Triplex inducer-directed self-assembly of single-walled carbon nanotubes: a triplex DNA-based approach for controlled manipulation of nanostructures

Abstract

As a promising strategy for artificially control of gene expression, reversible assembly of nanomaterials and DNA nanomachine, DNA triplex formation has received much attention. Carbon nanotubes as gene and drug delivery vector or as ‘building blocks’ in nano/microelectronic devices have been successfully explored. Therefore, studies on triplex DNA-based carbon nanotube hybrid materials are important for development of smart nanomaterials and for gene therapy. In this report, a small molecule directed single-walled carbon nanotubes (SWNTs) self-assembly assay has been developed by disproportionation of SWNTs–dT₂₂·dA₂₂ duplex into triplex dT₂₂·dA₂₂·dT₂₂ and dA₂₂ by a triplex formation inducer, coralyne. This has been studied by circular dichroism, light scattering (LS) spectroscopy, scanning electron microscopy (SEM), atomic force microscopy (AFM), electrophoretic mobility shift assay and supported by using DNA random sequence. In contrast, SWNTs do not aggregate under the same experimental conditions when the small molecules used can not induce dT₂₂·dA₂₂·dT₂₂ triplex formation. Therefore, this novel small molecule-directed SWNTs self-assembly assay has also been used for screening of triplex inducers in our studies.

INTRODUCTION

As the leading nanodevice candidate, single-walled carbon nanotubes (SWNTs) have great potential applications in electronics, optics, mechanics, thermal transportation and biosensing (1–4). The use of carbon nanotubes as ‘building blocks’ in nano/microelectronic devices could revolutionize the electronic industry in the same way that the microchips have revolutionized the computer industry. Individual SWNTs have been utilized to realize molecular-scale electronic devices such as single-electron (5) and field-effect transistors (6). Several SWNTs-based devices have been successfully integrated into logic circuits (7) and transistor arrays (8). A key challenge for the application of SWNTs is how to assemble them into desired large architectures, that has attracted much attention. Self-assembly based on molecular recognition establishes a promising approach for designing complex architectures from SWNTs that excludes the need for precise nanofabrication and mechanical manipulations (9). Biological molecules, such as DNA and RNA, possess highly specific and precise molecular recognition capability (10–12) that can be evolved through genetic engineering and can exert rational control over assembly and hierarchical pattern formation at a molecular scale. Integration of such unique capabilities of oligonucleotides with novel nanomaterials can offer many opportunities for bottom-up fabrication, including hierarchical assembly of 2D and 3D functional nanoarchitectures, molecular electronic and optoelectronic devices and molecular sensors (13).

Recently, functionalization of SWNTs with biomolecules has drawn much attention in nanotechnology because of their potential applications in molecular electronics, field-emission devices, biomedical engineering and biosensors (14). Several studies are particularly devoted to the generation of SWNTs–DNA complexes via two major routes: covalent construction (15,16) and non-covalent association (17,18). More recently, self-assembly strategies based on the biorecognition capability of single-stranded DNA (ssDNA) have been proposed (19–22). However, most of the multicomponent structures fabricated by self-assembling of multiple carbon nanotubes are directed by DNA hybridization.

Triple-helix DNA (triplex) is a well-documented structure in biology that is formed by the association of a target duplex and a homopyrimidinic ‘triplex-forming oligonucleotide’ (23). As a possible means to artificially control gene expression (24–27), reversible assembly of nanomaterials (28,29) and DNA nanomachine (30–34), DNA triplex formation has received much attention. However, for most triplex sequences the binding of the third strand (i.e. Hoogsteen strand) is considerably less stable than the Watson–Crick duplex (35). Previous studies have indicated that some small molecules with planar structures can facilitate triplex formation and stabilize triplex. Coralyne, a small crescent-shape molecule, can bind strongly to single-stranded homoadenine-containing nucleic acids. Our previous studies have shown that coralyne can induce poly(rA) tail in the 3′-end of mRNA self-structured that has a melting temperature of 60°C (36). Polak and Hud have demonstrated that coralyne can cause the strands of poly(dA)·poly(dT) to repartition into an equimolar amount of triplex poly(dT)·poly(dA)·poly(dT) and poly(dA) (37,38). Therefore, a small molecule that can bind to these kinds of DNA and/or RNA and trigger the formation of non-Watson–Crick secondary structures would be useful not only for design of potential new therapeutics but also for generating dynamic nucleic acid nanostructures.

