Double-Stranded DNA Matrix for Photosensitization Switching

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Sep 2021Double-Stranded DNA Matrix for Photosensitization Switching Yanying Wang, Hao Hu, Tianyu Dong, Hayam Mansour, Xinfeng Zhang, Feng Li and Peng Wu Yanying Wang Key Laboratory of Green Chemistry and Technology, Ministry of Education, College of Chemistry, Sichuan University, Chengdu 610064 , Hao Hu Analytical and Testing Center, Sichuan University, Chengdu 610064 , Tianyu Dong Department of Chemistry, Brock University, St. Catharines, ON L2S 3A1 , Hayam Mansour Department of Chemistry, Brock University, St. Catharines, ON L2S 3A1 Department of Cell Biology, National Research Centre, Cairo 12622 , Xinfeng Zhang College of Materials and Chemistry & Chemical Engineering, Chengdu University of Technology, Chengdu 610059 , Feng Li *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Key Laboratory of Green Chemistry and Technology, Ministry of Education, College of Chemistry, Sichuan University, Chengdu 610064 Department of Chemistry, Brock University, St. Catharines, ON L2S 3A1 and Peng Wu *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Key Laboratory of Green Chemistry and Technology, Ministry of Education, College of Chemistry, Sichuan University, Chengdu 610064 Analytical and Testing Center, Sichuan University, Chengdu 610064 https://doi.org/10.31635/ccschem.020.202000543 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail Photosensitization, originated from the activation of triplet states, is the basis of many photodynamic applications, but often competes with a series of nonradiative processes. Herein, we communicate a new function of double-stranded DNA (dsDNA) for label-free photosensitization switching. Up to ∼70-fold singlet oxygen generation boosting was observed for SYBR Green I (SG) upon binding with dsDNA. Detailed photophysical and theoretical studies have revealed the role of dsDNA as a matrix, which could efficiently suppress the nonradiative transitions of SG. Such photosensitization modulation is universal for a series of dsDNA-binding photosensitizers, including both base intercalators and minor groove binders. In conjunction with photochemical oxidation of chromogenic substrates, a simple and low-cost photosensitization-based colorimetric detection protocol has been developed, with sensitivity comparable to that of fluorescence detection. Through loop-mediated isothermal amplification (LAMP), colorimetric detection of hepatitis B virus (HBV) was achieved with a limit of detection (LOD) down to 1 aM, which is comparable with that of the standard quantitative polymerase chain reaction (PCR). To facilitate point-of-care testing, a simple and low-cost paper strip has been developed for distance-based detection of LAMP amplicons with a LOD of 100 aM for HBV DNA. Download figure Download PowerPoint Introduction Photosensitized singlet oxygen (1O2) generation has attracted significant research attention due to its intriguing applications in a myriad of photodynamic fields, such as photodynamic therapy (PDT),1–3 sunlight-activated insecticides,4 chromophore-assisted light inactivation of proteins,5 and analytical applications such as luminescent oxygen channeling immunoassay (LOCI)6 and in situ RNA localization assay.7,8 In a typical photosensitization process (Scheme 1a), a photosensitizer (PS) is first excited to its excited singlet state upon light irradiation (absorption). The excited-state energy can be either relaxed through emissions of prompt fluorescence or migrated to the triplet state (T1) via intersystem crossing (ISC). Subsequently, triplet-state PS may decay to the ground state (S0) to emit phosphorescence (anaerobic) or interact with dissolved oxygen to induce the generation of 1O2 (photosensitization). Therefore, modulating the excited state (S1) of PSs is an efficient strategy for photosensitization regulation,9–12 which complements common strategies such as heavy atom effect.13,14 Scheme 1 | Working mechanism of DNA hybridization-switched photosensitization in this work: (a) the Jablonski energy diagram of photosensitizer before and after matrix binding; and (b) the DNA matrix for