Skip to main content
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2003 Feb 6;100(4):1569–1573. doi: 10.1073/pnas.0335459100

DNA duplex–quadruplex exchange as the basis for a nanomolecular machine

Patrizia Alberti 1, Jean-Louis Mergny 1,*
PMCID: PMC149873  PMID: 12574521

Abstract

There is currently great interest in the design of nanodevices that are capable of performing linear or rotary movements. Protein molecular machines are abundant in biology but it has recently been proposed that nucleic acids could also act as nanomolecular machines in model systems. Several types of movements have been described with DNA machines: rotation and “scissors-like” opening and closing. Here we show a nanomachine that is capable of an extension–contraction movement. The simple and robust device described here is composed of a single 21-base oligonucleotide and relies on a duplex–quadruplex equilibrium that may be fueled by the sequential addition of DNA single strands, generating a DNA duplex as a by-product. The interconversion between two well defined topological states induces a 5-nm two-stroke, linear motor-type movement, which is detected by fluorescence resonance energy transfer spectroscopy.

Keywords: molecular machine‖FRET‖DNA structure‖G-quartet


A molecular-level machine may be defined as an assembly of a number of molecules that are designed to perform movements (1). Molecular self-assembly is an important route toward the construction of artificial molecular-level devices (2). These devices are characterized by the energy source, the nature of the movement, the way it may be controlled, its repeatability, its purpose, and the time scale of the nanometer-scaled conformational changes. Although protein machines are abundant (3), an attractive component for molecular recognition is DNA, because of its self-assembly properties (4, 5). Complex molecular objects (6) or quasicrystals may be constructed from simple double helices. However, DNA is prone to structural polymorphism (7), potentially expanding the repertoire of nanostructures that may be formed with this nucleic acid. Among these unusual DNA structures, G-quadruplexes are of special interest because they have well defined conformations, are relatively stable under physiological conditions, are highly polymorphic, and are likely to form higher order structures such as G wires (811). They may also be involved in telomere DNA structure (1214) and telomerase inhibition (15).

A motion can result from a reversible equilibrium between two conformational states. Several types of movements have been described with DNA machines: rotation (4, 16) and “scissors-like” opening and closing (5). The transition between these two states may be induced by a change in experimental conditions or by the addition of a “DNA fuel” (refs. 5 and 16; see Note) that provides the energy source for this change. A change in buffer composition or temperature may lead to a transition in the chirality of a double helix, which in turn may induce a movement (4). Although other dynamic chemical devices that undergo analogous structural modifications have recently been proposed (17), they are not nucleic acid-based and use a totally different chemical stimulus as a propellant. The system presented here has a large stroke amplitude, may easily be followed by fluorescence, and is DNA encoded. The principle of operation is illustrated in Fig. 1A and corresponds to a simple two-step cycle, with two structurally defined endpoints. Furthermore, our nanomechanical device is “robust” in the sense defined by Seeman et al. (16) because it oscillates between two well defined conformational states.

Figure 1.

Figure 1

Presentation of the system. (A) Switching between an intramolecular quadruplex (left) and a duplex (right). An intramolecular quadruplex is formed by the folding of a 21-base oligonucleotide that contains four blocks of three guanines mimicking the vertebrate telomeric motif. The schematic topology proposed here corresponds to the Na+ solution structure (19). The K+ motif would also lead to a closely similar 5′–3′ distance (18). The 5′ fluorescein and 3′ tetramethylrhodamine groups are depicted by green and orange triangles, respectively. The C-fuel strand is complementary to the F21T sequence, with six extra bases allowing duplex nucleation with the G-fuel strand. (B) Induced movement. Dethreading/rethreading of the DNA device is reminiscent of the movement of a piston in a cylinder. Double-stranded DNA has a persistence length of 100 bp or more (33), indicating that a 21-bp duplex should behave like a rigid rod.

