Application of the Luminescent luxCDABE Gene for the Rapid Screening of Antibacterial Substances Targeting Pseudomonas aeruginosa

Pseudomonas aeruginosa (P. aeruginosa) is a typical Gram-negative bacterium that can cause the spoilage of catered food products. Using a luminescent reporter gene (luxCDABE), this study sought to construct a cell-based biosensor (PAO1-CE) to rapidly screen antibacterial substances against P. aeruginosa. A total of six antibiotics belonging to five categories were used as the model test substances. The results of the bioluminescence detection method were verified using traditional antibacterial research assessments. The correlation coefficient of the regression equation fitting the data generated using this method was greater than 0.98, supporting the credibility of this approach. Additionally, the EC50 of each of the antibiotics assessed in this study was lower than the 1/2 MIC determined by conventional means. All six of the antibiotics caused varying degrees of damage to the cell membrane and cell wall of P. aeruginosa. Importantly, this novel method helped shorten the time necessary for active-compound detection and could be used for high-throughput detection, which would also help improve the detection efficiency. The application of this method towards the discovery of novel antibacterial compounds targeting P. aeruginosa holds substantial promise for greatly improving the efficiency of compound discovery.


Introduction
As spoilage bacteria, the Gram-negative Pseudomonas aeruginosa (P. aeruginosa) is a major contributor to the spoilage of aquatic products, meats, and cold ready-to-eat foods [1][2][3]. Simultaneously, P. aeruginosa is also a conditional pathogen and a common agent causing chronic infection in immunocompromised patients [4,5]. Moreover, these bacteria often cause respiratory, urinary tract, and intestinal infections, and provoke the formation of green pus in wounds [6][7][8][9]. Therefore, P. aeruginosa contamination has become an important factor in regard to human health [10,11]. However, a method to quickly screen and find substances that can inhibit the growth of P. aeruginosa is currently unavailable.
At present, the detection of P. aeruginosa antibacterial substances is generally carried out using traditional methods, such as minimal inhibit concentration (MIC) detection, which is used to determine if a particular substance can inhibit the growth of a given microorganism [12][13][14][15]. This would then be followed by assessing changes in the morphology (via transmission electron microscopy, scanning electron microscopy, laser confocal microscopy) [16][17][18][19][20], metabolism [21][22][23], and other characteristics at effective concentrations of the identified compound. However, it takes several days for each measurement of the MIC. As such, it is not possible to quickly and efficiently screen a large number of antibacterial substances via this approach, which has delayed the discovery of new antibacterial compounds. PAO1-CE is a P. aeruginosa strain PAO1 containing the pBBR1MCS-5-luxCDABE plasmid ( Figure 1). The whole sequence of the pBBR1MCS-5-luxCDABE plasmid is supplied in Supplementary Material S1. The gene cassette luxCDABE was obtained from the plasmid pBBR1MCS-4-luxCDABE. In this study, plasmid pBBR1MCS-5-luxCDABE was chosen because Amp could not screen out recombinant P. aeruginosa. The plasmids pBBR1MCS-5 (BioVector Science Lab Inc., Beijing, China) and pBBR1MCS-4-luxCDABE (BioVector Science Lab Inc., Beijing, China) were both cut with BamHI (G|GATTC) and XhoI (C|TCGAG; Thermo Scientific, Waltham, MA, USA), ligated using T4 DNA ligase (Thermo Scientific, USA), and transferred into E. coli TG1 competent cells. The plasmid pBBR1MCS-5-luxCDABE was then transferred into P. aeruginosa PAO1 (ATCC 15692, donated by Prof. Xihui Shen from College of Life Sciences, Northwest A&F University, China) using a Biorad electrotransformation unit (MicroPluser, Hercules, CA, USA) to obtain PAO1-CE. The detailed transformations procedures are supplied in Supplementary Material S2.

