TETRAMOLECULAR PARALLEL G-QUADRUPLEX-FORMING HYDROPHOBICALLY MODIFIED OLIGONUCLEOTIDES

Abstract

The present invention relates to tetramolecular parallel G-quadruplex-forming oligonucleotides. If G-quadruplexes are of prime importance in biology, their use is hampered by the propensity of G4-prone DNA molecules, in particular G4-prone DNA molecules of long size, to adopt many different G4 topological conformations or other alternative foldings. By introducing a lipid modification at the end of the oligonucleotide, the inventors succeeded in obtaining long tetramolecular parallel G-quadruplexes (tpG4). The present invention thus concerns an oligonucleotide modified by substitution at the 5 or the 3 end by a lipid moiety, wherein said oligonucleotide comprises a nucleic acid sequence of at least 10 nucleotides, said nucleic acid sequence including a series of at least 4 consecutive guanine residues located in the middle of said sequence. A tetramolecular parallel G-quadruplex comprising 4 identical modified oligonucleotides as defined above, wherein each of the 4 consecutive guanine residues included in the middle of the nucleic acid sequence of each oligonucleotide respectively form G-quartets with the corresponding guanine residues of the other 3 oligonucleotides, said G-quartets being stabilized by - staking and Hoogsteen hydrogen bonding, is also contemplated. The modified oligonucleotides have preferably the general formula (I) or (II), wherein the oligonucleotides are modified by substitution at the 5 or the 3 end by a lipid moiety, and said oligonucleotides comprise a nucleic acid sequence of at least 10 nucleotides, said nucleic acid sequence including a series of at least 4 consecutive guanine residues located in the middle of said sequence.

##STR00001##

Claims

1. An oligonucleotide modified by substitution at the 5 or the 3 end by a lipid moiety, wherein said oligonucleotide comprises a nucleic acid sequence of at least 10 nucleotides, said nucleic acid sequence including a series of at least 4 consecutive guanine residues located in the middle of said sequence.

2. The modified oligonucleotide according to claim 1, wherein the lipid moiety is selected from (i) a moiety comprising at least one saturated or unsaturated, linear or branched hydrocarbon chain comprising from 2 to 60 carbon atoms, and (ii) a moiety comprising at least one ketal functional group, wherein the ketal carbon of said ketal functional group bears two saturated or unsaturated, linear or branched, hydrocarbon chains comprising from 1 to 22 carbon atoms.

3. The modified oligonucleotide according to claim 1, of the general formula (I) ##STR00014## wherein: Oligo represents a nucleic acid sequence of at least 10 nucleotides, said nucleic acid sequence including a series of at least 4 consecutive guanine residues located in the middle of said sequence, wherein said nucleic acid sequence may be oriented 3-5 or 5-3 and/or may comprise modified nucleotides; X represents a divalent linker moiety selected from ether O, thio S, amino NH, and methylene CH.sub.2; R.sub.1 and R.sub.2 may be identical or different and represent: a hydrogen atom, a halogen atom, in particular fluorine atom, a hydroxyl group, or an alkyl group comprising from 1 to 12 carbon atoms; M.sub.1, M.sub.2 and M.sub.3 may be identical or different and represent: a hydrogen atom, a saturated or unsaturated, linear or branched hydrocarbon chain comprising from 2 to 30 carbon atoms, which may be substituted by one or more halogen atoms, notably be fluorinated or prefluorinated and/or be interrupted by one or more groups selected from ether O, thio S, amino NH, oxycarbonyl OC(O), thiocarbamate OC(S)NH, carbonate OC(O)O, carbamate OC(O)NH, phosphate OP(O)(O)O and phosphonate PO(O)(O) groups; and/or be substituted at the terminal carbon atom by an aliphatic or aromatic, notably benzylic or naphtylic ester or ether group; an acyl radical with 2 to 30 carbon atoms, or an acylglycerol, sphingosine or ceramide group, provided that at least one of M.sub.1, M.sub.2 and M.sub.3 is not a hydrogen atom.

