]M triplex-binding ligand produces stable complexes with oligonucleotides generating (G·TA)3, (T·CG)3 and (C·CG)3, with a further 30-fold reduction in affinity. No footprints are generated at targets containing four Ts, though the ligand induces a weak interaction with the oligonucleotide generating (T·CG)4.
INTRODUCTION
The generation of three-stranded nucleic acid complexes was first demonstrated over four decades ago (1 ,2 ). Interest in these unusual DNA structures has recently focused on the potential use of oligonucleotides as sequence specific DNA-binding agents (3 -5 ). Triplex formation has also been used for nucleic acid affinity capture (6 -8 ), cosmid restriction site mapping (9 ) and site-specific labelling of supercoiled DNA (10 ).
In these structures the third strand oligonucleotide lies in the major groove of duplex DNA forming specific hydrogen bonds to substituents on the purine bases. Two different types of triplex have been described which differ in the orientation of the third strand. Pyrimidine-rich oligonucleotides bind parallel to the duplex purine strand forming C+·GC and T·AT triplets (11 -13 ), while purine-rich third strands bind in an antiparallel orientation forming G·GC, A·AT and T·AT triplets (14 -16 ). Formation of the parallel motif usually requires conditions of low pH, necessary for protonation of the third strand cytosines.
Since the third strand base forms hydrogen bonds to substituents on the purine strand of the duplex, these structures are generally restricted to homopurine target sites. However, if the triplex strategy is to be used for recognising mixed DNA sequences it will be necessary to devise methods for recognising pyrimidine residues. Within the parallel motif other weaker triplets have also been described for recognition of TA and CG base pairs, including G·TA (17 -29 ) and T·CG (18 ,19 ,28 -31 ) (Fig. 1 ). Since the third strand base makes only one hydrogen bond contact with the duplex pyrimidine, and these triplets are not isomorphous with T·AT and C+·GC, they are less stable. To date most studies with these alternative combinations have used sequences containing only one G·TA (or T·CG) triplet, though two G·TA triplets were present in a complex formed on the 3' long terminal repeat of HIV DNA, separated by four canonical triplets (17 ). We have also shown that under certain conditions (AT)n can be targeted with (GT)n forming blocks of alternating G·TA and T·AT triplets (27 ). The stability of the G·TA triplet is affected by the nature of the surrounding bases and may be stabilized by adjacent T·AT triplets (22 ,25 ). NMR studies on the G·TA triplet have suggested that, as well as forming a hydrogen bond between the guanine 2-amino group and the O4 of thymine, there is potential for an additional weak interaction with a thymine on an adjacent base pair (22 ,25 ). In addition there appears to be a strong stacking interaction with a third strand thymine on the 5'-side.
MATERIALS AND METHODS
Chemicals and enzymes
Oligodeoxynucleotides were purchased from Oswel DNA Service. BamHI-cut alkaline-phosphatase treated pUC18 and DNA ligase were purchased from Pharmacia UK. Bovine DNase I was purchased from Sigma, and stored frozen at 7200 U/ml. Restriction enzymes and reverse transcriptase were purchased from Promega. Triplex-stabilising ligand 1 (Fig. 1 C) was stored at -20oC as a 20 mM stock solution in dimethylsulphoxide and was a gift from Dr L. Strekowski, Department of Chemistry, Georgia State University.
DNA sequences
Complementary oligonucleotides were treated with polynucleotide kinase and annealed prior to ligation into BamHI-cut pUC18. Following transformation of calcium permeabilised Escherichia coli strain TG2, successful ampR clones were picked from agar plates containing IPTG and X-Gal as white colonies, in the usual way. The sequences of successful clones were confirmed by sequencing using a T7 dideoxy sequencing kit (Pharmacia). The eight different clones obtained, containing pyrimidine interruptions within the oligopurine tract, are shown in Table 1 . The inserts were oriented so that labelling the 3'-end of the HindIII site visualised the purine-containing strands of AAGAAAAGAAGAAAA, AAGAAATTTAAGAAA and AAGAAACCCAAGAAA and the pyrimidine-containing strands of the other sequences.
DNA fragments
Plasmids containing the cloned inserts were digested with HindIII and SacI and labelled at the 3'-end of the HindIII site with [[alpha]-32P]dATP using reverse transcriptase. The fragments of interest were separated from the remainder of the plasmid DNA on a 6% (w/v) non-denaturing polyacrylamide gel. The radiolabelled DNA was eluted from the gel slice and dissolved in 10 mM Tris-HCl pH 7.5 containing 0.1 mM EDTA at a concentration of at least 10 c.p.s./[mu]l as measured on a hand held Geiger counter (~10 nM).
