TTVP was synthesized in compling with the literature mentioned (Wang et al., 2018b). Other chemicals and reagents were bought from J&K Scientific and Sigma-Aldrich. DMEM/F12 cell medium, penicillin, streptomycin and fetal bovine serum (FBS) were obtained from HyClone (United States). Yeasen (Shanghai, China) provided cell counting kit (CCK)-8. S. aureus was acquired from China General Microbiological Culture Collection Center. SD rats were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China).

Mother solution (5 mM) of TTVP was obtained through dimethyl sulfoxide (DMSO) initially, and then mixed with PBS to certain concentration. The aqueous dispersibility of TTVP in water, PBS and DMEM at the concentration of 5 μM was displayed excellently (Supplementary Figure S1). Ultraviolet-visible (UV-Vis) absorption and fluorescence (FL) spectra were measured by the Perkin Elmer Lambda 25 spectrophotometer and fluorescence spectrometry (Agilent Cary Eclipse), respectively.

A single S. aureus colony on solid agar plate was added to 10 ml liquid Luria broth (LB) culture medium. After incubation at 37°C overnight at 220 rpm, the bacteria were gathered by centrifugation (9,000 rpm, 3 min) and washed twice. After being resuspended with sterile normal saline, the concentrations were measured at 600 nm (1.0 OD = 10⁸ CFU ml⁻¹). Finally, the S. aureus suspensions were diluted with PBS for further use.

A standard plate counting method was carried out to evaluate sterilization effect of TTVP in vitro at concentrations of 0, 0.02, 0.05, 0.1 and 0.2 μM. The mixture was exposed to white light lamp for certain min (5, 10, 15 and 30 min), while the control group was exposed in dark. The treated bacterial suspension was diluted (1:5000) and incubated on the agar plate at 37°C for 24 h, followed by quantification and photo taking.

ROS accumulation of TTVP was assessed According to a given standard protocol. The mixture of TTVP (0.1 μM) and DCFH-DA was incubated under a white light lamp (20 mW/cm² power) for 15 min. The fluorescence intensity was recorded via a microplate reader before irradiation and every 1 min interval for 16 times (λex = 480 nm, λem = 525 nm). Meanwhile, OD values of DCFH-DA solution and PBS were recorded as control.

S. aureus was collected and resuspended with PBS, followed by incubating with different concentrations of TTVP (0, 0.02, 0.05, 0.1 and 0.2 μM) for 15 min. Then Zeta potential was measured via a Malvern Zetasizer.

S. aureus was co-cultured with human corneal epithelial cells (HCECs) to test the selective affinity of TTVP against bacteria. S. aureus suspension was added to pre-cultured adherent HCECs, along with different concentrations (0, 0.02, 0.05, 0.1 μM) of TTVP for 15 min. The slides were sealed and photographed using a fluorescence microscope.

The possible cytotoxicity of TTVP is a crucial issue to be examined before application. In general, cell viability, cell morphology and live/dead staining were performed to evaluate the biosafety of TTVP in vitro. In order to explore the biosafety of TTVP in keratopathy, research based on normal cells of ocular surface was necessary. Hence, HCECs were selected to evaluate the toxicity of TTVP.

Adherent HCECs were cultured in 96-well plates, being divided into irradiation group (20 mW/cm², group a) and dark group (group b). After co-cultured with TTVP (0–1.0 μM) for 15 min, HCECs were placed in light or dark environment during the following 15 min. HCECs were continued culturing in the incubator for 24 h after removal of TTVP. Also, the cells were incubated for 24 h at the same concentration of TTVP in dark (group c) to detect the intrinsic cytotoxicity. Afterward, CCK-8 reagent was incubated for 2 h to obtain the fluorescence intensity at 450 nm. In addition, the cellular morphology of HCECs after incubation for 24 h was also recorded via optical microscopy.

For the live/dead staining analysis, HCECs were intervened with TTVP (0, 0.05, 0.1, 0.2 μM) for 24 h. Live and dead cells were labeled with calcein AM and PI, which excited green and red fluorescence under microscope, respectively.

The hemocompatibility of TTVP was detected through hemolysis assay. Saline was added to the whole blood to wash several times until the supernatant was colorless. After diluting with saline to 2% red cell suspension, different TTVP concentrations (0, 0.02, 0.05, 0.1, and 0.2 μM) were added to the above red cell suspension, and heated at 37°C for 1 h. After centrifugation (3,000 rpm, 5 min), the images were captured and the ratio of hemolysis was calculated by the absorbance of supernatants at 540 nm recorded by a microplate reader. The erythrocytes treated with distilled water were regarded as positive control.

SPSS 23.0 software and GraphPad Prism 5 were utilized for statistical analyses. ImageJ version 1.48 was used for quantification of bacterial colonies. One-way ANOVA was carried out between multiple groups. p < 0.05 were identified as statistically significant.