Intramolecular folding of a guanine-rich oligonucleotide into an intramolecular G-quadruplex leads to a compact structure stabilized by monocations. It has recently been demonstrated that the potassium-based telomeric quadruplex (18) may be very different from the sodium-stabilized telomeric structure (19). Nevertheless, both structures are relatively compact as compared with a DNA duplex involving the same number of base pairs. The different conformations can be detected by fluorescence resonance energy transfer (FRET) spectroscopy. FRET has given valuable information on the structure of nucleic acids, because of its distance and orientation dependence. In this case, dye quenching was used to monitor the state of the machine, using fluorescein as a donor and tetramethylrhodamine as an acceptor, covalently attached to the 5′ and 3′ ends of the DNA, respectively (20). Folding of the oligonucleotide was confirmed by UV absorption, UV melting, and circular dichroism experiments (20). The sensitivity and dynamic response of fluorescence spectroscopy allows the monitoring of intramolecular folding over a wide range of concentrations [down to 100 pM (21)], and the presence of even a large excess of nonfluorescent fuel DNA does not interfere with FRET measurements. In previous experiments, a single-strand-quadruplex equilibrium was followed by this method (20). In this manuscript, we are interested in an equilibrium between an intermolecular duplex and an intramolecular quadruplex. The switching event should lead to an atomic displacement of ≈5 nm that would easily be demonstrated by FRET (Fig. 1B).

Methods

Oligonucleotides.

All oligonucleotides and their fluorescent conjugates were synthesized and purified by Eurogentec, Brussels, except for the 24-base, G-rich oligonucleotide (24Gmor) morpholino (Gene Tools, Philomath, OR). The primary sequences and names of the fluorescent oligonucleotides are given in Fig. 2.

Figure 2.

Figure 2

Sequence of the oligonucleotides used in this study. Names are given on the right side. (Top) G-quadruplex forming oligonucleotide. F, fluoresceine; T, tetramethylrhodamine. The 3′ overhang (6–12 bases long) present on the C-fuel strands is shown with lowercase letters. Mismatched bases (with respect to the F21T oligonucleotide) are underlined and boldfaced. The 5′ overhang on the G-fuel strand is shown with lowercase letters. For the 24Gmut control oligonucleotide, this overhang is not complementary to the last six bases of the C-fuel strand and is therefore presented with underlined and bold letters.

UV Absorption.

Spectra were obtained as described (21). A major issue in these experiments is the stoichiometry between the different strands. A small error in the concentration of C- or G-fuel strand would lead to the accumulation of undesired single-stranded species that would poison the machine, as already observed (5). Even with the help of nearest-neighbor extinction coefficients (22), it is relatively difficult to measure DNA strand concentration with >10% accuracy. This uncertainty may be partially relieved by performing isothermal titration curves or Job plots (i.e., using the continuous variation method) with two complementary strands to assess precise molar ratios between the two strands of a duplex.

UV Melting Curves.

Formation and stability of the different duplexes and quadruplexes were estimated by UV melting experiments, recording the absorbance at 245, 295, or 260 nm as a function of temperature on a Uvikon (Kontron, Zurich) spectrophotometer.

Fluorescence Studies.

Spectra were obtained as described (21).

Results

Quadruplex Formation.

F21T is a 21-base oligonucleotide that mimics the vertebrate telomeric repeats and contains four blocks of three guanines. Fluorescein and rhodamine were covalently attached to the 5′ and 3′ ends, respectively. Formation of the F21T quadruplex was first demonstrated by UV-absorbance (23) and fluorescence (20) thermal denaturation. Intramolecular folding was confirmed by the concentration independency of the melting temperature, which was determined to be 33, 44, and 57°C in 0.1 M LiCl, NaCl, and KCl, respectively (not shown). As expected (14), the nature of the monocation determines the stability of the quadruplex (K+ > Na+ > Li+). These values are slightly lower than the melting temperatures of the unmodified 21-base d(GGGTTA)3GGG oligonucleotide, showing that the attachment of fluorescent groups at both ends of the oligonucleotide slightly destabilizes the stability of the quadruplex. From these melting profiles, one can deduce that quadruplexes are the dominant (and often the exclusive) species under most of the experimental conditions tested, except at temperatures >37°C in 0.1 M LiCl. For example, it is possible to determine the Gibb's free energy (ΔG°) of quadruplex formation at 37°C in 0.1 M KCl for the F21T oligomer (20): this value is negative (−3.8 kcal⋅mol−1), indicating that in the absence of a fuel strand the preferred (>99%) conformational state is the folded conformation.

Quadruplex-to-Duplex Transition: The “Opening” Step.