. Target sites and third strand oligonucleotides used in this work
Target sequencesa
Third strand oligonucleotides
5'-AAGAAAAGAAGAAAA
5'-TTCTTTTCTTCTTTT
5'-AAGAAAATAAGAAAA
5'-TCCTTTTATTCTTTT
5'-AAGAAAACAAGAAAA
5'-TTCTTTTGTTCTTTT
5'-TTCTTTTTTTCTTTT
5'-AAGAAAATTAAGAAA
5'-TTCTTTTGGTTCTTT
5'-AAGAAAACCAAGAAA
5'-TTCTTTTCCTTCTTT
5'-TTCTTTTTTTTCTTT
5'-TTCTTTTAATTCTTT
5'-AAGAAATTTAAGAAA
5'-TTCTTTGGGTTCTTT
5'-AAGAAACCCAAGAAA
5'-TTCTTTCCCTTCTTT
5'-TTCTTTTTTTTCTTT
5'-TTCTTTAAATTCTTT
5'-AAGAATTTTAAGAAA
5'-TTCTTGGGGTTCTTT
5'-AAGAACCCCAAGAAA
5'-TTCTTTTTTTTCTTT
aOnly the purine-rich strand of each sequence is shown. The central pyrimidine bases in the duplex targets, and the corresponding positions on the third strand oligonucleotides are underlined.
DNase I footprinting
Radiolabelled DNA (1.5 [mu]l), oligonucleotide (1.5 [mu]l) and 1.5 [mu]l of buffer or triplex binding ligand (30 [mu]M) were mixed to give final third strand concentrations between 100 and 10-3 [mu]M. Both the oligonucleotide and the triplex-binding ligand were dissolved in 50 mM sodium acetate pH 5.5 containing 10 mM MgCl2. The complexes were left to equilibrate for at least 3 h at room temperature (20oC). Digestion was initiated by adding 2 [mu]l DNase I (0.4 U/ml, dissolved in 20 mM NaCl, 2 mM MnCl2 and 2 mM MgCl2), and stopped after 1 min by adding 4 [mu]l of 80% formamide, containing 10 mM EDTA and 0.1% (w/v) bromophenol blue.
Electrophoresis
The products of digestion were separated on 10% (w/v) denaturing polyacrylamide gels (National Diagnostics) containing 8 M urea. Electrophoresis conditions were 1500 V/41 W for ~2 h. Gels were soaked in 10% (v/v) acetic acid before drying at 80oC for 1 h, and exposed overnight to autoradiography at -70oC using an intensifying screen. Bands in the digestion pattern were assigned by comparison to Maxam-Gilbert sequencing lanes specific for adenine and guanine.ABC
RESULTS
Triplex formation at an oligopurine site
Before studying the effects of central pyrimidines on the formation of intermolecular triplexes we first examined the formation of complexes with a target site containing a guanine at the central position, using the 15mer oligopurine tract AAGAAAAGAAGAAAA. This should form a parallel triplex with the third strand TTCTTTTCTTCTTTT, generating a structure containing 12 * T·AT and 3 * C+·GC triplets. [All the other sequences used in this work generate complexes containing (13-N) * T·AT, 2 * C+·GC and N * YR triplets. All the experiments were performed at pH 5.5 to ensure protonation of the third strand cytosines.] The results with the fragment containing this simple oligopurine tract are presented in Figure 2 , showing the concentration dependence of the footprinting pattern in the presence or absence of 10 [mu]M naphthylquinoline triplex-binding ligand (1). Looking at Figure 2 A it can be seen that, in the absence of the ligand (left hand lanes), the oligonucleotide designed to form a perfect match at the target site, with a central C+·GC triplet, produces a clear footprint which persists to a concentration of ~30 nM. This footprint extends above (5') the target site by ~3-4 base pairs, and is accompanied by enhanced cleavage at the lower (3') end. Such enhanced DNase I cleavage is often observed on the duplex purine strand at the triplex-duplex junction. In the presence of 10 [mu]M triplex-binding ligand the footprinting pattern is very similar (though the enhancement at the 3'-end is slightly less pronounced) and persists to an oligonucleotide concentration of 10 nM, indicating a small (3-fold) increase in stability.
