Theoretical Investigations on the Sensing Mechanism of Phenanthroimidazole Fluorescent Probes for the Detection of Selenocysteine

The level of selenocysteine (Sec) in the human body is closely related to a variety of pathophysiological states, so it is important to study its fluorescence sensing mechanism for designing efficient fluorescent probes. Herein, we used time-dependent density functional theory to investigate the fluorescence sensing mechanism of phenanthroimidazole derivates A4 and B4 for the detection of Sec, which are proposed to be designed based on excited state intramolecular proton transfer (ESIPT) and intramolecular charge transfer (ICT) mechanisms. The calculation results show that the fluorescence quenching mechanism of A4 and B4 is due to the photo-induced electron transfer (PET) process with the sulfonate group acts as the electron acceptor. Subsequently, A4 and B4 react with Sec, the sulfonate group is substituted by hydroxyl groups, PET is turned off, and significant fluorescence enhancement of the formed A3 and B3 is observed. The theoretical results suggest that the fluorescence enhancement mechanism of B3 is not based on ICT mechanism, and the charge transfer phenomenon was not observed by calculating the frontier molecular orbitals, and proved to be a local excitation mode. The reason for the fluorescence enhancement of A3 based on ESIPT is also explained by the calculated potential energy curves.


Introduction
Selenocysteine (Sec) is an important member of the reactive selenium species and an important component of selenoprotein (SeP) [1] because the content of Sec is closely related to the catalytic ability of SeP, which further affects a variety of pathological and physiological states in the human body, such as cancer, diabetes, neurodegenerative diseases, cardiovascular disease, and male infertility [2][3][4]. Therefore, how to detect Sec quickly, efficiently, and accurately is a huge challenge faced by researchers. Compared with highperformance liquid chromatography, electrochemical assay, thin layer chromatography, mass spectrometry, capillary electrophoresis, titration analysis, Fourier-transform infrared spectroscopy, ultraviolet-visible spectroscopy [5][6][7][8][9][10][11], fluorescence detection technology can better protect the integrity of biological samples and has the advantages of fast response, low cost, simplicity, high sensitivity, huge potential, etc. [12]. That is to say, researchers need to develop fluorescent probes with high biocompatibility and selectivity to monitor Sec levels.
Excited state intramolecular proton transfer (ESIPT) reaction is one of the important fundamental reactions in photochemistry and photo physical, which is a four-level cyclic process and usually exhibits dual fluorescence characteristic from the normal form N* (or process and usually exhibits dual fluorescence characteristic from the normal form N* (or Enol*) and the tautomeric form T* (or Keto*) due to the proton transfer [13][14][15][16][17]. Based on this characteristic, the ESIPT system has made great contributions in the fields of UV filters, laser dyes, white light emitting materials, molecular switches, fluorescent probes, organic optoelectronic materials, etc. [18][19][20][21][22][23][24]. In particular, fluorescent probes developed based on the ESIPT system, also known as ESIPT-based fluorescent probes, can minimize self-absorption and reduce interference from auto-fluorescence due to their in vivo application [25,26]. Therefore, more and more scientists are designing new ESIPT-based fluorescent probes to detect Sec levels, and they are proving to have good application prospects. In addition, intramolecular charge transfer (ICT) processes also play an important role in the field of fluorescent probes because of their special photo physical properties [27]. Upon photo-excitation, ICT-based fluorescent probes typically occur on an ultrafast time scale and produce a large Stokes shift in the emission spectrum.
Recently, Wang et al. synthesized two phenanthroimidazole turn-on probes, A4 and B4 (isomers), as turn-on fluorescent probes for the detection of Sec [28]. They found that A4 and B4 have high selectivity and sensitivity to Sec and can be applied to fluorescence imaging of Sec in living cells. Among them, A4 and B4 were designed according to the ESIPT and ICT mechanisms, respectively. In other words, when A4 and B4 react with Sec, the sulfonate group is replaced by the hydroxyl group, photo-induced electron transfer (PET) is turned off, and significant fluorescence enhancement of the formed A3 and B3 is observed due to the turned-on ESIPT and ICT mechanisms (Scheme 1). It is worth mentioning that the PET from donor (D) to acceptor (A) leads to the formation of a charge separated state composed of donor free radical cations and acceptor free radical anions. (1,2) In the intramolecular PET process, D and A coexist in the same molecule, while in the intermolecular PET process, they correspond to different molecules. Generally, PET responsible for fluorescence quenching [29]. However, we found that Wang et al. did not provide a detailed explanation for the fluorescence quenching of A4 and B4, and also overlook the enhanced fluorescence of A3 after the occurrence of ESIPT [28]. Therefore, a detailed theoretical investigation of the fluorescence quenching mechanism of A4 and B4 and the fluorescence enhancement mechanism of A3 and B3 is necessary. In this work, we use the time-dependent density functional theory (TD-DFT) method to reveal the fluorescence quenching mechanism of A4 and B4 and the fluorescence enhancement mechanism of A3 and B3. The calculated electronic excitation energies, Scheme 1. Mechanism of Sec Sensing by A4 and B4.
