Hypochlorite-Activated Fluorescence Emission and Antibacterial Activities of Imidazole Derivatives for Biological Applications

The ability to detect hypochlorite (HOCl/ClO−) in vivo is of great importance to identify and visualize infection. Here, we report the use of imidazoline-2-thione (R 1 SR 2) probes, which act to both sense ClO− and kill bacteria. The N2C=S moieties can recognize ClO− among various typical reactive oxygen species (ROS) and turn into imidazolium moieties (R 1 IR 2) via desulfurization. This was observed through UV–vis absorption and fluorescence emission spectroscopy, with a high fluorescence emission quantum yield (ՓF = 43–99%) and large Stokes shift (∆v∼115 nm). Furthermore, the DIM probe, which was prepared by treating the DSM probe with ClO−, also displayed antibacterial efficacy toward not only Escherichia coli (E. coli) and Staphylococcus aureus (S. aureus) but also methicillin-resistant Staphylococcus aureus (MRSA) and extended-spectrum ß-lactamase–producing Escherichia coli (ESBL-EC), that is, antibiotic-resistant bacteria. These results suggest that the DSM probe has great potential to carry out the dual roles of a fluorogenic probe and killer of bacteria.


INTRODUCTION
Invasion of microorganisms such as bacteria and viruses can cause infectious diseases. Due to the worldwide increase in cases of severe bacterial diseases, scientists have attempted to develop technologies that serve as both fluorogenic probes for the identification of infection and antibacterial agents. Unfortunately, the continued overuse of antibiotics coupled with the rapid spread of resistance mechanisms has rendered many antibiotics inactive (Kardas et al., 2005). As a result, new pathogens have come into being that are multidrug resistant (MDR), such as methicillinresistant Staphylococcus aureus (MRSA) and extended-spectrum ß-lactamase-producing Escherichia coli (ESBL-EC), which have undermined most clinically useful antibiotics. Therefore, the emergence of drug-resistant bacteria is one of the growing challenges to anti-infection therapy. At the same time, emerging infectious diseases need very urgent and immediate treatment due to their rapid spread. In this regard, theragnostics, a treatment strategy that combines therapeutics with diagnostics, could be embraced by clinicians and patients (Pene et al., 2009). In recent years, material-based approaches have found preliminary use for the treatment of bacterial infections (Kurapati et al., 2016;Gupta et al., 2019) and for the image-guided treatment of bacterial infections Lee S. et al., 2020). Such methods have provided an approach that can produce the desired therapeutic effect with a reduced potential to develop drug-resistant bacteria Li et al., 2018a;Li et al., 2018b;Li et al., 2019). Among the various reactive oxygen species (ROS), hypochlorite (HOCl/ClO − ) acts as a powerful microbicidal agent in the innate immune system. ClO − is mainly produced by the myeloperoxidase (MPO)-catalyzed reaction of H 2 O 2 and Cl − in immunocytes (Domigan et al., 1995;Xu et al., 2013;Pak et al., 2018;Wu et al., 2019;Nguyen et al., 2020). The regulated production of microbicidal HOCl is required for the host to control invading microbes. On the other hand, OCl − reacts rapidly with a variety of biomolecules and is connected with various disorders (Winterbourn et al., 2000;Krasowska and Konat, 2004;Jeitner et al., 2005).
