Four [C_nmim]Br ILs, namely 1-butyl-3-methylimidazolium bromide ([C₄mim]Br), 1-hexyl-3-methylimidazolium bromide ([C₆mim]Br), 1-octyl-3-methylimidazolium bromide ([C₈mim]Br), and 1-decyl-3-methylimidazolium bromide ([C₁₀mim]Br) (Figure 1B), were purchased from InnoChem Science & Technology Co., Ltd. (Beijing, China). Kits for detecting inorganic phosphate, superoxide dismutase (SOD) activity, and catalase (CAT) activity were purchased from Solarbio Science & Technology Co., Ltd. (Beijing, China). R. sphaeroides was derived from the frozen samples in the laboratory. Other reagents were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). All reagents used were analytical grade.

Bacterial cultivation and chromatophore preparation have been described (Hunter and Turner, 1988). Briefly, frozen R. sphaeroides were taken out of−80°C, inoculated into M22 + liquid medium, and anaerobically cultivated under a 60-W tungsten lamp for 3–5 days after dark adaptation for 12 h. Bacterial cells were harvested by centrifugation and resuspended in 20 mM Tris-HCl (pH 8.0). The chromatophore was collected by centrifugation (254 000 × g, 4°C, 1 h) after the cells were disrupted by ultrasonication. The surface layer of the sediment was gently brushed and homogenized with 20 mM Tris-HCl (pH 8.0), and the chromatophore was diluted to a final concentration of OD₈₅₀ = 50 cm⁻¹.

Bacteria were suspended in M22 + medium with 10⁵–10⁶ colony-forming units per mL (CFU·mL^-1) and exposed to 0–128 mM of ILs. After 4 h of dark adaptation, the bacteria were cultured under illumination. The optical density of the bacterial solution at 660 nm (OD660) was recorded after incubation for 96 h. The half maximal inhibitory concentration (IC₅₀) of ILs to PNSB was fitted according to survival rate-concentration curves using Origin 2021 software.

Samples for SEM were prepared as described (Xiao et al., 2017). Briefly, R. sphaeroides was incubated with 5 mM ILs for 4 h, collected by centrifugation, and washed three times with PBS. Bacterial cells were fixed with 2.5% glutaraldehyde, dehydrated with a 20%–100% ethanol gradient for 0.5 h each time, and air-dried. The surface morphology of bacteria was imaged using SEM (S-4800, Hitachi, Tokyo, and Japan).

The exogenous fluorescent probe propidium iodide (PI) was used to evaluate cell membrane integrity. Live bacteria were incubated in 10 mM PBS (pH 7.4), 20 μM PI, and 5 mM ILs for 30 min in the dark at 25°C. The unabsorbed dye was removed by washing and centrifugation (15,000 × g, 3 min) three times. The bacterial cells were resuspended in PBS buffer. PI fluorescence was determined by the FLS 980 fluorescence spectrometer (Edinburgh Instruments, Edinburgh, United Kingdom) at λ _ex = 538 nm and λ _pr = 617 nm.

The Q_y absorption bands of BChl a in LHs can be used as a spectral indicator to evaluate changes in the supercomplex structure of the cell membrane. The UV-Vis NIR absorption spectrum of the chromatophore incubated with 5 mM [C_nmim] Br for 4 h under dark was recorded using the Cary-60 UV-Vis spectrophotometer (Agilent, California, United States).

The measurement of the electrochromic absorption band shift of Crt is described (Zhou et al., 2018). Briefly, the chromatophore from R. sphaeroides was diluted to a final concentration of OD₈₅₀ = 0.55 cm⁻¹ in 10 mM MOPS (pH 7.3) buffer containing 20 mM ascorbate sodium (Vc-Na), 2 μM phenazine methyl sulfate (PMS), and 5 mM IL. The test system was based on the Cary-60 UV-Vis absorption spectrometer and modified using a homemade accessory (Supplementary Figure S1). The absorption spectra of the chromatophore, measured under excitation “without” or “with” an additional continuous actinic laser, were termed “Dark” and “Light,” respectively. The electrochromic absorption band shift was characterized by the differential spectra of “Light”-minus-“Dark.” All tests were repeated five times.

The ATP synthesis assays were performed as described (Kim et al., 2017). Briefly, the chromatophore isolated from R. sphaeroides was diluted to a final concentration of OD₈₅₀ = 5 cm⁻¹ in working buffer (20 mM Tris-HCl (pH 8.0), 20 mM Vc-Na, 2 μM PMS, 5 mM K₂HPO₄, 5 mM ADP, and 10 mM MgSO₄). [C_nmim]Br was added to the working solution at a final concentration of 5 mM. ATP synthesis was performed under illumination with a xenon lamp (15 W/m²) at 30°C for 15 min, and then terminated by adding 4% trichloroacetic acid. A tissue inorganic phosphorus content detection kit was used to measure the consumption of inorganic phosphate in the reaction system. The amount of ATP generated was calculated based on ADP + Pi → ATP under illumination. The reference experiment was performed by adding 0.2 mM N, N′-dicyclohexylcarbodiimide (DCCD), an inhibitor of F₀F₁-ATP synthase (Zürrer et al., 1983). ATPase activity was calculated compared with the control group under illumination.