In this work, we use the time-dependent density functional theory (TD-DFT) method to reveal the fluorescence quenching mechanism of A4 and B4 and the fluorescence enhancement mechanism of A3 and B3. The calculated electronic excitation energies, corresponding oscillator intensities, corresponding compositions, and frontier molecular orbitals (FMOs) well reveal that the fluorescence quenching of A4 and B4 is due to the PET mechanism, and the sulfonate group was found to be the main cause of PET formation. Meanwhile, the calculated energy and potential energy curves indicate that the forward occurrence of the ESIPT process in A3 is easier and more structurally stable than the reverse, which is responsible for the enhanced fluorescence after the occurrence of ESIPT in A3. In particular, according to the analysis of FMOs and ∆r index, it was confirmed that B3 did not undergo ICT processes, but only normal local excitation characteristics. Thus, we provide a reasonable sensing mechanism that will help scientists develop new fluorescent probes for the detection of Sec.

Theoretical Methods
In this paper, all theoretical calculations are carried out in Gaussian 16 [30]. The ground state (S 0 ) and the first excited state (S 1 ) of A3, A4, B3, and B4 are optimized by using the B3LYP functional and TZVP basis set in DFT and TDDFT methods, respectively [31][32][33][34][35]. For considering dispersion force, all the simulations have been performed by adopting the D3 version of Grimme's dispersion [36]. Exploring the effects brought by surrounding environments, this work adopts the integral equation formal variables of the polarizable continuum model (IEFPCM) model for all simulations [37][38][39], using water (ε = 78.3553) solvent. By further checking the stability of all structures via infrared (IR) vibrational analysis, we confirmed that all optimized geometries are local minima without imaginary frequency. In addition, the calculated electron data and FMOs are the same as those of optimization and are used for the study of fluorescence properties. To interpret the ESIPT mechanism by the PECs, we construct the PECs of S 0 and S 1 states along with the forward direction of the reaction path. In particular, the PECs construction is relaxation scanning. Importantly, since the CAM-B3LYP functional can better describe charge transfer [40], we also used the CAM-B3LYP-D3/TZVP/IEFPCM level to calculate the excitation and emission energies and their related oscillator strengths, components and FMOs. We also provide some important data in the supporting information (Tables S1-S4, Figures S1 and S2, and all the optimized geometries).

A3 and A4
In Figure 1, we provide the front and side views of the A4 geometry in the S 0 and S 1 states, respectively. It can be seen that the change of the sulfonate groups in A4 is more obvious, and we assume that this is caused by the strong electron absorption ability of this group. Here, we suspect that the cause of A4 fluorescence quenching may be caused by the sulfonate group, which we also verify in the later calculation content. To better compare with the A4 geometry, we also calculated the A3 geometry, and since the ESIPT mechanism is involved, we provide the A3−Enol and A3−Keto in the S 0 and S 1 states, respectively ( Figure 2a). We have labeled the most important atoms in intramolecular hydrogen bonds with numbers 1-3 for better understanding. It can be seen that the A3 geometry does not change significantly in the S 0 and S 1 states, because the hydroxyl group replaces the sulfonic group and the structural rigidity is enhanced. The primary bond parameters (lengths and angles) of the A3−Enol and A3−Keto forms in water solvents are shown in Figure 2a. Calculation of A3−Enol geometric parameters shows that the O 1 −H 2 bond length increases from 0.996 Å (S 0 state) to 0.998 Å (S 1 state) and the δ (O 1 −H 2 ···O 3 ) bond angle expands from 147.9 • to 148.7 • , indicating that the intramolecular hydrogen bonding is enhanced in the S 1 state. In particular, we found that the H 2 −N 3 bond length increased in the S 1 state, which would be detrimental to the occurrence of ESIPT.