Figure 2 B shows the results of similar experiments using the oligonucleotide TTCTTTTATTCTTTT, which should generate an unstable A·GC triplet across the central base pair in place of C+·GC. Although this oligonucleotide causes a general reduction in DNase I cleavage throughout the target no clear footprint is evident. On adding the triplex-binding ligand a footprint is evident at the highest oligonucleotide concentrations (>10 [mu]M). Although this oligonucleotide does not produce a clear footprint it does generate enhanced cleavage at the 3'-(lower) end of the target, in the same position as that seen in Figure 2 A, which presumably indicates a transient interaction with the target sequence. Examples of enhanced DNase I cleavage in the absence of clear footprints have been observed in other studies (19 ,26 ,34 ). These results confirm that the 15mer oligonucleotide, designed to form only C+·GC and T·AT triplets forms a very strong interaction (Kd [approx] 30 nM), and that changing the central base of the third strand from C to A causes a large decrease in affinity.
Since it might be argued that some of the oligonucleotides described below might bind by covering only half the target site, avoiding the central pyrimidine base(s), we examined the interaction of the 7mer sequence TTCTTTT with this target site. This oligonucleotide corresponds to the 5' and 3' regions of the 15mer studied above, and has the correct sequence to form a short triplex with either the 5'-(upper) or 3'-(lower) portions of the target site. The results of these experiments are shown in Figure 2 C. It can be seen that although this oligonucleotide alone does not affect the cleavage pattern (left hand lanes), bands within the target site are attenuated with the highest oligonucleotide concentrations (100 and 30 [mu]M) in the presence of the triplex-binding ligand. This is not accompanied by any enhanced DNase I cleavage. It therefore appears that the interaction of this short oligonucleotide with half the target site is much weaker (at least 1000-fold) than that of the full length 15mer.
One central X·TA triplet
One central X·CG triplet
Figure 4 shows the results of similar experiments using the target site containing a C in the centre of the purine tract. It can be seen that, in the absence of the ligand, all four oligonucleotides produce footprints which cover the entire target. For the third strand containing a central T (generating a T·CG triplet) the footprint persists to a concentration of ~1 [mu]M, though bands within the target site are still attenuated with 0.3 [mu]M oligonucleotide. In addition two bands are evident at the centre of the target site at all oligonucleotide concentrations. We can be sure that these are genuine DNase I cleavage products, and not artifacts, as they are no longer present after adding the triplex-binding ligand (see below). The oligonucleotides containing central C or A also footprint to 1 [mu]M, but more bands can be seen within the target site, while the oligonucleotide containing a central G requires slightly higher concentrations. It seems that, at this target site, the rank order of triplet stability is T·CG > C·CG = A·CG > G·CG. In the presence of 10 [mu]M triplex-binding ligand all four oligonucleotides produce footprints which persist to 0.3 [mu]M. The triplex-binding ligand has removed the weak bands at the centre of the target site, and as seen with the target containing a central T, the ligand-induced footprints are 3-4 bases shorter and now coincide with the 5'-(upper) end of target site where there is enhanced DNase I cleavage.
Two central X·TA triplets
Figure 5 shows DNase I footprinting patterns for the fragment containing two Ts at the centre of the oligopurine tract. It can be seen that the oligonucleotide containing two central Gs (forming two adjacent G·TA triplets) causes attenuated DNase I cleavage at concentrations >10 [mu]M. Similarly, oligonucleotides with two Ts or two As at the centre only show reduced cleavage at the highest concentration (100 [mu]M) while very little change is observed with two central Cs. In this instance it appears that (G·TA)2 is the most stable combination; (T·TA)2, (A·TA)2 and (C·TA)2 are much weaker. In the presence of 10 [mu]M triplex-binding ligand all four oligonucleotides produce clear footprints which cover the entire target site. These are in a similar location to those seen with the targets containing single pyrimidine residues (Figs 3 and 4 ); the footprints extend by 3-4 bases beyond the lower (3') edge of the target site, but are coincident with the upper (5') edge where there is enhanced DNase I cleavage. These footprints require oligonucleotide concentrations of 0.3, 1, 3 and 10 [mu]M for central G, C, T and A, respectively, suggesting that the rank order of stability is (G·TA)2 > (C·TA)2 > (T·TA)2 > (A·TA)2. In the presence of the ligand the stability of the complex with (G·TA)2 is very similar to that with a single G·TA (Fig. 3 ).