R. sphaeroides was incubated with 5 mM [C_nmim]Br for 4 h under light. The bacterial cells were collected after centrifugation and lysed ultrasonically for 15 min. The concentration of total soluble cellular protein in suspension was determined using the Bradford method (Bradford, 1976). Commercial kits were used to determine the activity of SOD and CAT.

The effect of C_nmim]Br ILs on the growth of R. sphaeroides is shown in Figures 2A–D. All four ILs exhibited a significant inhibitory effect on bacterial growth. In the absence of ILs, bacteria grew to the logarithmic growth phase within 24 h and reached the stationary phase after 72 h. After the addition of ILs, taking [C₄mim]Br as an example (Figure 2A), bacterial growth was retarded at low concentrations (2 mM). Growth inhibition became more pronounced with increasing IL concentration, and the bacteria ceased to grow in the presence of 128 mM [C₄mim]Br. ILs with different alkyl chain lengths showed the same trend—a positive correlation between concentration and inhibition rate. Comparing the effect of alkyl chain length on bacterial growth revealed that the longer the carbon chain, the more significant the inhibition. Bacterial growth exposed to 2 mM [C_nmim]⁺ is shown in Supplementary Figure S2. The IC₅₀ values of [C₄mim]Br, [C₆mim]Br, [C₈mim]Br, and [C₁₀mim]Br at 96 h were 35.68 ± 2.21, 8.25 ± 0.49, 1.08 ± 0.03, and 0.25 ± 0.09 mM, respectively (Table 1), which are comparable to those reported for other Gram-negative bacteria (Ghanem et al., 2015). An excellent linear positive correlation between the logarithmic concentration of IC₅₀ and the corresponding alkyl chain length of IL was observed (Figure 2E). Given that the logP (oil–water partition coefficient) of the four ILs gradually increased from [C₄mim]Br to [C₁₀mim]Br (Supplementary Table S1) (Fan et al., 2016), the toxicity of ILs to R. sphaeroides was positively correlated with their hydrophobicity. This result was consistent with others on IL toxicity in bacteria (Pernak et al., 2007; Łuczak et al., 2010).

The morphology of R. sphaeroides was examined by SEM (Figure 3). The untreated bacterial cells were typically long and rod-shaped, approximately 1.5 × 0.5 μm in size, with an intact envelope and smooth surface and no adhesion (Figure 3A). After IL treatment (Figures 3B–E), the bacteria gradually lost their rod-like shape and shrank in size to a minimum of 0.9 μm × 0.5 μm, with a rough and concave surface. In scaled-up figures (Figures 3G–J), pores and cavities are visible on the surfaces of bacteria treated with longer alkyl chain ILs, such as [C₁₀mim]Br (Figure 3J). These results clearly indicate that ILs damaged the bacterial cell wall and cell membrane. This structural damage inhibited cell growth. The degree of morphological change is positively dependent on the alkyl chain length of [C_nmim]Br, consistent with other reports (Jeong et al., 2012; Mehmood et al., 2015).

PI is used to stain nuclei. PI exhibits enhanced fluorescence after intercalating with DNA. As a water-soluble dye, PI is used as an exogenous probe to evaluate cell membrane integrity because it cannot get through the intact membrane of living cells (Petkovic et al., 2012). The fluorescence intensity of PI up-taken by bacteria that were incubated with 5 mM ILs is shown in Supplementary Figure S3. PI fluorescence increase corresponded with the increase in alkyl chain length. This result is consistent with the results of SEM analysis. Surface potential measurements of bacteria revealed that ILs with longer chains were adsorbed more easily on the surface of the bacterial cell wall, because surface potential increased from−17−13 mV with the increase in alkyl chain length (Supplementary Figure S4).