The synthesis and purification of TTVP were under the guidance to the previous study (Wang et al., 2018b). As represented in Figure 1, UV—Vis and FL spectra were performed to verifythe optical characteristics of TTVP. After incubating TTVP in S. aureus for 15 min, red fluorescence was obviously visible under 365 nm UV irradiation, indicating that TTVP had a remarkable bioimaging effect on S. aureus (Figure 1A). Afterward, as shown in Figure 1B, TTVP solutions alone photoexcited by 489 nm showed no obvious difference in FL intensity under different concentrations. After a mixture with S. aureus at room temperature for 15 min, FL intensity showed the emission maximum wavelength at around 620 nm under concentration-dependent manner. The phenomenon might be due to the change of TTVP aggregation states of TTVP in response to S. aureus. TTVP has a strong affinity for S. aureus and may become a valuable visualization reagent for bacterial research.

The antibacterial property in vitro was tested through traditional spread plate method, and the bacteriostasis was evaluated intuitively by comparing the number of viable bacteria (Zhang et al., 2022a). The bacterial growth in the plate was recorded and the antimicrobial rate of TTVP was calculated. First of all, as shown in Figure 2A, TTVP exerted negligible toxicity without light irradiation process after various concentrations of TTVP intervened. Next, the S. aureus incubated with TTVP was irradiated in an intensity of 40 mW/cm² for 15 min, showing good bactericidal ability at the concentration of 0.02 μM. In order to optimize the illuminance, we halved the light intensity of 20 mW/cm² and performed at different durations including 5, 10, 15 and 30 min. As shown in Figure 2A, the antibacterial effect of TTVP exhibited a dose-dependency feature, and enhanced gradually along with the increased concentrations ranging from 0 to 0.2 μM. Amazingly, S. aureus was killed effectively with irradiation (20 mW/cm²) for 15 min, along with 0.1 μM TTVP co-cultured (Figure 2B). In order to better bactericidal effect, this intervention method was applied for further experiments.

Under light irradiation (20 mW/cm²), we further explored the mechanisms of TTVP killing S. aureus by measuring the ROS generation ability within 15 min. As shown in Figure 3, Supplementary Figure S2, abundant ROS was generated by TTVP at the concentration of 0.1 μM under white light irradiation, which demonstrated that AIEgens could generate efficient ROS in an aggregation state. The possible mechanism was the block of energy consumption by the way of non-radiative channels and aggregation-induced intersystem crossing (AI-ISC), inducing the transfer of facilitated energy state between singlet (S1) and triplet (T1) (Xu et al., 2015b; Yang et al., 2016). What’s more, our previous research has shown a terrific generation of singlet oxygen in the aggregation state, inducing the antibacterial ability of TTVP (Lee et al., 2020). Thus, TTVP was expected to be a meaningful photosensitizer for PDT evolvement.

The zeta potential value of pure S. aureus was −25.31 ± 0.54 mV, while the values were increased to −21.76 ± 1.04, −21.06 ± 1.13, −20.71 ± 0.90 mV, and −18.99 ± 1.50 mV after incubated with TTVP (0.02, 0.05, 0.1 and 0.2 μM), which verified the affinity of TTVP to S. aureus (Supplementary Figure S3). The co-culture method of HCECs and S. aureus was further employed to verify the selective killing ability of TTVP. As shown in Supplementary Figure S4, in a co-culture system of HCECs and S. aureus, red fluorescence was observed in S. aureus (area circled) and extremely weak in HCECs. TTVP could bind S. aureus specifically under certain concentrations for the surface charge discrepancy. Such a high selectivity reduced the damage on the normal corneal cells, thus implying promising application in BK treatment.

The biocompatibility of AIEgens should be taken into serious consideration prior to biological application. Since TTVP was administrated on the cornea, HCECs were performed to evaluate the biosafety. In Figure 2C, no significant difference was observed in cell viability of HCECs incubated with TTVP at concentrations under 0.1 μM in dark and light condition. Meanwhile, the cellular morphology and density were observed visually by optical microscope. In Supplementary Figure S5, no obvious cell morphology change and density decrease were observed after TTVP incubation. What’s more, live/dead staining of HCECs was performed to evaluate the cytotoxicity of TTVP. As shown in Figure 2D, the results showed no meaningful differences between live and dead fluorescence after being treated with TTVP (0–0.2 μM) under dark or light condition (20 mW/cm², 15 min), indicating low cytotoxicity of TTVP. In general, a hemolysis ratio no more than 10% was considered safe and acceptable for biological application (Xu et al., 2015a). It was found that the hemolysis rates were less than 1% after TTVP was added to blood, and calculated to be 0.27% ± 0.001% at the concentration of 0.2 μM (Supplementary Figure S6). Therefore, TTVP in our study was expected to be a promising option in vitro for further application.

In order to assess the biosafety of TTVP in vivo, H&E staining was conducted on heart, liver, spleen, lung and kidney to detect the structural damage, while several biochemical indexes were applied to test the major functions of organs. H&E staining and biochemical indexes are extensively used to visually appraise the structural and functional change respectively (Abdelhalim et al., 2018; Zhang et al., 2018; Nguyen et al., 2019). As shown in Figure 6, no histomorphological changes were revealed in main organs in vivo, which proved the safety of TTVP in organ structure. Meanwhile, no obvious discrepancy was shown in the content of AST, ALT, TG, T-CHO, CRE, K⁺, Ca²⁺ and Cl⁻ among all groups, which indicated excellent liver and renal functions, lipid levels, and electrolyte stabilization in this study (Figure 7). TTVP was thus expected to be a promising option in vivo for BK application.