4. The modified oligonucleotide according to claim 1, of the general formula (II) ##STR00015## wherein: Oligo represents a nucleic acid sequence of at least 10 nucleotides, said nucleic acid sequence including a series of at least 4 consecutive guanine residues located in the middle of said sequence, wherein said nucleic acid sequence may be oriented 3-5 or 5-3 and/or may comprise modified nucleotides; Y represents a divalent linker moiety selected from ether O, thio S, amino NH, and methylene CH.sub.2; R.sub.3 and R.sub.4 may be identical or different and represent: a hydrogen atom, a halogen atom, in particular fluorine atom, a hydroxyl group, or an alkyl group comprising from 1 to 12 carbon atoms; L.sub.1 and L.sub.2 may be identical or different and represent a saturated or unsaturated, linear or branched hydrocarbon chain comprising from 1 to 22 carbon atoms; B is an optionally substituted nucleobase, selected from the group consisting of purine nucleobases, pyrimidine nucleobases, and non-natural monocyclic or bicyclic heterocyclic nucleobases wherein each cycle comprises from 4 to 7 atoms.

5. The modified oligonucleotide according to claim 1, wherein said oligonucleotide consists of a DNA sequence of 19 nucleotides.

6. The modified oligonucleotide according to claim 1, wherein said oligonucleotide consists of the sequence 5-TTAGTTGGGGTTCAGTTGG-3 (SEQ ID NO: 1).

7. A tetramolecular parallel G-quadruplex comprising 4 identical modified oligonucleotides as defined in claim 1, wherein each of the 4 consecutive guanine residues included in the middle of the nucleic acid sequence of each oligonucleotide respectively form G-quartets with the corresponding guanine residues of the other 3 oligonucleotides, said G-quartets being stabilized by rrT-r staking and Hoogsteen hydrogen bonding.

8. The tetramolecular parallel G-quadruplex according to claim 7, wherein said G-quadruplex further comprises a divalent cation which coordinate said G-quartets.

9. The tetramolecular parallel G-quadruplex according to claim 8, wherein said divalent cation is a Mg.sup.2+ cation.

10. A composition comprising tetramolecular parallel G-quadruplexes according to claim 7, wherein the tetramolecular parallel G-quadruplexes self-assembled into micelles.

11. The composition according to claim 10 further comprising a hydrophobic active principle hosted in said micelles.

12-15. (canceled)

16. A method for administrating a medicinally and/or pharmaceutically active substance, said comprising the use of the composition according to claim 10 as a vehicle for said medicinally and/or pharmaceutically active substance.

17. A method for treating a subject in need thereof, comprising administering to a subject in need thereof a therapeutically effective amount of a composition according to claim 10.

18. A method of manufacture of nanotechnology devices, said method comprising the use of a tetramolecular parallel G-quadruplex according to claim 7.

19. An artificial implant comprising a tetramolecular parallel G-quadruplex according to claim 7.

20. Vehicle comprising the composition according to claim 10.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0125] FIG. 1: Native PAGE of the different (L)ONs synthesized in the example in the presence of 1X salts (except lane 3). The absence of band for k-LON.sup.G4 in lane 7 results from the formation of stable micellar aggregates (the other micellar aggregates do not survive the PAGE conditions). 1X salts=50 mM Nacl, 5 mM KCl, 5 mM MgCl.sub.2.

[0126] FIG. 2: CD/NMR signatures of 1-LON.sup.G4, 1-LON.sup.SC, ON.sup.G4 and ON.sup.SC of the example (Conditions: [LON]=5 M, diluted from original 30 M solutions, 1X salts, phosphate buffer pH 6.9, 20 mM).

[0127] FIG. 3: Gel retardation of 1-LON.sup.G4. Lane 1: no added salt; Lane 2: 1X; Lane 3: LiCl: 55 mM LiCl, 5 mM MgCl.sub.2; Lane 4: 4X; Lane 5: LiCl 4X; Lane 6: 4X; Lane 7: KCl 0.3M; Lane 8: KCl 1.2M; Lane 8: AcONH.sub.4 0.05M; Lane 9: AcONH.sub.4 0.2M; Lane 10: control intramolecular 77-mer G4.

[0128] FIG. 4: Native agarose gel of the different (L)ONs studied in the example (see FIG. 1 for salt conditions).

[0129] FIG. 5: DLS/TEM analysis of 1-LON.sup.G4 of the example.

[0130] FIG. 6: Effect of temperature cycles form 1-LON.sup.G4

[0131] FIG. 7: Conceptual scheme showing the lipid-driven assembly of large tetramolecular parallel G-quadruplexes.