In this article, we demonstrate that small molecule coralyne-induced triplex formation can direct SWNTs assembly through disproportionation of SWNTs–dT₂₂·dA₂₂ duplex into triplex dT₂₂·dA₂₂·dT₂₂ and dA₂₂. This novel small molecule-directed SWNTs self-assembly assay has also been used for screening of triplex inducers in our studies. To the best of our knowledge, this is the first example to use SWNTs assambly to detect triplex structure formation.

MATERIALS AND METHODS

Materials

SWNTs (φ = 1.1 nm, purity > 90%) were purchased from Aldrich (St Louis, MO, USA), purified as described previously by sonicating SWNTs in a 3:1 v/v solution of concentrated sulfuric acid (98%) and concentrated nitric acid (70%) for 10 h at 35–40°C, and washed with water, leaving an open hole in the tube side and functionalizing the open end of SWNTs with carboxyl group (39–43).

DNA oligomers, dT₂₂, its corresponding complementary strand dA₂₂, the amino-end-functionalized (dT)₂₂ [5′-(dT)₂₂-3′-NH₂], and the control DNA (5′-CCA ACC CCC CAG AAA GAA-3′) were purchased from Sangon (Shanghai, China) and used without further purification. 1-Ethyl-3-(3-dimethyl aminopropyl) carbodiimide hydrochloride (EDC) was purchased from BBI (Bio Basic Inc.), Sulfo-N-hydroxy succinimide (Sulfo-NHS) was purchased from Pierce. 2-[N-morpholino] ethanesulfonic acid (MES), sodium dodecyl sulfate (SDS), Triton X-100 and coralyne chloride were purchased from Sigma and used as received. Coralyne chloride was dissolved in water and its concentration was determined by UV–Vis spectrometry using the extinction coefficient of ε_420 nm = 14500 M⁻¹ cm⁻¹. The others are of analytical or biochemical grade reagents.

Preparation of SWNTs-(dT)₂₂ conjugates

A 0.5 mg portion of oxidized, shortened SWNTs was suspended in 5 ml 100 mM MES buffer (pH 6.0) containing 0.2% Triton X-100 and sonicated for 1 h at room temperature. Add 20 mM EDC, 20 mM sulfo-NHS (set to pH 6.0) and stirred for another 1 h. Following the activation step, the pH was raised to 8.5, and the amino-modified (dT)₂₂ was added to a final concentration of 2 μM. The reaction mixture was stirred for ∼12 h at room temperature. The samples were then centrifuged for 30 min at 13 000 rpm. The upper supernatant were recovered and continued to centrifuge for 30 min at 13 000 rpm. The upper supernatants were recovered again and added NaCl to a final concentration of 0.5 M and incubated for a few hours. Then, samples were continued to centrifuge for 5 h at 13 000 rpm. The sediments were washed with double distilled water and 2 × SSPE/0.2% SDS buffer (consisting of 2.5 mM EDTA, 7 mM SDS, 300 mM NaCl and 20 mM NaH₂PO₄ with pH 7.4) by several centrifugation cycles at 13 000 rpm in order to remove non-specifically adsorbed DNA (43,44). The samples were dispersed in 0.5 ml double distilled water and dialyzed against water for 1 day while changing water every 3 h. The molecular weight cutoff size of the dialysis membrane is 8000. Then, the sample was dialyzed one more day against sodium cacodylate buffer (100 mM NaCl, 10 mM cacodylic acid, pH 6.8). Samples were centrifuged for 30 min at 13 000 rpm, and the supernatants were collected. Thus, the SWNTs–dT₂₂ conjugates were obtained.

