Novel supramolecular assemblies of repulsive DNA–anionic porphyrin complexes based on covalently modified multi-walled carbon nanotubes and cyclodextrins

Abstract

We have prepared the water-soluble anionic 5-(p-aminophenyl)-10,15,20-tri(p-sulfonatophenyl)porphyrin (ATSPP) for constructing novel supramolecular assemblies of DNA–anionic porphyrin complexes. ATSPP was covalently linked to multi-walled carbon nanotubes (MWCNTs) to give a suspension of a conjugate (ATSPP–MWCNTs) that was stable for >2 weeks. On the addition of cyclodextrins (CDs), the corresponding complex (ATSPP–MWCNTs–CDs) was formed. Absorption and fluorescence spectra displayed that 1 : 1 inclusion complexes were formed with either α-CD or β-CD, but that virtually no interaction occurred with γ-CD. A comparative study on the interactions of DNA with ATSPP–MWCNTs and the ATSPP–MWCNTs–CDs complexes was carried out using fluorescence spectroscopy, resonance light-scattering, and transmission electron microscopy. The results showed that negatively-charged ATSPP–MWCNTs can interact with the negatively-charged DNA, indicating that covalently modified CNTs contribute to the counterintuitive binding of anionic porphyrins with DNA. This can possibly be attributed to π-stacking interactions between the sidewalls of the CNTs and the bases in DNA. Secondly, ATSPP–MWCNTs–CDs complexes interact better with DNA than the ATSPP–MWCNTs, indicating that CDs can promote the binding of DNA to anionic porphyrin-covalently modified MWCNTs. This is probably due to the fact that the amphiphilic CDs can greatly improve the solubility and dispersity of ATSPP–MWCNTs by their effective inclusion complexation with ATSPP. The successful construction of the supramolecular assemblies ATSPP–MWCNTs–CDs–DNA provides a new approach to the binding anionic porphyrins to negatively-charged DNA.

---

Introduction

Interactions of porphyrins with DNA have been widely reported and porphyrin–DNA complexes have been applied in various biological and medical fields, such as DNA detection and recognition,¹ and tumor therapy.² With negative charges and grooves on its main chain, DNA preferentially interacts with substrates bearing positive charges or highly π-conjugated groups through electrostatic attractions or π-stacking interactions.³ Owing to their structural motifs, such as positively-charged peripheral groups and a macrocyclic aromatic core, the representative of cationic porphyrins meso-tetra(N-methylpyridinium-4-yl)porphyrin (TMpyP) and its derivatives have been verified to exhibit good affinity with DNA.^1–7 In contrast, anionic porphyrins are generally believed not apt to “naturally bind” DNA due to the inherent electrostatic repulsions caused by the negative charges,⁸ and research in this field is scarce. Nevertheless, interactions between anionic porphyrins and DNA still attract much attention, based on the following points. Firstly, anionic porphyrins have the same macrocyclic aromatic core (porphyrin ring) as that of cationic porphyrins, and the π–π interactions which are similar to those between cationic porphyrins and DNA may also occur between anionic porphyrins and DNA. These π–π interactions may offer a possibility to bind anionic porphyrins with DNA successfully, because they can be strengthened to overcome the electrostatic repulsions. On the other hand, anionic porphyrins have negatively-charged groups (e.g., nido-carboranyl, SO₃⁻ or COO⁻)^9–11 that are quite different from those of the cationic porphyrins, and so may show different properties and behaviors when interacting with DNA, which will give an opportunity to develop novel porphyrin–DNA complexes for interesting and promising applications.

In recent years, researchers have been focused on developing easy and effective methods for achieving counterintuitive but feasible interactions between anionic porphyrins and DNA. Purrello reported that DNA could interact well with anionic nido-carboranyl porphyrins, attributed to both porphyrin inner core protonation to reduce electrostatic repulsions and a non-covalent interaction caused by a chirality match between them.¹² Besides, a protonated reagent like spermine was used as a positively-charged mediator to remarkably shield electrostatic repulsions between a sulfonated nickel(II) porphyrin and Z-DNA, forming a stable anionic porphyrin–Z-DNA complex that could be applied for spectroscopic discrimination of Z-DNA.^13,14 It is necessary to point out that there was another synergetic interaction contributing to the stability of that complex, which was the axial coordination between nitrogen N7 of the guanine in Z-DNA and the central metal, nickel(II), in the middle of the porphyrin ring. Thus, as demonstrated by these findings, the reducing of inherent electrostatic repulsions along with the increasing of non-electrostatic interactions should be considered as a feasible strategy to facilitate anionic porphyrin interactions with DNA.

It is well known that carbon nanotubes (CNTs) can bind to DNA through π-stacking interactions,¹⁵ and can be modified by porphyrins to improve their dispersity for application.^16–18 Thus, by connecting CNTs to anionic porphyrins, the former will endow the latter with a very large π-conjugated structure, which will probably enable the originally repelled latter to interact counterintuitively but well with DNA resulting in a great increase in π–π interactions. On the other hand, with a special amphiphilic structure, cyclodextrins (CDs) can mediate DNA interactions with organic functional molecules via supramolecular inclusion complexation¹⁹ and may include peripheral substituents at the porphyrin ring inside their hydrophobic cavities.^20–23 Thus, through the introduction of CDs, the CNTs modified with anionic porphyrins will probably interact better with DNA, due to a possible decrease in electrostatic repulsions, attributed to the fact that CDs can include the phenyl rings of the porphyrin into their hydrophobic cavities to keep the sulfonato groups away from DNA and so reduce electrostatic repulsions between them. Therefore, in the present study, we could follow a strategy of selecting CNTs and CDs as two fascinating candidates helpful for forming stable DNA–anionic porphyrin complexes. The main route could be described as follows (Scheme 1): from the initial anionic porphyrin to form anionic porphyrin–CNTs dyads, and then to form anionic porphyrin–CNTs–CDs triads, and finally transform these into the targeted anionic porphyrin–CNTs–CDs–DNA tetrads that have not yet been reported to the best of our knowledge. This strategy would have high feasibility since it is based on our previous work in the field of porphyrin interactions with CNTs^16,17 or CDs.^20,21 Moreover, the tetrad “anionic porphyrin–CNTs–CDs–DNA” would combine the outstanding merits of the “two fascinating candidates”, such as (i) non-toxicity, (ii) good biocompatibility, (iii) an unusual capability to penetrate cell membranes to serve as gene delivery systems (CNTs),²⁴ (iv) extraordinary amphiphilicity used to improve the solubility of “guest molecules” and overcome their aggregation in solution (CDs).²⁵ Therefore, the current work will surely enrich the research on porphyrin–DNA complexes and may lay novel foundations for promising biological or medical applications.

Scheme 1 The strategy for constructing novel DNA–anionic porphyrin supramolecular assemblies by the “fixation” of CNTs and the “inclusion” of CDs.

Experimental

Reagents and apparatus

MWCNTs were purchased from Shenzhen Bill Technology Development Ltd. (Shenzhen, China) and used without further purification. α-CD and γ-CD were purchased from Shanghai Sangon Biotech Co., Ltd. (Shanghai, China), and β-CD was purchased from Shanghai Chemical Reagent Factory (Shanghai, China). Calf thymus DNA (ct-DNA) was purchased from Livzon Pharmaceutical Group Inc (Zhuhai, China). A buffer solution (pH 7.04) of KHPO₄–KH₂PO₄ was freshly prepared. N,N-Dimethyl formamide (DMF) was freshly distilled over anhydrous calcium sulfate. All chemicals were of analytical grade and the water was double distilled.

The ¹H NMR spectrum was recorded using a DMSO-d₆ solution at 25 °C on a Varian Mercury Plus 400 (Varian, USA). FT-IR spectra were obtained with a BRUKER TENSOR 27 instrument (Bruker Optics, Germany), and the samples were prepared with the use of KBr of spectroscopic grade. The absorption spectra were measured on a La 25 UV/vis spectrometer (PerkinElmer, USA), while the fluorescence and RLS spectra were measured on a Ls-55 fluorescence spectrometer (PerkinElmer, USA), and all the experiments were carried out in aqueous solution. Thermogravimetric analyses were conducted on a TG-60 instrument (Shimadzu, Japan), under flowing air at a scanning rate of 20 °C min⁻¹ from room temperature to 800 °C. The TEM images were obtained with a JEM-3100 transmission electron microscope (JEOL, Japan).

Preparation of 2

Porphyrin 1 was prepared according to the literature.²⁶ Then ATSPP–MWCNTs 2, a complex of a porphyrin and covalently modified multi-walled carbon nanotubes, was synthesized using a procedure described in Scheme 2. MWCNTs were purified and oxidized with a mixture of concentrated sulfuric and nitric acids (3 : 1, 98% and 70%, respectively) at 80 °C under sonication for 8 h. The resulting MWCNT–COOH (20 mg) was refluxed in 20 mL of thionyl chloride at 70 °C for 72 h. After removing the excess thionyl chloride under vacuum, the residue was re-dispersed in 10 mL of DMF, then 20 mg of 1 was added, and the resulting mixture was stirred at 100 °C for 72 h under an argon atmosphere. After the reaction finished, the DMF was removed and 20 mL of water was added. Then, through filtration, the sediment was washed with 60 mL of water to remove the unreacted porphyrin. The crude product was purified by centrifugation, and dried at 40 °C for 10 h under vacuum. Finally, the desired black solid of 2 was obtained. ATSPP–π–π–MWCNTs 2-non, the complex of porphyrin and non-covalently modified multi-walled carbon nanotubes, was obtained also as a black solid according to a procedure similar to the above, using pristine MWCNTs without any treatment, at room temperature and without argon protection (Scheme 2).

Scheme 2 Synthesis of ATSPP–MWCNTs 2 and ATSPP–π–π–MWCNTs 2-non.

Interaction investigations

To investigate the inclusion interactions of 2 with CDs, 1.0 mL of a stock solution (9.0 × 10⁻² g L⁻¹) of 2 was transferred into a 10 mL volumetric flask, and then a CD solution (1.5 × 10⁻² mol L⁻¹) was added in a volume ranging from 0 μL to 80 μL. The mixed solution was diluted to the final volume with double distilled water, and the pH was fixed at 7.04 using a 0.2 mol L⁻¹ phosphate buffer solution. After being shaken thoroughly, the interactions were determined in 10 min at 25 ± 1 °C.

To investigate the interactions of 2 with DNA and the influence of CDs during this process, firstly, 1.0 mL of the stock solution (9.0 × 10⁻² g L⁻¹) of 2 was transferred into a 10 mL volumetric flask in the absence or presence of 80 μL of a CD solution (1.5 × 10⁻² mol L⁻¹). Then, an appropriate amount of 2.0 × 10⁻³ mol L⁻¹ ct-DNA solution was added in a volume ranging from 0 μL to 72 μL. The mixed solution was diluted to the final volume with double distilled water, and the pH was fixed at 7.04 using a 0.2 mol L⁻¹ phosphate buffer solution. After being shaken thoroughly, the interactions were determined in 10 min at 25 ± 1 °C.

Results and discussion

Characterization of 1, 2 and 2-non

The structure of ATSPP 1 was confirmed by ¹H NMR (Fig. S1, ESI†). Both ATSPP–MWCNTs 2 and ATSPP–π–π–MWCNTs 2-non were soluble in H₂O, DMF and ethanol, and the formed black solutions of 2 or 2-non could remain stable for more than two weeks or about six days, respectively, which was favourable for the following reactions and characterization. These results indicated that the solubility or dispersity of the MWCNTs can be improved by either covalent or non-covalent modification with water-soluble ATSPP; however, the covalent method might be preferable since 2 was found to exhibit better dispersity and stability than 2-non. In the FT-IR spectrum of 2, characteristic absorptions at 1642 and 1050 cm⁻¹ could be assigned to the stretching vibrations of the amide C=O bond and its C–N bond, respectively (Fig. S2, ESI†). This indicated that 1 was successfully covalently attached to the surface of the MWCNTs with an amide bond by the method shown in Scheme 2. In the spectrum of 2-non, the small shoulder peak on the left side of 1642 cm⁻¹ (another characteristic absorption for the C=O bond) could not be observed (Fig. S2, ESI†), which indicated that there wasn’t a carbonyl group present in 2-non and that 1 was absorbed onto the MWCNT’s surface in a non-covalent way (π–π interactions) as shown in Scheme 2.

In the UV-vis spectra, the absorption of 2 in H₂O showed a Soret band at 423 nm (for 2-non, 415 nm) and a broad signal with a monotonical decrease in the range of 300–800 nm, corresponding separately to the attached porphyrin and the multi-walled carbon nanotubes (Fig. 1a). Compared to the absorption spectrum of 1 (the Soret band: 412 nm), the spectrum of 2 showed a red-shift of the Soret band as large as 11 nm (for 2-non, as small as 3 nm), without seeing the Q-bands due to the overwhelmingly broad absorption of the MWCNTs. The phenomenon observed here gave strong support for the covalent attachment of porphyrins to MWCNTs in complex 2 and the non-covalent absorption of porphyrins on MWCNTs in complex 2-non, since non-covalently attached porphyrins on CNTs should not cause a large shift of the Soret band, as previously reported.²⁷ In the fluorescence spectra, upon excitation of the porphyrin moiety at the Soret band (423 nm), the solution of 2 exhibited about 80–90% quenching of the emission bands at 654 and 716 nm, as compared to that of 1 at a matching absorption (Fig. 1b). The fluorescence spectrum of 2-non was similar to that of 2, but with a slightly stronger quenching. This might be explained by the different distances from the porphyrin to the CNTs in the different complexes.¹⁷ In 2-non, the porphyrin molecule could be kept very close to the CNTs surface by π–π stacking interactions such as “face-to-face”, which was favourable for the electron-transfer process and so caused a stronger fluorescence quenching; in 2, the amide bond means a relatively longer distance from porphyrin to the CNTs, which led to a more difficult electron-transfer process and finally a relatively smaller quenching degree. Nevertheless, all of these results support the presence of an electron-transfer process from ATSPP to the MWCNTs, which proved that the MWCNTs could be modified by ASTPP successfully in either a covalent or a non-covalent way. Besides, the TGA curve of 2 showed no clear “major” weight loss but only a “gradual” one over the temperature range corresponding to the porphyrin (Fig. S3, ESI†), reflecting the absence of free porphyrins (herein, 1) in the sample (herein, 2), and in turn providing evidence for the covalent linkage of 1 with the MWCNTs.

Fig. 1 (a) The absorption spectra of ATSPP 1, ATSPP–MWCNTs 2 and ATSPP–π–π–MWCNTs 2-non (C_ATSPP, 1.0 × 10⁻⁶ mol L⁻¹; C_{ATSPP–MWCNTs}, 5.0 × 10⁻² g L⁻¹; C_{ATSPP–π–π–MWCNTs}, 5.0 × 10⁻² g L⁻¹). (b) The fluorescence spectra of ATSPP 1, ATSPP–MWCNTs 2 and ATSPP–π–π–MWCNTs 2-non (C_ATSPP, 1.0 × 10⁻⁷ mol L⁻¹; C_{ATSPP–MWCNTs}, 3.0 × 10⁻³ g L⁻¹; C_{ATSPP–π–π–MWCNTs}, 3.0 × 10⁻³ g L⁻¹).

Construction and confirmation of the inclusion complexes 3

In order to investigate the influence of CDs on the interactions between DNA and 2, two kinds of complex 3 (ATSPP–MWCNTs–α-CD 3a and ATSPP–MWCNTs–β-CD 3b) were constructed. The inclusion abilities of different CDs with 2 and their inclusion constants were further determined by fluorescence spectroscopy. Fig. 2 shows changes in the fluorescence characteristics of 2 in a KH₂PO₄–K₂HPO₄ buffer solution (pH 7.04) at various concentrations of α-CD and β-CD. The fluorescence excitation bands were fixed at 423 nm. Along with the increasing concentration of α-CD or β-CD, the fluorescence intensity of the emission bands at 654 nm decreased gradually, which illustrated the formation of the inclusion complexes 3.

Fig. 2 The fluorescence spectra of 2 (9 × 10⁻³ g L⁻¹) in a pH 7.04 phosphate buffer solution containing various concentration of (A) α-CD, and (B) β-CD at 25 °C (excitation wavelength: 423 nm). The concentrations of the CDs were: (1) 0 M; (2) 1.5 × 10⁻⁵ M; (3) 3.0 × 10⁻⁵ M; (4) 4.5 × 10⁻⁵ M; (5) 6.0 × 10⁻⁵ M; (6) 7.5 × 10⁻⁵ M; (7) 9.0 × 10⁻⁵ M; (8) 1.05 × 10⁻⁴ M; (9) 1.2 × 10⁻⁴ M. Inset: the linear plot of 1/(F − F₀) versus 1/[CD].

The inclusion constant (K) is an important parameter, which represents the inclusion capacity. K can be obtained from fluorescence data by the modified Benesi–Hildebrand equation:²⁸

Herein, F and F₀ represent the fluorescence intensities of 2 in the presence and absence of CDs, respectively. [P]₀ denotes the initial concentration of 2 and [CD]₀ denotes that of the CDs. k is the instrument constant, and Q is the fluorescence quantum yield of the inclusion complex. n is the inclusion ratio (the stoichiometry of CD versus the corresponding guest molecule 2). K, the inclusion constant of 3, was determined using a doubled reciprocal method. 1/(F − F₀) versus 1/[CD] was plotted and K was obtained from the ratio of the intercept to the slope. According to this formula, the K values for the inclusion complexes of α-CD and β-CD with 2 were calculated as 5.3 × 10³ M⁻¹ and 3.1 × 10³ M⁻¹, respectively, with the same stoichiometry of 1 : 1. However, when similar experiments were performed under the same conditions using γ-CD, few changes in the fluorescence spectra could be observed, indicating that there was no interaction between γ-CD and 2. These results showed that α-CD and β-CD could include ATSPP–MWCNTs 2 while γ-CD could not, and that the inclusion ability of α-CD was stronger than that of β-CD. As is well known, these three kinds of CDs have no differences in the molecular structure except for the number of glucose units, which results in different molecular dimensions. Among them, α-CD has the smallest size of hydrophobic cavity (α-CD: 13.7/5.7 Å; β-CD: 15.3/7.8 Å; γ-CD: 16.9/9.5 Å), which is closest to that of the phenyl group (about 5.8 Å). Meanwhile, “size matching” is known to be the first important factor for determining the stability of CD inclusion complexes.²² Since the peripheral substituent of 2 (sulphonatophenyl) is too small for the cavity of γ-CD, it would easily pass in and out the cavity with little bonding. Thus, in the present work, α-CD was found to best include 2 and form the most stable complex, ATSPP–MWCNTs–α-CD 3a.