Inhibition Ratio (IR) of the Six Antibiotics on Luminescence
A single colony of PAO1-CE was inoculated into 50 mL of Luria-Bertani (LB) liquid medium (containing 50 μg/mL kanamycin and 100 μg/mL gentamicin). The bacterial culture was grown overnight (37 °C, 180 rpm) to stationary phase (OD600nm = 1.8). The bacterial culture was then transferred to fresh LB medium with an inoculum size of 1% (v/v)

Inhibition Ratio (IR) of the Six Antibiotics on Luminescence
A single colony of PAO1-CE was inoculated into 50 mL of Luria-Bertani (LB) liquid medium (containing 50 µg/mL kanamycin and 100 µg/mL gentamicin). The bacterial culture was grown overnight (37 • C, 180 rpm) to stationary phase (OD 600nm = 1.8). The bacterial culture was then transferred to fresh LB medium with an inoculum size of 1% (v/v) and incubated under the same conditions (37 • C, 180 rpm) for 7 h until the bioluminescence intensity of the bacterial culture reached the maximum value. The culture was centrifuged at 5000× g for 5 min and resuspended with fresh LB medium to OD 600nm = 1.0 before cytotoxicity testing. A total of 200 µL of the above culture medium containing the different antibiotic concentrations (with an equal volume of distilled water as the negative control) were added into each well of a black 96-well plate. The bioluminescence intensity was measured with a Victor X3 luminescence detector (PerkinElmer, Waltham, MA, USA) at Foods 2023, 12, 392 4 of 18 37 • C. Each treatment was carried out in triplicate, and the IR after a 2 h exposure was calculated as follows: where IR: inhibition ratio of luminescence (%); L CK : luminescence intensity of the negative controls (distilled water) (count per second (CPS)); and L: luminescence intensity of the samples (CPS).

Traditional
Methods for Verifying the Sensitivity of Antibacterial Substances 2.5.1. MIC Tryptic soybean peptone agar (TSA) was aseptically transferred into sterile 24-well plates containing the different antibacterial substances. The contents (the final volume was 1 mL) of each well were gently mixed. The final concentrations of each antibacterial substance were between 4 mg/mL to 0.5 µg/mL. After being solidified, the agar in each well was spotted with 2 µL (approximately 10 4 CFU) of the test bacterium, and the plates were then incubated at 37 • C for 24 h. The lowest concentration of an antibacterial substance that resulted in no visible growth of the test bacterium was considered the MIC.

Growth Curve and Influence of the Dynamic Growth Model
The growth curves were constructed using the method of Shi et al. [20] with minor modifications. An overnight seed culture of PAO1-CE was centrifuged at 5000× g for 5 min to collect the cells. The OD 600nm was adjusted to 0.5 (approximately 1.5 × 10 8 CFU/mL) with fresh LB medium. Each antibacterial substance was then dissolved with the bacterial solution above (OD 600nm = 0.5) and added to the cultures to obtain final concentrations of 0 MIC, 1/32 MIC, 1/16 MIC, 1/8 MIC, 1/4 MIC, 1/2 MIC, and 1 MIC. The 0 MIC was used as the positive control (CK group), and uninoculated LB served as the negative control. Then, 300 µL of the culture was transferred into each well on a honeycomb plate (Bioscreen C, Helsinki, Finland). Cell growth was monitored using an automatic growth curve analyzer (Bioscreen C; Oy Growth Curves Ab, Helsinki, Finland) at 37 • C every 1 h at 600 nm. All analyses were conducted in triplicate. With time (h) as the abscissa and the suspension OD 600nm as the ordinate, a growth curve was constructed. The modified Gompertz model was selected to fit the growth conditions and characterize the growth parameters of this strain. The modified Gompertz's model can be expression as: where OD t : the concentration of the bacterial liquid at t; B: the maximum bacterial concentration; A: the initial bacterial concentration; M: the time (h) used for the strain to reach the stationary phase; µ: relative growth rate during the exponential phase (∆OD 600nm /h); λ: lag phase (h); and µmax: the maximum growth rate (OD 600nm /h). The sigmoidal function available in the SigmaPlot software was used to fit the parameters of A, B, M, µ, λ, and M. The correlation coefficient R 2 was used as the evaluation factor to gauge the fit of the model.