[0132] FIG. 8: Chemical scheme of a G-quartet.

EXAMPLE

[0133] This example shows that the lipid modification of a G4 prone oligonucleotide sequence with lipids drastically increases the probability of forming tetramolecular parallel G-quadruplexes over other unspecific oligomers.

Materials and Methods

1. Automated DNA Synthesis and Purification of LONs

[0134] LONs were synthesized using the phosphoramidite methodology on an automated Expedite 8909 DNA synthesizer at the mol scale on 1000 primer support (loading: 30-100 mol/g, Link technologies, Synbase Control Pore Glass). Prior to use, the phosphoramidites 1 of formula (III)

##STR00010##

(to obtain 1-LONs) and 2 of formula (IV)

##STR00011##

(to obtain k-LONs) were dried over P.sub.2O.sub.5 overnight and then dissolved in dry CH.sub.2Cl.sub.2/CH.sub.3CN 1/1 to a 0.1 M concentration (lipidic phosphoramidites did not dissolve in pure acetonitrile). N-benzylthiotetrazole was used for activation of the phosphoramidite prior to coupling. The phosphoramidites 1 and 2 were manually coupled last on the solid support by passing (via syringes) the activator and the phosphoramidite solution (0.25 mL) back and forth several times for 7 min. Deblocking and detachment from the solid support was achieved using 1 mL of a saturated aqueous NH.sub.4OH solution for 12 h at 55 C. The supernatant was collected and the CPG beads were washed 3 times with 0.25 mL of EtOH/CH.sub.3CN/H.sub.2O 3/1/1 (vol). The solutions were pooled and evaporated (speed vac). The crude LONs were dissolved in 0.3 mL of water and purified (except the k-LONs) on an analytical C4-reverse phase HPLC using buffer A (0.1 M triethylammonium acetate, pH 6.5) with 20% of buffer B (0.1 M triethylammonium acetate, pH 7.0, 80% acetonitrile) over 2 min, followed by buffer B over 50 min (flow rate: 2 ml/min). The product oligoamphiphiles eluted after ca. min. Product containing fractions were pooled and evaporated to dryness and dissolved in autoclaved milliQ water. The (L)ONs were then dialyzed against a 10 mM LiCl solution (not to favor G4 formation) followed by water. Yields of recovery were acceptable following this protocol (15-30% yield after purification of the crude material). Yields of final LON.sup.G4 (=182800) and LON.sup.SC (=184600).

[0135] k-LONs were purified by preparative PAGE using conventional protocols with 20*20*0.2 cm 20% polyacrylamide gels at a limiting power=15W. Importantly, the inventors found that the quantity of k-LON.sup.G4 that could be loaded on the gel did not exceed ca. 100 nmoles of crude material in ca. 300 L of loading sample. Higher amounts of crude LON led to substantial trailing of the LON band probably because of the formation of the micelles in that case even in the presence of heat+7M urea in the running buffer. Using smaller quantities, the k-LON.sup.G4 band is well defined (UV-shadow) and cut out directly with a clean scalpel. The gel slab was chopped into fine particles to elute the LON. The inventors have been unsuccessful at eluting k-LON.sup.G4 from the gel using different eluting buffers with additional heating (up to 90 C.) and/or sonication and/or freeze and thaw protocols. Small quantities of k-LON.sup.G4 were systematically obtained for each of these tests. The inventors therefore developed an original electroelution protocol for the purification of these LONs. In short, the electrical wire that was plugged in the negative pole of the generator was manually modified by wrapping a platinum wire around the naked copper wire. The latter was immersed in the eluting buffer contained in a plastic pipette that was chopped off at both edges to 1) facilitate pouring of the eluting buffer on top and 2) increase the cross section at the bottom to minimize resistance to the current flow. The bottom of the syringe was blocked by polymerizing 0.5 mL of a 8% polyacrylamide solution (the bottom of the pipette being temporarily blocked by wrapping a parafilm foil around). After polymerization, the gel was prerun to remove any unpolymerized materials prior to loading the crushed acrylamide gel slab containing the LON (the TBE eluting buffer from the pipette can be withdrawn beforehand for practical reasons). A dialysis tubing (with a cutoff of 2 kDa) was adapted and wrapped with a parafilm foil around the bottom of the pipette to recover the sample after electroelution. This allowed the dialysis tubing to be immersed in a large quantity of eluting buffer at the bottom in the electrophoresis tank. The elution was carried out with an electrical power of 7-10 W to allow enough heat dissipation in the gel to favor denaturation of the LONs. The LONs were finally dialyzed in a similar manner as described above for the other (L)ONs.