Hybridization of dA₂₂ with SWNTs–dT₂₂ conjugates

Hybridization was carried out in sodium cacodylate buffer (100 mM NaCl, 10 mM cacodylic acid, pH 6.8) at 37°C for 2 h. dA₂₂ was added to the SWNTs–dT₂₂ solution to a final concentration of 22 μM in nucleotide. The hybridized products were designed SWNTs–dT₂₂·dA₂₂ conjugates.

Coralyne induces triplex formation between SWNTs–dT₂₂·dA₂₂ conjugates

Add coralyne to the SWNTs–dT₂₂·dA₂₂ conjugates solution, heating to 45°C and cooling to 4°C slowly. The concentration of coralyne was 11 μM (0.5 M equivalents/bp for dT₂₂ dA₂₂), and all the experiments were performed in sodium cacodylate buffer (100 mM NaCl, 10 mM cacodylic acid, pH 6.8).

Physical measurements

UV–Vis absorbance experiments were carried out on a Cary 300 UV–Vis spectrophotometer equipped with a Peltier temperature control accessory. All the spectra were measured in 1.0 cm path-length cell with the same concentration of SWNTs aqueous solution accordingly as the reference solution.

CD spectra experiments were measured on a JASCO J-810 spectropolarimeter equipped with a temperature controlled water bath. The optical chamber of CD spectrometer was deoxygenated with dry purified nitrogen (99.99%) for 5 min at a speed of 5 l min⁻¹ before use and kept the nitrogen atmosphere during experiments. Three scans were accumulated and automatically averaged. CD melting experiments of SWNTs–dT₂₂·dA₂₂ conjugates in the presence of coralyne were monitored at 274 nm with a heating rate of 1.0°C min⁻¹. Primary data were transferred to the graphics program Origin for plotting and analysis.

Light scattering (LS) measurements were carried out on a JASCO FP-6500 spectrofluorometer. The LS spectra were obtained by synchronously scanning the excitation and emission monochromators of the fluorescence spectrophotometer from 250 to 700 nm (namely, Δλ = 0) with the slit width for the excitation and emission of 5 nm.

SEM images of DNA–SWNTs were obtained on a HITACHI S-4800 scanning electron microscope. Samples were prepared by pipetting 5 μl of colloid solution onto a silicon substrate pretreated with piranha etch solution (4:1 concentrated H₂SO₄/30% H₂O₂) for 1 h at room temperature. After evaporating the solvent, the substrate was dried overnight under vacuum.

An AFM (Nanoscope IIIa, Digital Instruments, Santa Barbara, CA, USA) was used to image SWNTs–dT₂₂ or SWNTs–dT₂₂·dA₂₂ conjugates in the presence or absence of coralyne. The sample solution was deposited onto a piece of freshly cleaved mica. After rinsed with water, samples were dried before measurements. Tapping mode was used to acquire the images under ambient condition.

Denaturing PAGE was used to characterize the formation of SWNTs–dT₂₂ conjugate and to examine its purity. The products of reaction solution at each purifying step were loaded onto a 10% denaturing polyacrylamide gel (containing 7 M urea) using a loading buffer (30% glycerol, 25 mM EDTA, 0.01% bromophenol blue and 7 M urea). Each reaction was heated to 100°C for 3 min before loading on the gel to denature any non-specific DNA binding. The gel was run for 1 h at 100 V on a Bio-Rad vertical electrophoresis unit using TBE buffer (Tris–HCl, borate, EDTA, pH 8.0). The gels were silver-stained.