Two central X·CG triplets
Figure 6 shows DNase I footprinting patterns for the fragment containing two Cs at the centre of the oligopurine tract. In the absence of the ligand none of the oligonucleotides produce a clear footprint, though there is attenuated cleavage with the highest concentration (100 [mu]M) of the oligonucleotide generating two T·CG triplets. Comparison with Figure 5 suggests that two consecutive T·CG triplets are marginally less stable than consecutive G·TAs. In the presence of 10 [mu]M triplex-binding ligand all four oligonucleotides produce footprints in a similar location to those seen with the other sequences shown in Figures 3 , 4 and 5 . These footprint to concentrations of 0.1, 0.3, 1 and 1 [mu]M for central T, C, G and A, respectively, suggesting that the rank order of stability is (T·CG)2 > (C·CG)2 > (G·CG)2 = (A·CG)2. In the presence of the ligand the stability of the complex with (T·CG)2 is very similar to that with a single T·CG (Fig. 4 ).
Three central X·TA triplets
Figure 7 (left hand panel) shows DNase I footprinting patterns for the fragment containing three consecutive Ts at the centre of the oligopurine tract. In the absence of the ligand none of the oligonucleotides affect the cleavage pattern. However, in the presence of 10 [mu]M triplex-binding ligand the oligonucleotide containing three central guanines produces a clear footprint at the target site which persists to a concentration of 3 [mu]M. Although this is a specific interaction, the complex is less stable than that seen with two consecutive G·TA triplets (Fig. 5 ). No enhanced cleavage is evident at either end of the target site, even though we are observing the purine-containing strand. None of the other oligonucleotides affects the cleavage pattern in the presence of the triplex-binding ligand, even at the highest concentration (100 [mu]M). In this case it is clear that (G·TA)3 provides the most stable means of recognising (TA)3.
Three central X·CG triplets
Figure 7 (right hand panels) shows DNase I footprinting patterns for the fragment containing three consecutive Cs at the centre of the oligopurine tract, in the presence of oligonucleotides containing three consecutive Cs or Ts. No interaction is seen with oligonucleotides containing central G or A (not shown) in either the presence of absence of the triplex-binding ligand. In the absence of the ligand neither the T- nor C-containing oligonucleotides produce DNase I footprints, though both oligonucleotides produce enhanced DNase I cleavage at the 3'-(lower) end of the target site in a concentration dependent fashion. In the presence of the triplex-binding ligand a clear footprint is observed at the target site with the T-containing oligonucleotide which persists to a concentration of 3 [mu]M, which is also accompanied by enhanced cleavage at the triplex-duplex junction. This enhancement persists to lower oligonucleotide concentrations. A similar footprint is also observed with the oligonucleotide containing three central Cs, but this requires higher oligonucleotide concentrations. In this case it is clear that (Y·CG)3 provides the most stable means of recognising (CG)3 and that T·CG is more stable than C·CG.
Four central X·TA triplets
In similar experiments (not shown) we used DNase I footprinting to examine triplex formation at the target containing four consecutive Ts at the centre of the oligopurine tract, using an oligonucleotide containing four central Gs. In this case no footprints were evident in either the presence or absence of the triplex-binding ligand. It appears that (G·TA)4 cannot be stabilised by the surrounding 11 T·AT and C+·GC triplets, even in the presence of the triplex-binding ligand.
Four central X·CG triplets
Further experiments (not shown) examined triplex formation at the target containing four consecutive Cs at the centre of the oligopurine tract, in the presence of the oligonucleotide containing four consecutive Ts. In this case although no footprint was seen with the oligonucleotide alone (even at 100 [mu]M), the cleavage pattern was altered in the presence of the triplex-binding ligand. In this case bands within the target site were attenuated and there was enhanced cleavage at the 5'-(upper) end of the target site, suggestive of some specific interaction. It appears that although (T·CG)4 is a poor combination, it is better than (G·TA)4.
DISCUSSION
The results presented in this paper demonstrate that it is possible to recognise oligopurine tracts containing multiple pyrimidine interruptions by triple helix formation. Although previous studies have examined the effect of single pyrimidine insertions, the present work addresses the effects of several (up to four) consecutive pyrimidine bases on triplex stability.