A fundamentally important yet challenging feature of studies in this area is the design of new chemorecognition processes. Imidazolium salts with good water solubility and stability have been used as fluorescence sensors in aqueous solution (Xu et al., 2010;Kim et al., 2012;Xu et al., 2015). To take advantage of the properties of imidazolium salts, we designed imidazoline-2thione (R 1 SR 2 ) probes to be ClO − fluorescent probes capable of inducing bacterial growth inhibition. Furthermore, the photophysical properties of R 1 IR 2 and R 1 SR 2 were examined via not only experimental results but also time-dependent DFT (TD-DFT) calculation. Bacterial growth was significantly reduced by the imidazolium moieties (R 1 IR 2 ) that were generated by the treatment of R 1 SR 2 with ClO − . Among the pairs of R 1 IR 2 and R 1 SR 2 , the DSM probe showed excellent selectivity and sensitivity toward ClO − . Also, the DIM probe showed antibacterial efficacy toward not only E. coli and S. aureus but also methicillin-resistant S. aureus (MRSA) and extendedspectrum ß-lactamase-producing E. coli (ESBL-EC). The DIM probe initially induces electrostatic interactions between the cationic imidazolium salts and the negatively charged bacterial surface, followed by structural perturbation, resulting in bacterial cell death. Similar membrane disruption through interactions of cationic imidazolium groups has been suggested by previous reports (Riduan and Zhang, 2013). Overall, this report demonstrates the importance and benefits of the new fluorogenic probe DSM for anti-pathogenic diagnostic and therapeutic applications.

EXPERIMENTAL DESIGN
For the synthesis of R 1 SR 2 , a mixture of R 1 IR 2 (0.1 mmol), sulfur (1.0 mmol), and sodium methoxide (1.0 mmol) in anhydrous methanol (20 ml) was stirred overnight at room temperature. After the solvent was removed, the crude product was dissolved in DW and extracted by MC 3 times. The organic phase was collected and dried over Na 2 SO 4 . It was purified by silica gel column chromatography, using H/MC (9/1) as the eluent to get a white solid as the product (yield ∼90%).

Molecular Design, Synthesis, and Characterization
As shown in Scheme 1, R 1 IR 2 was synthesized from 2,3-Diaminonaphthalene in a 3-step process as follows: imidazole cyclization, alkylation, and imidazolium salt formation. Several bromide and carbazole derivatives have shown antibacterial activity (Yaqub et al., 2013;Gottardi et al., 2014;Bashir et al., 2015;Liu et al., 2015;Salih et al., 2016;Popescu et al., 2021). Thus, the introduction of bromobenzyl, dibromobenzyl, and carbazole groups is expected to increase the antibacterial effect of the probes. Then, the R 1 IR 2 salt was treated with sulfur and CH 3 ONa as a catalyst in ACN, leading to the formation of the corresponding molecule R 1 SR 2 . All synthetic processes and collected structures are detailed in the experimental section and the supporting information. Several products were characterized not only by 1 H NMR, 1 C NMR, and HRMS spectra (Supplementary Material) but also by crystallization structures (Figure 1; Supplementary Figures S37, S38; Supplementary Tables S1, S2), which have not been reported in previous studies on similar N 2 C S structures (Xu et al., 2015;Xu et al., 2016). In particular, the length of the thioketone in the N 2 C S type was found to be 1.657-1.671 A ( Table 1), which is longer than that in other thioketone types such as thiobenzophenone, thioformaldehyde, thioacetone, etc. (∼1.63-1.64 A) due to the presence of C-N π-bonds (Mullen and Hellner, 1978;Allen et al., 1987). Thus, the N 2 C S bond can make the compound more active and sensitive toward ROS/RNS. SCHEME 1 | Synthesis of imidazoline-2-thiones (R 1 SR 2 ). To better understand not only the molecular structures but also the molecular orbitals and energy levels of R 1 IR 2 and R 1 SR 2 , geometrical optimization was performed through theoretical DFT calculations in the Gaussian 09 program package using the B3LVPs functional with the 6-31+g(2d,p) basis set (Pham et al., 2020). The optimized structures without imaginary frequencies were similar to the crystallization structures in terms of several critical bond lengths and angles ( Table 1). Their molecular orbitals and energy levels from HOMO+2 to LUMO+2 are shown in Supplementary Tables S1-S5. The HOMO of CIR 2 is located in the carbazole moiety, and the LUMO is located in the naphthalene-imidazolium salt center core. However, the HOMO and LUMO of BIR 2 and DIR 2 are concentrated in the naphthalene-imidazolium salt core ( Figure 2). Similarly, the HOMO of CSR 2 is located in the carbazole moiety, whereas the HOMO and LUMO of BSR 2 and DSR 2 are located in the naphthalene and imidazoline-2thione moieties. The difference originates from the introduction of the carbazole moiety, which is known as a strong donor and fluorescence quencher (Ledwon, 2019;Rehmat et al., 2020;Saritha et al., 2020). Thus, the energy gap between the LUMO and the HOMO (E g ) of CIR 2 (1.80-2.03 eV) is significantly lower than those of BIR 2 and DIR 2 (3.77-3.86 eV). The E g of CSR 2 (3.79-3.86 eV) is lower than those of BIR 2 and DIR 2 (4.08-4.14 eV). The reduction in difference is assigned to the strong electron acceptor ability of imidazolium salts, which enhances electron transfer from the carbazole electron donor to the imidazolium salt electron acceptor (vs. from the carbazole to the naphthalene and imidazoline-2-thione moieties).

