Assembling aggregation‐induced emission with natural DNA to maximize donor/acceptor ratio for efficient light‐harvesting antennae

Arranging dense donors around a single acceptor for the assembly of efficient light‐harvesting antennas is a long‐standing challenge due to the intractable aggregation‐caused quenching of dense donors. Herein, we designed a cationic aggregation‐induced emission (AIE) amphiphile to self‐assemble with natural DNA duplexes. As an efficient donor, the as‐prepared cationic AIE amphiphile could be densely attached to the phosphate groups of natural DNA duplexes by using the smaller cationic trimethylammonium. The long alkyl chain between the cationic trimethylammonium and the AIE fluorophore allowed for avoiding the insufficient binding caused by the steric hindrance of the AIE fluorophore, resulting in a remarkably high donor/acceptor ratio comparable to that of the widely developed custom DNA assemblies. The proposed self‐assembly strategy provided novel flexible avenues for the assembling of finely controlled and efficient light‐harvesting systems into natural DNA with little synthetic modifications and low cost.


INTRODUCTION
The natural light-harvesting antennas enable efficient collection and conversion of solar energy for natural photosynthesis by pairing more than 200 energy donors with a single energy acceptor. [1][2][3][4] Nowadays, numerous artificial light-harvesting antennas have been developed to convert light energy into chemical energy for applications in photocatalysis, sensors, imaging, and photodynamic therapy. [5][6][7][8][9] Amplifying the donor/acceptor ratio within the effective distance of energy transfer is highly required. [10][11][12] The widely developed custom DNA duplexes have exhibited great potential in arranging a high-density distribution of energy donors through the precise design and synthesis of specific DNA sequences. [13][14][15] Generally, multiarmed DNA junctions were designed to contain multiple DNA arms and one hub, where the energy donors and acceptors can be integrated into the DNA arm and the hub. [16][17][18][19][20][21] With the increase of DNA arms, the increase in donor/acceptor ratio can lead to higher light-harvesting efficiency. For example, a maximum donor/acceptor ratio of 48:1 was obtained when the number of DNA arms and donors per DNA arm reach 8 and 6, respectively. [21] It is worth noting that custom DNA is usually sophisticated, costly, and small-scale. Therefore, there is a long-standing desire to explore a new strategy to fabricate efficient light-harvesting antennas using simple and low-cost natural DNA produced by methods other than oligo synthesis.
Natural DNA inherently owns a series of binding interactions (e.g., electrostatic, hydrogen bonding, and hydrophobic) due to the presence of phosphate groups, nitrogenous bases, and deoxyribose sugar. [22][23][24][25][26][27] For a helical turn consisting of 10 base pairs (∼3.4 nm), there are 20 negatively charged phosphate groups, a hydrophobic major groove of ∼1.2 nm wide, and a hydrophobic minor groove of ∼0.6 nm F I G U R E 1 Schematic illustration of dense binding of aggregationinduced emission (AIE) donors into natural DNA for constructing an efficient light-harvesting system.
wide. [26] In principle, the effective distance of Förster resonance energy transfer (FRET) is ∼10 nm, and thus 30 base pairs (∼60 phosphate groups) could be extended in the opposite directions along the DNA backbone with a minor groove as the center. [28,29] If the donor chromophores are ideally attached to each phosphate group by electrostatic interaction, and the energy acceptor only binds to the minor groove through hydrophobic interaction, efficient light-harvesting antennas could be self-assembled with a donor/acceptor ratio of 120:1. However, such a close packing of donor chromophores would induce intractable aggregation-caused quenching (ACQ) problem. [30][31][32][33][34] To mitigate these barriers, the larger donors with a binding site size of 2 base pairs or base-specific donors were favored to avoid aggregation. [34] However, these strategies could significantly decrease the binding number of donor chromophores, resulting in a lower donor/acceptor ratio. Therefore, it seems to be a great challenge to assemble natural DNA with ACQ donors to maximize the donor/acceptor ratio for efficient light harvesting.
Aggregation-induced emission (AIE) is an exciting photophysical phenomenon that has contributed significantly to the development of luminescent materials in the past two decades. [35,36] Unlike conventional dyes, AIE materials usually have nonplanar configurations, which exhibit poorly emissive features in dilute solutions but intense emission in the aggregated or solid states. In this contribution, tetraphenylethylene (TPE) with AIE property was covalently attached to the tail of decyltrimethylammonium bromide (C 10 TAB) to synthesize the cationic AIE amphiphile (TPE-C 10 TAB). Trimethylammonium cation of the smaller size allowed TPE-C 10 TAB to densely bind phosphate groups in the natural DNA duplexes, while the long carbon chain between trimethylammonium cation and TPE enabled the larger TPE to aggregate away from the phosphate groups and grooves. The obtained DNA-TPE-C 10 TAB complex showed efficient luminescence without ACQ problems. On the other hand, hydrophobic Nile red (NR), which had the potential to bind to the minor groove of natural DNA duplexes, was used as an acceptor model. By tuning the TPE-C 10 TAB/NR ratio, an efficient light-harvesting antenna with a considerable antenna effect (28.1) was constructed without any modification of natural DNA duplexes ( Figure 1). The versatility of the proposed strategy was verified by replacing NR with another hydrophobic acceptor. The success of this work could be widely applied to various DNA-based nanostructures due to the inherent presence of electrostatic and hydrophobic interactions. Moreover, the exclusion of covalent modifications in assembling efficient light-harvesting systems offers the benefits of more flexibility, low cost, and easy operation.