Single G·TA and T·CG triplets
The complexes formed across single pyrimidine residues within an oligopurine tract are less stable than those containing only T·AT and C+·GC triplets, requiring ~30-fold higher oligonucleotide concentrations than at the simple oligopurine site. Although this is a large decrease in stability, it is less than that produced by a single A·GC mismatch. It may be a general feature that recognition of pyrimidines within a homopurine target is less destabilising than targeting a purine with the wrong third strand base (e.g. A·GC). This may be due to the larger size of the purine base which will impose a greater distortion on the phosphodiester backbone. The equivalence of single G·TA, T·TA and C·TA triplets within this structure was surprising, but suggests that the stability of these complexes is dominated by the blocks of T·AT and C+·GC triplets. However, the lower stability of A·TA, taken together with the different affinities of the various oligonucleotides at multiple TA and CG sites, demonstrates that the T·AT and C+·GC triplets are not the only factors controlling the stability of these complexes. As shown in other publications T·CG is the best triplet for recognition of a CG base pair, though C·CG is only slightly less stable. Since the T·CG triplet involves a hydrogen bond between the exocyclic amino group of C and O2 of T, a similar structure can be envisaged using O2 of C as the hydrogen bond acceptor. Indeed it is difficult to see why T·CG should be more stable than C·CG.
Multiple pyrimidines
For sequences containing more than one pyrimidine the discrimination between the various triplet combinations becomes more pronounced. The stability for third strand recognition of two consecutive Ts is G > C > T> A, while for recognition of two Cs the order is T > C > G = A. The different order in these two cases emphasises the specificity of the G·TA and T·CG triplets and suggests that these triplets are not merely the best tolerated on account of the base stacking (which would be the same for recognition of both TA and CG). These results with unmodified oligonucleotides differ from those in a recent study which incorporated an acridine in the third strand adjacent to the mismatched base (28 ). These showed that G·CG generated the most stable complex, an interaction which may be affected by the stacking of the base against the intercalating acridine.
It might be postulated that the footprints arise from binding of the third strand oligonucleotide to one or other half of the oligopurine target without interacting with the central pyrimidine(s). However, we can dismiss this suggestion for two reasons. Firstly the 7mer oligonucleotide, which could in theory bind to either half of the target site, failed to produce clear DNase I footprints (Fig. 1 C). Secondly models which involved half-site recognition by the 15mers do not explain the variations in affinity displayed by oligonucleotides with different central bases.
The concentrations of best oligonucleotide required to generate a footprint at the various target sites are ~0.03, 1 and 30 [mu]M for 0, 1 and 2 pyrimidines, respectively, and 0.01, 0.3, 0.3 and 3 [mu]M for 0, 1, 2 and 3 pyrimidines in the presence of the triplex-binding ligand. There appears to be a ~30-fold decrease in third strand affinity for each successive pyrimidine substitution in the oligopurine tract. Since G·TA and T·CG are not isomorphous with each other or with the canonical T·AT and C+·GC triplets there must be a distortion in the DNA backbone at each of these triplets. Although successive G·TA or T·CG triplets impart a further decrease in stability due to the loss in hydrogen bonding, their effects may not be strictly additive as there will be no further backbone distortions since successive numbers of these triplets will be isomorphous with each other. Nonetheless, within the range of concentrations employed in this work, two adjacent pyrimidines is the upper limit for triplex recognition; this is extended to three pyrimidines in the presence of the triplex-binding ligand. Although these triplets may not provide a practical means of recognising CG and TA inversions they indicate the least destabilising combinations which might be used for `skipping' these bases when using longer triplex-forming oligonucleotides.
Triplex-binding ligand
The results presented in this paper demonstrate that the naphthylquinoline triplex-binding ligand can be used to stabilise complexes which contain unusual or weak triplets. This increased binding strength is accompanied by some loss of stringency, especially for complexes which contain only one base pair inversion. However, in each case the third strand still binds to the specific target site and does not affect DNase I cleavage in the remainder of the fragment.