To further prove our idea, we calculated the IR vibrational spectra of A3−Enol forms for the analysis of the intramolecular hydrogen bond (Figure 2b). For A3−Enol forms in a water solvent, the calculated O 1 −H 2 stretching vibrational frequency of the A3−Enol forms was changed from 3137 cm −1 (S 0 state)→3102 cm −1 (S 1 state). We can see that the O 1 −H 2 stretching modes of the 2P3HBQ−Enol and 2HP3HBQ−Enol forms both exhibit a redshift of 25 cm −1 , indicating enhanced intramolecular hydrogen bonding in the S 1 state.
Although the IR vibration spectrum shows that the redshift is in the S 1 state, the change in the redshift value is relatively small, so it is also easy to judge whether the ESIPT process occurs. To more intuitively judge whether the ESIPT process occurs, we constructed the PECs of A3 in the S 0 and S 1 state, with a function of O 1 −H 2 bond length in the step of 0.05 Å (Figure 2c). For A3 fluorophore, the energy potential barrier of A3−Enol→A3−Keto is 7.32 kcal/mol in the S 0 state and 5.35 kcal/mol in the S 1 state. The energy potential barrier of A3−Keto→A3−Enol is 0.83 kcal/mol in the S 0 state and 6.13 kcal/mol in the S 1 state. These results show that Keto→Enol is more likely to occur in the S 0 state, and Enol→Keto is more likely to occur in the S 1 state. In addition, we can see that A3−Keto is more stable in the S 1 state than A3−Enol (Figure 2a). These results directly prove the main reason for A3−Keto* emission and reasonably explain the experimental phenomenon.  To further prove our idea, we calculated the IR vibrational spectra of A3−Enol forms for the analysis of the intramolecular hydrogen bond (Figure 2b). For A3−Enol forms in a water solvent, the calculated O1−H2 stretching vibrational frequency of the A3−Enol forms was changed from 3137 cm −1 (S0 state)→3102 cm −1 (S1 state). We can see that the O1−H2 stretching modes of the 2P3HBQ−Enol and 2HP3HBQ−Enol forms both exhibit a redshift  To further prove our idea, we calculated the IR vibrational spectra of A3−Enol forms for the analysis of the intramolecular hydrogen bond (Figure 2b). For A3−Enol forms in a water solvent, the calculated O1−H2 stretching vibrational frequency of the A3−Enol forms was changed from 3137 cm −1 (S0 state)→3102 cm −1 (S1 state). We can see that the O1−H2  To understand the excitation and emission processes of A3 and A4, we calculated the electronic excitation energy and emission energy (Tables 1 and 2), as well as the corresponding oscillator intensities and compositions. It can be seen that the S 0 →S 1 transition is allowed for A3, so the oscillator intensity is 0.5677, which indicates that the S 1 state is a bright state. The calculated electron excitation energy is 3.58 eV (Table 1), which is consistent with the experimental results [28]. In addition, the calculated emission energies of A3−Enol and A3−Keto are 2.99 eV and 2.76 eV (Table 2), respectively, which are also in good agreement with the experimental results [28], indicating that our calculation method is reasonable. However, it can be seen that the S 0 →S 1 transition is forbidden for A4, so the oscillator intensity is 0.0013. We find that the S 0 →S 7 transition is allowed because the oscillator intensity of the S 0 →S 7 transition is 0.6518. Meanwhile, the calculated emission energy is 0.98 eV, and the oscillator intensity is 0, so the S 1 →S 0 transition undergoes a non-radiative pathway.  Based on the above results, we added FMOs analysis to further elaborate the sensing mechanism of A3 and A4 (Figure 3). The calculation results show that the S 0 →S 1 transition of A3 is mainly H→L, and the S 1 →S 0 transition is mainly L→H, so we provide HOMO and LUMO. It can be seen that H→L is ππ * features, which provides evidence that the S 1 state is a bright state. Therefore, the sensing mechanism of A3 can be summarized: Firstly, the A3 excitation transitions to the S 1 state, then vibrationally relaxes to a stable A3−Enol* in the S 1 state. One part of the energy emits fluorescence back to the S 0 state (A3−Enol), the other part of the energy overcomes the energy barrier and stabilizes to A3−Keto* through the ESIPT process, then emits fluorescence back to the S 0 state (A3−Keto), and finally returns to stable A3-Enol. Besides, the S 0 →S 1 and S 0 →S 7 transition of A4 are mainly H→L and H→L+2, respectively, and the S 1 →S 0 transition is mainly L→H, so we provide HOMO, LUMO, and LUMO+2. In Figure 3, it can be seen that the H→L+2 is ππ * features with the ICT process. However, the H→L shows a classic PET process. It is well known that the PET process is a non-radiative transition process, which is also the main reason for A4 fluorescence quenching. Therefore, the sensing mechanism of A4 can be summarized: After photoexcitation, A4 transitions from the S 0 state to the S 7 state and then undergoes an ultrafast internal conversion to reach the S 1 state accompanied by the PET, and nonradiative return to the S 0 state. returns to stable A3-Enol. Besides, the S0→S1 and S0→S7 transition of A4 are mainly H→L and H→L+2, respectively, and the S1→S0 transition is mainly L→H, so we provide HOMO, LUMO, and LUMO+2. In Figure 3, it can be seen that the H→L+2 is ππ * features with the ICT process. However, the H→L shows a classic PET process. It is well known that the PET process is a non-radiative transition process, which is also the main reason for A4 fluorescence quenching. Therefore, the sensing mechanism of A4 can be summarized: After photoexcitation, A4 transitions from the S0 state to the S7 state and then undergoes an ultrafast internal conversion to reach the S1 state accompanied by the PET, and nonradiative return to the S0 state.