Photophysical Properties and Theoretical Calculations
The UV-vis absorption and fluorescence emission spectroscopy of R 1 IR 2 and R 1 SR 2 were examined in various solvents ( Supplementary Figures S39-S47). At the same time, timedependent DFT (TD-DFT) calculations were carried out in the optimized structures using a hybrid functional method, a gradient-corrected method, and a popular local method (Adamo and Jacquemin, 2013;Pham et al., 2021a) to better understand their photophysical properties. The optical excitation energies of R 1 IR 2 and R 1 SR 2 were determined using the CAM-B3LYP functional with the Def-2-TZVP basis set and the TPSSTPSS functional with the 6-31+G (2 days, p) basis set, respectively. The results corresponded well to the experimental data. R 1 SR 2 showed absorption peaks at approximately 350 nm, with a high molar absorption coefficient (ε 31.4-62.8 × 10 3 ) and weak emission (Փ F 0.1-1.8%) ( Figure 3B; Table 2). On the other hand, R 1 IR 2 exhibited absorption peaks at about 325 nm, with a lower molar absorption coefficient (ε 7.2-13.3 ×10 3 ) ( Figure 3A; Table 2). R 1 IR 2 without a carbazole moiety showed a strong emission peak at approximately 440 nm (Փ F 26-63%) and a large Stokes shift (Δv ∼ 115 nm). In sharp contrast, R 1 IR 2 with a carbazole moiety exhibited weak emission (Փ F 3.8-8.2%) due to photoinduced electron transfer (PET) from the carbazole donor to the naphthalene-imidazolium salt acceptor (Sun et al., 2019). The absorption of R 1 IR 2 is assigned to the S 0 → S 1 transition and its emission is assigned to the S 1 ′ → S 0 transition, with the orbital contribution located in the naphthalene-imidazolium salt core (Supplementary Table  Frontiers in Chemistry | www.frontiersin.org July 2021 | Volume 9 | Article 713078 S12). R 1 SR 2 exhibited S 0 → S 3 absorption and S 2 ′ → S 1 emission in the absence of carbazole groups, while it showed S 0 ↔ S n /S n ′ transition at the higher level of S n /S n ′ in the presence of carbazole groups. The absorption band is contributed by other orbitals in addition to the HOMO and the LUMO (Supplementary Table  S13), but natural transition orbitals (NTOs) showed a similar electronic transition in the naphthalene and imidazoline-2-thione moieties of R 1 SR 2 (Supplementary Table S14), which demonstrates that the two have the same UV/vis absorption spectra ( Figure 3B). The fluorescence emission quantum yield of BIM and DIM (Փ F 59.2-62.9%) is significantly greater than that of BIB, DIB, and DID (26.4-32.6%) in DMSO, which is attributable to the absence or presence of the second (di)bromobenzyl group. The S 1 absorption energy of BIB and BIM vs. DIB, DID, and DIM is similar, whereas the S 1 ′ emission energy of BIM and DIM is higher than that of BIB and DIB or DID, respectively. Thus, the energy gap (ΔE) between the S 1 absorption and the S 1 ′ emission of R 1 IM (R 1 B or D) is lower than that of R 1 IR 2 (R 1 , R 2 B or D) (Figure 4). Moreover, the energy relaxation wastage of R 1 IM (R 1 B or D) from the S 1 absorption level to the S 1 ′ emission level is less than that of R 1 IR 2 (R 1 , R 2 B or D), leading to the increase in fluorescence emission quantum yield of BIM and DIM.