An unusual observation is that, in the presence of the ligand, several of the footprints are shorter by 3-4 bases at the 5'-(upper) end of the target, where they are accompanied by enhanced DNase I cleavage at the triplex-duplex junction. This effect has been noted previously for the effect of the naphthylquinoline compound on the binding of C5T5 to the target sequence G5A5·T5C5, but was not seen for the interaction of T5C5 with A5G5·C5T5 (34 ). It is worth remembering that these footprints do not arise from steric occlusion of the enzyme by the third strand since DNase I cuts from the DNA minor groove, whereas the third strand oligonucleotide is located in the major groove. Triplex footprints must therefore arise from oligonucleotide-induced changes in DNA structure and/or rigidity. Examination of these cleavage patterns reveals that, in the presence of the ligand, the upper (5') edge of the footprints is located at the boundary of the target site, whereas with the oligonucleotide alone the footprint extends by a further 3-4 bases. Since DNase I footprints are usually staggered in the 3' (not 5') direction we would expect the upper edge of the footprint to lie close to the end of the target site. It therefore appears that the unusual pattern is produced by the oligonucleotide alone, generating a footprint which is longer than expected; this reverts to the predicted size in the presence of the ligand. One explanation for this phenomenon is that the oligonucleotide-induced changes continue beyond the actual target site, whereas in the presence of the triplex-binding ligand these are restricted to the actual site of interaction. Previous studies have suggested that these changes are only observed when Ts are located at the 3'-end of the third strand oligonucleotide (34 ) consistent with the present studies.
ACKNOWLEDGEMENTS
This work was supported by grants from the Cancer Research Campaign and the Medical Research Council.
REFERENCES
1 Felsenfeld, G., Davies, D.R. and Rich, A. (1957) J. Am. Chem. Soc.79, 2023-2024.
2 Felsenfeld, G. and Rich, A. (1957) Biochim. Biophys. Acta26, 457-468.
3 Postel, E.H., Flint, S.J., Kessler, D.J. and Hogan, M.E. (1991) Proc. Natl. Acad. Sci. USA88, 8227-8231.MEDLINE Abstract
4 Alunni-Fabbroni, M., Manfioletti, G., Manzini, G. and Xodo, L.E. (1994) Eur. J. Biochem.226, 831-839.MEDLINE Abstract
5 Lavrovsky, Y., Mastyugin, V., Stoltz, R.A. and Abraham, N.G. (1996) J. Cell. Biochem.61, 301-309.MEDLINE Abstract
6 Kiyama, R., Nishikawa, N. and Oishi, M. (1994) J. Mol. Biol.237, 193-200.MEDLINE Abstract
7 Johnson, A.F., Wang, R., Ji, H., Chen, D., Guilfoyle, R.A. and Smith, L.M. (1996) Anal. Biochem.234, 83-95.MEDLINE Abstract
8 Nishikawa, N., Kanda, N., Ioshi, M. and Kiyama, R. (1997) Nucleic Acids Res.25, 1701-1708.MEDLINE Abstract
9 Ji, H.M., Francisco, T., Smith, L.M. and Guilfoyle, R.A. (1996) Genomics31, 185-192.
10 Pfannschimdt, C., Schaper, A., Heim, G., Jovin, T.M. and Langowski, J. (1996) Nucleic Acids Res.24, 1702-1709.
12 Le Doan, T., Perrouault, L., Praseuth, D., Habhoub, N., Decout, J.L., Thuong, N.T., Lhomme, J. and Hélène, C. (1987) Nucleic Acids Res.15, 7749-7760.MEDLINE Abstract
13 Radhakrishnan, I. and Patel, D.J. (1994) Structure2, 17-32.MEDLINE Abstract
14 Beal, P.A. and Dervan, P.B. (1991) Science251, 1360-1363.MEDLINE Abstract
15 Broitman, S.L., Im, D.D. and Fresco, J.R. (1987) Proc. Natl. Acad. Sci. USA84, 5120-5124.MEDLINE Abstract
16 Radhakrishnan, I. and Patel, D.J. (1993) Structure1, 135-152.MEDLINE Abstract
17 Griffin, L.C. and Dervan, P.B. (1989) Science245, 967-971.MEDLINE Abstract
18 Yoon, K., Hobbs, C.A., Koch, J., Sardaro, M., Kutny, R. and Weis, A. (1992) Proc. Natl. Acad. Sci. USA89, 3840-3844.MEDLINE Abstract
D. A. Rusling, V. E. C. Powers, R. T. Ranasinghe, Y. Wang, S. D. Osborne, T. Brown, and K. R. Fox Four base recognition by triplex-forming oligonucleotides at physiological pH
Nucleic Acids Res.,
May 23, 2005;
33(9):
3025 - 3032.
[Abstract][Full Text][PDF]
CiteULike 
Connotea 
Del.icio.us What's this?