B3 and B4
In Figure 4, the B3 and B4 geometries are also optimized in the S0 and S1 states. The calculation results show that the geometry of B3 does not change significantly in the S0 and S1 states, while the geometry of B4 changes significantly in the S0 and S1 states, which is also due to the sulfonate group. Therefore, we believe that the fluorescence quenching mechanism of B4 is also related to the sulfonate group. In order to prove our point of view, the excitation and emission energies, as well as their related oscillator intensities and compositions are also calculated (Tables 3 and 4). The S0→S1 transition is allowed for B3, so the oscillator intensity is 0.2500, which indicates that the S1 state is a bright state. The calculated electron excitation energy is 3.57 eV (Table 3), which is consistent with the experimental results [28]. In addition, the calculated emission energies of B3 is 2.81 eV (Table  4), which are also in good agreement with the experimental results [28]. Similarly, it can be seen that the S0→S1 transition is forbidden for A4, so the oscillator intensity is 0.0085, which indicates that the S1 state is a dark state. We also found that the S7 state is a bright state, because the oscillator intensity of the S0→S7 transition is 0.5826. Meanwhile, the calculated emission energy is 0.97 eV, and the oscillator intensity is 0.0243, so the S1→S0 transition undergoes a non-radiative pathway.

B3 and B4
In Figure 4, the B3 and B4 geometries are also optimized in the S 0 and S 1 states. The calculation results show that the geometry of B3 does not change significantly in the S 0 and S 1 states, while the geometry of B4 changes significantly in the S 0 and S 1 states, which is also due to the sulfonate group. Therefore, we believe that the fluorescence quenching mechanism of B4 is also related to the sulfonate group. In order to prove our point of view, the excitation and emission energies, as well as their related oscillator intensities and compositions are also calculated (Tables 3 and 4). The S 0 →S 1 transition is allowed for B3, so the oscillator intensity is 0.2500, which indicates that the S 1 state is a bright state. The calculated electron excitation energy is 3.57 eV (Table 3), which is consistent with the experimental results [28]. In addition, the calculated emission energies of B3 is 2.81 eV (Table 4), which are also in good agreement with the experimental results [28]. Similarly, it can be seen that the S 0 →S 1 transition is forbidden for A4, so the oscillator intensity is 0.0085, which indicates that the S 1 state is a dark state. We also found that the S 7 state is a bright state, because the oscillator intensity of the S 0 →S 7 transition is 0.5826. Meanwhile, the calculated emission energy is 0.97 eV, and the oscillator intensity is 0.0243, so the S 1 →S 0 transition undergoes a non-radiative pathway.