The baseline-corrected near-infrared region absorption spectrum of the chromatophore, the debris of the cytoplasmic membrane of R. sphaeroides, exposed to 5 mM IL solution for 4 h is shown in Figure 4A (normalized at 850 nm). The absorption bands near 800 and 850 nm originate from BChl a bound to LH2 and the one near 880 nm from BChl a bound to LH1-RC (Figure 1A). LH2, as the main light-harvesting complex, forms aggregate in the cell membrane (Westerhuis et al., 1998; Kangur et al., 2008; Sumino et al., 2013). Treatment with ILs, with increasing alkyl chain lengths, increased the blue-shift of the B850 band, accompanied with a decrease in the intensity of B880 (Figure 4A and the corresponding insert graph). This result indicates that IL insertion decreased LH2 aggregation and disrupted the lipid bilayer structure. More significant variation of spectra was seen as differential spectra between the [C_nmim]Br-treated or untreated control (Figure 4B). The positive peak at 840 nm and the negative valley at 860 nm were generated in pairs due to the blue-shift of B850 bands in treated samples. The amplitude of ΔOD detected at 840 nm was plotted as a positively linear function of n (insert graph in Figure 4B). These results imply that the BChl a bound to the chromatophore can be used as a quantitative spectral indicator to evaluate the effects of [C_nmim]Br on PNSB.

In the PNSB chromatophore, Crt is considered a good endogenous intramembrane voltmeter, because its absorption spectra shift responds to the change in membrane potential, i.e., the Stark effect. Thus, Crt has been widely used as a spectral indicator to evaluate the effect of ionophores antibiotics, detergents, and ILs on membrane permeability (Malferrari et al., 2015; Zhou et al., 2018). In this study, an 808-nm continuous laser was used to initiate the photoinduced “proton-coupled electron transfer” (PCET) in the chromatophore (Figure 1A), which resulted in a transmembrane proton gradient (Δμ_H ⁺). Then the electrochromic absorption band shift of Crt was observed due to the transmembrane ΔΨ accompanying Δμ_H ⁺.

The absorption spectra (420–545 nm) of the Crt S₀→S₂ transition in the R. sphaeroides chromatophore under dark and actinic light is shown in Supplementary Figure S5. The Crt absorption bands red-shifted slightly under actinic excitation at 808 nm compared with those under dark. The “light”-minus-“dark” differential spectra, representing the electrochromic absorption band shift of Crt is shown in Figure 5A. The differential spectra had typical oscillating signal peaks at 526, 493, 460, and 431 nm, consistent with a previous report (Zhou et al., 2018). The amplitude of the differential spectra (ΔOD_amp) decreased linearly with alkyl chain length, i.e., n (Figure 5B). To obtain better S/N ratios, ΔOD_amp was calculated as ΔOD₅₂₆ − ΔOD₅₁₀. Considering that ILs can decrease the integrity of cell membranes, Δμ_H ⁺ and the accompanying transmembrane ΔΨ would consequently decrease. The corresponding decrease of ΔOD_amp upon IL treatment was reasonable. The linearly dependent relationship between ΔOD_amp and n implied that Crt can be developed as a quantitative and endogenous spectral probe to evaluate the effects of other effectors on biomembranes. Furthermore, this method could be developed for other species of purple bacteria.

To further investigate the mechanism of IL toxicity in R. sphaeroides, ATP synthesis and antioxidant enzyme activity were evaluated. Chromatophores isolated from R. sphaeroides contain the entire photosynthesis unit; therefore, they can utilize light to synthesize ATP. The ATPase activity of chromatophores treated without and with ILs is shown in Figure 6A. Under light, the ATPase activity of the IL-free group was considered 100%. Four ILs exhibited remarkable repressive activity of ATPase in the order of [C₄mim]Br < [C₆mim]Br < C₈mim]Br < [C₁₀mim]Br. Notably, for the IL containing the longest alkyl chain ([C₁₀mim]Br), ATPase activity drastically decreased to 20%, which is comparable to the effect of the typical proton pump inhibitor DCCD. The result that ATPase activity decreased upon treatment with ILs was consistent with those shown above, where Δμ_H ⁺ decreased upon IL addition. This is because the inability to establish a proton gradient between the cytoplasm and periplasm leads to a blockade of ATP synthesis in the chromatophore.

Antioxidant enzymes are highly expressed in organisms, including PNSB, under environmental stressors, such as light, oxygen, and pollutants, to eliminate damage caused by excess reactive oxygen species production (Zhang et al., 2012; Su et al., 2020). In this study, the activities of intracellular antioxidant enzymes, namely SOD and CAT, were determined in PNSB treated with 5 mM [C_nmim]Br. The activities of SOD and CAT were 2.90 and 92.31 U/mg without ILs, respectively (Figures 6B, C). However, after IL treatment, the activities of SOD and CAT significantly increased, with a positive correlation with alkyl chain length. Notably for [C₁₀mim]Br, the activities of SOD and CAT were the highest, that is 10.7-and 4.3-fold higher than that of the control, respectively. Combined with the results of bacterial morphology analysis, a reasonable conclusion can be drawn that hydrophilic ILs insert and disturb cell membrane structure, leading to membrane perforation, which consequently causes oxidative stress, and eventually, cell death.