Agarose gel electrophoresis was also used to confirm the triplex formation. Electrophoresis was carried out in 0.5 % agarose gel in 0.5 × TBE (45 mM Tris, 45 mM boric acid, 1 mM EDTA, pH 8.0) as an electrophoresis buffer for 1 h at 60 V. Samples were loaded in electrophoresis buffer supplemented with glycerol to a final concentration of 20 w/v%.

RESULTS

Design of triplex-based SWNTs assembly

First, we prepared DNA–SWNTs conjugates by commonly used covalent functionalization of SWNTs with 3′-amino-modified homo-thymine ssDNA dT₂₂ via carbodiimide-mediated amidation process. Then, the conjugates were hybridized with homo-adenine ssDNA dA₂₂ to form SWNTs–dT₂₂·dA₂₂ conjugates. The working principle in our design follows: a small molecule that can cause the strands of dA₂₂·dT₂₂ to repartition into an equimolar amount of triplex dT₂₂·dA₂₂·dT₂₂ and dA₂₂ might lead SWNTs–dT₂₂ conjugates to form a cross-linked network of nanotubes. Coralyne, a strong triplex formation inducer, should disproportionate duplex dA₂₂·dT₂₂ into triplex dT₂₂·dA₂₂·dT₂₂ and ss dA₂₂. Coralyne-enhanced and -directed third strand, SWNTs–dT₂₂, binding to SWNTs–dT₂₂·dA₂₂ can subsequently induce SWNTs aggregation. However, when a small molecule used can not induce dT₂₂·dA₂₂·dT₂₂ triplex formation, no SWNTs aggregation would occur under the same experimental conditions. Therefore, we can use this novel small molecule-directed SWNTs self-assembly assay to screen other triplex inducers. This design has been illustrated in Scheme 1.

Scheme 1.

(a) Schematic representation of the fabrication of DNA–SWNTs conjugates and coralyne induced self-assembly of SWNTs based on triplex formation. (b) SWNTs-assembled nanostructure directed by coralyne-induced triplex formation and this assay used for screening of triplex inducers. Small molecules used are typical DNA binders, including neomycin, EB and DAPI.

Preparation and characterization of SWNTs–DNA conjugates

Carbodiimide-mediated amidations of SWNTs is a common method that was widely used to prepare DNA or protein–SWNTs conjugates (16,43,45,46). Oxidized SWNTs (0.1 mg ml⁻¹) and ssDNA (2 μM) solutions were mixed in the presence of EDC (20 mM) and Sulfo-NHS (20 mM), then incubated for 12 h. After separation and purification steps, the soluble fraction obtained from coupling to (dT)₂₂–NH₂ consists of a clear, transparent, blackish, homogeneous suspension as shown in Figure 1a. Denaturing PAGE and UV–Vis spectroscopy were used to characterize the purified SWNTs–(dT)₂₂ conjugates. After the sample was dialyzed and centrifuged as described in the materials section, dT₂₂ alone was not observed (Figure 1b, lane 2). Since the average length of SWNT is about several hundred nanometers long, the major SWNT–DNA conjugate hardly moved in denatured PAGE although the short SWNT-formed DNA conjugate may run faster than the long-SWNT–DNA conjugate. These results suggest that we obtained purified SWNT–DNA conjugates. Compared the UV–Vis spectra of SWNTs–dT₂₂ conjugates with SWNTs alone, there is an apparent peak ∼260 nm in the UV–Vis spectrum of SWNT–dT₂₂ conjugates, while no peaks at this range existed in the UV–Vis spectrum of SWNTs alone that clearly indicate that DNA molecules have been successfully attached to SWNTs (Figure 1c). Scanning electron microscopy (SEM) and atomic force microscopy (AFM) images (Figure 1d and e) also confirm the conjugation of ssDNA with SWNTs.

Figure 1.