Starting from a folded quadruplex, the two flexible 5′ and 3′ ends of this device may be pushed apart by hybridization with the complementary strand, called the C-fuel strand (Fig. 1A, upper part of the cycle). The important point of this study was to determine whether a cytosine-rich complementary telomeric strand could disrupt a preformed quadruplex and hybridize with its complementary strand under various experimental conditions. This hypothesis was tested at four different temperatures in a pH 7.2 sodium cacodylate buffer supplemented with different types of monocations. The formation of the duplex was followed using the emission of the fluorescein group of F21T as a marker of quadruplex–duplex formation. In the duplex form, the emission at 520 nm is seven to nine times higher than in the intramolecular quadruplex form because of the large increase in distance between the fluorescence donor and acceptor groups. Fig. 3 presents the experimental results in the presence of 0.1 M lithium (Middle Left), sodium (Middle Right), or potassium chloride (Bottom). In lithium, on addition of d(CCCTAA)3CCC to the quadruplex oligonucleotide, duplex formation is fast; one should note that at 37 and 45°C the starting oligonucleotide is not fully folded into a quadruplex, making duplex formation easy, because no secondary structure needs to be disrupted for hybridization to occur. This result is in agreement with previous UV melting studies which showed that the human telomeric repeat did not form a stable structure in the presence of lithium (23). In contrast, in the presence of sodium or potassium, the F21T oligonucleotide was fully folded at all temperatures (4–45°C) before C-fuel addition. The addition of C-fuel induced a large increase in fluorescence emission at 520 nm, and a plateau was reached after 1 h in potassium at 37°C. Duplex formation was significantly faster in sodium than in potassium, as a likely result of a more stable quadruplex formed with the latter ion.

Figure 3.

Figure 3

The opening step. (Top) Principle of the experiment. Quadruplex-to-duplex conversion is monitored by an increase in fluorescence emission at 520 nm. The 5′ fluorescein and 3′ tetramethylrhodamine groups are depicted by green and orange triangles, respectively. (Middle and Bottom) Effect of temperature on quadruplex-to-duplex conversion. All experiments were performed at four different temperatures in a 10 mM sodium cacodylate pH 7.2 buffer containing 0.1 M LiCl (Middle Left), 0.1 M NaCl (Middle Right), or 0.1 M KCl (Bottom). The predominant starting structure (t = 0 s) before C-fuel strand addition is always an intramolecular quadruplex, except at 45°C in the presence of Li+, where F21T is mainly single-stranded.

These results indicate that, at the chosen oligonucleotide concentrations (in the μM range), the duplex is the thermodynamically favored species and that folding of the F21T oligonucleotide into a quadruplex delays, but does not prevent, formation of a Watson–Crick duplex (24). A similar conclusion was reached in the case of other quadruplex sequences with DNA or PNA (peptide nucleic acids; ref. 25) complementary sequences. The increased stability of the duplex as compared with the quadruplex was confirmed by UV-melting analysis of the F21T/C-fuel duplexes under a variety of conditions (not shown). In contrast to the quadruplex case, the nature of the monocation plays little, if any, role in the stability of the duplex.

To promote quadruplex opening, a 21-base d(CCCTAA)3CCC oligonucleotide is sufficient. However, the reversal step requires a longer C-fuel strand length, to initiate C-fuel/G-fuel duplex formation on a single-stranded overhang (see below). We therefore analyzed F21T quadruplex opening by longer C-fuel strands, which contain 6, 9, or 12 extra bases at their 3′ end. The presence of a 6- to 12-base overhang does not hamper F21T/C-fuel duplex formation and does not significantly alter the kinetics of quadruplex-to-duplex conversion (data not shown).

We next investigated whether quadruplex-specific ligands could poison the opening step. When added to a preformed quadruplex, a quadruplex ligand (BisA) is able partially to inhibit duplex formation (26). In contrast, the addition of an inactive molecule [Mono A (26)] to the G-quadruplex before 21C addition has little, if any, effect on duplex formation. This is likely to be the result of the higher affinity of BisA toward G-quadruplexes that locks the F21T oligonucleotide into a conformation that is no longer recognized by the complementary strand (26). Therefore, the proper functioning of the device may be altered by structure-specific DNA ligands.

Duplex-to-Quadruplex Transition: The “Closing” Step.