2. Mass Spectra Measurements

[0136] Mass spectra were recorded on a MALDI-Tof-ToF mass spectrometer (Ultraflex, Bruker Daltonics, Bremen, Germany). Best results were obtained in the linear mode with positive-ion detection. Mass spectra were acquired with an ion source voltage 1 of 25 kV, an ion source voltage 2 of 23.5 kV, a lens voltage of 6 kV, by accumulating the ion signals from 1000 laser shots at constant laser fluence with a 100 Hz laser. External mass calibration was achieved using a mixture of oligonucleotides dT.sub.12-dT.sub.18 (Sigma). 1:1 mixture of samples of LONs (20-50 M) and matrix was spotted on a MALDI target and air-dried before analysis. The performance of the following matrices was evaluated: 2,5-dihydroxybenzoic acid (DHB), 2,4,6-trihydroxy-acetophenone (THAP), 3-hydroxypicolinic acid (3-HPA) and 2,6-dihydroxyacetophenone (DHA). THAP at a concentration of 20 mg/mL in a 4:1 mixture of ethanol and 100 mM aq ammonium citrate was shown to yield optimal mass spectral results.

[0137] MALDI-TOF mass analyses:

[0138] 1-LON.sup.G4: [M-H] calculated MW: Da, found: Da (0.0% error);

[0139] 1-LON.sup.SC: [M-H] calculated MW: Da, found: Da (0.0% error).

[0140] k-LON.sup.G4: [M-H] calculated MW: Da, found: Da (0.0% error);

[0141] k-LON.sup.SC: [M-H] calculated MW: Da, found: Da (0.0% error).

[0142] ON.sup.SC: [M-H] calculated MW: 5920.9 Da, found: 5924.29 Da (0.056% error).

[0143] ON.sup.G4: [M-H] calculated MW: 5920.9 Da, found: 5924.22 Da (0.057% error).

3. PAGE/Aqarose Electrophoresis

[0144] Electrophoresis experiments were performed according to standard procedures with 1% agarose gels. PAGE were carried out with 17% polyacrylamide gels and run with a 100V limiting tension for native PAGE experiments.

4. Dynamic Light Scattering (DLS)

[0145] Particle size was determined using a Zetasizer 3000 HAS MALVERN. Experiments were realized with samples containing different concentration of LONs dissolved in deionized water or phosphate buffer. Measurements were performed at 25 C.

5. Taylor Dispersion Analysis (TDA, Discosizinq)

[0146] TDA analyses were recorded on a Viscosizer TD (Malvern Instruments Ltd., Malvern, UK equipped with a 254 nm UV filter close to the .sub.max of oligonucleotides. Prior to analysis, the non-coated capillary has been prepared by injecting 1M NaOH during 30 min at 3000 mbar followed by 10 min rinse with water at 3000 mbar. The cellulose coated capillary has been prepared by injecting water during 30 min at 3000 mbar.

[0147] Internal Material (non-coated): fused silica, internal diameter: 75 m, outer diameter: 360 m, L1: 45 cm, L2: 85 cm, Total Length: 130 cm.

[0148] The samples were injected (pressure: 50 mbar), analyzed at 25 C. (Mobilization pressure: 140 mbar). Taylorgramms were recorded and analyzed with the viscosizer TD software 2.01 with a one component fit. Washing: 1 min of water between each sample, pressure: 3000 mbar.

[0149] The coated cellulose capillary was in general necessary for the analysis of LONs with the noticeable exception of k-LON.sup.G4. Unless the LON forms stable micellar assemblies as in the case of k-LON.sup.G4, unspecific adsorption was observed onto the uncoated capillary with the other LONs as evidenced by a trailing in the absorption curve of the chromatogram.