Measurement of Intracellular ATP Concentrations
ATP is the most basic carrier for energy conversion in living systems, and its content is directly related to energy metabolism in most organisms. As a critical energy molecule, ATP plays a role in various physiological and pathological processes of the cell, and many functions of the cell are affected by changes in ATP levels [24]. Usually, when cells are apoptotic, necrotic, or in a toxic state, ATP levels decrease, while high glucose stimulation can increase intracellular ATP levels in some cells. The intracellular ATP concentration of PAO1-CE was analyzed according to the method described by Tian et al. [29], with some modifications. An overnight culture of PAO1-CE was harvested by centrifugation (5000× g, 5 min), and the cells were then resuspended in DDW to achieve an OD 600nm of 0.5 (approximately 4 × 10 8 CFU/mL), with 2 mL of cell solution placed into Eppendorf tubes. Each antibiotic was then added to a tube at final concentrations of 0 MIC, 1 MIC, and 2 MIC, respectively. The samples were then incubated at 37 • C for 30 min. ATP was extracted via ultrasound (SCIENTZ-IID; Ningbo Scientz Biotechnology Co., Ningbo, China) on ice for 15 min. Then, the samples were centrifuged for 5 min at 5000× g, and the supernatant was taken and stored on ice to prevent ATP loss. ATP was measured using an ATP assay kit according to the manufacturer's instructions (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

Measurement of Cell Wall Permeability via AKP Activity
AKP is localized between the cell wall and cell membrane, and AKP activity cannot be detected in the culture medium under normal conditions [30]. As such, the content of AKP activity in the cells can be used as a measure of the cell wall permeability of a bacterium [21]. AKP assays were used to evaluate the effects of each antibiotic on the integrity of the cell wall. The alkaline phosphatase (AKP) concentration of PAO1-CE was analyzed using an AKP assay kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The pretreatment and active concentrations of the antibacterial substances used were the same as the measurement of ATP, except the ultrasonic step was omitted. The protein contents were determined using a total protein quantitative assay kit according to the manufacturer's instructions (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

Measurement of Cell Membrane Damage via Integrity
The bacterial cell membrane is a structural component of bacteria that serves as a physical barrier for the cell and also performs as an important structure with complex functions in the life of cells [31]. If the cell membrane is damaged, both small molecules and some macromolecules, such as DNA and RNA, can escape from the cell. Therefore, the extravasation of intracellular substances can be used as a good indicator for evaluating cell membrane integrity. Since DNA and RNA have strong absorptions at A 260nm , the determination of this indicator can estimate the integrity of the cell membrane. Damage to the cell membrane was determined using the method described by Lin et al. [32] with some modifications. Pretreatment of the bacteria was the same as for the method of the growth curve mentioned above. The concentration levels of the antibacterial substances were 0 MIC, 1 MIC, and 2 MIC. Then, 200 µL of the sample was transferred into each well of a 96-well UV microtiter plate (Corning, NewYork, NY, USA). The absorbance at 260 nm was recorded via a multilabel plate reader (Victor X3; PerkinElmer, Waltham, MA, USA) at an interval of 1 h. All analyses were conducted in triplicate.

Morphological Observation via FESEM
Observations were performed using a field emission scanning electron microscope (FESEM), as described, with some modifications [33]. The cells (OD 600nm = 0.5) were treated independently with each antibacterial substance at 0 MIC and 2 MIC. After incubation at 37 • C for 8 h, the cells were harvested by centrifugation (5000× g, 5 min) and washed twice with PBS (0.1 M, pH 7.0). Then, the cells were resuspended in 2.5% glutaraldehyde and kept at 4 • C for 24 h. After centrifugation, the cells were dehydrated in a water-alcohol gradient at various alcohol concentrations (30%, 50%, 70%, 80%, 90%, 100%, and 100%) for 10 min, respectively. Subsequently, 2 µL samples were dropped onto the smooth surface of a single throw wafer (5 mm × 5 mm × 0.1 mm; Harbin Tebo Technology Co., Harbin, China) and then dried using an automatic Critical Point Dryer (Leica EM CPD300; Leica Microsystems, Wetzlar, Germany) for 4 h. Finally, the wafers were fixed on a FESEM support and sputter-coated with gold under vacuum (Q105TS; Quorum Technologies Ltd., Laughton, UK); this was followed by microscopic examination using a FESEM (Nova Nano SEM-450; FEI Company, Hillsboro, OR, USA).