For contrast, two kinds of inclusion complex 3-non (ATSPP–π–π–MWCNTs–α-CD 3a-non and ATSPP–π–π–MWCNTs–β-CD 3b-non) together with two kinds of inclusion complex 3-por (ATSPP–α-CD 3a-por and ATSPP–β-CD 3b-por) were constructed in the same way as above-mentioned. The changes in the fluorescence characteristics of 2-non or ATSPP in the KH₂PO₄–K₂HPO₄ buffer solution (pH 7.04) at various concentrations of α-CD and β-CD are shown in Fig. S4 or S5 (ESI†), respectively. Similarly, no change could be observed when using γ-CD. The K values for the inclusion complexes 3a-non, 3b-non, 3a-por and 3b-por were calculated as 1.4 × 10³ M⁻¹, 8.9 × 10² M⁻¹, 7.2 × 10³ M⁻¹ and 6.4 × 10³ M⁻¹, respectively, with the same stoichiometry of 1 : 1. Compared to the interactions between the CDs and 2, there were a lot of similarities in the interactions between the CDs and 2-non or the free porphyrin ATSPP, such as: (i) the changes in the fluorescence intensity shown in Fig. S4 or S5 (ESI†) decreased gradually in the same way as shown in Fig. 2, and the emission bands also peaked at 654 nm; (ii) 2-non or ATSPP could be included by α-CD and β-CD, but not by γ-CD; (iii) α-CD could include 2-non or ATSPP better than β-CD, and all of their stoichiometries were the same at 1 : 1. Besides, the value of their inclusion constants K changed regularly with the included (guest) compounds accordingly: K_3a-por > K_3a > K_3a-non; K_3b-por > K_3b > K_3b-non. Obviously, as guest compounds for CDs, the free porphyrin ATSPP could be the best one, the covalent complex (ATSPP–MWCNTs) was the second best, while the non-covalent complex (ATSPP–π–π–MWCNTs) was the worst. The reasons might be as follows: (i) once the water-soluble ATSPP was complexed with the insoluble MWCNTs, the originally good compatibility between the free porphyrin and the CDs would decrease significantly, and meanwhile between them a steric hindrance caused by the bulky MWCNTs would arise immediately. So the free porphyrin ATSPP could interact better with the CDs than ATSPP–MWCNTs or ATSPP–π–π–MWCNTs; (ii) in the non-covalent complex the porphyrin molecule could be kept very close to the CNTs surface, while in the covalent complex there might be a relatively longer distance between ATSPP and the MWCNTs due to the amide bond, so the steric hindrance between ATSPP–π–π–MWCNTs and the CDs would be greater than that between ATSPP–MWCNTs and the CDs. Moreover, ATSPP–π–π–MWCNTs showed poorer stability and dispersity than ATSPP–MWCNTs. Thus, the covalent complex ATSPP–MWCNTs could interact better with CDs than the non-covalent complex ATSPP–π–π–MWCNTs.

Construction of supramolecular assemblies based on 2 or 3

Based on the investigation of the interactions between 2 and CDs, comparative studies on DNA interactions with 2 and 3 were also carried out, respectively. To the best of our knowledge, no research concerning DNA condensation with anionic porphyrin-covalently modified CNTs has been reported. Thus, it is valuable to investigate whether DNA can interact with complex 2, in which the anionic porphyrin has been “fixed” onto the surface of the π-conjugated MWCNTs via a covalent bond. In the current experiments, upon the addition of various concentrations of DNA (0–72 μL) into an aqueous solution of 2, the fluorescence spectrum of 2 exhibited about 25 ± 0.26% (n = 6) quenching of the emission bands at 654 and 716 nm (Fig. 3A), indicating that the interaction between DNA and 2 indeed happened. Additionally, contrasting experiments were performed using ATSPP 1 or 2-non and few changes in the fluorescence spectra were observed, indicating that there was no interaction between the DNA and 1 (or 2-non) (Fig. S4 or S5, ESI†: dotted line 1 in the left inset). These results attested that CNTs could facilitate the binding of the negatively-charged porphyrin with the negatively-charged DNA, as expected, but only when they were “covalently” modified by anionic porphyrins. This probably occurred since (i) the anionic porphyrin in complex 2 was “fixed” firmly on the surface of the carbon nanotubes via a stable covalent bond, resulting in less contact with the DNA and so less repulsions, (ii) in complex 2, the CNTs bearing the anionic porphyrin could non-covalently interact with DNA owing to π-stacking interactions between the sidewalls of the CNTs and the bases in DNA.¹⁵ But for complex 2-non, the CNTs had already non-covalently interacted with the anionic porphyrin by π–π stacking and some of their surfaces were “covered” with “free” anionic porphyrin molecules, so that they had less space for DNA and their “anionic-porphyrin-covered” surface would even show great resistance against DNA due to the electrostatic repulsions. Herein, by covalently linking 1 with MWCNTs to form 2, the inherent electrostatic repulsions between 1 and DNA might be reduced, and the π-stacking interactions existing between the CNTs and DNA could cause the successful binding of 2 with DNA.

Fig. 3 The fluorescence spectra of 2 (9 × 10⁻³ g L⁻¹) in a pH 7.04 phosphate buffer solution containing various concentrations of DNA in the presence of (A) 0 μL CD, (B) 80 μL α-CD (1.2 × 10⁻⁴ M), and (C) 80 μL β-CD (1.2 × 10⁻⁴ M) at 25 °C (excitation wavelength: 423 nm). The concentrations of the DNA were: (1) 0 M; (2) 1.8 × 10⁻⁶ M; (3) 3.6 × 10⁻⁶ M; (4) 5.4 × 10⁻⁶ M; (5) 7.2 × 10⁻⁶ M; (6) 9.0 × 10⁻⁶ M; (7) 1.08 × 10⁻⁵ M; (8) 1.26 × 10⁻⁵ M; (9) 1.44 × 10⁻⁵ M. Inset: the linear plot of 1/(F − F₀) versus 1/[DNA].

Further experiments were performed to investigate the interactions between DNA and 3, as well as to study the influence of CDs on this process. Fig. 3B and C display the changes in the fluorescence spectra when a DNA solution (0–72 μL) was gradually added into the aqueous solution of 2 in the presence of α-CD and β-CD, respectively, under the same experimental conditions. In Fig. 3B and C, a gradual decrease in the fluorescence intensity of the emission bands peaked at 654 nm could be observed, with a quenching of about 74 ± 0.32% for 3a and 46 ± 0.29% for 3b, respectively. The difference was significant (n = 6, p ≤ 0.05, t test, two sides). This suggested that both 3a and 3b could interact with DNA to construct the corresponding supramolecular assemblies ATSPP–MWCNTs–α-CD–DNA 4a and ATSPP–MWCNTs–β-CD–DNA 4b. Thus, 3a as well as 3b exhibited a relatively stronger affinity for DNA than 2 (the fluorescence quenching was about 25%), indicating that these two kinds of CDs could promote the interaction between DNA and 2 as expected. At the same time, 3a exhibited a greater degree of emission quenching than 3b, implying its relatively stronger binding affinity for DNA in line with its stronger inclusion capability of 2. However, when control experiments were performed under the same conditions to investigate the interactions between DNA and 3-non or 3-por, few changes in the fluorescence spectra could be observed, indicating that there was no interaction between DNA and 3-non or 3-por (including both 3a-por and 3b-por) (Fig. S4 or S5, ESI†: dotted lines 8 and 9 in the left inset). The results indicated that CDs could promote the binding of DNA with ATSPP only when ATSPP was covalently modified onto the MWCNTs surface. Obviously, in the inclusion complex 3-por, although the CDs could include phenyl rings inside their hydrophobic cavities and might isolate the sulfonato groups from the DNA to reduce their repulsions to a certain extent, the electrostatic repulsions between DNA and 3-por were still strong enough to prevent them from binding. But once the anionic porphyrin was “fixed” onto the MWCNTs, as mentioned, based on a strong π-stacking interaction between the CNTs and DNA, a stable supramolecular assembly including both the anionic porphyrin and negative DNA would be successfully constructed; and in this case, amphiphilic CDs could effectively improve the solubility and dispersity of ATSPP–MWCNTs and the stability of ATSPP–MWCNTs–DNA and so promote the final formation of the complex ATSPP–MWCNTs–CDs–DNA.

Recently, resonance light-scattering (RLS) has been regarded as a useful tool for the investigation of supramolecular complexes.²⁹ The amount of scattered light is directly proportional to the volume of the particles, and monomeric molecules or small oligomers do not show enhanced scattering.³⁰ Herein, RLS technology was used to detect the interactions between DNA and 3. Prior to the addition of DNA, the RLS of ATSPP–MWCNTs–α-CD 3a was mono-exponential, and its special profile was determined (Fig. 4, line 1, black). Upon the addition of different amounts of DNA (4–12 μL, 2 × 10⁻³ M), the RLS intensity changed from 450 to 990, implying the formation of the supramolecular assembly ATSPP–MWCNTs–α-CD–DNA 4a (Fig. 4, line 2–4).

Fig. 4 The RLS spectra of 3a in solution with (1) 0 μL, (2) 4 μL, (3) 8 μL, and (4) 12 μL of DNA (2 × 10⁻³ M).

The morphologies of the MWCNTs, 2, 3a and 4a were observed by TEM, providing direct evidence for the whole formation process of 4a. Fig. 5A shows the unmodified MWCNTs, and we could see that the MWCNT’s wall was very smooth. In contrast, the surfaces of dyad 2 showed a thick coverage (about 2.7 nm) of porphyrin 1 on the sidewalls of the MWCNTs, indicating that the MWCNTs were modified by ATSPP (Fig. 5B). A similar observation was found for 3a in Fig. 5C, and the thickness of the layer attached on the carbon nanotubes was about 3.4 nm. From 2 to 3a, the thickness increase of the layer as shown in Fig. 5B and C was basically in conformity with the size of α-CD.²² In Fig. 5D, there is clearly a thicker layer (about 5 nm) attached on the MWCNTs, which suggested a thick coverage of DNA onto the sidewalls of the MWCNTs moiety of 3a.

Fig. 5 TEM photomicrographs of (A) MWCNTs 1; (B) ATSPP–MWCNTs 2; (C) ATSPP–MWCNTs–α-CD 3a; and (D) ATSPP–MWCNTs–α-CD–DNA 4a.

Conclusions

In this paper, based on water-soluble anionic 5-(p-aminophenyl)-10,15,20-tri(p-sulphonatophenyl)-porphyrin (ATSPP) 1, two nanocomposites, ATSPP–MWCNTs dyad 2 and ATSPP–MWCNTs–CDs triad 3 were prepared. Using these two nanocomposites to interact with DNA respectively, the corresponding supramolecular assemblies, ATSPP–MWCNTs–DNA and the targeted ATSPP–MWCNTs–CDs–DNA tetrad 4, were successfully constructed, and the effects of the CNTs or CDs on the binding of DNA with anionic porphyrins were studied in detail through UV/vis, fluorescence, RLS and TEM. For contrast, a series of complexes, ATSPP–π–π–MWCNTs 2-non, ATSPP–π–π–MWCNTs–CDs 3-non and ATSPP–CDs 3-por, were prepared and control experiments were performed to investigate their interactions with DNA. Results showed that both CNTs and CDs are favourable for the construction of supramolecular assemblies containing the negatively-charged anionic porphyrin 1 and the negatively-charged DNA. However, CNTs could facilitate the anionic porphyrin binding with DNA by increasing π–π interactions only when they were covalently modified by the anionic porphyrin; while CDs could only promote the interaction between DNA and ATSPP–MWCNTs, probably by greatly improving the solubility and dispersity of the ATSPP–MWCNTs based on the effective inclusion complexation of CDs with ATSPP. Our present work developed a new route for the construction of anionic porphyrin–DNA complexes via supramolecular assembly, which is significant for the advancement of porphyrin–DNA chemistry and for their promising applications in drugs and clinical therapies. Further studies on the interaction mechanisms between anionic porphyrins and repulsive DNA are in progress in our laboratory.

Acknowledgements

The authors gratefully acknowledge financial support from the National Natural Science Funds for Young Scholar of China (no. 31301564), the Projects in the National Science & Technology Pillar Program during the Twelfth Five-Year Plan Period (no. 2012BAD31B08), Hunan Provincial Natural Science Foundation (no. 2015JJ2011) and Hunan Provincial Engineering Research Center for Food Processing of Aquatic Biotic Resources (no. 2015GCZX004), Changsha University of Science & Technology, PR China.

References

1. K. Zhu, X. Y. Hu, Q. Y. Ge and Q. J. Sun, Anal. Chim. Acta, 2014, 812, 199.
2. S. Ghosh, K. B. Ucer Jr, R. D’Agostino, K. Grant, J. Sirintrapun, M. J. Thomas, R. Hantgan, M. Bharadwaj and W. H. Gmeiner, Nanomed. Nanotech. Biol. Med., 2014, 10, 451.
3. B. Jin, H. M. Lee, Y. A. Lee, J. H. Ko, C. Kim and S. K. Kim, J. Am. Chem. Soc., 2005, 127, 2417.
4. R. J. Fiel, J. Biomol. Struct. Dyn., 1989, 6, 1259.
5. M. A. Bork, C. G. Gianopoulos, H. Y. Zhang, P. E. Fanwick, J. H. Choi and D. R. McMillin, Biochemistry, 2014, 53, 714.
6. J. Manono, P. A. Marzilli and L. G. Marzilli, Inorg. Chem., 2009, 48, 5636.
7. G. P. Liu, Y. M. Wu, Y. L. Yuan, Y. Q. Chai, S. Q. Wei and D. J. Zhang, RSC Adv., 2014, 4, 52117.
8. Y. F. Li, C. R. Geyer and D. Sen, Biochemistry, 1996, 35, 6911.
9. R. C. Haushalter and R. W. Rudolph, J. Am. Chem. Soc., 1978, 100, 4628.
10. J. Winkelman, Cancer Res., 1962, 22, 589.
11. F. Niedercorn, New J. Chem., 1988, 12, 897.
12. R. Lauceri, R. Purrello, S. J. Shetty, M. Gracüa and H. Vicente, J. Am. Chem. Soc., 2001, 123, 5835.
13. A. D’Urso, A. Mammana, M. Balaz, A. E. Holmes, N. Berova, R. Lauceri and R. Purrello, J. Am. Chem. Soc., 2009, 131, 2046.
14. J. K. Choi, A. D’Urso and M. Balaz, J. Inorg. Biochem., 2013, 127, 1.
15. M. Zheng, A. Jagota, M. S. Strano, A. P. Santos, P. Barone, S. G. Chou, B. A. Diner, M. S. Dresselhaus, R. S. McLean, G. B. Onoa, G. G. Samsonidze, E. D. Semke, M. Usrey and D. J. Walls, Science, 2003, 302, 1545.
16. H. B. Zhao, H. K. Wang, H. Chang, S. J. Qiu, B. Y. Deng and J. X. Liao, Chin. J. Chem., 2011, 29, 1901.
17. H. Zhang, H. B. Zhao, X. J. Lu, W. L. Wang, H. Chang and Y. F. Wang, Acta Chim. Sin., 2011, 69, 316.
18. D. Baskaran, J. W. Mays, X. P. Zhang and M. S. Bratcher, J. Am. Chem. Soc., 2005, 127, 6916.
19. Y. Wang and A. H. Zhou, J. Photochem. Photobiol., A, 2007, 190, 121.
20. W. L. Wang, H. B. Zhao, J. H. Ning, Y. Hu and X. J. Lu, J. Anal. Sci., 2011, 27, 183–186.
21. Y. Hu, H. B. Zhao, J. H. Ning, M. X. Ning, Y. J. Xu, S. J. Qiu and J. X. Liao, Huaxue Xuebao, 2008, 66, 2391.
22. S. Li and W. G. Purdy, Chem. Rev., 1992, 92, 1457.
23. R. H. Yang, K. A. Li, K. M. Wang, N. Li, F. Zhao and F. Liu, Anal. Chem., 2003, 75, 612.
24. D. Pantarotto, R. Singh, D. McCarthy, M. Erhardt, J. P. Briand, M. Prato, K. Kostarelos and A. Bianco, Angew. Chem., 2004, 116, 5354.
25. J. Carlstedt, D. Lundberg, R. S. Dias and B. Lindman, Langmuir, 2012, 28, 7976.
26. W. J. Kruper, T. A. Chamberlin and M. Kochanny, J. Org. Chem., 1989, 54, 2753.
27. D. M. Guldi, G. M. A. Rahman, N. Jux, N. Tagmatarchis and M. Prato, Angew. Chem., Int. Ed., 2004, 43, 5526.
28. G. C. Catena and F. V. Bright, Anal. Chem., 1989, 61, 905.
29. R. F. Pasternack, C. Bustamante, P. J. Collings, A. Giannetto and E. J. Gibbs, J. Am. Chem. Soc., 1993, 115, 5393.
30. R. F. Pasternack and P. J. Collings, Science, 1995, 269, 935.

---

Footnote

† Electronic supplementary information (ESI) available: ¹H NMR spectrum of 1 in DMSO-d₆. FT-IR spectra of 1, 2 and 2-non. Thermogravimetric curve of 2. The fluorescence spectra of 2-non interacting with (A) α-CD, (B) β-CD and then with DNA. The fluorescence spectra of 1 interacting with (A) α-CD, (B) β-CD and then with DNA. See DOI: 10.1039/c4ra15741a

---