Characterization of SWNTs–dT₂₂ conjugates. (a) Photograph of the supernatant fraction obtained from the coupling steps with dT₂₂. (b) Denaturing PAGE image of (lane 1) dT₂₂ alone; (lane 2) purified SWNTs–dT₂₂ conjugates. (c) UV–Vis absorption spectra of SWNTs (black line) and SWNTs–dT₂₂ conjugates (red line); (d) SEM image of SWNTs–dT₂₂ conjugates. After the sample was purified as described in ‘Materials and Methods’ section, SEM shows that SWNTs are dispersed upon DNA conjugation. (e) AFM image of SWNTs–dT₂₂ conjugates. AFM indicates that dT₂₂ are conjugated to SWNTs and the formed conjugates are dispersed.

Then, SWNTs–dT₂₂ was hybridized with its complementary strand, dA₂₂ (22 μM in nucleotide) to form duplex in sodium cacodylate buffer (100 mM NaCl, 10 mM cacodylic acid, pH 6.8) at 37°C. After separation and purification, the SWNTs–dT₂₂·dA₂₂ conjugates were characterized by circular dichroism (CD) spectroscopy (Figure 2a). CD spectrum of SWNTs-conjugated dA₂₂·dT₂₂ sample is similar to duplex dA₂₂·dT₂₂ alone. Figure 2b shows that SWNTs-conjugated dA₂₂·dT₂₂ has an apparent melting transition at 46.7°C observed in its melting curve, indicating that SWNTs–dT₂₂ forms duplex in the presence of dA₂₂.

Figure 2.

(a) CD spectra of the duplex dA₂₂·dT₂₂ (red line) and SWNTs–dT₂₂·dA₂₂ conjugates (black line). (b) CD melting profile of the SWNTs–dT₂₂·dA₂₂ conjugates.

Coralyne directed self-assembling of SWNTs and screening of triplex inducers

The coralyne-directed SWNTs self-assembling was accomplished by adding coralyne (11 μM, 0.5 M equivalents/base pair for dT₂₂·dA₂₂) to the SWNTs–dT₂₂·dA₂₂ conjugates in sodium cacodylate buffer (100 mM NaCl, 10 mM cacodylic acid, pH 6.8), and then the sample was heated to 45°C and cooled to 4°C slowly. The solution was then analyzed by LS spectroscopy. In the absence of coralyne, LS spectra of SWNTs–dT₂₂·dA₂₂ conjugates (Figure 3a, black and blue lines) did not show significant difference before and after heating, that indicates that aggregation does not occur. However, in the presence of coralyne, the LS intensity at 555 nm (Figure 3a, red line) increases remarkably, suggesting that the coralyne induced triplex formation and caused SWNTs aggregation. In contrast, this change can not be observed for a random selected DNA sequence (control DNA: 5′-CCA ACC CCC CAG AAA GAA-3′), that was used instead of dA₂₂ to hybridize with SWNTs–dT₂₂ conjugates under the same conditions (Figure 3b). To go further, we decide to explore if this assay can be used to screen triplex inducers. Typical duplex binders (chemical structures are shown in Scheme 2) are used in this assay. Ethidium bromide (EB; a typical DNA intercalator) and 4′,6-diamidino-2-phenylindole (DAPI; a typical minor groove binder) were tested using SWNTs–dT₂₂·dA₂₂ conjugates under the same experimental conditions, no intensity change was observed in their LS spectra (Figure 3c). Neomycin, a typical major groove binder, has been recently reported as a triplex stabilizing compound (47,48). However, due to its weak inducing capability (Supplementary Figure S1), neomycin can not drive SWNTs–dT₂₂·dA₂₂ to form triplex under the same coralyne experimental conditions and no LS signal change was observed either (Figure 3c). These results indicate that this novel small molecule-directed SWNTs self-assembly assay can also be used for screening of triplex inducers. The control experiments using random sequence and none triplex formation inducers confirm that the observed spectral change is due to coralyne-directed third strand SWNTs–dT₂₂ binding to SWNTs–dT₂₂·dA₂₂ that subsequently induces SWNTs aggregation. This is further supported by our next CD, SEM, AFM and electrophoresis results.

