Pseudomonas aeruginosa Phosphate Transporter PitA (PA4292) Controls Susceptibility to Aminoglycoside Antibiotics by Regulating the Proton Motive Force

ABSTRACT

Pseudomonas aeruginosa is an opportunistic Gram-negative bacterium that causes nosocomial infections in immunocompromised patients. β-lactam and aminoglycoside antibiotics are commonly used in the treatment of P. aeruginosa infections. Previously, we found that mutation in a PA4292 gene increases bacterial resistance to β-lactam antibiotics. In this study, we demonstrated that mutation in PA4292 increases bacterial susceptibility to aminoglycoside antibiotics. We further found enhanced uptake of tobramycin by the ΔPA4292 mutant, which might be due to an increase of proton motive force (PMF). Sequence analysis revealed PA4292 is homologous to the Escherichia coli phosphate transporter PitA. Mutation of PA4292 indeed reduces intracellular phosphate concentration. We thus named PA4292 as pitA. Although the PMF is enhanced in the ΔpitA mutant, the intracellular ATP concentration is lower than that in the isogenic wild-type strain PA14, which might be due to lack of the ATP synthesis substrate phosphate. Overexpression of the phosphate transporter complex genes pstSCAB in the ΔpitA mutant restores the intracellular phosphate concentration, PMF, ATP synthesis, and aminoglycosides resistance. In addition, growth of wild-type PA14 in a low-phosphate medium resulted in higher PMF and aminoglycoside susceptibility compared to cells grown in a high-phosphate medium. Overall, our results demonstrate the roles of PitA in phosphate transportation and reveal the relationship between intracellular phosphate and aminoglycoside susceptibility.

INTRODUCTION

Pseudomonas aeruginosa is a versatile Gram-negative pathogen that causes acute and chronic infections in humans (1, 2). The bacterium is intrinsically resistant to a variety of types of antibiotics, including aminoglycosides, β-lactams, macrolides, quinolones, etc. (3). The resistance mechanisms mainly include low membrane permeability, β-lactamase, multidrug efflux systems, and pyocyanin production (4, 5).

Aminoglycosides are broad-spectrum cationic antibiotics that bind to the ribosomal 30S subunit, leading to codon misreading and cell death (6). Tobramycin and amikacin are the most commonly used aminoglycosides in the treatment of pulmonary P. aeruginosa infections in cystic fibrosis (CF) patients (6, 7). The initial step of aminoglycoside antibiotics to enter Gram-negative bacterial cell is to bind to the negatively charged LPS, followed by displacement of Mg²⁺, which may disrupt the outer membrane integrity, allowing themselves to enter the periplasm (8). Crossing the cytoplasmic membrane by aminoglycosides requires the proton motive force (PMF) that is generated by the electron transport chain (ETC) (9). The intracellular aminoglycosides then interfere with protein synthesis and cause mistranslation. Insertion of the mistranslated proteins into the cytoplasmic membrane impairs the integrity, which facilitates entrance of additional aminoglycoside molecules, accelerating cell death (10).

Given the critical role of the PMF in aminoglycoside uptake, alteration of central metabolism and cellular respiration may affect the bactericidal effect of the antibiotics. It has been demonstrated that carbon sources influence bacterial susceptibility to aminoglycosides (11 to 13). Meylan et al. demonstrated that fumarate enhances the tricarboxylic acid (TCA) cycle, respiration, and PMF, which sensitize P. aeruginosa stationary-phase cells to tobramycin (11). In contrast, glyoxylate promotes the glyoxylate shunt, which represses cellular respiration and decreases tobramycin susceptibility of P. aeruginosa (11). By using a three-dimensional lung epithelial cell model, Crabbé et al. demonstrated that the host metabolites potentiate bactericidal activities of aminoglycosides by stimulating the PMF of P. aeruginosa (14). We previously found that mutation in a triosephosphate isomerase gene tpiA sensitizes P. aeruginosa to aminoglycosides by enhancing the TCA cycle and subsequently the PMF (15).

In bacteria, PMF is the driving force of ATP formation by the F₁F₀ ATP synthase using ADP and inorganic phosphate (16). We previously identified that mutation in a phosphate transporter homologous gene PA4292 increases β-lactams resistance in P. aeruginosa (5). Here, we found that mutation of PA4292 increases bacterial susceptibility to aminoglycosides by enhancing PMF. We further verified the role of PA4292 in phosphate uptake and elucidated the relationships among phosphate deficiency, ATP synthesis, PMF increase, and aminoglycoside sensitization.

RESULTS

Mutation of the PA4292 gene increases bacterial susceptibilities to aminoglycoside antibiotics.

To examine the overall role of PA4292 in antibiotics resistance, we tested the MICs of various classes of antibiotics for the ΔPA4292 mutant. Deletion of the whole open reading frame of PA4292 did not affect the resistance to ciprofloxacin, ofloxacin, azithromycin, colistin, or polymyxin B (Table 1). However, the ΔPA4292 mutant was more susceptible to aminoglycoside antibiotics, with a 4-fold decrease in the MIC of amikacin and a 2-fold decrease in the MICs of tobramycin and gentamicin. Complementation with the PA4292 gene restored the resistance to these antibiotics (Table 1). Time-kill assays verified the increased susceptibilities to tobramycin and amikacin of the ΔPA4292 mutant (Fig. S1A, B in the supplemental material).

TABLE 1.

MICs of antibiotics (μg/mL)^a

Strain	MIC
	Ciprofloxacin	Ofloxacin	Azithromycin	Gentamicin	Tobramycin	Amikacin	Colistin	Polymyxin B
PA14	0.25	1	128	2	1	2	0.125	0.25
ΔPA4292	0.25	1	128	1	0.5	0.5	0.125	0.25
ΔPA4292/PA4292	0.25	1	128	2	1	2	0.125	0.25

^aData represent results from three independent experiments.

To understand the mechanism of the increased susceptibilities, we determined the intracellular accumulation of tobramycin by using a Texas Red-labeled tobramycin (designated Tob-TR) (11). Mutation of PA4292 increased the intracellular amount of Tob-TR, which was restored to wild-type level by complementation with the PA4292 gene (Fig. 1A). As a control, mutation of PA4292 did not affect the intracellular accumulation of the free Texas Red (Fig. 1A). To further verify the intracellular amounts of tobramycin, we determined the mRNA levels of the heat shock protein genes ibpA, groES, and hlsV, whose expression is induced by intracellular tobramycin (17). Mutation of PA4292 resulted in higher mRNA levels of these genes in the ΔPA4292 mutant than those in the wild-type PA14 after tobramycin treatment (Fig. 1B), indicating a higher intracellular tobramycin amount.

FIG 1.

Mutation of PA4292 increases the uptake of tobramycin. (A) Wild-type PA14, the ΔPA4292 mutant, and the complemented strain were incubated with Texas Red (TR) or tobramycin Texas Red (TbTR) for 30 min. The fluorescence was normalized by the total protein amount. (B) The bacteria were treated with or without 0.25 mg/L tobramycin for 30 min, followed by RNA isolation. The mRNA levels of groES, ibpA, and hslV were determined by real-time PCR. Data represent the mean ± standard deviation of the results from three samples. *, P < 0.05; **, P < 0.01, by Student's t test.

The increased uptake of tobramycin is due to enhanced PMF in the ΔPA4292 mutant.

The intracellular accumulation of aminoglycoside antibiotics is determined by both influx and efflux. The expression levels of the major aminoglycoside efflux pump gene mexX were similar between the ΔPA4292 mutant and wild-type PA14 in the absence and presence of a subinhibitory concentration of tobramycin (Fig. S2) (18). We then examined the uptake of aminoglycosides, which depends on the PMF (10, 11). BCECF-AM staining revealed a higher PMF in the ΔPA4292 mutant than that in the wild-type PA14 (Fig. 2A). Furthermore, the PMF-dependent membrane potential was determined by staining with the potential-sensitive membrane dye DiSC3 (5). The lower fluorescent signal of the ΔPA4292 mutant represented enhancement of the membrane potential (Fig. 2B). An alamarBlue assay revealed enhanced respiratory rate in the ΔPA4292 mutant (Fig. 2C), which might lead to the enhanced PMF.

FIG 2.

Mutation of PA4292 increases the bacterial PMF, membrane potential, and respiratory rate. (A) The relative PMF levels of wild-type PA14, the ΔPA4292 mutant, the complemented strain, treated with or without CCCP (10 mg/L) or DCCD (50 mg/L), were measured after incubation with BCECF-AM. (B) The membrane potential dye DiSC3(5) was incubated with wild-type PA14, the ΔPA4292 mutant, and the complemented strain. The fluorescence was monitored for 30 min. (C) The respiratory rates of PA14, the ΔPA4292 mutant, and the complemented strain were quantified by alamarBlue. *, P < 0.05; **, P < 0.01; ***, P < 0.001, by Student's t test.