photosensitization switching. Download figure Download PowerPoint As shown in Scheme 1a, photosensitization efficiency (1O2 generation efficiency, ΦΔ) as well as prompt fluorescence, both compete with a series of nonradiative processes, such as intramolecular motions and intermolecular collisions.15 Therefore, suppression of nonradiative transitions represents an efficient strategy for photosensitization modulation. Previously, we reported an interesting phenomenon that double-stranded DNA (dsDNA) could modulate the photosensitization of SYBR Green I (SG, a typical dsDNA-staining dye),16,17 but the mechanism was unclear. Herein, we have performed detailed photophysical and theoretical studies to reveal the role of dsDNA as a matrix in fixing PS SG, allowing activation of T1 by suppressing nonradiative transitions (Scheme 1b). For the first time, O2-free aqueous room-temperature phosphorescence (RTP) of SG was observed, together with up to a ∼74-fold increase of 1O2 generation efficiency. The fixing effect of dsDNA originates from the double-helix structure (Scheme 1b), which allows distinct host–guest chemistry over single-strand DNA (ssDNA) for dye inclusion.15,18 Upon binding with dsDNA, SG can be isolated from its surrounding environment by the rigid double helix to exclude the intermolecular (collisions) and intramolecular (rotations) nonradiative processes. We also confirmed that the matrix effect of dsDNA is universal for several PSs, including both base intercalators and minor groove binders. To explore the practical applicability of dsDNA-switched photosensitization, loop-mediated isothermal amplification (LAMP) was introduced to construct a nucleic acid testing (NAT) platform. NAT is the benchmark technology for disease diagnosis and monitoring,19,20 but is mostly performed in centralized laboratories with high-end instrumentation and highly skilled personnel,21,22 that is, not suitable for point-of-care testing (POCT). Therefore, NAT methods with simple operation and low-cost instrumentation are highly desired. Here, through exploration of a simple light-emitting diode (LED)-assisted photochemical oxidation protocol (1O2-induced chromogenic reaction of 3,3',5,5'-tetramethylbenzidine (TMB)), a colorimetric LAMP method was developed for the analysis of the hepatitis B virus (HBV) gene. Since photosensitization-mediated colorimetric signals can accumulate through prolonged light irradiation, the newly developed colorimetric LAMP assay features comparable analytical performances as that of a SG fluorescence assay. To facilitate POCT, we also developed a simple and low-cost paper strip for distance-based detection of LAMP amplicons with a limit of detection (LOD) of 100 aM for HBV DNA. Experimental Section Materials SG was primarily studied in this work, and several other dsDNA-binding dyes were also included. Detailed information about the materials and oligonucleotides is given in Supporting Information Tables S1 and S2, respectively. The concentration of SG (10,000×) was calculated to be 3.95 mM according to Vitzthum's work.23 Oligonucleotides were provided by Shanghai Sangon Biotech Co., Ltd. (Shanghai, China). Study on the photosensitization performance of SG The detailed instrumental information are given in Supporting Information Table S3. To study the photosensitization performance of SG, oligonucleotides and SG were mixed and diluted to 1 mL with citrate-phosphate buffer (pH 4.5, 10 mM MgCl2). The final concentrations of dsDNA and SG were 100 nM and 0.8 μM (2×), respectively. The mixture was incubated for 10 min, followed by the addition of TMB (0.2 mg/mL), and then irradiated with a cyan LED (495 nm, 3 V, 3 W) for another 2 min. The 1O2 quantum yields (ΦΔ) of SG were evaluated using Ru(bpy)32+ as the tandard (see Section 3 for details in Supporting Information Figures S8–S12). To confirm the matrix effect of DNA, radiative (kr) and nonradiative transitions (knr) were calculated through the following equations 24: ∅ = k r k r + k nr (1) τ = 1 k r + k nr (2) k nr = 1 − ∅ τ (3)where Φ and τ are the quantum yield and lifetime of fluorescence and phosphorescence, respectively. Specifically, SG and dsDNA were mixed and diluted to 2 