At the different volume fractions of PBS buffer with a pH value of 7.4 (0-99.5%), the fluorescence emission of DIM and DID in DMF is maintained owing to the high water solubility of  The molar absorption coefficient (ε) (M −1 cm −1 ), stock shift (Δv), and fluorescence quantum yield (Փ F ) were measured in DMSO and toluene for R 1 IR 2 and R 1 SR 2 , with 9,10-Diphenylanthracene (Փ F 0.90 in cyclohexane) being used as a reference; the oscillator strength (f), the energy gap (E g ) between the HOMO and LUMO levels, and the energy gap (ΔE) between S 0 and S n /S n ′ were not observed.
Frontiers in Chemistry | www.frontiersin.org July 2021 | Volume 9 | Article 713078 5 imidazolium salt groups ( Figure 5B; Supplementary Figures S48,  S49). We further examined their fluorescence emission in the aggregate state. Interestingly, the emission peak of DIM at 450 nm decreased, whereas a 335-350-nm emission band slightly increased with the increasing of toluene (Tol) concentration (0-99.5%) ( Figure 5A). DIM showed the ACQ effect on its core structure  -not calculated.
Frontiers in Chemistry | www.frontiersin.org July 2021 | Volume 9 | Article 713078 6 in DMF being quenched in the high concentration of toluene (Supplementary Figure S49); at the same time, it showed blueshift emission assigned from the rotation of bromobenzyl groups around the imidazolium salt core.
On the other hand, the incorporation of dibromobenzyl groups enhanced the antibacterial effect toward Gram-positive bacteria such as S. aureus and MRSA. The introduction of a carbazole moiety increased the strength of the antibacterial effect toward S. aureus (CFU 50 1.9-4.2 µM, p > 37.6), whereas the antibacterial ability was moderate toward EC-GFP and MRSA and weak toward E. coli and ESBL-EC (CFU 50 > 128.0 µM). The negative amino group of carbazole affects the antibacterial ability of a positively charged imidazolium salt toward Gram-negative bacteria and enhances the antibacterial ability toward Gram-positive bacteria. In sum, the imidazolium salts DIM and DID showed strong antibacterial effects (p > 5.8) compared to the imidazoline-2-thiones (R 1 SR 2 ) DSM and DSD, respectively. Thus, DSM and DSD were potentially selected for OFF-ON antibacterial fluorescent probes.