6. TEM TEM analyses were performed at the Bordeaux Imaging Center (BIC) of the University of Bordeaux using an Hitachi H7650 at a voltage of 80 kV. For sample preparation a drop of 100 nM solution of 1-LON.sup.G4, 2X selex salts was placed on a carbon film 200 Mesh copper grid and left to dry for 10 min. A drop of 1% uranyl acetate solution as a negative stain for 1 min was then added to the copper grid and left to dry.

Results

[0150] The oligonucleotide sequence used by the inventors was chosen to embark a G-tract of 4 consecutive guanines in the middle of a 19-mer DNA sequence (LON.sup.G4).

[0151] The DNA molecule was modified at the 5-end with different lipid phosphoramidites (Table 1).

TABLE-US-00001 TABLE 1 Phosphoramidites used in the example Lipid Name moiety Lipidic phosphoramidite used ON None None (5-OH) 1-LON n-C.sub.18H.sub.37 [00012] k-LON Ketal bis-C.sub.15 [00013]

[0152] The phosphoramidites 1 and 2 were synthesized according to literature procedures and then coupled to the 5-end of the DNA using an automated solid phase DNA synthesizer. The randomized LON sequences (scramble-LON.sup.SC) were also synthesized as controls wherein the guanines were evenly distributed among the whole oligonucleotide sequence to minimize the chance of forming undesired foldings.

[0153] In summary, the following optionally lipid-modified oligonucleotides were synthesized:

TABLE-US-00002 (SEQIDNO:1) ON.sup.G4:5-TTAGTTGGGGTTCAGTTGG-3 (SEQIDNO:2) ON.sup.SC:5-TGTAGTAGGTTGTGTCTGG-3 (SEQIDNO:1) 1-LON.sup.G4:n-C.sub.18H.sub.37-5-TTAGTTGGGGTTCAGTTGG-3 (SEQIDNO:2) 1-LON.sup.SC:n-C.sub.18H.sub.37-5-TGTAGTAGGTTGTGTCTGG-3 (SEQIDNO:1) k-LON.sup.G4:ketal-5-TTAGTTGGGGTTCAGTTGG-3 (SEQIDNO:2) k-LON.sup.SC:ketal-5-TGTAGTAGGTTGTGTCTGG-3

[0154] The purification of these LONs offered a first insight at their self-assembling properties. When the lipid is attached at the 5-end of the oligonucleotide, the capping step during the ON synthesis precludes abortive sequences from reacting with the lipid phosphoramidite. Consequently, the purification of these LONs is theoretically straightforward as only the desired full length ON is coupled to the lipid and somewhat interacts with the reverse stationary phase. While the control k-LON.sup.SC was readily purified by RP (C.sub.4)HPLC, the majority of the k-LON.sup.G4 eluted in the dead volume after the first injection of the crude mixture. Surprisingly, more k-LON.sup.G4 was eluted in the second HPLC run when water alone was injected. The di-alkylated LONs likely form stable aggregates whose lipid segments are buried inside the aggregate preventing them from interacting with the stationary phase (vide infra). An original polyacrylamide gel electrophoresis (PAGE) protocol for the purification of k-LON.sup.G4 was therefore developed.

[0155] The nature of the supramolecular assemblies formed from these different LONs was investigated by non-denaturing PAGE. Interestingly, both the sequence (G4 or scramble) and the nature of the lipid were shown to impact the self-assemblies formed from these LONs (FIG. 1).

[0156] As expected, no clear foldings were observed for the different scramble LONs (lanes 1, 4 and 6). Surprisingly, no G4 structures were visible as well with the unmodified ON.sup.G4 (lane 2), probably because of the length of the oligonucleotide that precludes the correct folding of the G-quadruplex. In contrast, the presence of the octadecyl chain in 1-LON.sup.G4 resulted in the appearance of a retarded band (lane 5) corresponding to a tetramolecular parallel G-quadruplex as judged by its CD/NMR signature and its retardation in the gel (FIGS. 2 and 3). A similarly retarded faint band was visible with the unmodified ON.sup.G4 only when the salt concentration was increased (lane 3 of FIG. 1). Of note, the inventors clearly ruled out the possibility of a kinetic control over the equilibrium of formation of the tpG4. The ratio between the monomer and the tpG4 for both 1-LON.sup.G4 and ON.sup.G4 were clearly time independent (up to 4 weeks) as judged by PAGE and CD.