mL with phosphate-buffered saline (PBS) buffer (pH 7.4, 10 mM). The final concentrations of dsDNA and SG were 1.0 μM and 0.8 μM (2×), respectively. Detailed measurements of ΦFL, ΦPHOS, τFL, and τPHOS are given in the Supporting Information. Theoretical calculations For density functional theory (DFT) and time-dependent DFT (TD-DFT) calculations, Gaussian 0925 was applied. The ground- and excited-state geometries were optimized at the B3LYP/6-31g (d) level. The Solvent Model based on Density (SMD) was applied for all calculations, and water was used as the solvent here. A potential energy scan of the S0was based on a ground-state optimized molecular structure, and the S1 potential energy surface calculation was based on an excited-state optimized molecular structure. The potential energy surface calculation was carried out using relaxed potential energy surface scanning at the B3LYP/6-31g (d) level. On the basis of optimized geometries in the S1, the excitation energies were both calculated using TD/B3LYP/def-TZVP for the electronically excited 1O2 and T1. At the same level, the spin–orbit coupling (SOC) matrix elements between the singlet and T1 were given by ORCA 4.1.1 (free access from the Max-Planck-Institute). Molecular docking Molecular docking was performed with Surflex-Dock available in Sybyl-X 2.0 program (Tripos Inc., St Louis, USA). A dsDNA structure (sequence: 5′-ACA GAC ACC T-3′ and its complementary strand) was built in the biopolymer builder module, and all the hydrogen atoms were added to define the correct configuration and tautomeric states. After adding the charge with the Gasteiger–Marsili charges, the model structure was energy minimized using the Powell energy minimization algorithm with a MMFF94 force field to obtain the DNA molecule for next docking analysis. The SG structure was designed in the Sybyl program and energy minimized to reach a reasonable three-dimensional (3D) conformation utilizing the Tripos force field with distance-dependent dielectric and Powell energy minimization algorithm. The maximum number of iterations performed in the minimization was set as 1000, and other parameters were set as default. Docking pockets were generated based on "Automatic mode", and the binding site was determined by a protomol probe. An additional starting conformation was set as 10 with a minimum of 0.05 Root Mean Square Deviation (RMSD) between poses. The spin-alignment method was used with moderate accuracy (density of search, 6.0) and 12 spins per alignment, and other parameters were set as defaults. Docking results were ranked based on Surflex-Dock's scoring function (an empirical scoring function), in terms of hydrophobic, polar, repulsive, entropic, and solvation. The lowest binding energy conformation was searched out of 20 different conformations and used for further analysis. We selected the top two structures in terms of score. All binding models were visualized in Pymol (version 2.4.0, Schrödinger, NY, USA). Fluorescence versus photosensitization To compare the photosensitization-based colorimetry and fluorescence, oligonucleotides of different lengths (10–130 bps) and PSs were mixed and diluted to 1 mL with citrate-phosphate buffer (pH 4.5, 10 mM, MgCl2) for photosensitization and PBS buffer (pH 7.4, 10 mM) for fluorescence, respectively. The concentrations of dsDNAs of different lengths were 10 nM, and SG was 0.8 μM (2×). To obtain the LOD of the photosensitization-based colorimetric assay and fluorescence for DNA quantification, concentrations of the probe ssDNA were fixed (30–60 bps, 100 nM; 70–130 bps, 50 nM), while the final concentration of SG was 0.8 μM (2×). The mixture was incubated for 10 min, followed by the addition of TMB, and then irradiated with a cyan LED for another 2 min. LAMP design of paper strips for photosensitization-based colorimetric assays The designed patterns onto cellulose chromatographic paper used a XEROX ColorQube 8580 solid ink printer and then heated on a hot plate at 150 °C for 40 s. The strip was then fabricated by stacking the patterned paper and a layer of paraffin film on a microscope glass slide. This sandwiched device was then bonded by heating on the hot plate at 110 °C for 30 s. Finally, TMB was coated onto the test zone of the strip. To do so, TMB was first dissolved in acetonitrile and then deposited onto cellulose paper upon rapid solvent evaporation. Results and Discussion Confirmation of the dsDNA-switched photosensitization switching of SG First, 1O2 generation from the dsDNA–SG complex was confirmed with standard electron paramagnetic resonance (EPR) characterization. Upon mixing of SG with dsDNA (dsDNA-1; ,Figure 1a, Supporting Information Figures S1 and S2 and Table S2), a distinct EPR signal of 1O2 from photosensitization was observed from the typical 1:1:1 triplet peaks of 4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl (4-hydroxy-TEMPO; Figure 1b).26 In addition, a very weak 1275 nm characteristic peak of 1O2 was also collected ( Supporting Information Figure S3). However, for SG alone or the mixture of ssDNA and SG, no appreciable EPR signal was observed. Therefore, only binding of SG with dsDNA resulted in a type-II photosensitization process (1O2, Supporting Information Figures S4–S6).1 Quantitatively, the quantum yields (ΦΔ) of SG for producing 1O2 were evaluated using Ru(bpy)32+ as a standard (see Section 3 in Supporting Information).27,28 As shown in Figure 1c, ΦΔ of free SG was extremely low (∼0.3%), but increased ∼74-fold after binding with dsDNA (∼22.2%). Treating the dsDNA–SG complex with NaN3 (a specific 1O2 scavenger) further resulted in a sharp decrease of ΦΔ (Figure 1c). Besides dsDNA-1, it was found that other synthetic and natural dsDNAs could also switch the 1O2 generation from SG ( Supporting Information Figure S7), indicating that the double helix is responsible for regulating SG photosensitization (Scheme 1b). Figure 1 | Verification of singlet oxygen generation from the dsDNA–SG complex: (a) scheme of the formed dsDNA–SG complex with photophysical properties turned on; (b) EPR for identification of the photosensitization products with TEMPO (100 mM, specific for 1O2) as the probe; (c) singlet oxygen generation efficiency (ΦΔ) of SG and the dsDNA–SG complex; (d) fluorescence emission spectra; (e) fluorescence lifetime; (f) O2-free phosphorescence spectra (delay time: 2 ms); and (g) O2-free phosphorescence lifetime. Experimental conditions: SG concentration, 0.8 μM (2×); dsDNA, 1.0 μM; light irradiation, cyan LED for 2 min. Download figure Download PowerPoint SG binding with dsDNA not only switched on the photosensitization of SG, but also the fluorescence and phosphorescence. As shown in Figure 1d, the SG fluorescence increased over 1000-fold, accompanied by a largely increased lifetime (Figure 1e). Meanwhile, distinct O2-free RTP emerged (Figure 1f) with a lifetime of ∼3.07 ms (Figure 1g). Compared with fluorescence, SG phosphorescence shows another Stokes shift of ∼100 nm but the same excitation ( Supporting Information Figure S13). For free SG, no significant phosphorescence was observed. This is the first report of RTP collection from pure organic PSs upon interaction with dsDNA in a label-free manner. Therefore, SG binding with dsDNA could activate the T1 of SG, thereby switching on the photosensitization. To study the mechanism of fluorescence and phosphorescence modulation, the radiative (kr) and nonradiative (knr) rate constants of SG before and after binding with dsDNA were derived with eqs 1–3 (see Experimental Section). As shown in Table 1, although the radiative rate constants of SG only doubled for fluorescence and tripled for phosphorescence, the corresponding nonradiative rate constants decreased by ∼391-fold (fluorescence) and ∼5850-fold (phosphorescence), respectively. Therefore, the dsDNA-induced luminescence lighting up and photosensitization switching of SG can be ascribed to dsDNA matrix-restricted nonradiative transitions (Scheme 1b). Herein, the inhibition of the nonradiative transition of the T1 is much more significant, probably because the long-lived phosphorescence is prone to deactivation. Table 1 | Photophysical Parameters of SG before and after Binding with dsDNA Parameters SG (1) SG + dsDNA (2) (2)/(1) ΦFL (%) 0.6 66.5 110.8 τFL (ns) 0.086 4.3 50 kr (FL, 108 s−1) 0.7 1.5 2 knr (FL, 109 s−1) 21.5 0.055 1/391 ΦPHOS (%) <0.001 0.057 ∼60 τPHOS (ms) 0.05 3.07 600 kr (PHOS, s−1) 0.06 0.18 3 knr (PHOS, 104 s−1) 193 0.033 1/5850 Abbreviations: FL, fluorescence; PHOS, phosphorescence Origin of the nonradiative process of SG Next, the origin of the nonradiative transitions of SG was investigated. Structurally, SG is similar to the typical molecular motor thioflavin T (ThT),15 which contains electron acceptor (benzothiazole) and electron donor (1,4-dihydroquinoline) moieties. Thus, DFT and TD-DFT calculations were employed to confirm the above prediction. Results of the coordinate-driving potential surface scanning indicates that when excited, a relaxation process with a barrierless rotation of SG starts until the dihedral angle (θ) between the 1,4-dihydroquinoline and benzothiazole moieties reaches 90° (S′1, the most stable (S1)) (Figure 2a). In such a process, a nonradiative twisted intramolecular charge-transfer (TICT) process will occur,29 with extremely low oscillator strength (f) of almost zero (Figure 2b). Such relaxation is fastest due to the small energy gap (S′1 → S′0, Figure 2c). Therefore, the radiative processes (fluorescence and phosphorescence) and photosensitization are expected to be off at this stage. Figure 2 | Study on the photophysical properties of free SG: (a) potential energy (eV) versus dihedral angle (θ) at the ground and the S1; (b) oscillator strength versus dihedral angle (θ); (c) scheme of the rotational mobility of SG between dihedral angle of 90° and <90°, with Jablonski diagram of the plausible mechanism showing inset; (d) temperature-dependent fluorescence and phosphorescence (delay time: 2 ms) emission spectra; and (e) temperature-dependent fluorescence and phosphorescence lifetime. The temperature-dependent investigations were carried out in 2Me-THF media and the concentration of SG was 0.8 μM. Download figure Download PowerPoint To verify the above transitions, temperature-dependent fluorescence and phosphorescence of SG were collected. As shown in Figures 2d and 2e, the fluorescence and phosphorescence of SG were extremely weak at room temperature [298 K, 2Me-tetrahydrofuran (THF) as solvent], which agrees well with the results in the absence of dsDNA (Figures 1d and 1f). With decreasing temperature, appreciable fluorescence (∼525 nm) and phosphorescence (∼625 nm, delay time of 2 ms) were observed only after freezing (136 K, the ice point of the solvent 2Me-THF), with spectral features similar to those in the presence of dsDNA (Figures 1d and 1f). Further temperature decreases from 136 K to 77 K resulted in increased emission intensity and lifetime, but was not as significant as freezing (from 298 K to136 K). Solution viscosity-dependent investigations further verified the above deactivation process of SG ( Supporting Information Figure S14). These phenomena confirmed the TICT nature of SG, which causes significant nonradiative transitions. Upon freezing or viscosity increasing, the bond rotation at the S1 was restricted (θ < 90°, together with significantly enhanced f), making the relaxation pathway (S1 → S′1 → S′0) less probable. Therefore, fluorescence (S1 → S0) and phosphorescence (S1 → T1 → S0) would be switched ON (Figure 2c). On the basis of the above spectral features, it is expected that SG binding with dsDNA would result in the restriction of TICT in the S1, leading to the activation of the excited T1 of SG and, in turn, photosensitization (Scheme 1). of dsDNA binding on the process of SG Besides nonradiative also to the activation of the excited T1. In it was that PSs with intramolecular charge-transfer state may be from the donor and acceptor structure in excited Therefore, the dihedral angle of SG after binding with dsDNA may to the of SG. The binding conformation of SG with dsDNA was first with molecular docking As shown in Figure groove binding of SG with the double helix can be which is with the In such a restricted the most dihedral of SG were in the of in Figure Figure 3 | The of the dihedral angle of SG on its (a) binding conformation of SG with dsDNA by molecular docking the most dihedral angle of SG was in the of and (b) the oscillator molecular and of the energy and of SG at dihedral angle of most dihedral angle after