We are now interested in the second part of the machine cycle, which consists of the reverse reaction, starting from the C-fuel/F21T duplex and liberating the F21T strand, allowing intramolecular quadruplex reformation. The assembled duplex may be reopened with another DNA strand called the G-fuel strand. The isothermal denaturation of the F21T/C-fuel duplex (bottom part of the cycle shown in Fig. 1A) involves the disassembling of a thermodynamically stable duplex. The destruction of the noncovalent interaction between the two strands (F21T and C-fuel) may be obtained by the formation of a thermodynamically more stable C-fuel/G-fuel competing duplex.

A 24G, which lacks the three terminal guanines (and is therefore unable to form a competing intramolecular quadruplex), was initially tested for its ability to bind to the C-fuel strand. Hybridization between the G- and C-fuel strands is expected to start first at the exposed six-base overhang and to proceed by branch migration. Sequence complementarity in that region is mandatory, because a mutated G-fuel strand (24Gmut) cannot properly unfold the F21T/C-fuel duplex (not shown). However, the 27C/24G duplex is not more stable than the F21T/27C duplex (Tm = 60°C). Complete displacement of this duplex is therefore unlikely, prompting us to use different pairs of C-fuel/G-fuel strands that should favor the C-fuel/G-fuel duplex over the F21T/C-fuel duplex.

Increasing the length of the overhang by using longer C- and G-fuel oligonucleotides should lead to the formation of a longer (and more stable) duplex. 30C/27G and 33C/30G hybrids have a higher melting temperature than the F21T/C-fuel duplex; however, very slow kinetics of F21T/C-fuel duplex disruption are obtained; increasing the length of the overhang to 9 or 12 bases does not accelerate this step (data not shown). We then reasoned that any modification that stabilizes the G-fuel/C-fuel duplex without affecting the stability of the F21T/C-fuel duplex would facilitate the closing step. Unfortunately, the use of chemical modifications in the G-fuel strand [2′-O-methyl groups, morpholino groups, locked nucleic acids (LNA), etc.] does not accelerate the closing step (data not shown).

To speed up this closing step, we chose to introduce a few mismatched bases in the F21T/C-fuel duplex. The C-fuel strand 27Cm3C is no longer perfectly complementary to the F21T sequence. The corresponding duplex that involves three mismatches (out of 21 base pairs) may still be formed, but with a lower stability (Tm = 44°C in 100 mM KCl/10 mM cacodylate, pH 7.2). On the other hand, the G-fuel strand 24Gm3G is fully complementary to the C-fuel strand 27Cm3C, leading to a 24-bp perfect duplex. As a further bonus, the G-fuel strand no longer contains several blocks of guanines and is unlikely to form a stable quadruplex. The monovalent ion concentration is kept at 0.1 M and different concentrations of multivalent ions are tested. Efficient closing of the quadruplex was obtained in the absence of multivalent ions, but the presence of 10–20 mM MgCl2 or 0.2–0.5 mM spermine significantly improved the kinetics (Fig. 4A). All further experiments were therefore performed in the presence of 20 mM MgCl2. As bimolecular association was considered, we verified that increasing all strand concentrations led to faster switching kinetics (Fig. 4B). A 10-fold increase in the concentrations of the three different strands had a significant impact on the opening and closing steps. At the highest strand concentration tested (2, 2.5, and 2.5 μM for F21T, C-fuel, and G-fuel, respectively), a switching half time <5 s could be estimated for closing, as compared with 20 s for opening. As expected for a bimolecular phenomenon, switching to the closed state was ≈10 times slower at lower strand concentrations (0.2, 0.25, and 0.25 μM).

Figure 4.

Figure 4

Optimization of the device. (A) Effect of magnesium. Only one cycle is presented. The closing step is much faster in the presence (solid line) than in the absence (dashed line) of 20 mM MgCl2. The experiment was performed at 37°C in a 100 mM NaCl/10 mM sodium cacodylate (pH 7.2) buffer. F21T, 33C, and 30G oligonucleotide concentrations were 0.2, 0.25, and 0.25 μM, respectively. (B) Concentration dependency. All experiments were performed at 45°C in a 20 mM MgCl2/100 mM KCl/10 mM sodium cacodylate (pH 7.2) buffer. On the dotted line, the concentrations for F21T, C-fuel (3′-TGCAATGCCAATCGCAATCGCAATCCC-5′ 27Cm3C), and G-fuel (5′-ACGTTACGGTTAGCGTTAGCGTTA-3′ 24Gm3G) are 0.2, 0.25, and 0.25 μM, respectively. The full line presents a similar experiment with a 10-fold increase in all oligonucleotide concentrations (normalized fluorescence intensity).