Figure 3.

LS spectra of the SWNTs–DNA conjugates under different conditions. (a) SWNTs–dT₂₂·dA₂₂ conjugates in the absence or presence of coralyne. (b) SWNTs–dT₂₂ with control DNA (random sequence) conjugates in the absence or presence of coralyne. (c) SWNTs–dT₂₂·dA₂₂ conjugates in the absence or presence of different DNA binders: EB, DAPI and neomycin.

Scheme 2.

Chemical structures of coralyne, EB, DAPI and neomycin.

Demonstration of triplex formation

CD spectroscopy was used to monitor coralyne-induced conformational change of duplex dA₂₂·dT₂₂ and SWNTs–dT₂₂·dA₂₂.

Coralyne can cause disproportionation of duplex dA₂₂·dT₂₂ at 45°C

The CD spectrum of duplex dA₂₂·dT₂₂ at 4°C changes dramatically upon the addition of 0.5 M equivalents of coralyne per DNA base pair (Supplementary Figure S2a). These changes include the appearance of a substantial positive CD bands near 300 nm and negative CD bands near 350 nm. However, typical CD bands of triplex DNA with intercalated coralyne at ∼340 and ∼440 nm are virtually non-existed (36,37,48). This indicates that coralyne does not cause the complete and spontaneous formation of triplex DNA at 4°C.

Heating of dA₂₂·dT₂₂ sample with added coralyne to 45°C dramatically reduces the magnitude of the duplex-specific coralyne CD bands at 300 and 350 nm (Supplementary Figure S2b); small positive bands at ∼340 and ∼440 nm (Supplementary Figure S2b) are observed, similar to CD bands observed for coralyne intercalated in the dT₂₂·dA₂₂·dT₂₂ triplex (Supplementary Figure S2d). Additionally, the melting profile of the dA₂₂·dT₂₂ sample in the presence of coralyne exhibits a transition at 66.1°C (Supplementary Figure S3a), which shows the same transition temperature of coralyne-intercalated triplex dT₂₂·dA₂₂·dT₂₂ (Supplementary Figure S3b). The magnitude of this transition in the CD melting profile of the duplex sample with added coralyne is half the transition in the triplex sample with added coralyne (Supplementary Figure S3a and b), which is also consistent with the dA₂₂·dT₂₂ sample being disproportioned by coralyne into 0.5 M equivalents of triplex dT₂₂·dA₂₂·dT₂₂ and 0.5 M equivalents of dA₂₂. Thus, the 66.1°C transition in the dA₂₂·dT₂₂ sample with coralyne can be assigned to the melting of coralyne-intercalated triplex dT₂₂·dA₂₂·dT₂₂ in the disproportioned sample.

The process of duplex disproportionation by coralyne in the dA₂₂·dT₂₂ sample during the first time heating has broad transition (Supplementary Figure S3a), showing that coralyne-induced duplex disproportionation is cooperative. The central transition is at 39.5°C (Supplementary Figure S3a). During the second time heating of the same sample, this broad transition is absent (Supplementary Figure S3a). This indicates that the intercalated triplex and dA₂₂ of a coralyne-disproportioned duplex sample do not immediately revert back to the duplex state when the sample is returned to 4°C. This lack of reversion from the disproportioned state is also supported by the fact that the CD spectrum of the coralyne-disproportioned dA₂₂·dT₂₂ at 4°C after heat cycling (from 4°C to 80°C and back to 4°C) is radically different from the CD spectrum of the sample prior to heating (Supplementary Figure S2a and c). Furthermore, there is an excellent match between the CD spectrum of coralyne-disproportioned dA₂₂·dT₂₂ sample and a composite CD spectrum generated by the summation of a CD spectrum of coralyne-intercalated triplex dT₂₂·dA₂₂·dT₂₂ and the CD spectrum of dA₂₂ in the presence of coralyne (Supplementary Figure S2c).

Coralyne can also cause disproportionation of duplex SWNTs–dT₂₂·dA₂₂ at 45°C

As shown in Figure 4a, coralyne can not cause duplex SWNTs–dT₂₂·dA₂₂ to form triplex at 4°C (the same CD spectrum characteristics with that of dA₂₂·dT₂₂/coralyne (Supplementary Figure S2). However, heating SWNTs–dT₂₂·dA₂₂ sample with added coralyne to 45°C, CD spectrum shows typical features of DNA triplex with intercalated coralyne, together with appearance of small positive CD bands at ∼340 and ∼440 nm (Figure 4b) (36,37,49), similar to the case of dA₂₂·dT₂₂/coralyne in the absence of SWNTs (Supplementary Figure S2). After cooling to 4°C, the CD spectrum is significantly different from that of the same sample at 4°C prior to being heated (Figure 4a and c), showing that the transition to a sample of complete coralyne-intercalated triplex is not reversed upon cooling. This is also illustrated by the fact that the CD melting profile for the first heating of the SWNTs–dT₂₂·dA₂₂ sample with coralyne has a transition at ∼35°C, that is assigned to DNA strand reorganization (36,37). That transition is disappeared during the second heating of the sample and one single transition at ∼61°C is observed, that is attributed to the intercalated triplex melts into single strands (Figure 4d).

Figure 4.

CD spectra of SWNTs–dT₂₂·dA₂₂. (a) SWNTs–dT₂₂·dA₂₂ in the absence or presence of coralyne at 4°C prior to heating. (b) Spectra of SWNTs–dT₂₂·dA₂₂ in the presence of coralyne at 45°C. (c) Heat-cycled spectrum of disproportioned SWNTs–dT₂₂·dA₂₂ sample in the presence of coralyne at 4°C. (d) CD melting profiles for SWNTs–dT₂₂·dA₂₂ samples at 274 nm to show structural transitions: first time (black line) and second time heating (red line) of SWNTs–dT₂₂·dA₂₂ after the addition of coralyne.

Morphology of dA₂₂·dT₂₂–SWNTs–coralyne assembly

To directly study the morphology of dA₂₂·dT₂₂–SWNTs–coralyne assemblies, we deposited a mixture of dA₂₂·dT₂₂–SWNTs and coralyne on a silicon substrate and observed the aggregates (Figure 5) by SEM and AFM (40,43). As shown in Figure 5a, in the presence of coralyne, SEM results clearly show that the aggregates form large DNA-linked 3D nanostructure. The SWNTs are linked to each other due to the triplex dT₂₂·dA₂₂·dT₂₂ formation via bound coralyne (Figure 5b), however, this is not observed for dA₂₂·dT₂₂–SWNTs conjugates in the absence of coralyne (Figure 5c). AFM images also support that the SWNTs are cross-linked and form large nanostructure in the presence of coralyne (Figure 5d). Electrophoretic mobility shift assay (41,43,50) was also used to compare the size between coralyne-induced SWNTs aggregates and SWNTs–dT₂₂·dA₂₂ duplex (Supplementary Figure S4). In combination with CD, LS spectroscopy, gel mobility shift, SEM and AFM results, coralyne can induce triplex formation and cause SWNTs cross-linked and form large assembly.

Figure 5.

(a and b) SEM images of heat-cycled SWNTs–dT₂₂·dA₂₂ in the presence of coralyne. (c) SEM image of heat-cycled SWNTs–dT₂₂·dA₂₂ without coralyne. (d) AFM image of heat-cycled SWNTs–dT₂₂·dA₂₂ sample in the presence of coralyne.