SG binding with and 90° (free molecular lowest molecular Download figure Download PowerPoint The of is determined by the energy between S1 and T1 and the T | | k T | | 2 2 S1 molecular and structures of SG dihedral were further calculated with As shown in Figure Supporting Information Figures and the dihedral angle results in of the constants and decrease of the energy gap which is with However, the oscillator strength of SG at 90° is which is not for the Therefore, may be a design for highly efficient PSs from the structure of but the oscillator strength also be Here, the S1 of SG after binding with dsDNA may be a between and oscillator the role of the dsDNA matrix in the photosensitization activation of SG (Scheme 1). of the dsDNA matrix for photosensitization switching The matrix effect of dsDNA for photosensitization switching was not only to SG, but also a series of other As shown in Supporting Information Figures typical groove and and base and were also found to activation upon binding to dsDNA. Further photophysical studies that nonradiative rate constants of PSs decrease after binding with dsDNA ( Supporting Information Figures and Table Therefore, dsDNA may be a universal matrix for photosensitization On the other based on the above and also are photosensitization of the dsDNA-binding (1) to dsDNA allowing efficient binding; (2) TICT structures for efficient of photosensitization before binding with and structures benzothiazole or moieties to photosensitization which can be switched on after dsDNA Therefore, the studied PSs, SG, and performances over and Analytical performance of colorimetric photosensitization and fluorescence detection Previously, we determined that 1O2 generated from photosensitization could the chromogenic TMB with colorimetric signal (Figure for DNA Herein, we the colorimetric signal from dsDNA-switched photosensitization with isothermal DNA amplification to a new and simple NAT platform. SG was as the model PS for this due to its photosensitization performance over other PSs ( Supporting Information Figure As shown in Figure photosensitization-based colorimetric signals of the dsDNA–SG increased the irradiation time, the of 1O2 could be through further photosensitization. Such photochemical (1O2 → efficient signal which is for the of photosensitization-based colorimetric detection time: 2 was comparable with that of fluorescence 10 nM; dsDNA Figure when with phosphorescence the dsDNA matrix could also induce O2-free the sensitivity of the colorimetric detection was ( Supporting Information Figure and Table Since the same relaxation the signal photosensitization can be an of the weak phosphorescence This assay is different from more colorimetric methods with much fluorescence, leading to comparable through photochemical signal ( Supporting Information Figures and and Tables and Figure | of the analytical performances of fluorescence and photosensitization for DNA (a) scheme of energy diagram of fluorescence and photosensitization (b) of the irradiation time-dependent signal of fluorescence and photosensitization and (c) of the of fluorescence and photosensitization colorimetry for detection of dsDNA of different lengths (10–130 to the potential of TMB over the base and interaction of dsDNA and TMB, 1O2 generated from photosensitization will first TMB the base ( Supporting Information Figure of the assay sensitivity are in the Supporting Information. Experimental conditions: the concentration of SG, 0.8 μM (2×); media for fluorescence, media for photosensitization, irradiation time for cyan 2 min. The of the used are given in Supporting Information Table Download figure Download PowerPoint NAT through dsDNA-switched photosensitization with LAMP On the basis of the DNA detection performance of the photosensitization we further it with LAMP for NAT LAMP is an isothermal amplification protocol that performance in DNA amplification for LAMP is also of dsDNA products after which can be as the matrix for photosensitization switching. As a a set of were designed to a at on the HBV (Figure To the operation of photosensitization-based colorimetric a series of a simple LED or based on the plate were

Topics & Concepts