Statistical Analysis
All experiments were performed in triplicate. Statistical analyses were performed using SPSS software (version 20.0; SPSS, Inc., Chicago, IL, USA). The data were presented as the mean values ± standard error (SE, n = 3), and differences between means were tested using least significant difference (LSD) tests (p < 0.05, p < 0.01). The data were plotted using SigmaPlot 10.0 (Systat Software Inc., San Jose, CA, USA).

Cytotoxicity of the Six Test Antibiotics Evaluated by PAO1-CE
A self-bioluminescence biosensor, PAO1-CE, was established using the host strain P. aeruginosa harboring the plasmid pBBR1MCS-5-luxCDABE. PAO1-CE was then applied to detect antibiotics targeting P. aeruginosa. A total of six antibiotics representing five different categories were used as the model test substances. The IR% values of PAO1-CE increased with the increasing concentrations of Str, Car, Amp, Ter, Tmp, and Cip, which exhibited a significant correlation with the concentrations presented in Figure 2 and Table 1. With these concentrations as the abscissa and the luminescence inhibition rate as the ordinate, a mathematical model was fitted to the data. As Figure 2 shows, the maximum concentration of Str/Car/Amp/Ter/Tmp/Cip used in this experiment is 50 µg/mL, 47.5 µg/mL, 1800 µg/mL, 100 µg/mL, 2000 µg/mL, and 0.7 µg/mL. At these concentrations, the IR% of PAO1-CE was 98.52%, 97.25%, 98.14%, 99.64%, 99.37%, and 95.04%, respectively. The sensitivity of the PAO1-CE biosensor was higher than that determined via MIC detection. The minimum concentration used in this experiment of Str (1 µg/mL), Car (1 µg/mL), Amp (56.25 µg/mL), Ter (6.75 µg/mL), Tmp (500 µg/mL), and Cip (0.1 µg/mL) also resulted in the inhibition of PAO1-CE bioluminescence ( Figure 2). The R 2 correlation coefficients of the mathematical models were both greater than 0.98, and detailed information regarding the six fitted curves is listed in Table 1. According to the EC 50 (concentration for 50% of maximal effect) of the six measured antibiotics towards PAO1-CE, their cytotoxicity was in the order of Cip > Str > Car > Ter > Amp > Tmp. The six antibiotics showed different inhibitory effects against PAO1-CE. The MICs of Str, Car, Amp, Ter, Tmp, and Cip against PAO1-CE were 200.00 µg/mL, 190.00 µg/mL, 1800.00 µg/mL, 100.00 µg/mL, 2000.00 µg/mL, and 1.00 µg/mL respectively.

Growth Curves and the Influence of the Dynamic Growth Model
When grown in media with one of the six antibiotics at MIC concentrations, the growth rate of P. aeruginosa was significantly decreased and completely inhibited ( Figure 3). The modified Gompertz equation was applied to fit the growth curves and characterize the growth parameters, and the R 2 of each fitting equation was greater than 0.98, indicating that this equation can accurately depict the growth of PAO1-CE. The growth parameters of the six antibiotics are listed in Table 2. Except for Ter, the µ max of the other five antibacterial substances was significantly reduced (p < 0.05). The OD max was also significantly reduced, except for Amp and Ter. The results of the µ max and OD max measurements indicate that concentration-dependent effects were clearly present.    3). The modified Gompertz equation was applied to fit the growth curves and characterize the growth parameters, and the R 2 of each fitting equation was greater than 0.98, indicating that this equation can accurately depict the growth of PAO1-CE. The growth parameters of the six antibiotics are listed in Table 2. Except for Ter, the μmax of the other five antibacterial substances was significantly reduced (p < 0.05). The ODmax was also significantly reduced, except for Amp and Ter. The results of the μmax and ODmax measurements indicate that concentration-dependent effects were clearly present.