Cycling the Device.

The device described here oscillates between two well defined states, a folded intramolecular quadruplex and an intermolecular duplex. Each device cycle leads to the creation of a single 24-bp duplex as a “waste product.” The loss of cycling efficiency due to the accumulation of the waste product will depend on the equilibrium constant between the closed and the opened state. Under the chosen conditions (Fig. 5A), the closed state is strongly favored and the accumulation of the C-fuel/G-fuel duplex does not hamper proper cycling of the machine for at least 11 successive cycles, provided that the C-fuel/G-fuel strand ratio is precisely controlled. Because strand concentrations are usually determined with a 10% or more uncertainty, using calculated extinction coefficients, it is necessary to preestablish molar equivalence between these two strands by preliminary UV-absorbance titration profiles. Otherwise, the accumulation of a slight excess of the C- or G-fuel strand eventually leads to the poisoning of the machine.

Figure 5.

Figure 5

Cycling the device. (A) With DNA strands. By alternatively adding stoichiometric amounts of the C-fuel (27Cm3C; 2.5 μM) and G-fuel (24Gm3G; 2.5 μM) strands, F21T may be opened and closed repeatedly (at least 11 times); for purposes of clarity, the times of C- and G-fuel additions are indicated for only one cycle. Each addition of the C- or G-fuel strand results in a 0.5% dilution of the reactants, which is mathematically corrected on this graph. Experimental conditions: 100 mM KCl/20 mM MgCl2/10 mM sodium cacodylate (pH 7.2) at 45°C. η, the average cycling efficiency (3% loss per cycle). (B) Cycling with a modified oligonucleotide. The G-fuel strand is a chemically modified oligonucleotide (morpholino, with a neutral backbone, synthesized by Gene Tools). Nine successive cycles are shown. Despite a careful assessment of strand concentrations, each successive cycle leads to a significant loss in signal. The experiment was performed in a 100 mM KCl/10 mM sodium cacodylate (pH 7.2) buffer at 45°C (no magnesium). Oligonucleotide sequences are shown in C.

The switching time for this machine may be modulated by a number of factors, such as temperature, nature of monovalent cation, ionic strength, presence of divalent cations, sequence and chemical modification of the strand(s), and strand concentration. Fig. 3 illustrates the importance for the opening step of temperature and the nature of the monovalent ion. It is obvious from these profiles that, to keep a short cycling time, higher temperatures (37–45°C) should be used. Fig. 4 illustrates the effect of concentration: using strand concentrations in the micromolar range leads to acceptable kinetics. From these data, and from the comparison of different oligonucleotide sequences, it is possible to design an experimental system in which the machine has a relatively fast cycle (Figs. 4 and 5). Interestingly, these conditions of moderate ionic strength (10 mM sodium cacodylate/100 mM KCl/20 mM MgCl2) and mild temperature (37 or 45°C) are not very distant from physiological conditions. Oligonucleotides of different chemistry (morpholino group) may also be used, alleviating the need for high magnesium or spermine concentrations (Fig. 5B).