RESULTS AND DISCUSSION
The chemical structure and synthetic route of TPE-C 10 TAB were shown in Figure 2A and Figure S1, respectively. Briefly, hydroxyl groups of para-hydroxy-tetraphenylethylene (TPE-OH) were activated using K 2 CO 3 in acetone and reacted with 1,10-dibromodecane (BrC 10 Br) to form the bromofunctionalized TPE-C 10 Br. [37] The purified TPE-C 10 Br was verified by 1 H nuclear magnetic resonance (NMR) spectroscopy ( Figure S2) and was easily converted to TPE-C 10 TAB by reacting with trimethylamine for 72 h. [38] The corresponding 1 H NMR ( Figure 2B), 13 C NMR ( Figure S3), and positive-ion mode mass spectra ( Figure S4) confirmed the successful synthesis of TPE-C 10 TAB. TPE-C 10 TAB is a typical cationic amphiphile and its critical micelle concentration (CMC) in water was determined by the conductivity method. The measured conductivity (κ) showed two linear relationships with the TPE-C 10 TAB concentration, where the breaking point was around 46 µM, indicating the CMC ( Figure 2C). [39] The corresponding optical properties at concentrations below and above CMC were investigated by measuring the absorption and fluorescence spectra. As shown in Figure 2D, the absorption peaks of TPE-C 10 TAB aqueous solution appeared at about 247 and 314 nm, [40] displaying no changes with concentration ( Figure  S5). The absorbance at 314 nm increased linearly with increasing the TPE-C 10 TAB concentration in the range of 10-100 µM ( Figure S6). On the other hand, the fluorescence emission wavelength of TPE-C 10 TAB aqueous solutions is located at around 475 nm ( Figure 2D). As the concentration of TPE-C 10 TAB increased above CMC, the emission intensity continued to increase without the ACQ phenomenon ( Figure S7). These results demonstrated that cationic TPE-C 10 TAB amphiphile can form micelles in aqueous solutions and emit intense fluorescence in the aggregated state.
Theoretically, cationic amphiphiles are able to bind to negatively charged DNA in the form of molecules or micelles through electrostatic interaction. [41] In consideration of the CMC of TPE-C 10 TAB, 20 µM, and 60 µM were used for assembling with calf thymus DNA (ctDNA) to obtain two types of supramolecular assemblies (i.e., the DNA-TPE-C 10 TAB and the DNA-TPE-C 10 TAB micelle). Fluorescence emission spectra of the DNA-TPE-C 10 TAB ( Figure S8) and the DNA-TPE-C 10 TAB micelle ( Figure S9) were studied at different DNA concentrations. They showed the same emission peak as TPE-C 10 TAB with enhanced emission intensity. The fluorescence enhancement could be attributed to the restriction of the intramolecular movement of TPE-C 10 TAB bound to the DNA duplexes. [42] Through plotting the emission intensity against the DNA concentration, it was found that the DNA-TPE-C 10 TAB ( Figure 3A) and the DNA-TPE-C 10   of 10 and 20 µg/mL, respectively. As the DNA concentration continued to increase, the fluorescence intensity of the DNA-TPE-C 10 TAB and the DNA-TPE-C 10 TAB micelle remained constant. The resulting DNA-TPE-C 10 TAB consisting of 20 µM of TPE-C 10 TAB and 10 µg/mL of DNA, and the DNA-TPE-C 10 TAB micelle consisting of 60 µM of TPE-C 10 TAB and 20 µg/mL of DNA were used as the light-harvesting models. Furthermore, the binding ratio of TPE-C 10 TAB to DNA in these two light-harvesting models was investigated by absorption spectroscopy. As shown in Figure S10, the absorbance A of DNA at 260 nm was 0.167 (10 µg/mL) and 0.324 (20 µg/mL), respectively. The concentration of phosphate groups in the solution was calculated by the Lambert-Beer law A = εbc as 25 µM and 49 µM, respectively. [43] These results indicated that the ratio of the number of bases to the number of TPE-C 10 TAB in the DNA−TPE-C 10 TAB assembly was approximately 1.25. On the other hand, the positive charge in the DNA−TPE-C 10 TAB micelle assemblies was in excess, which may be due to the fact that only some of the positive charge can be bound to the phosphate groups of DNA after the formation of micelle aggregates.