Intracellular ATP Concentrations
The effect of the six test antibiotics on the intracellular ATP concentrations is shown in Figure 4A. The intracellular ATP concentration in the CK group was 264.65 ± 0.10 µmol/L, while the Str/Car/Amp/Ter/Tmp/Cip at 1 MIC decreased to 70.44 ± 2.71 µmol/L, 170.34 ± 0.49 µmol/L, 174.96 ± 0.06 µmol/L, 140.31 ± 0.11 µmol/L, 59.26 ± 3.05 µmol/L, and 14.62 ± 1.99 µmol/L. Compared with the CK group, the six treatment groups showed a significant inhibition (p < 0.01) towards intracellular ATP concentrations in PAO1-CE. The influence of the six antibiotics on the intracellular ATP concentration of P. aeruginosa at 1 MIC from high to low is as follows: Cip > Tmp > Str > Ter > Car > Amp. As the concentration of action increased to 2 MIC, the intracellular ATP concentration of P. aeruginosa directly fell below 10% of the CK group and significantly decreased (p < 0.01).

Cell Wall Permeability
The AKP activity results are presented in Figure 4B. The AKP activity of the CK group was 33.95 ± 3.09 U/gprot. At the MIC, the destructive effect of Str, Car, Tmp, and Cip on the cell wall was the same as that measured in the CK group (p > 0.05). However, the AKP activity of the Amp/Ter group increased to 70.14 ± 6.59 U/gprot and 130.71 ± 9.70 U/gprot, respectively, which was significantly higher than the CK group (p < 0.05). When the concentration increased to 2 MIC, the AKP activity of the Str/Car/Amp/Ter/Tmp/Cip group was significantly increased. The antibiotic that had the greatest influence on AKP activity was Ter, whose AKP activity increased 6.9 times.

Cell Membrane Integrity
The leakage of nucleotides from the bacterial cells treated with each antibiotic was significantly increased (p > 0.01) compared to the CK ( Figure 5), indicating that these antibiotics could cause membrane permeabilization in PAO1-CE. As the treatment concentration increased from 1 MIC to 2 MIC, the leakage of nucleic acids in the Str ( Figure 5A), Car ( Figure 5B), Amp ( Figure 5C), and Ter ( Figure 5D) groups was significantly heightened, while the difference between 1 MIC and 2 MIC of the Tmp ( Figure 5E) and Cip ( Figure 5F) groups was not significant (p > 0.01). For Amp and Ter, as the concentration of action increased, the degree of destruction of the cell membrane of P. aeruginosa cells increased, and the effect of the treatment concentration of the two groups was significantly different (p < 0.01).

Morphological Analysis via FESEM
Untreated cells are shown in Figure 6A. Comparing the control with treated cells ( Figure 6B-G), the differences in morphology can be easily seen. The cells of the CK group showed a smooth surface and typical Gram-negative morphology ( Figure 6A). In contrast, the cell surface of each treatment group presented with varying degrees of collapse. These phenomena demonstrate that treatment with Str, Car, Amp, Ter, Tmp, and Cip led to morphological changes in the cells. Comparing the degree of damage in each treatment group, Amp, Ter, and Tmp presented with more extensive visual damage to the cell compared to that of Str, Car, and Cip.
12, x FOR PEER REVIEW 10 of 19 MIC from high to low is as follows: Cip > Tmp > Str > Ter > Car > Amp. As the concentration of action increased to 2 MIC, the intracellular ATP concentration of P. aeruginosa directly fell below 10% of the CK group and significantly decreased (p < 0.01).

Cell Wall Permeability
The AKP activity results are presented in Figure 4B. The AKP activity of the CK group was 33.95 ± 3.09 U/gprot. At the MIC, the destructive effect of Str, Car, Tmp, and Cip on the cell wall was the same as that measured in the CK group (p > 0.05). However, the AKP activity of the Amp/Ter group increased to 70.14 ± 6.59 U/gprot and 130.71 ± 9.70 U/gprot, respectively, which was significantly higher than the CK group (p < 0.05). When the concentration increased to 2 MIC, the AKP activity of the Str/Car/Amp/Ter/Tmp/Cip   Untreated cells are shown in Figure 6A. Comparing the control with treated cells ( Figure 6B-G), the differences in morphology can be easily seen. The cells of the CK group showed a smooth surface and typical Gram-negative morphology ( Figure 6A). In contrast, the cell surface of each treatment group presented with varying degrees of collapse. These phenomena demonstrate that treatment with Str, Car, Amp, Ter, Tmp, and Cip led to morphological changes in the cells. Comparing the degree of damage in each treatment group, Amp, Ter, and Tmp presented with more extensive visual damage to the cell compared to that of Str, Car, and Cip.

Comparison of PAO1-CE with other Whole-Cell Recombinant Bacterial Bioreporters
At present, the application of whole-cell biosensors in P. aeruginosa has been used for evaluating the mechanism of substance inhibition on P. aeruginosa growth, as well as con-

Comparison of PAO1-CE with Other Whole-Cell Recombinant Bacterial Bioreporters
At present, the application of whole-cell biosensors in P. aeruginosa has been used for evaluating the mechanism of substance inhibition on P. aeruginosa growth, as well as contaminant detection in food, drugs, and cosmetics. These research studies adopted different plasmids or transposons to obtain their whole-cell biosensors of P. aeruginosa, including pKD-201/202/204/205/207 [34], pGLITE [35], pME4510-lux [36], pKD-algU/pslM/pelA/algA/ppyR/-bdlA [37], pUTminiTn5luxCDABEKm2 [38], and pUC18-miniTn7TGm-luxCDABE [39]. The specific information regarding these studies is listed in Table 3. However, there is no effective mathematical model fitted to the biosensor data generated in the above studies. In this current study, we compared the relative changes in PAO1-CE luminescence inhibition under different concentrations of model antibiotics and obtained a strong positive correlation between the growth and luminescence expression of PAO1-CE (Table 1). These results help provide a solid foundation for the use of this biosensor in various P. aeruginosa research studies. Table 3. Summary of lux-tagged P. aeruginosa that have been presented in recent studies.

Cytotoxicity of the Six Test Antibiotics Verified by Traditional Methods
Based on the fact that PAO1-CE can evaluate the toxic effects of various model substances on P. aeruginosa, in order to clarify the ways in which various substances have toxic effects on them, the authors measured them using traditional microbial bacteriostatic research methods.
The MIC values of Cip, Car, and Ter in this study were higher than those presented in previous studies (with MIC values of 0.06 µg/mL, >128 µg/mL, and 12 µg/mL, respectively) [40][41][42], which suggests the possibility that the resistance of P. aeruginosa towards these compounds has increased in recent years. According to the results of Figure 4A, six antibiotics had a significant effect on the intracellular ATP level of P. aeruginosa, while the expression of luxCDABE was closely related to the intracellular ATP level [43]. This proves that PAO1-CE could be used to evaluate the toxicity of compounds of P. aeruginosa. The AKP activity of the CK group and treatment groups was invariant between 1 MIC and 2 MIC, according to Figure 4B, which indicates an impaired function regarding cell wall permeability. However, as the antibiotic concentrations increased, the cell walls of all the treatment groups indicated significant damage compared with the CK group (p < 0.01). The results in Figure 5 show that the A 260nm of each group has almost no changes within 120 min after treatment with six antibiotics, indicating that the six antibiotics have damaged the integrity of the P. aeruginosa cell membranes, and this damage is caused by the action of the antibiotics themselves.
ATP is used for various cellular functions and is the basic carrier of energy conversion in all organisms. Its content is directly related to energy metabolism [44]. Usually, intracellular ATP levels decrease when cells are apoptotic, necrotic, or in a toxic state, and high glucose stimulation increases intracellular ATP levels in some cells. According to Figure 4A, six antibiotics significantly affected the intracellular ATP level of P. aeruginosa, and the expression of luxCDABE was inseparable from ATP [45]. This theoretically confirms that PAO1-CE could be used to evaluate the toxic effect of the test substance on P. aeruginosa.
The AKP enzyme is an enzyme that can dephosphorylate the corresponding substrate by hydrolyzing phosphomonoesters and producing PO 4 3+ and free OH − . Such substrate molecules include nucleic acids, proteins, alkaloids, etc. AKP is usually located between the cell wall and the cell membrane, and AKP enzyme activity is very low in the medium under normal conditions [32]. Therefore, AKP activity can be used as a measure of bacterial cell wall permeability [46]. According to the experimental results in Figure 4B, the presence of six antibiotics led to a rapid increase in AKP enzyme activity, indicating that all six antibiotics changed the cell wall permeability of P. aeruginosa.
The six antibiotics selected in this research belong to aminoglycosides (Str), β-lactams (Car; Amp), tetracyclines (Ter), sulfonamides (Tmp), and quinolones (Cip). Aminoglycoside antibiotics are glycoside antibiotics formed by the connection of amino sugars and aminocyclic alcohols through an oxygen bridge. Aminoglycoside, β-lactam, and tetracycline drugs all exert antibacterial effects by inhibiting the synthesis of bacterial proteins [47][48][49]. The protein synthesis in bacteria is related to many life activities and involves a large amount of ATP release. Therefore, Str, Car, Amp, and Ter have a significantly greater effect on the intracellular ATP level of P. aeruginosa than other antibiotics ( Figure 4A). The principle of the antibacterial action of sulfonamides is to interfere with the folate metabolism of bacteria. It inhibits the activity of bacterial dihydrofolate reductase selectively, so that dihydrofolate cannot be reduced to tetrahydrofolate. As a coenzyme of one-carbon unit transferase, tetrahydrofolate is involved in the synthesis of nucleic acid precursors (purine, pyrimidine). The nucleic acid is an essential component for bacterial growth and reproduction [50]. However, the A 260nm of Tmp was higher than that of Cip, which means the amount of intracellular nucleic acid leakage of Tmp was higher than that of the Cip treatment group. The reason may be due to the inconsistency of the acute degree of inhibition of P. aeruginosa by Tmp and Cip. As the third-generation quinolone drugs, Cip has a nitrogen (hetero) bicyclic ring structure, which acts on bacterial cell DNA helicases, inhibits the synthesis and replication of bacterial DNA, and causes bacterial death [51]. This is consistent with the results of the A 260nm experiment that measures the integrity of the cell membrane in Figure 5. Cip destroys the integrity of the P. aeruginosa cell membrane, but its A 260nm value change is less than 0.2. This is because Cip directly inhibits the DNA synthesis of P. aeruginosa, making the treatment group's nucleic acid leakage significantly lower than other antibiotics.
According to the results utilizing the abovementioned antibiotics, it is clear that each of these compounds can cause damage to P. aeruginosa in regard to cell metabolism, cell membrane integrity, and cell wall permeability, suggesting that PAO1-CE could be used as a self-bioluminescence biosensor for rapidly detecting effective antibacterial substances targeting P. aeruginosa.

Analysis of Cytotoxicity Evaluation Mathematical Model Using PAO1-CE
At present, mathematical models are widely used in scientific research [32,[52][53][54]. For different research samples, choosing an appropriate model to analyze and obtain a highly reliable research model can provide a certain degree of theoretical guidance for practical production applications. For the bioluminescence method to evaluate the toxicity of substances, the current research is placed more in environmental science, using various wild luminescent bacteria or recombinant luminescent bacteria to evaluate the toxicity of toxic and harmful substances in water or soil. Researchers used various wild luminescent bacteria or recombinant luminescent bacteria to evaluate the toxicity of toxic and harmful substances in water or soil [55,56]. Bello-López et al. [38] used the pUT miniTn5luxCDABEKm2 transposon to construct three recombinant luminescent bacteria to investigate the pollution of bagged platelet concentrates. The results showed that, in the logarithmic growth period, the linear correlation between the luminous intensity of the three bacteria and bacterial biomass R 2 was 0.985, 0.976, and 0.981, respectively. This indicated that using the luxCDABE system to quantify luminescent activity at this stage is a fast and sensitive alternative method to study the propagation and automatic sterilization of bacterial contaminants in platelet concentrates. Shah and Naseby [36] determined the antiseptic efficacy of BKC by recombinant luminescent bacteria P. aeruginosa ATCC9027 tatH5-pMElux. The obtained correlation coefficient of the bioluminescence evaluation method was higher than 0.9, which was not significantly different from the results of the colony-forming units (CFUs) count. It also proved that lux+-tagged P. aeruginosa is the best construct for testing various antimicrobial agents.
The fitting model of Str and Car obtained in this study looks similar to the Gompertz model, but according to the comparison in Table 4, it is found that the decision coefficient of the Gompertz model is less than the exponential model: y = a(1 − exp(−bx)), which shows that the exponential model is more suitable for Str and Car samples. In the fitted exponential model, parameter a represents the limit IR% of the substance to P. aeruginosa, and b represents the relative growth rate of IR%. By comparing the parameter changes of Str and Car, it is concluded that, within the corresponding linear range, PAO1-CE is slightly more sensitive to Str dose changes than Car. The fitting models of Amp, Ter, Tmp, and Cip are linear models: y = b + ax. Parameter a is the relative growth rate of IR%, which represents the speed of the change in IR% with the acting dose; parameter b is a constant, and the value is the IR% of the PAO1-CE of the same volume of control when no dose of drug is added.
According to the values of parameter a of each fitting model formula in Table 1, it could be determined that the sensitivity of PAO1-CE to Amp, Ter, Tmp, and Cip dose changes from high to low was Cip > Amp > Ter > Tmp within the corresponding linear range. Compared with the studies of Bello-López et al. [38] and Visaggio et al. [57], the detection method established in this study has a higher correlation. This also proves that PAO1-CE can be used to quantitatively evaluates the sensitive inhibition of P. aeruginosa. Recombinant luminescent PAO1-CE can be used to quickly evaluate whether P. aeruginosa has a toxic response to chemical substances. At the same time, it can also be used to study the distribution of P. aeruginosa in food systems, soil, or water bodies. It can intuitively obtain the migration and distribution of bacteria in infected objects without any damage in the whole process, which is helpful for developing effective control methods. For example, changes in luminous intensity play a role in the rapid evaluation of P. aeruginosa biofilm formation and fungicide efficiency.

Conclusions
In this study, the luxCDABE gene was used to construct the fluorescent biosensor PAO1-CE to rapidly screen six antibiotics that inhibit P. aeruginosa. The results of the bioluminescence detection method show that their cytotoxicity was in the order of Cip > Str > Car > Ter > Amp > Tmp. The six antibiotics all damaged the membrane and cell wall of P. aeruginosa to varying degrees, causing the extravasation of intracellular substances. AKP activity increased significantly, while intracellular ATP content decreased significantly. The FESEM results confirmed that all six antibiotics significantly changed the morphological appearance of P. aeruginosa. In addition, the EC 50 of each antibiotic evaluated in this study was lower than the 1/2 MIC measured by conventional methods. The mathematical models fitted in this study also provide a reference for the subsequent discovery of other classes of bacteriostatic substances. As such, the luminescent bacteria test method using PAO1-CE appears to be promising for the evaluation of antibacterial effects targeting P. aeruginosa and the discovery of new bacteriostatic substances.  Data Availability Statement: All data generated or analyzed during this study are included in this published article (and its supplementary information files).