To verify whether the increased PMF in the ΔPA4292 mutant is responsible for the aminoglycosides sensitization, we used the oxidative phosphorylation uncoupler carbonyl cyanide meta-chlorophenyl (CCCP) to decrease PMF (Fig. 2A). At the concentration of 10 μg/mL, CCCP did not affect the MICs of tobramycin and amikacin for wild-type PA14, but restored the MICs for the ΔpitA mutant (Table 2). Previous studies demonstrated that Na₂ATP reduces bacterial membrane potential by inhibiting the TCA cycle (15, 19). Treatment with Na₂ATP increased the MICs of tobramycin and amikacin for the ΔpitA mutant and wild-type PA14 to the same levels (Table 2). Collectively, these results demonstrate that mutation of PA4292 enhances PMF, which subsequently increases bacterial susceptibility to aminoglycosides.

TABLE 2.

MICs of aminoglycosides (μg/mL)^a

MIC	PA14	ΔpitA	PA14 + CCCP	ΔpitA + CCCP	PA14 + DCCD	ΔpitA + DCCD	PA14 + Na2ATP	ΔpitA + Na2ATP	PA14/pMMB67	ΔpitA/pMMB67	PA14/pMMB67-pstSCAB	ΔpitA/pMMB67-pstSCAB
Tobramycin	1	0.5	1	1	0.5	0.5	32	32	1	0.5	1	1
Amikacin	2	0.5	2	2	0.5	0.5	32	32	2	0.5	2	2

^aData represent results from three independent experiments.

Defective phosphate uptake enhances PMF in the ΔPA4292 mutant.

To investigate the mechanism of PMF enhancement in the ΔPA4292 mutant, we analyzed the physiological role of PA4292. PA4292 is predicted as a potential phosphate transporter (20) that is homologous to the E. coli phosphate transporter PitA (Fig. S3) (21). Deletion of PA4292 indeed reduced the intracellular phosphate concentration, which was restored by complementation with the PA4292 gene (Fig. 3A), indicating a role of PA4292 in phosphate transportation. We thus named PA4292 as pitA. To verify whether the reduced intracellular phosphate concentration contributes to the enhanced PMF and aminoglycosides susceptibility, we cloned the phosphate transporter complex genes pstSCAB on the plasmid pMMB67EH, where transcription of the genes is driven by a tac promoter. Overexpression of the pstSCAB genes restored the intracellular phosphate concentration, PMF, and aminoglycosides resistance (Fig. 3B, C, Table 2).

FIG 3.

PitA (PA4292) affects intracellular phosphate level and PMF. (A) The intracellular phosphate concentration of PA14, the ΔpitA mutant, and the complemented strain. (B) Overexpression of pstSCAB restored the intracellular phosphate level of the ΔpitA (PA4292) mutant. (C) Overexpression of pstSCAB restored the PMF level of the ΔpitA (PA4292) mutant. *, P < 0.05; **, P < 0.01; ns, no significance, by Student's t test.

Given that PMF and phosphate are the driving force and substrate for ATP synthesis, respectively, we suspected that the increase of PMF in the ΔpitA mutant might be due to defective ATP synthesis. Indeed, the intracellular ATP level was lower in the ΔpitA mutant (Fig. 4A), which was restored to the wild-type level by overexpression of pstSCAB (Fig. 4B). To verify the effect of defective ATP synthesis on PMF level, we treated bacteria with the F₁F₀-ATP synthase inhibitor N,N′-dicyclohexylcarbodiimide (DCCD), which increased the bacterial PMF and aminoglycoside susceptibilities (Fig. 2A, Table 2). Collectively, these results demonstrate that mutation of pitA reduces intracellular phosphate level, leading to defective ATP synthesis, increase of PMF, and enhanced aminoglycosides susceptibility.

FIG 4.

PitA affects the intracellular ATP level. (A) The intracellular ATP concentrations of PA14, the ΔpitA mutant, and the complemented strain. (B) Overexpression of pstSCAB restored the intracellular ATP level of the ΔpitA mutant. **, P < 0.01, by Student's t test.

Phosphate depletion enhances PMF and aminoglycosides susceptibility in wild-type PA14.

We next examined whether phosphate depletion can increase the bacterial PMF and susceptibilities to aminoglycosides in wild-type P. aeruginosa. Growth of wild-type PA14 in a low-phosphate medium reduced the intracellular phosphate and ATP concentrations and increased the PMF (Fig. 5A, B, C). The MICs of tobramycin and amikacin were 4-fold lower in the low-phosphate medium than those in the high-phosphate medium (Table 3). Addition of CCCP (10 μg/mL) in the low-phosphate medium increased the MICs by 4-fold, whereas the MICs in the high-phosphate medium were not affected by CCCP (Table 3). These results demonstrate that phosphate depletion might increase bacterial susceptibility to aminoglycoside antibiotics by enhancing PMF.

FIG 5.

Phosphate limitation increases PA14 susceptibility to aminoglycosides through increasing the PMF. Wild-type PA14 was grown in the high-Pi or low-Pi medium to an OD₆₀₀ of 0.5, followed by determination of intracellular concentrations of phosphate (A), ATP (B), and PMF (C). *, P < 0.05; **, P < 0.01; ***, P < 0.001, by Student's t test.

TABLE 3.

MICs of aminoglycosides in low and high phosphate media (μg/mL)^a

MIC	PA14 in low Pi	PA14 in high Pi	PA14 in low Pi + CCCP	PA14 in high Pi + CCCP
Tobramycin	1	4	4	4
Amikacin	2	8	8	8

^aData represent results from three independent experiments.

DISCUSSION

In this study, we demonstrate that mutation of pitA sensitizes P. aeruginosa to aminoglycoside antibiotics by enhancing PMF, which enhances bacterial uptake of aminoglycosides (22, 23). The cellular respiration and ETC are enhanced in the ΔpitA mutant, thereby increasing PMF. We further verified the role of PitA in phosphate uptake, even though with the higher PMF, the intracellular ATP level is lower in the ΔpitA mutant, likely owing to the reduced uptake of phosphate that serves as the substrate for ATP synthesis. It has been demonstrated that increased intracellular ADP/ATP ratio stimulates cellular respiration through respiratory control, which might further increase PMF in the ΔpitA mutant (24, 25). A recent study in E. coli demonstrated that dissipation of membrane voltage gradients by proton ionophore cyanide m-chlorophenyl hydrazine (CCCP) did not affect kanamycin uptake. However, inhibition of ribosomal translation by kanamycin causes dysregulated membrane potential, which contributes to the bactericidal effect (26). Based on previous results and ours, we summarized the physiological role of PitA and its relationship with aminoglycoside susceptibility in Fig. 6.

FIG 6.

The schematic diagram shows the proposed roles of PitA in phosphate uptake and aminoglycosides susceptibilities. PitA is a potential cation-phosphate symporter (19). The intracellular phosphate serves as the substrate for ATP synthesis by the F₁F₀ ATP synthase driven by the PMF. Mutation of pitA decreased the intracellular phosphate concentration, which impairs ATP synthesis, leading to increase of PMF. The increased PMF promotes aminoglycoside uptake through a putative transporter AmT. In addition, the dysregulated membrane potential may directly contribute to the bactericidal effect.

Here, we found that phosphate depletion increases aminoglycoside susceptibility by enhancing PMF in wild-type PA14. Compared to the ΔpitA mutant grown in Luria-Bertani (LB) broth, growth of wild-type PA14 in the low-Pi medium reduced the intracellular ATP concentration and increased PMF to lesser extents (Fig. 3A, 3C, 4A, 5B, and 5C). However, the low Pi condition resulted in a 4-fold increase in bacterial susceptibility to tobramycin (Table 3), which is a bigger effect than the mutation of pitA (2-fold increase in the susceptibility). We suspected that the difference might be due to different causes of Pi starvation. Mutation of pitA reduced the intracellular Pi concentration due to transportation defect in a complete medium, whereas the low-Pi medium imposes an extracellular Pi starvation stress. The bacterium might respond differently to the external and internal signals. In addition, extracellular Pi starvation might have a stronger impact on membrane protein expression and conformation. Further studies are warranted to fully understand the difference in bacterial response to internal and external Pi starvation signals.

In the in vivo environment of human health, the phosphate availability is low. Thus, phosphate limitation is a common stress P. aeruginosa faces during infection (27 to 31). In vitro studies demonstrated that phosphate depletion induces production of multiple virulence factors in P. aeruginosa, including phospholipases, the quorum sensing-systems and downstream pyocyanin, the virulence regulator σ factor VreI, etc. (31 to 35). In combination with our results in this study, the in vivo phosphate starvation condition might potentiate bactericidal effects of aminoglycoside antibiotics. However, previous studies demonstrated adaptive aminoglycoside resistance in P. aeruginosa in vivo (36, 37). The human defence peptide LL-37 was found to stimulate aminoglycoside resistance in P. aeruginosa (38). Carbon sources alter PMF by impacting respiration and ETC, which affect bacterial susceptibility to aminoglycosides (11, 12, 19). We suspected that the in vivo aminoglycoside resistance level might be determined by the combination of various environmental signals and bacterial responsive pathways, such as induction of the MexXY-OprM pump, phosphate starvation, carbon sources, immune cells, host-generated antimicrobial substances, etc.

We previously found that mutation of pitA enhances pyocyanin production, which increases bacterial resistance to β-lactam antibiotics (5). We suspected that overproduction of the redox-active pyocyanin might alter outer membrane permeability, which reduces influx of β-lactam antibiotics into the periplasm, thus increasing bacterial resistance. Here, we demonstrated that mutation of pitA reduces intracellular phosphate concentration. Overexpression of the phosphate transporter complex genes pstSCAB in the ΔpitA mutant restored the resistance to aminoglycoside antibiotics. In P. aeruginosa, phosphate starvation activates the two-component regulatory system PhoR/PhoB, which activates the quorum-sensing systems (33). Therefore, intracellular phosphate deficiency might be the common mechanism that alters resistance to β-lactam and aminoglycoside antibiotics by affecting the quorum-sensing systems and PMF, respectively. Currently, we are making an effort to explore the mechanism of pyocyanin overproduction in the ΔpitA mutant. Overall, we revealed the role of PitA in phosphate uptake and the relationship between phosphate deficiency and aminoglycoside resistance.

MATERIALS AND METHODS

Bacterial strains, plasmids, and primers.

The bacterial strains, plasmids, and primers used in this study are listed in Table S1. The bacteria were cultured in LB broth (10 g/L tryptone, 5 g/L yeast extract, 5 g/L NaCl, pH 7.4) or the phosphate minimal medium (0.3% [wt/vol] proteose peptone containing 100 mM HEPES, 20 mM NH₄Cl, 20 mM KCl, 3.2 mM MgCl₂, and 0.4% [wt/vol] glucose [pH 7.2], with [high Pi] or without [low Pi] 10 mM KH₂PO₄) at 37°C with shaking at 200 rpm (27). We used100 μg/mL ampicillin and 10 μg/mL gentamicin for E. coli and 150 μg/mL carbenicillin and 50 μg/mL gentamicin for P. aeruginosa.

MIC determination.

The MICs were determined using broth microdilution in accordance with the Clinical and Laboratory Standards Institute guidelines (CLSI 2021). Briefly, individual antibiotics were serially diluted by 100 μL of cation-adjusted Mueller Hinton broth (CA-MHB) medium (Bio Faith, catalog number 11865) and then mixed with 100 μL bacterial culture (1 × 10⁵ CFU) in each well of a 96-well microtiter plate. After incubation at 37°C for 20 h, the MICs were defined as the lowest antibiotic concentrations with no visible growth. Each strain was tested in triplicate.

Tobramycin Texas Red uptake assay.

The tobramycin Texas Red (TbTR) was prepared by slowly adding 50 μL anhydrous N,N-dimethyl formamide containing 1 μg Texas Red sulfonyl chloride into 2.3 mL K₂CO₃ (0.1 M, pH 8.5) containing 10 mg/mL tobramycin. The TbTR (40 mg/L) or Texas Red was added to the late-logarithmic phase bacterial culture (OD₆₀₀ = 0.8 to 1; the growth curves are shown in Fig. S4). After incubation at 37°C for 30 min, the bacterial pellets were collected and washed three times with PBS. The bacterial cells were then suspended in 1 mL PBS and lysed by sonication. The fluorescence of Texas Red was measured at excitation/emission 595/615 nm by using a fluorometer (Varioskan Flash; Thermo Scientific). The relative uptake of TbTR was normalized by total protein concentrations of each bacterial sample.

RNA extraction and real-time PCR (RT-PCR).

The total RNA of bacteria was extracted by using the RNAprep pure cell/bacteria kit (Tiangen Biotec, Beijing, China). The cDNA was synthesized by using random primers and a PrimeScript Reverse Transcriptase (TaKaRa, Dalian, China). The RT-PCR was performed by using a CFX Connect Real-Time system (Bio-Rad, USA) and the SYBR II green supermix (Bio-Rad, Beijing, China). The ribosomal protein gene rpsL was used as the internal control (39). The cycle threshold (Ct) values are listed in Table S2.

Proton motive force measurement.

The bacterial membrane PMF was measured as previously described, with minor modifications (40). One milliliter of overnight cultured bacterial cells was collected and washed twice by a potassium phosphate buffer (50 mM K₃PO₄, pH 6.0) and once by the potassium phosphate buffer containing 5 mM EDTA. After resuspension in 1 mL of the phosphate-EDTA buffer, 10 μL of BCECF-AM (2 mM, US Everbright) was added, followed by incubation at room temperature for 1 h. The bacterial cells were collected by centrifugation and suspended in 120 μL phosphate-EDTA buffer, then incubated for 4 h on ice. Two microliters of the cell suspension were added to 200 μL phosphate-EDTA buffer in each well of a 96-well plate. The fluorescence was monitored at excitation of 500 nm and emission of 522 nm with a fluorometer (Varioskan Flash; Thermo Scientific).

Membrane potential assay.

The late-logarithmic phase bacteria (OD₆₀₀ = 0.8 to 1) were collected and washed three times by a GHEPES buffer (5 mM HEPES, 5 mM glucose) (41). All the bacterial suspensions were adjusted to the concentration of 5 × 10⁸ CFU/mL by the GHEPES buffer. 3,3-Dipropylthiadicarbocyanine iodide [DiSC3(5)] was added to the mixture at a final concentration of 2 μM. The fluorescence intensity was monitored every 5 min for half an hour at the excitation/emission of 622/670 nm with a fluorometer (Varioskan Flash; Thermo Scientific).

Respiratory rate assay.

The bacterial respiratory rates were quantified by alamarBlue as previously described (42). Bacteria were cultured to an OD₆₀₀ of 1.0. We mixed 180 μL of the bacterial culture and 20 μL alamarBlue dye and added them to a well of a black 96-well plate. The fluorescence was monitored every 10 min for 1 h. Data represent the mean ± standard deviation of the results from three biological replicates.

Intracellular phosphate concentration determination.

The bacterial intracellular inorganic phosphate concentration was determined using a Fluorimetric Phosphate assay kit (AAT Bioquest, USA). Briefly, 1-mL overnight cultures of indicated bacteria were collected and washed three times in deionized water. The bacteria were suspended in 1 mL deionized water and lysed by ultrasound sonication. After centrifugation at 12,000 rpm for 5 min, the supernatants were collected. Forty microliters of the supernatant were mixed with 40 μL of the assay buffer and 20 μL of the phosphate sensor working solution (provided by the kit) in each well of a 96-well plate. After incubation at room temperature for 45 min, the fluorescence was monitored with a fluorometer (Varioskan Flash; Thermo Scientific) at the excitation/emission of 540/590 nm. The phosphate concentrations were calculated by a standard curve (Y = 2.993X-4.0117) and normalized by the total proteins of each samples following the manufacturer's instructions. The fluorescence and total protein concentrations are listed in Table S3.

Bacterial ATP assay.

The bacterial intracellular ATP concentrations were determined by using an ATP assay kit (Beyotime Biotec, Shanghai, China). Briefly, 3 mL late-logarithmic growth phase bacteria (OD₆₀₀ = 0.8 to 1) were collected and resuspended in the provided lysis buffer, followed by cell lysis by ultrasound sonication. After centrifugation, 20 μL of the supernatant was mixed with 100 μL of the ATP-detection working fluid and incubated for 5 min at room temperature. The luminescence was detected with a fluorometer (Varioskan Flash; Thermo Scientific). The concentration of intracellular ATP was calculated with standard curves prepared using the provided ATP samples. All samples were normalized by total protein concentrations quantified by the BCA analysis.