ClO − Response
Recognition of ROS/RNS by DSM was observed through UV-vis absorption and fluorescence emission spectroscopy in PBS buffer with a pH value of 7.4 (0.5% DMF). The absorbance band (300-375 nm) of DSM (5 µM) is decreased, and its blue fluorescence emission is significantly enhanced after 30 min of incubation in ClO − (50 µM). In sharp contrast, the UV-vis absorption spectra of DSM are slightly changed, and its fluorescence emission is quenched upon exposure to other types of ROS/RNS, even at higher concentrations ( Figure 6A; Supplementary Figure S51). On the other hand, when DSD (5 µM) was treated with ClO − (0-160 µM), its fluorescence emission was also inhibited (Supplementary Figure S50A). DSD cannot react to ClO − owing to steric hindrance between the two dibromobenzyl groups. Thus, DSM showed a highly selective response to ClO − among various ROS/RNS, relative to other R 1 SR 2 ( Figure 6A). Upon the gradual addition of ClO − (0-65 µM) to DSM (5 µM) in PBS buffer with a pH value of 7.4 (0.5% DMF), the absorption band at 300-375 nm decreased, and an absorption peak appeared at about 325 nm ( Figure 7A). Furthermore, the fluorescence intensity of DSM (5 µM) at ∼445 nm was significantly increased in the presence of ClO − (50-65 µM) ( Figure 7A), which is attributed to the appearance of DIM via desulfurization ( Figure 6B). Thiourea, a basic form of N 2 C S, occurs in two tautomeric isomers, a thione form and a thiol form. The thiol form of DSM can react with reactive ClO − , leading to the breakage of the C-S bond and the formation of DIM ( Figure 6B). The reaction of this fused imidazolium salt was referred to in several previous reports (Xu et al., 2015;Xu et al., 2016). The conversion of DSM to imidazolium salt was examined via the reaction of DSM and ClO − under similar conditions. The obtained main product (DSM′) was confirmed with DIM by 1 H NMR, 13 C NMR, and ESI-HRMS (Supplementary Figures S35,  S36B). Furthermore, DSM was highly sensitive to ClO − , with a detection limit (LOD) of 0.13 µM (Supplementary Figure S50B), which is lower than that of many ClO − probes in previous reports (Xiao et al., 2015;Shen et al., 2017;Wang et al., 2018;Lee S. C. et al., 2020;Nguyen et al., 2020).
Owing to the high fluorescence emission quantum yield and antibacterial activity of DIM, DSM can act as a potential OFF-ON In contrast, their CFU percentages decreased under simultaneous treatment with ClO − and DSM. At ClO − concentrations of 20 and 80 μM, the lowest CFU rates were achieved at DSM concentrations of 2 and 8 μM, respectively. At a higher concentration of ClO − (140 µM), the lowest CFU rates were observed to be 59.6 and 55.9% at 16 µM of DSM toward ESBL-EC and MRSA bacteria, respectively. This demonstrates that DSM is converted to DIM upon ClO − treatment and subsequently inhibits bacterial growth. Therefore, DSM can be applied as a potential ClO − -activated fluorophore for enhanced fluorescence emission and antibacterial activity. This finding is unprecedented with regard to previous reports of ClO − fluorescent probes (Jiao et al., 2018;Pham et al., 2021b;Kwon et al., 2021).

CONCLUSION
A series of imidazolium salts (R 1 IR 2 ) and imidazoline-2-thiones (R 1 SR 2 ) (R 1 , R 2 methyl, dibromobenzyl, and carbazole groups) were synthesized and characterized by 1 HNMR, 13 CNMR, mass spectra, X-ray crystal structures, and DFT calculation-based molecular orbital analysis. The excitation wavelengths of the molecules were theoretically examined via TD-DFT with various functions and basis sets. Imidazolium salts (R 1 IR 2 ) without a carbazole moiety showed high fluorescence emission, whereas imidazoline-2-thiones (R 1 SR 2 ) exhibited weak emission. The antibacterial activities of these compounds against E. coli, S. aureus, ESBL-EC, EC-GFP, and MRSA were studied. Among these structures, DSM/DIM and DID/DSD showed high antibacterial activity ratios that would be useful for the design of OFF-ON antibacterial probes. However, DSD rarely reacted with ClO − due to steric hindrance, whereas DIM showed a high fluorescence emission quantum yield (Փ F 62.9%) that would be useful for bacterial imaging, and DSM exhibited a highly selective ClO − response with an LOD of  Frontiers in Chemistry | www.frontiersin.org July 2021 | Volume 9 | Article 713078 0.13 µM. Finally, the DSM probe was converted to DIM upon ClO − treatment and inhibited bacterial growth.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material; further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
TP conceived and designed the probes and calculated the DFT in the Gaussian 09 package, V-NN designed the probes, YC synthesized the probes, DK and O-SJ calculated the X-ray crystal structures, DL and HK performed the antibacterial test, ML performed the computational study, and JY, HK, and SL designed the probes and wrote the manuscript.