We also calculated FMOs to better explain the sensing mechanism of B3 and B4 ( Figure 5). Among them, the S 0 →S 1 transition of B3 is mainly H→L, and the S 1 →S 0 transition is mainly L→H, so we provide HOMO and LUMO. It can be seen that H→L is ππ * features, which provides evidence that the S 1 state is a bright state. It is worth mentioning that we have not found the ICT process in H→L, but a local excitation (LE) process, which is inconsistent with the conclusion given by Wang et al. [28]. In order to further prove whether ICT process occurs in B3, we calculated the ∆r index. The ∆r index is a new method based on the hole-electron distance proposed by Guido et al. in 2013, which can be a good way to determine whether an ICT process is occurring or not [41]. The method has been widely accepted by researchers. Guido et al. suggested the use of 2.0 Å (∆r index) as a criterion to distinguish between LE and CT excitation. The ∆r index value of B3 calculated by B3LYP functional is 1.13 Å, which can further indicate that B3 has no ICT process. Therefore, the sensing mechanism of B3 can be summarized: Firstly, B3 excitation from the S 0 state to the S 1 state, which is an LE process, followed by stabilization in the S 1 state and finally back to the S 0 state. Similarly, the S 0 →S 1 and S 0 →S 7 transition of B4 are mainly H→L and H→L+2, respectively, and the S 1 →S 0 transition is mainly L→H, so we provide HOMO, LUMO, and LUMO+2. In Figure 5, the H→L+2 is ππ * features, also accompanied by the ICT process, which is similar to the A4 result. Meanwhile, the H→L shows a classic PET process. It is well known that the PET process is a non-radiative transition process, which is also the main reason for B4 fluorescence quenching. Therefore, the sensing mechanism of B4 can be summarized: Upon photo excitation, the B4 transition from the S 1 state to the S 7 state then undergoes an internal conversion to reach the S 1 state, and finally undergoes a PET process from the S 1 state back to the S 0 state. Finally, because the B3LYP functional may underestimate the energy of CT state, we chose CAM−B3LYP functional to recalculate the excitation and emission energies and their related oscillator intensities, components, and FMOs (supporting information). The calculation results show that the trend is consistent with the results obtained by B3LYP functional calculation, further demonstrating the reliability and reasonableness of our chosen calculation method.    We also calculated FMOs to better explain the sensing mechanism of B3 and B4 ( Figure  5). Among them, the S0→S1 transition of B3 is mainly H→L, and the S1→S0 transition is mainly L→H, so we provide HOMO and LUMO. It can be seen that H→L is ππ * features, which provides evidence that the S1 state is a bright state. It is worth mentioning that we   S7 state then undergoes an internal conversion to reach the S1 state, and finally undergoes a PET process from the S1 state back to the S0 state. Finally, because the B3LYP functional may underestimate the energy of CT state, we chose CAM−B3LYP functional to recalculate the excitation and emission energies and their related oscillator intensities, components, and FMOs (supporting information). The calculation results show that the trend is consistent with the results obtained by B3LYP functional calculation, further demonstrating the reliability and reasonableness of our chosen calculation method.

Conclusions
In summary, this work mainly investigates the fluorescence quenching mechanism of A4 and B4, and the fluorescence enhancement mechanism of A3 and B3. Compared to the A3 and B3 geometries, the A4 and B4 geometries are significantly more variable in the S0 and S1 states by geometric calculations, which is caused by the greater flexibility of the sulfonate groups on A4 and B4 than the hydroxyl groups on A3 and B3. Meanwhile, the calculated geometries and IR spectra demonstrate the enhanced intramolecular hydrogen bonding of A3−Enol, indicating susceptibility to the ESIPT processes. The PECs indicate that A3−Enol→A3−Keto is more likely to occur than A3−Keto→A3−Enol, and A3 Keto is more stable, which reasonably explains the enhanced fluorescence of ESIPT emission. The calculated electronic excitation energies, corresponding oscillator intensities, corresponding compositions, and FMOs well reveal that the fluorescence quenching of A4 and B4 is due to the PET mechanism. In addition, the S1 of B3 is a LE state and does not involve ICT characteristics, based on FMOs and Δr index. Our work not only elucidates that the fluorescence quenching pathways of A4 and B4 are caused by the PET process, but also demonstrates the cause of A3 and B3 fluorescence enhancement.

Conclusions
In summary, this work mainly investigates the fluorescence quenching mechanism of A4 and B4, and the fluorescence enhancement mechanism of A3 and B3. Compared to the A3 and B3 geometries, the A4 and B4 geometries are significantly more variable in the S 0 and S 1 states by geometric calculations, which is caused by the greater flexibility of the sulfonate groups on A4 and B4 than the hydroxyl groups on A3 and B3. Meanwhile, the calculated geometries and IR spectra demonstrate the enhanced intramolecular hydrogen bonding of A3−Enol, indicating susceptibility to the ESIPT processes. The PECs indicate that A3−Enol→A3−Keto is more likely to occur than A3−Keto→A3−Enol, and A3 Keto is more stable, which reasonably explains the enhanced fluorescence of ESIPT emission. The calculated electronic excitation energies, corresponding oscillator intensities, corresponding compositions, and FMOs well reveal that the fluorescence quenching of A4 and B4 is due to the PET mechanism. In addition, the S 1 of B3 is a LE state and does not involve ICT characteristics, based on FMOs and ∆r index. Our work not only elucidates that the fluorescence quenching pathways of A4 and B4 are caused by the PET process, but also demonstrates the cause of A3 and B3 fluorescence enhancement.