[0157] While the proportion of 1-LON.sup.G4 molecules that form tpG4 fold quickly (t<5 mins for CD experiments), the other 1-LON.sup.G4 molecules are trapped in undesired foldings (vide infra). Again, the k-LON.sup.G4 LON modified with a double chain lipid exhibited a peculiar behavior. The micellar aggregates that are formed from this LON are stable and large enough to survive the electrophoresis conditions and to prevent migration within the reticulated acrylamide polymer (lane 7 of FIG. 1). This result clearly illustrated the importance the oligonucleotide sequence has on the self-assembling process as the control k-LON.sup.SC migrated normally in the gel (lane 6 of FIG. 1).

[0158] Given the relatively small size and volume of the lipid segment of the LONs compared to the large DNA polar heads, the LONs used for these investigations were expected to form micellar aggregates. Indeed, dynamic light scattering (DLS) and Taylor Dispersion Analysis (TDA) experiments confirmed these expectations. These results suggest that the G4-prone forming sequence of k-LON.sup.G4 greatly stabilizes the micelles. The kinetic stability of k-LON.sup.G4 micelles over those formed from k-LON.sup.SC was further evidenced by the partitioning and the lack thereof of k-LON.sup.SC and k-LON.sup.G4 monomers respectively into micelles of the neutral triton detergent as well as the absence of any micelle disassembly in the presence of acetone. On the other hand, the presence of the G4-prone forming sequence in the LON sequence did not necessarily translate into stable micellar aggregates as evidenced by the migration observed with 1-LON.sup.G4 (FIG. 1, lane 5). The reticulated sieving acrylamide matrix is prone to disassemble loose micellar aggregates especially in the presence of an efficient divalent cation scavenger like EDTA present in the PAGE experiment. Micelles of 1-LON.sup.G4 were nevertheless present in solution as evidenced by agarose gel (FIG. 4, lane 3), DLS and TEM (FIG. 5). Interestingly, no micellar aggregates were observed with the control 1-LON.sup.SC (FIG. 4, lane 4). Consequently, tpG4-mediated micelle stabilization was still at play with 1-LON.sup.G4 The tpG4 monomer of 1-LON.sup.G4 partitioned with the ones in the micelle in that case. Agarose gel results also confirmed the weak tendency of unmodified ON.sup.G4 to form G4 structures, only a faint retarded band being visible in the gel (FIG. 4, first lane). Interestingly, k-LON.sup.G4 exhibited well-defined micellar bands compared to their scramble analogs (FIG. 1, lanes 5 and 6 respectively). No voluminous aggregates being detectable by DLS and Taylor dispersion analyses, the prominent trailing shoulders observed with k-LON.sup.SC more likely resulted from unspecific adsorptive interactions between the highly concentrated free scramble DNA at the micelle surface and the agarose gel matrix during electrophoresis. These unspecific interactions have been shown to increase with DNA sizes and decrease with increasing salt concentrations, in line with the present observations. Conversely, favorable intra-micellar G4 formation in k-LON.sup.G4 micelles may prevent unspecific adsorption of the LON micelles with the agarose matrix.

[0159] Furthermore, the inventors observed a noticeable retardation with k-LON.sup.G4 compared to k-LON.sup.SC. The G4-prone sequences may indeed be poised to adopt a brush-like regime at the surface of the micelle upon formation of the extended and rigid intra-micellar and parallel G-quadruplexes with a concomitant increase in micellar size and/or change in morphology, or aggregation number compared to the looser scramble sequences. DLS and Taylor Dispersion Analysis (TDA) of these samples fully supported native PAGE and agarose observations. For instance, micelles of 1-LON.sup.SC were only visible by DLS at concentration above 50 M compared to the 20 M used in the PAGE experiments and the size of k-LON.sup.G4 micelles was superior to k-LON.sup.SC. It is worth mentioning that even though no discrete parallel G-quadruplexes were observed with k-LON.sup.G4, quadruplex formation at the surface of the micelle still occurred as judged by their CD signature and the K.sup.+-dependence observed in agarose gel.

[0160] All tetramolecular G-quadruplexes reported in the literature so far were parallel and were formed from short oligonucleotide sequences, most probably because of polymorphism issues. Given their molecularity of four, their kinetics of formation are extremely slow at micromolar concentrations. Yet, once formed, these aggregates are very stable: no melting is observed at T>95 C. provided a minimum amount of K.sup.+ is present in solution. To the inventor's knowledge, the tpG4 from 1-LON.sup.G4 is the first tpG4 of moderate size reported. Besides, the inventors found that this G4 melted at a reasonable temperature (Tm=79 C.). The decrease in the melting temperature of 1-LON.sup.G4 tpG4s compared to shorter ones may result from the long flanking probably disordered DNA sequences. The thermal reversibility of tpG4 formation constitutes a clear advantage for the design of switchable nano-devices for biotechnological applications.

[0161] Although, a single lipid chain was effective at promoting tpG4 formation from 1-LON.sup.G4, monomers were still apparently visible (FIG. 1, lane 5). Given the high thermal stability of 1-LON.sup.G4 tpG4 and the absence of change in the ratio 1-LON.sup.G4 monomer/tpG4 with time (see above), no equilibrium exists between these 2 entities. In fact, the inventors found that the alleged monomers of 1-LON.sup.G4 were not free to equilibrate in solution with the tpG4. Given its high thermal stability, the 1-LON.sup.G4 tpG4 corresponds to an energetic minimum in the different energetic pathways that lead to other undesired foldings in solution. This was first checked by running a bi-dimensional native PAGE with 1-LON.sup.G4: the tpG4 band showed no trace of equilibrium in the second dimension (no salt). Instead, the remaining monomers of 1-LON.sup.G4 in solution were kinetically trapped in less stable, undesired foldings and/or aggregates. This hypothesis was confirmed by heating the solution above the melting temperature of the undesired foldings but below the Tm of the tpG4. The freed monomers were again available to form more tpG4s upon cooling. Several temperature cycles were required to enrich the solution in the desired tpG4 (FIG. 6).

[0162] Very importantly, while temperature cycles applied to 1-LON.sup.G4 in the absence of salt or in the presence of KCl had no or little influence on the formation of the tpG4 (FIG. 6, 2 first lanes), Mg.sup.2+ was found to be very important for the correct folding of 1-LON.sup.G4 into tpG4 (FIG. 6, last lane). This result came as a real surprise as K.sup.+ is the cation of choice for the stabilization of G-quadruplexes. Instead, the inventors found that Mg.sup.2+ remarkably stabilized all the LON micellar assemblies irrespective of their oligonucleotide sequence (G4 or scramble). For instance, Mg.sup.2+ was necessary to observe 1-LON.sup.G4 micelle in agarose gel (FIG. 4) and no more migration of the k-LON.sup.SC was observed in PAGE with increased amounts of Mg.sup.2+ as a result of the higher stability of the micelles, just as what was observed in the absence of salts with k-LON.sup.G4 (FIG. 1, lane 7). These explanation of the inventors for this impressive magnesium effect is that only the desired tpG4 is capable of forming stable micellar assemblies that are further stabilized by Mg.sup.2+ bridging between quadruplexes. Following the Le Chatelier's principle, the tpG4s are withdrawn from the equilibrium to form the stable micelle and provided the unspecific other aggregates are heated above their melting temperature, more tpG4s are formed at each heating and cooling cycle.

[0163] In conclusion the inventors have shown that the modification of a G4-prone oligonucleotide sequence with lipids drastically increases the probability of forming tetramolecular parallel G-quadruplexes over other undesired foldings or oligomers. These results indicate that lipids constitute key elements in the supramolecular organization of G4-prone LONs. They first serve as a guide for the correct folding of the LON into the desired G4 over other undesired aggregates. The G4 segment in return (together with Mg.sup.2+) are crucial for the stabilization of the micellar systems that eventually lead to the enrichment in tpG4 provided heating and cooling cycles are applied to the solution. This lipid-driven assembly of tpG4 is schemed on FIG. 7. The stability of the micellar aggregates can be modulated playing around 1) the lipid, 2) the nature of the salts present in solution (with an emphasis on Mg.sup.2+) and 3) the presence of a G4-prone segment within the oligonucleotide sequence of the LON. These data highlight the potential of LON for the construction of parallel G-quadruplexes of unprecedented length. This provides a means to design and construct promising potentially switchable supramolecular architectures for nanotechnology and nucleic acid-based therapeutics.