DISCUSSION

Self-assembly has received much attention aimed at integrating nanoscale building blocks into functioning devices. SWNTs have already been considered as ‘building blocks’ in nano/microelectronic devices. However, since SWNTs lack chemical recognition, SWNT-based electronic devices and sensors are strictly related to the development of a bottom-up self-assembly technique. The specific molecular recognition of DNA coupled with SWNTs and the hybridization of these macromolecular wires make the DNA molecule an ideal candidate for this task. Here we report for the first time that a small triplex formation inducer (coralyne) can cause the self-assembly of SWNTs by disproportionation of duplex DNA into triplex and ssDNA.

Coralyne, a small crescent-shaped molecule, can intercalate into both duplex and triplex DNA, but binds stronger to triplex DNA (51–55). Polak and Hud have demonstrated that coralyne can cause complete and irreversible disproportionation of duplex poly(dT)·poly(dA) into a 1:1 mixture of coralyne-intercalated triplex poly(dT)·poly(dA)·poly(dT) and poly(dA), with each resulting structure being at one half the concentration of the original duplex (37,38). We have also observed that coralyne can cause duplex dA₂₂·dT₂₂ to repartition into a 1:1 mixture of coralyne-intercalated triplex dT₂₂·dA₂₂·dT₂₂ and dA₂₂. CD spectra and melting experiments indicate that this unique property of coralyne is also applicable to the duplex dA₂₂·dT₂₂ that is attached to the surface of SWNTs. This is intriguing not only for its property, but also for its application to the controllable self-assembly of SWNTs. Coralyne can cross-link two separate SWNTs–dT₂₂·dA₂₂ conjugates by forming a coralyne-intercalated triplex SWNT–dT₂₂·dA₂₂·dT₂₂–SWNT and dA₂₂, that leads to SWNTs aggregation and forms large DNA-linked 3D nanostructure (Scheme 1). This has been studied by using LS spectroscopy, SEM, AFM and electrophoretic mobility shift assay.

In order to further verify our design principle, we also used other DNA duplex binders to see whether these small molecules can direct SWNTs self-assembly. EB can intercalate into DNA. DAPI is a fluorescent dye that can bind strongly to DNA. Two DNA-binding modes have been suggested for DAPI: one is minor groove-binding mode, preferentially interacting with AT-rich regions (56,57); the other is intercalation into GC or mixed GC and AT DNA sequences (58–60). Neomycin is one of the most effective aminoglycoside groove binders to stabilize DNA triple helix. Neomycin selectively stabilizes triplex DNA and hardly influence duplex DNA (47,48). Under the same coralyne experimental conditions, all these small molecules can not drive dT₂₂·dA₂₂ to form triplex. These results indicate that although neomycin can enhance triplex DNA stability it can not induce duplex dT₂₂·dA₂₂ disproportionation to form triplex dT₂₂·dA₂₂·dT₂₂ (Figure 3 and Supplementary Figure S1), neomycin is a triplex DNA stabilizer but not an effective triplex inducer.

CONCLUSION

In summary, DNA triplex formation has been considered a promising strategy for realizing artificially control of gene expression, reversible assembly of nanomaterials and DNA nanomachine. Here we design a small molecule-directed SWNTs self-assembly assay by disproportionation of duplex into triplex by triplex formation inducer, coralyne. This design has been studied by CD, LS spectroscopy, SEM, AFM, electrophoretic mobility shift assay and supported by using DNA random sequence. For small molecules that can not induce triplex formation, SWNTs do not aggregate under the same experimental conditions. Besides, specific small molecule-fueled cross-linking of SWNTs networks can be useful in the area of nanobiotechnology. For example, this small molecule-directed SWNTs assembling can offer a potential method to facilitate construction of desired SWNTs nanoscale multicomponent/multifunctional architectures for electrical and biosensing applications.

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.

FUNDING

973 Project 2011CB936004; NSFC (20831003, 90813001, 20833006, 90913007); Chinese Academy of Sciences. Funding for open access charge: NSFC; 973 Project.

Conflict of interest statement. None declared.