Hydrophobic acceptor chromophores have the potential to be self-assembled into the hydrophobic grooves of DNA duplexes. [44] Hydrophobic NR was chosen as the acceptor model because its absorption spectrum overlapped well with the emission spectrum of TPE-C 10 TAB ( Figure S11A). The gray overlap area between the excitation spectrum of the DNA-TPE-C 10 TAB assembly/NR at the emission wavelength of NR (625 nm) and the absorption spectrum of TPE-C 10 TAB indicated that the excitation light (330 nm) was absorbed by TPE-C 10 TAB, leading to a subsequent energy transfer from the donor to the acceptor ( Figure S11B). The fluorescence spectra of the DNA-TPE-C 10 TAB ( Figure 3C) and the DNA-TPE-C 10 TAB micelle ( Figure 3D) were measured at different NR concentrations under 330 nm excitation. With the increase of NR concentration, the emission intensity of TPE-C 10 TAB at 475 nm decreased, whereas the emission intensity of NR at 625 nm increased. The corresponding antenna effects were calculated according to the known formula. [45] As expected, the antenna effects of the DNA-TPE-C 10 TAB ( Figure S12) were generally better than those of the DNA-TPE-C 10 TAB micelle ( Figure S13). In addition, we performed control experiments in the absence of DNA ( Figure S14). Similarly, a decrease in donor emission intensity can be noted accompanied by an increase in acceptor emission intensity. The calculated antenna effects (Figures S15 and S16) were much lower than those of the DNA-TPE-C 10 TAB and DNA-TPE-C 10 TAB micelle systems, confirming the potential role of DNA as a lightharvesting framework. Therefore, the DNA-TPE-C 10 TAB offered a better possibility of maximizing the number of binding donors.
Fluorescence decays of the DNA-TPE-C 10 TAB were measured in the presence of different concentrations of NR to further investigate the light-harvesting behavior. As shown in Figure 4A and Table S1, the decay trace of each system was fitted well to the sum of three exponential functions, where the shorter τ 1 may represent TPE-C 10 TAB in the looser aggregation state and the longer τ 2 and τ 3 may represent TPE-C 10 TAB in the tighter aggregation state. [46] The fluorescence lifetimes of the DNA-TPE-C 10 TAB gradually decreased from 4.26 ns to 2.26 ns with the increase of the TPE-C 10 TAB/NR ratio, indicating an effective energy transfer between them. [5] Based on the fluorescence lifetime of the time-resolved data of DNA-TPE-C 10 TAB in Figure 4A, the FRET efficiency (Ф ET ) was calculated to be 46.9%, 39.5%, 19.0%, and 16.2%, [47] corresponding to the TPE-C 10 TAB/NR ratios of 100:1, 250:1, 500:1, and 1000:1, respectively ( Figure 4B). In comparison, the antenna effects were 21.9, 27.6, 28.1, and 26.2, respectively. The maximum antenna effect was achieved at TPE-C 10 TAB/NR ratios between 250:1 and 500:1, which coincided with the ratios in nature. [48,49] Compared with other reported energy transfer results, this system had a moderately high Ф ET (Table S2). These results confirmed the great potential of DNA-TPE-C 10 TAB to mimic natural light-harvesting systems in aqueous environments.
To elucidate the possible structure of the proposed lightharvesting system, the DNA-TPE-C 10 TAB was investigated by transmission electron microscopy (TEM), dynamic light scattering, zeta potential measurements, and circular dichroism (CD) spectroscopy. As shown in Figure S17, the zeta potentials of ctDNA, DNA-TPE-C 10 TAB, and DNA/NR were −53.3, −31.8, and −54.2 mV, respectively. DNA/NR and ctDNA had nearly the same zeta potential, indicating that electrically neutral NR had no effect on the surface potential of negatively charged DNA. [50] In contrast, the significant increase in the zeta potential of DNA-TPE-C 10 TAB implied that cationic TPE-C 10 TAB was electrostatically attracted to the surface of negatively charged DNA. [51] Notably, the negative potential exhibited by DNA−TPE-C 10 TAB corresponded to the presence of a small amount of unbound phosphate groups in the assembly. Subsequently, their interactions were analyzed in detail by comparing the CD signals of ctDNA duplex in the absence and presence of NR and TPE-C 10 TAB. As shown in Figure 5A, the CD spectrum of ctDNA duplex had a positive peak at 276 nm (attributed to base stacking) and a negative peak at 247 nm (attributed to helicity), indicating that ctDNA was the right-handed B form DNA. [52] Intercalation resulted in simultaneous changes in the base stacking and helicity, whereas groove binding and electrostatic interaction caused little or no perturbation of the base stacking. [53] Apparently, the intensity of negative and positive peaks of ctDNA did not change after binding to NR or TPE-C 10 TAB, which suggested that NR and TPE-C 10 TAB were not assembled with ctDNA by intercalation. Further converting the CD signal to the molecular delta epsilon (Figure S18), the spectra remained unchanged significantly, indicating that the introduction of TPE-C 10 TAB or NR did not change the secondary structure of DNA.
UV-visible absorption spectroscopy is a simple and effective method of detecting interaction between small molecules and DNA. [54] Compared to groove binding and electrostatic interaction, intercalation generally produces hypochromism along with a redshift of the absorption spectra. [55] Because intercalative interaction is a kind of stacking interaction between base pairs of DNA and small chromophores, the π orbital of base pairs of DNA can couple with the π* orbital of small chromophores and form π-π* conjugation. The reduced difference of the π-π* transition energy causes a redshift of the absorption band. The distance between chromophores and base pairs will also decrease when small chromophores intercalated into DNA, resulting in hypochromism of the absorption band. [56,57] As shown in Figure S19, NR showed a maximum absorption band at around 595 nm. With the addition of ctDNA, neither NR nor ctDNA showed redshift and hypochromism, which indicated the formation of a complex between NR and ctDNA different from the intercalation process. The observed increase in the fluorescence intensity of NR with increasing ctDNA concentration indicated the binding of NR to ctDNA ( Figure   S20). [58] Moreover, TEM images showed that the DNA-TPE-C 10 TAB ( Figure 5B) had the same morphology as ctDNA ( Figure S21), which possessed an elongated coiled conformation. [59] The corresponding hydrodynamic diameter was determined to be about 62.04 nm (inset of Figure 5B). Therefore, TPE-C 10 TAB was electrostatically attracted to the ctDNA surface to form the DNA-TPE-C 10 TAB, in which the structure and morphology of natural DNA remained unchanged.
The versatility of the proposed light-harvesting system was examined using an alternative acceptor, 4,7-dithien-2-yl-2,1,3-benzothiadiazole (DBT). In comparison with NR, DBT has a smaller molecular size and stronger hydrophobicity, which would be easier to bind to the minor groove. [60][61][62] The absorption spectrum of DBT overlapped well with the emission spectrum of TPE-C 10 TAB ( Figure S22A), and the excitation spectrum of the DNA−TPE-C 10 TAB assembly/DBT at the emission wavelength of DBT (540 nm) overlapped well with the absorption spectrum of the donor TPE-C 10 TAB ( Figure S22B). The fluorescence spectra of the DNA-TPE-C 10 TAB were measured in the presence of different concentrations of DBT under 330 nm excitation ( Figure 6A). With the increase in DBT concentration, the emission intensity of TPE-C 10 TAB at 475 nm decreased, whereas the emission intensity of DBT at 540 nm increased. The fluorescence decay measurements validated the energy transfer process ( Figure S23 and Table S3). When the mixing molar ratio of TPE-C 10  fluorescence lifetime and FRET efficiency of the system were estimated to be 2.93 ns and 31.1%. The corresponding antenna effect can reach 27.6 ( Figure S24), strongly supporting that the DNA-TPE-C 10 TAB offered superior versatility to yield efficient light-harvesting antennas. Additionally, the DNA-TPE-C 10 TAB showed good potential in fabricating white light-emitting systems by simply tuning the concentrations of NR ( Figure S25). As shown in Figure 6B, the corresponding Commission Internationale de l'Eclairage (CIE) coordinate was found to be (0.33, 0.30), which was close to that of the pure white (0.33, 0.33). [63]

CONCLUSION
In summary, we have designed a simple and highly flexible method to use natural DNA as a scaffold for the creation of specific binding sites. The as-obtained DNA-TPE-C 10 TAB was capable of achieving a high donor/acceptor ratio by conveniently noncovalent self-assembly, rather than traditionally covalent modifications or sophisticated design of custom DNA sequences. The dense binding of TPE-C 10 TAB donors to phosphate groups of natural DNA duplexes could emit strong luminescence without ACQ problems. As the hydrophobic acceptors had the potential to approach the minor grooves, efficient light-harvesting antennas were successfully constructed with a high antenna effect. The interaction between AIE and phosphate groups was independent of the DNA sequence, providing novel ideas and considerable flexibility in the design of DNA-based light-harvesting antennas. The proposed light-harvesting antennas would be utilized in various photonic applications by varying different types of acceptors.

A C K N O W L E D G M E N T S
This work was supported by the National Natural Science Foundation of China (U22A20397 and 21974008), the Beijing Natural Science Foundation (2212013), and the Fundamental Research Funds for the Central Universities (buctrc201820).

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflict of interest.