Conclusion

Several reports suggest that DNA may be used as a building block for novel nanosized objects (6), including 2D crystals (27), and as controlling elements in nanorobotics (28) and DNA computers (29). DNA is an excellent molecule for the design of such devices, because of its sequence-specific recognition properties and easy chemical synthesis. In this manuscript, we report the construction of a DNA machine in which DNA is used both as a structural material and as a fuel. The “chameleon tongue” movement performed by such a type(s) of device(s) complements the rotational and angular movements performed by previously discovered DNA-based machines. The use of a DNA fuel allows the precise control of movements. The 5′–3′ distance oscillates between 1.5 nm (quadruplex) and 7 nm, with a calculated force of ≈8 pN. In many examples, both ends of the device are coupled to fluorescent dyes to monitor conformational transitions. However, at least one end of the device may be linked to another organic or inorganic component, adding novel functionalities or reactivities to be useful in biotechnology. For example, biotin labeling of the G-rich telomeric strand does not prevent quadruplex formation or duplex conversion after addition of the C-fuel strand (J.-L.M., unpublished data). Furthermore, the telomeric sequence motif chosen for these experiments may be replaced by other quadruplex forming motifs, the essential sequence restriction being the presence of four blocks of two to four guanines, i.e., GaNxGaNyGaNzGa (with 2 ≤ a ≤ 4, 3 ≤ x, y, z ≤ 6). As a result, variable lengths of the extended duplex may be obtained, leading to duplexes involving 17–42 base pairs (see Note). Even longer devices could be designed with strands containing eight or more guanine repeats, with several intramolecular quadruplexes stacking on each other, forming higher-order DNA structures with little topological difficulty (18). Specific sequence recognition may be achieved by DNA-encoded complementarity in the three variable loops and on the 3′ extension, allowing the simultaneous and specific control of different G4-based nanomechanical devices. Each G-rich oligonucleotide may then be individually addressed using a specific C-fuel/G-fuel pair. This system could be used to approach alternatively or separate chemically reactive groups (30), or to obtain precise control of movements on the nanometer scale.

Acknowledgments

We thank L. Lacroix, L. Guittat, P. B. Arimondo, T. Garestier, and C. Hélène (MNHN, Paris) for helpful discussions. This work was supported by Institut National de la Santé et de la Recherche Médicale, Centre National de la Recherche Scientifique, Association pour la Recherche sur le Cancer Grant 4321 (to J.-L.M.), and an Aventis Pharma S.A. research grant.

Abbreviations

FRET

fluorescence resonance energy transfer

24G

24-base, G-rich oligonucleotide

Note

After submission of this manuscript, we became aware of a recent paper reporting a duplex–quadruplex equilibrium based on a different sequence motif (31). A very recent review summarizes findings concerning nanomechanical devices based on DNA (32).

Footnotes

This paper was submitted directly (Track II) to the PNAS office.

References

  • 1.Balzani V V, Credi A, Raymo F M, Stoddart J F. Angew Chem Int Ed Engl. 2000;39:3348–3391. doi: 10.1002/1521-3773(20001002)39:19<3348::aid-anie3348>3.0.co;2-x. [DOI] [PubMed] [Google Scholar]
  • 2.Balzani V, Credi A, Venturi M. Proc Natl Acad Sci USA. 2002;99:4814–4817. doi: 10.1073/pnas.022631599. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Alberts B. Cell. 1998;92:291–294. doi: 10.1016/s0092-8674(00)80922-8. [DOI] [PubMed] [Google Scholar]
  • 4.Mao C D, Sun W Q, Shen Z Y, Seeman N C. Nature. 1999;397:144–146. doi: 10.1038/16437. [DOI] [PubMed] [Google Scholar]
  • 5.Yurke B, Tuberfield A J, Mills A P, Jr, Simmel F C, Neunmann J L. Nature. 2000;406:605–608. doi: 10.1038/35020524. [DOI] [PubMed] [Google Scholar]
  • 6.Chen J H, Seeman N C. Nature. 1991;350:631–633. doi: 10.1038/350631a0. [DOI] [PubMed] [Google Scholar]
  • 7.Rich A. Gene. 1993;135:99–109. doi: 10.1016/0378-1119(93)90054-7. [DOI] [PubMed] [Google Scholar]
  • 8.Aboul-ela F, Murchie A I H, Lilley D M J. Nature. 1992;360:280–282. doi: 10.1038/360280a0. [DOI] [PubMed] [Google Scholar]
  • 9.Williamson J R. Annu Rev Biophys Biomol Struct. 1994;23:703–730. doi: 10.1146/annurev.bb.23.060194.003415. [DOI] [PubMed] [Google Scholar]
  • 10.Marsh T C, Henderson E. Biochemistry. 1994;33:10718–10724. doi: 10.1021/bi00201a020. [DOI] [PubMed] [Google Scholar]
  • 11.Batalia M A, Protozanova E, MacGregor R B, Jr, Erie D A. Nano Lett. 2002;2:269–274. [Google Scholar]
  • 12.Sen D, Gilbert W. Nature. 1988;334:364–366. doi: 10.1038/334364a0. [DOI] [PubMed] [Google Scholar]
  • 13.Sundquist W I, Klug A. Nature. 1989;342:825–829. doi: 10.1038/342825a0. [DOI] [PubMed] [Google Scholar]
  • 14.Williamson J R, Raghuraman M K, Cech T R. Cell. 1989;59:871–880. doi: 10.1016/0092-8674(89)90610-7. [DOI] [PubMed] [Google Scholar]
  • 15.Neidle S, Parkinson G. Nat Rev Drug Des. 2002;1:383–393. doi: 10.1038/nrd793. [DOI] [PubMed] [Google Scholar]
  • 16.Yan H, Zhang X, Shen Z, Seeman N C. Nature. 2002;415:62–65. doi: 10.1038/415062a. [DOI] [PubMed] [Google Scholar]
  • 17.Barboiu M, Lehn J M. Proc Natl Acad Sci USA. 2002;99:5201–5206. doi: 10.1073/pnas.082099199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Parkinson G N, Lee M P H, Neidle S. Nature. 2002;417:876–880. doi: 10.1038/nature755. [DOI] [PubMed] [Google Scholar]
  • 19.Wang Y, Patel D J. Structure (London) 1993;1:263–282. doi: 10.1016/0969-2126(93)90015-9. [DOI] [PubMed] [Google Scholar]
  • 20.Mergny J L, Maurizot J C. Chembiochem. 2001;2:124–132. doi: 10.1002/1439-7633(20010202)2:2<124::AID-CBIC124>3.0.CO;2-L. [DOI] [PubMed] [Google Scholar]
  • 21.Mergny J L. Biochemistry. 1999;38:1573–1581. doi: 10.1021/bi982208r. [DOI] [PubMed] [Google Scholar]
  • 22.Cantor C R, Warshaw M M, Shapiro H. Biopolymers. 1970;9:1059–1077. doi: 10.1002/bip.1970.360090909. [DOI] [PubMed] [Google Scholar]
  • 23.Mergny J L, Phan A T, Lacroix L. FEBS Lett. 1998;435:74–78. doi: 10.1016/s0014-5793(98)01043-6. [DOI] [PubMed] [Google Scholar]
  • 24.Phan A T, Mergny J L. Nucleic Acids Res. 2002;30:4618–4625. doi: 10.1093/nar/gkf597. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Datta B, Armitage B A. J Am Chem Soc. 2001;123:9612–9619. doi: 10.1021/ja016204c. [DOI] [PubMed] [Google Scholar]
  • 26.Alberti P, Ren J, Teulade-Fichou M P, Guittat L, Riou J F, Chaires J B, Hélène C, Vigneron J P, Lehn J M, Mergny J L. J Biomol Struct Dyn. 2001;19:505–513. doi: 10.1080/07391102.2001.10506758. [DOI] [PubMed] [Google Scholar]
  • 27.Seeman N C. Curr Opin Struct Biol. 1996;6:519–526. doi: 10.1016/s0959-440x(96)80118-7. [DOI] [PubMed] [Google Scholar]
  • 28.Seeman N C, Belcher A M. Proc Natl Acad Sci USA. 2002;99:6451–6455. doi: 10.1073/pnas.221458298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Adleman L M. Science. 1994;266:1021–1024. doi: 10.1126/science.7973651. [DOI] [PubMed] [Google Scholar]
  • 30.Gartner Z J, Kanan M W, Liu D R. Angew Chem Int Ed Engl. 2002;41:1796–1800. doi: 10.1002/1521-3773(20020517)41:10<1796::aid-anie1796>3.0.co;2-z. [DOI] [PubMed] [Google Scholar]
  • 31.Li J J, Tan W. Nano Lett. 2002;2:315–318. [Google Scholar]
  • 32.Niemeyer C M, Adler M. Angew Chem Int Ed Engl. 2002;41:3779–3783. doi: 10.1002/1521-3773(20021018)41:20<3779::AID-ANIE3779>3.0.CO;2-F. [DOI] [PubMed] [Google Scholar]
  • 33.Manning G S. Biopolymers. 1981;20:1751–1755. [Google Scholar]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES