Highly Sensitive Luminescent Bioassay Using Recombinant Escherichia coli Biosensor for Rapid Detection of Low Cr(VI) Concentration in Environmental Water

In this study, we constructed a recombinant Escherichia coli strain with different promoters inserted between the chromate-sensing regulator chrB and the reporter gene luxAB to sense low hexavalent chromium (Cr(VI)) concentrations (<0.05 mg/L); subsequently, its biosensor characteristics (sensitivity, selectivity, and specificity) for measuring Cr(VI) in various water bodies were evaluated. The luminescence intensity of each biosensor depended on pH, temperature, detection time, coexisting carbon source, coexisting ion, Cr(VI) oxyanion form, Cr(VI) concentration, cell type, and type of medium. Recombinant lux-expressing E. coli with the T7 promoter (T7-lux-E. coli, limit of detection (LOD) = 0.0005 mg/L) had the highest luminescence intensity or was the most sensitive for Cr(VI) detection, followed by E. coli with the T3 promoter (T3-lux-E. coli, LOD = 0.001 mg/L) and that with the SP6 promoter (SP6-lux-E. coli, LOD = 0.005 mg/L). All biosensors could be used to determine whether the Cr(VI) standard was met in terms of water quality, even when using thawing frozen cells as biosensors after 90-day cryogenic storage. The SP6-lux-E. coli biosensor had the shortest detection time (0.5 h) and the highest adaptability to environmental interference. The T7-lux-E. coli biosensor—with the optimal LOD, a wide measurement range (0.0005–0.5 mg/L), and low deviation (−5.0–7.9%) in detecting Cr(VI) from industrial effluents, domestic effluents, and surface water—is an efficient Cr(VI) biosensor. This unprecedented study is to evaluate recombinant lux E. coli with dissimilar promoters for their possible practice in Cr(VI) measurement in water bodies, and the biosensor performance is clearly superior to that of past systems in terms of detection time, LOD, and detection deviation for real water samples.


Introduction
Hexavalent chromium (Cr(VI)) is widely used in metal refining, leather tanning, wood preservation, and chemical and refractory processing; it is also applied to produce stainless steel, textile dyes, mimeographs, and plastics and as an anticorrosion agent [1,2]. If wastewater containing Cr(VI) is not treated properly, it may directly or indirectly pollute the water body or soil. Cr(VI) has been listed as a hazardous substance by the Agency for Toxic Substances and Disease Registry since 2011. Cr(VI) can damage DNA and cause varied toxicity, mutagenicity, and carcinogenicity, because it can quickly pass through the cell membrane and enter the cell given its high solubility, bioavailability, and mobility [3][4][5]. In Taiwan, environmental protection agencies have set 0.5 mg/L as the maximum allowable level (MAL) for Cr(VI) in industrial and domestic effluents and 0.05 mg/L as the MAL in surface water and fishery water. Thus, the need for sensitive Cr(VI) detection is high, but the design of such sensors is challenging. 2 of 14 Conventional analytical techniques such as atomic absorption spectroscopy, inductively coupled mass spectrometry, and ultraviolet-visible spectrophotometry are sensitive and reliable for Cr(VI) detection but are expensive, laboratory-bound, and timeconsuming [6]. Additionally, they only measure total metals but cannot distinguish between bioavailable and biounavailable metal concentrations or analyze the toxicity of metals [7,8]. By contrast, biological methods are useful alternatives because they can detect bioavailability and are low cost, easy to use, small, portable, highly sensitive, and highly selective [5,9]. Of the biological methods, bacterial biosensors are suitable for application as environmental sensors and early warning devices, even for on-field measurements [10].
Some Cr(VI) measurement techniques based on biological reactions have been developed. The performance of enzyme-based biosensors (e.g., amperometric enzyme and urease) for Cr(VI) detection is easily affected by the environment [11,12]. V79 cell biosensors have a wide Cr(VI) measurement range (0.97-19.4 mg/L), but their cost is high compared with that of microbe biosensors [13]. The performance of microbial-based biosensors (e.g., sulfur-oxidizing bacteria and Enterobacter aerogenes T2) for Cr(VI) detection is easily affected by coexisting metal ions, and these biosensors cannot measure Cr(VI) concentrations ≤0.05 mg/L [14,15]. The limit of detection (LOD) of microbial fuel cell (MFC)-based biosensors (inoculated with Exiguobacterium aestuarii YC211) is 2.5 or 5 mg/L depending on the system configuration [16,17]; however, the LOD is insufficiently sensitive to detect low Cr(VI) concentrations (e.g., ≤0.05 mg/L).
To detect low Cr(VI) concentrations, an engineered biosensor (whole-cell biosensor) should be considered. Many strains that can remove Cr(VI) have been discovered, such as Pseudomonas spp., Streptococcus lactis, Stenotrophomonas maltophilia, Pannonibacter phragmitetus, Cupriavidus metallidurans, and Ochrobactrum sp., and a key gene-the chr gene (responsible for Cr(VI) reduction)-has been identified [18][19][20][21]. However, considerable differences exist in the sequences of chr genes from different strains, which results in disparate reduction activities [21]. Among such strains, the genus Ochrobactrum has high Cr(VI) reduction ability; thus, the chr gene from Ochrobactrum spp. should be used to measure Cr(VI) [19,22].
A biosensor can be genetically engineered by placing a reporter gene, such as lacZ, gfp, luc, or lux, under the control of a transcriptional activator [23]. Of these reporter genes, lux and luc can rapidly obtain detectable signals, whereas gfp requires a long induction period for obtaining environmental responses [10], and the LOD of the biosensor using gfp as the reporter gene is often inferior to that of the biosensor using lux or luc as the reporter gene [24]. Thus, using the reporter gene from luminescent bacteria to construct a recombinant bacterium has considerable application potential for achieving satisfactory LOD. To increase the sensitivity, signal intensity, and response speed of wholecell biosensors, a suitable promoter should be selected. The common promoters SP6, T3, and T7 have similar but distinct specificities [25]. T7 is a strong promoter that actuates gene expression tuned to the highest level, thus amplifying the detection signal and producing a high LOD [26]. By contrast, relatively weak promoters (SP6 and T3) may adapt to environmental variation, which produces different signal characteristics and distinct detection ranges [27]. Thus, the LOD and measurement range of the biosensor would be improved or expanded with the insertion of suitable promoters into recombinant luminescent bacteria for moderate control of the expression of reporter genes [28,29].
Recombinant luminescent bacterial biosensors have been constructed to detect Hg 2+ , Cu 2+ , Pb 2+ , Ag + , Cd 2+ , Ni 2+ , Zn 2+ , As 3+ , and As 5+ [7,30]. However, few recombinant biosensors have been developed for Cr(VI) detection and for measuring Cr(VI) in real water samples. Smutok et al. (2011) constructed Hansenula polymorpha recombinant cells to detect Cr(VI), but the LOD (0.52 mg/L) was poor [31]. Branco et al. (2013) constructed two whole-cell biosensors, namely pCHRGFP1 Escherichia coli and pCHRGFP2 O. tritici, including the chrB regulator gene and gfp reporter gene to detect bioavailable Cr(VI). The recombinant E. coli biosensor was more sensitive than the recombinant O. tritici biosensor, and its LOD was 0.0194 mg/L in the 3-h detection time; however, complete and detailed evaluation in real water samples was lacking. To detect ≤0.05 mg/L Cr(VI) (MAL as per regulations), a strong promoter (e.g., T7 promoter) fused with a sensitive reporter gene (e.g., luxAB gene) should be used to shorten the detection time, enhance the luminescence intensity, and improve the LOD.
In this study, recombinant E. coli strains were engineered to carry the chrB plasmid from O. anthropi YC211 and contain various promoters (T7, T3, or SP6) to regulate luxAB expression. After the optimization of the promoter and the regulation of the luxAB expression level in E. coli, the recombinant luminescent biosensors could identify bioavailable Cr(VI) in real water samples of various source.

Bacterial Growth
E. coli with pCHR-T7-luxAB (T7-lux-E. coli), pCHR-T3-luxAB (T3-lux-E. coli), and pCHR-SP6-luxAB (T3-lux-E. coli) (all initial concentrations of 10 7 cfu/mL) were cultivated in LB broth containing a final ampicillin concentration of 50 mg/L, and Cr(VI) concentration of 0.05 mg/L at 37 °C at 200 rpm on an orbital shaker. After 18 h, cultures were diluted 100-fold into modified Tris minimal salts medium (TMM) containing 6.06 g/L Tris, 4.68 g/L NaCl, 1.49 g/L KCl, 1.07 g/L NH4Cl, 0.43 g/L Na2SO4, 0.2 g/L MgCl2.6H2O, 0.03 g/L CaCl2, 0.23 g/L Na2HPO4, 50 mg/L ampicillin, and 0.3% glucose, as per the previously described method of Mergeay et al. (1985) but with a slight modification [33]. In this study, E. coli was used as a control. To evaluate the feasibility of using frozen cells instead of fresh cells, the overnight culture grown in modified TMM was prepared as bacterial glycerol stocks and stored at −80 °C. To evaluate the relationship between bacterial growth and the luminescence intensity of recombinant E. coli, the cultures were incubated at 37 °C at 200 rpm on an orbital shaker, and optical density (OD) at 600 nm and the luminescence intensity were measured at specific intervals. Sampling 200 μL of the culture to a 96-well microplate, and then placing it under a microplate luminometer (Titertek-Berthold, Pforzheim, Germany), the luminescence intensity (expressed as relative light units (RLU)) was measured. All chemicals utilized in the experiment were analytical grade.

Determination of Optimal Conditions
After 18 h of cultivation in modified TMM, 1 mL of culture containing 10 8 cfu/mL T7lux-E. coli, T3-lux-E. coli, or SP6-lux-E. coli was inoculated into 100 mL of modified TMM with 0.5 mg/L Cr(VI) and incubated at 37 °C for 5 h. Cr(VI) was added as the oxyanion form of CrO4 2-, unless otherwise stated. The luminescence intensity was measured every 20 min. The effects of temperature (15 °C-45 °C) and pH value (4-9) on the luminescence intensities of the three recombinant E. coli strains were evaluated separately. Temperature was controlled using a constant-temperature incubator, and pH was adjusted using 1 N HCl or NaOH solution. To assess the effects of coexisting carbon sources on the luminescence intensities of the three recombinant E. coli strains, formic acid, acetic acid, citric acid, acetone, or fructose was added to the modified TMM, respectively. The final concentrations of these carbon compounds in the medium were 300 μM. To evaluate the effects of Cr(VI) oxyanion forms on the luminescence intensities of the three recombinant E. coli strains, CrO4 2− or Cr2O7 2− was introduced to the modified TMM.
The Cr(VI) concentrations added were 0.05, 0.5, and 5 mg/L. Coexisting cations (Ni(II), Co(II), Cd(II), Zn(II), Cu(II), and Cr(III)) and coexisting anions/similar structural configurations of CrO4 2− (SO4 2− , PO4 3− , and AsO4 2− ) were added to modified TMM to evaluate their effects on the luminescence intensities of the three recombinant E. coli strains. The concentrations of coexisting ions were 0.5-5 mg/L, and the Cr(VI) concentrations evaluated were 0.02, 0.5, and 5 mg/L. To evaluate the performance of frozen cells for detecting

Bacterial Growth
E. coli with pCHR-T7-luxAB (T7-lux-E. coli), pCHR-T3-luxAB (T3-lux-E. coli), and pCHR-SP6-luxAB (T3-lux-E. coli) (all initial concentrations of 10 7 cfu/mL) were cultivated in LB broth containing a final ampicillin concentration of 50 mg/L, and Cr(VI) concentration of 0.05 mg/L at 37 • C at 200 rpm on an orbital shaker. After 18 h, cultures were diluted 100-fold into modified Tris minimal salts medium (TMM) containing 6.06 g/L Tris, 4.68 g/L NaCl, 1.49 g/L KCl, 1.07 g/L NH 4 Cl, 0.43 g/L Na 2 SO 4 , 0.2 g/L MgCl 2 .6H 2 O, 0.03 g/L CaCl 2 , 0.23 g/L Na 2 HPO 4 , 50 mg/L ampicillin, and 0.3% glucose, as per the previously described method of Mergeay et al. (1985) but with a slight modification [33]. In this study, E. coli was used as a control. To evaluate the feasibility of using frozen cells instead of fresh cells, the overnight culture grown in modified TMM was prepared as bacterial glycerol stocks and stored at −80 • C. To evaluate the relationship between bacterial growth and the luminescence intensity of recombinant E. coli, the cultures were incubated at 37 • C at 200 rpm on an orbital shaker, and optical density (OD) at 600 nm and the luminescence intensity were measured at specific intervals. Sampling 200 µL of the culture to a 96well microplate, and then placing it under a microplate luminometer (Titertek-Berthold, Pforzheim, Germany), the luminescence intensity (expressed as relative light units (RLU)) was measured. All chemicals utilized in the experiment were analytical grade.

Determination of Optimal Conditions
After 18 h of cultivation in modified TMM, 1 mL of culture containing 10 8 cfu/mL T7-lux-E. coli, T3-lux-E. coli, or SP6-lux-E. coli was inoculated into 100 mL of modified TMM with 0.5 mg/L Cr(VI) and incubated at 37 • C for 5 h. Cr(VI) was added as the oxyanion form of CrO 4 2− , unless otherwise stated. The luminescence intensity was measured every 20 min. The effects of temperature (15 • C-45 • C) and pH value (4-9) on the luminescence intensities of the three recombinant E. coli strains were evaluated separately. Temperature was controlled using a constant-temperature incubator, and pH was adjusted using 1 N HCl or NaOH solution. To assess the effects of coexisting carbon sources on the luminescence intensities of the three recombinant E. coli strains, formic acid, acetic acid, citric acid, acetone, or fructose was added to the modified TMM, respectively. The final concentrations of these carbon compounds in the medium were 300 µM. To evaluate the effects of Cr(VI) oxyanion forms on the luminescence intensities of the three recombinant E. coli strains, CrO 4 2− or Cr 2 O 7 2− was introduced to the modified TMM. The Cr(VI) concentrations added were 0.05, 0.5, and 5 mg/L. Coexisting cations (Ni(II), Co(II), Cd(II), Zn(II), Cu(II), and Cr(III)) and coexisting anions/similar structural configurations of CrO 4 2− (SO 4 2− , PO 4 3− , and AsO 4 2− ) were added to modified TMM to evaluate their effects on the luminescence intensities of the three recombinant E. coli strains. The concentrations of coexisting ions were 0.5-5 mg/L, and the Cr(VI) concentrations evaluated were 0.02, 0.5, and 5 mg/L. To evaluate the performance of frozen cells for detecting 0.05 and 5 mg/L Cr(VI), frozen cells after 30-, 60-, 90-, and 120-day storage were examined. The luminescence emitted by T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli was stable and the highest was observed at 1.5-2.0 h, 1-1.5 h, and 0.5-1.0 h after incubation, respectively. Therefore, after 1.5-h incubation for T7-lux-E. coli cells, 1-h incubation for T3-lux-E. coli cells, and 0.5-h incubation for SP6-lux-E. coli cells, 200 µL of the cultures was sampled, and the luminescence intensity of these biosensors was measured immediately. After the completion of incubation, luminescence intensity was measured every 5 min over 20 min, and the five measured values were then averaged. The effects (luminescence intensity and stability time) of medium types (LB and modified TMM) and the Cr(VI) concentration (0.01-10 mg/L Cr(VI)) on luminescence characteristics of the three recombinant E. coli strains were evaluated. The cultures were incubated at 37 • C for 5 h, and the luminescence intensity was periodically measured. Incubation temperature, pH, and Cr(VI) concentration were controlled at 37 • C, pH 7, and 0.5 mg/L, respectively, unless otherwise stated. Measurements from at least three independent experiments were obtained, each performed at least in triplicate.

Establishment of Calibration Curve and Measurement of Real Water Samples
To establish the relationships between the Cr(VI) concentration and the luminescence intensity of the three recombinant E. coli biosensors, we mixed 2 mL deionized water containing Cr(VI) at different concentrations and 2 mL culture containing recombinant luminescent E. coli cells (final concentration after mixing: 5 × 10 6 cfu/mL). The culture was prepared using the thawed cells after 90-day cryogenic storage. The optimal incubation time and conditions were according to the results in previous experiments. Calibration curves were plotted on the basis of the linear regression of the luminescence intensity at each corresponding Cr(VI) concentration. That region (point) of the calibration, where there is a significant change in sensitivity (i.e., a break in the slope of the calibration), was defined as LOD concentration.
To verify that the established curves and methods were valid and feasible, similar solutions were prepared as mentioned in the earlier text, but industrial effluents (from Guishan Industrial Sewage Treatment Plant, Taoyuan City, Taiwan) and domestic effluents (from Dihua Sewage Treatment Plant, Taipei City, Taiwan) and surface water (from Tamsui River, New Taipei City, Taiwan) were used instead of pure Cr(VI) solution. Considering practical application, the performance of recombinant E. coli cells was examined after cryogenic storage. Thus, thawed cells after 90-day cryogenic storage were used instead of fresh cells for Cr(VI) measurement in real water samples. The Cr(VI) concentration in the real water samples was separately measured using the colorimetric 1,5-diphenylcarbazide (DPC) method [34] and the developed recombinant E. coli biosensors. Experimental data were collected from at least three independent experiments.

Time-Dependent Induction of Three Recombinant E. coli Biosensors with Cr(VI)
In the preliminary experiment, all the logarithmic growth phases of three recombinant E. coli cells were occurred at 8-20 h incubation. During this period, the luminescence intensities of the three recombinant E. coli strains were proportional to the bacterial growth. Therefore, we set the inoculation time of the three recombinant E. coli strains at 18 h after incubation in the subsequent experiment. Figure 2 indicated a comparison of the time-dependent induction of luminescence from the three recombinant E. coli biosensors (i.e., T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli) with 0.5 mg/L Cr(VI). The results indicated that the luminescence intensity of three recombinant E. coli biosensors rapidly increased, plateaued at 0.5 h, and then decreased during incubation. The luminescence intensity emitted from E. coli (control) maintained zero. However, this intensity variation was different from the luminescence caused by the gfp reporter gene, which slowly increased and plateaued [10]. This disparity was potentially due to the biochemical nature of the reporter gene luxAB [35].
was different from the luminescence caused by the gfp reporter gene, which slowly increased and plateaued [10]. This disparity was potentially due to the biochemical nature of the reporter gene luxAB [35].

Effects of Culture Conditions on Luminescence Intensity
The effects of pH and incubation temperature on the luminescence intensities of the biosensors for Cr(VI) detection were evaluated (considering practical aspects). Relative intensities were calculated using luminescence intensities at pH 7 (pH effect study) and 37 °C (temperature effect study). Figure 3A illustrates the effects of pH on the luminescence intensity induced with 0.5 mg/L Cr(VI) for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors. The results demonstrated that the optimal pH range for the luminescence intensities of T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors was 6-7, 5-7, and 5-8, respectively, with nonsignificant differences (p > 0.05). The SP6-lux-E. coli biosensor exhibited the highest pH adaptability among the biosensors. The T7-lux-E. coli biosensor was relatively unstable under pH change; nevertheless, the relative intensity of T7-lux-E. coli cells at pH 4 and 9 remained high at >93% (93.8% ± 0.41% and 95.2% ± 0.6%, respectively) compared with that at pH 7. Figure 3B displays the effects of incubation temperature on the luminescence intensities induced with 0.5 mg/L Cr(VI) for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors. The results indicated that the optimal temperature range for the luminescence

Effects of Culture Conditions on Luminescence Intensity
The effects of pH and incubation temperature on the luminescence intensities of the biosensors for Cr(VI) detection were evaluated (considering practical aspects). Relative intensities were calculated using luminescence intensities at pH 7 (pH effect study) and 37 • C (temperature effect study). Figure 3A illustrates the effects of pH on the luminescence intensity induced with 0.5 mg/L Cr(VI) for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors. The results demonstrated that the optimal pH range for the luminescence intensities of T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors was 6-7, 5-7, and 5-8, respectively, with nonsignificant differences (p > 0.05). The SP6-lux-E. coli biosensor exhibited the highest pH adaptability among the biosensors. The T7-lux-E. coli biosensor was relatively unstable under pH change; nevertheless, the relative intensity of T7-lux-E. coli cells at pH 4 and 9 remained high at >93% (93.8% ± 0.41% and 95.2% ± 0.6%, respectively) compared with that at pH 7. Figure 3B displays the effects of incubation temperature on the luminescence intensities induced with 0.5 mg/L Cr(VI) for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors. The results indicated that the optimal temperature range for the luminescence intensities of T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors was 20-37 • C, 20-37 • C, and 20-40 • C, with nonsignificant differences (p > 0.05). The SP6-lux-E. coli biosensor exhibited the highest temperature adaptability, which may be due to the short incubation time at the set temperature. The incubation temperature had considerable effects on T7-lux-E. coli cells. At the temperature of 15 • C or 45 • C, the relative intensity of T7-lux-E. coli cells decreased to 81.6% ± 1.2% and 86.2% ± 0.82%, respectively. This inconsistent result was presumed to be related to the promoter (T3, T7, SP6) structure and composition, which determine the strength of promoter-target DNA bonds and adaptabil-ity to environment changes [38]. However, the pattern of the incubation periods (T7-lux-E. coli > T3-lux-E. coli > SP6-lux-E. coli) was another possible factor.
The effects of medium types (LB and modified TMM) and different Cr(VI) concentrations on the luminescence intensity and incubation time for the three recombinant E. coli biosensors were evaluated. The results indicated that the effects of medium types on the stability time for the three recombinant E. coli biosensors were not obvious when Cr(VI) at 0.01-5 mg/L was used. The optimal time of luminescence induction was still maintained at 1.5 h, 1 h, and 0.5 h for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors, respectively; however, the luminescence intensities of the biosensors in modified TMM decreased by 9.2%, 7.5%, and 4.6% for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors, respectively, in comparison with that in LB as medium. In LB containing the high Cr(VI) concentration of 10 mg/L, the optimal induction time for all biosensors was postponed to 2-2.5 h. However, modified TMM containing 10 mg/L Cr(VI) did not affect the optimal induction time. This delay in the induction time may be attributed to the LB composition being more complex and nutritious than modified TMM [10]. In conclusion, modified TMM is a suitable medium for biosensor application.
intensities of T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors was 20-37 °C, 20-37 °C, and 20-40 °C, with nonsignificant differences (p > 0.05). The SP6-lux-E. coli biosensor exhibited the highest temperature adaptability, which may be due to the short incubation time at the set temperature. The incubation temperature had considerable effects on T7-lux-E. coli cells. At the temperature of 15 °C or 45 °C, the relative intensity of T7lux-E. coli cells decreased to 81.6% ± 1.2% and 86.2% ± 0.82%, respectively. This inconsistent result was presumed to be related to the promoter (T3, T7, SP6) structure and composition, which determine the strength of promoter-target DNA bonds and adaptability to environment changes [38]. However, the pattern of the incubation periods (T7-lux-E. coli > T3-lux-E. coli > SP6-lux-E. coli) was another possible factor.
The effects of medium types (LB and modified TMM) and different Cr(VI) concentrations on the luminescence intensity and incubation time for the three recombinant E. coli biosensors were evaluated. The results indicated that the effects of medium types on the stability time for the three recombinant E. coli biosensors were not obvious when Cr(VI) at 0.01-5 mg/L was used. The optimal time of luminescence induction was still maintained at 1.5 h, 1 h, and 0.5 h for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors, respectively; however, the luminescence intensities of the biosensors in modified TMM decreased by 9.2%, 7.5%, and 4.6% for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors, respectively, in comparison with that in LB as medium. In LB containing the high Cr(VI) concentration of 10 mg/L, the optimal induction time for all biosensors was postponed to 2-2.5 h. However, modified TMM containing 10 mg/L Cr(VI) did not affect the optimal induction time. This delay in the induction time may be attributed to the LB composition being more complex and nutritious than modified TMM [10]. In conclusion, modified TMM is a suitable medium for biosensor application.  Figure 4 illustrates the effects of coexisting carbon sources at 300 μM on the luminescence intensities of the three recombinant E. coli biosensors. The different coexisting carbon sources may potentially improve or weaken the luminescence intensity; thus, the results may deviate considerably relative to the theoretically expected effects using modified TMM alone [10]. The results demonstrated that the coexistence of fructose or acetic acid with modified TMM induced a high luminescence intensity in the T7-lux-E. coli biosensor compared with modified TMM alone. Fructose or acetic acid synergistically increased luminescence by 8.1% ± 0.84% or 5.2% ± 0.51%, respectively. Only the coexistence  Figure 4 illustrates the effects of coexisting carbon sources at 300 µM on the luminescence intensities of the three recombinant E. coli biosensors. The different coexisting carbon sources may potentially improve or weaken the luminescence intensity; thus, the results may deviate considerably relative to the theoretically expected effects using modified TMM alone [10]. The results demonstrated that the coexistence of fructose or acetic acid with modified TMM induced a high luminescence intensity in the T7-lux-E. coli biosensor compared with modified TMM alone. Fructose or acetic acid synergistically increased luminescence by 8.1% ± 0.84% or 5.2% ± 0.51%, respectively. Only the coexistence of acetic acid with modified TMM induced a high luminescence intensity (4.2% ± 0.56%) in the T3-lux-E. coli biosensor compared with modified TMM alone. However, all coexisting carbon sources had negligible effects (p > 0.05) on Cr(VI) detection by the SP6-lux-E. coli biosensor, and this may be because the genetic assembly of the SP6-lux-E. coli biosensor is relatively less susceptible to environmental interference [27]. The elevated luminescence intensity may be attributed to be the easy biodegradation of fructose and acetic acid, which improved the related physiological activity of recombinant E. coli [39].

Effects of Thawing Time on Luminescence Intensity
Regarding ease of use, constructing or preparing a recombinant biosensor when required is impractical. Thus, the luminescence intensities of the recombinant E. coli biosensors after cryogenic storage (30-120 days) were examined. The significant effects (p < 0.05) on the luminescence intensities of all thawing frozen cells after 120-day cryogenic storage were examined relative to fresh cells (0-day thawing time; Figure 5). After 90-day cryogenic storage, thawing recombinant T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli cells maintained luminescence intensities of 99.1% ± 0.51%, 99.2% ± 0.38%, and 99.5% ± 0.25%, respectively. Similar results were found for the luminescence intensities of the three recombinant E. coli biosensors induced with 0.05 mg/L Cr(VI). Thus, the subsequent experiment used thawing frozen cells after 90-day cryogenic storage.  [10]. Taken together, these results clearly illustrate that our recombinant luminescent biosensors possess favorable environmental adaptability, selectivity, and specificity under specific conditions.

Effects of Thawing Time on Luminescence Intensity
Regarding ease of use, constructing or preparing a recombinant biosensor when required is impractical. Thus, the luminescence intensities of the recombinant E. coli biosensors after cryogenic storage (30-120 days) were examined. The significant effects (p < 0.05) on the luminescence intensities of all thawing frozen cells after 120-day cryogenic storage were examined relative to fresh cells (0-day thawing time; Figure 5). After 90day cryogenic storage, thawing recombinant T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli cells maintained luminescence intensities of 99.1% ± 0.51%, 99.2% ± 0.38%, and 99.5% ± 0.25%, respectively. Similar results were found for the luminescence intensities of the three recombinant E. coli biosensors induced with 0.05 mg/L Cr(VI). Thus, the subsequent experiment used thawing frozen cells after 90-day cryogenic storage. Biosensors 2021, 11, x FOR PEER REVIEW 9 of 14

Relationship of Cr(VI) Concentration with Luminescence Intensity
According to the experimental results, we established the relationships between the Cr(VI) concentration and the luminescence intensity of the three recombinant E. coli biosensors using thawing frozen cells after 90-day cryogenic storage under optimal operating conditions. In this experiment, E. coli was used as a control. Figure 6A presents a set of regression equations for the Cr(VI) concentration and the luminescence intensity of the T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors when the Cr(VI) concentration was 0-0.5 mg/L: y = 107,464x + 999.6, y = 76,292x + 607.8, and y = 24,875x + 243.8, respectively. Coefficients of determination (R 2 ) of these regression equations indicated statistically significant results (>0.98), but the significance was insufficient for trace analysis of Cr(VI). Moreover, two obvious linear intervals between the Cr(VI) concentration and the luminescence intensity were noted. Thus, two other sets of relationships between the Cr(VI) concentration and luminescence intensity at various concentration ranges were further evaluated. Figure 6B presents another set of regression equations for the T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors when the Cr(VI) concentration was ≤0.075 mg/L: y = 168,421x − 31.2, y = 115,788x + 23.9, and y = 38,893x + 96.1, respectively. R 2 values of these regression equations were high (>0.999), indicating their reliability even if applied in trace analysis of Cr(VI). Figure 6C indicated another set regression equations for the T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors while the Cr(VI) concentration ranged from 0.075 mg/L to 0.5 mg/L. These regression equations for these given data were y = 98,746x + 4019.1, y = 71,770x + 2161.6, and y = 23,480x + 717.9, respectively. R 2 values for these equations were quite high (>0.999), representing their high reliability. The concentration-dependent differences in these linear relationships may be due to differences in promoter characteristics [9,40]. In addition, the LOD values of Cr(VI) measurement by the T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors were calculated as 0.0005, 0.001, and 0.005 mg/L, respectively. Therefore, the T7-lux-E. coli biosensor was the most sensitive, and the SP6-lux-E. coli biosensor was the least sensitive. Table 2

Relationship of Cr(VI) Concentration with Luminescence Intensity
According to the experimental results, we established the relationships between the Cr(VI) concentration and the luminescence intensity of the three recombinant E. coli biosensors using thawing frozen cells after 90-day cryogenic storage under optimal operating conditions. In this experiment, E. coli was used as a control. Figure 6A presents a set of regression equations for the Cr(VI) concentration and the luminescence intensity of the T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors when the Cr(VI) concentration was 0-0.5 mg/L: y = 107,464x + 999.6, y = 76,292x + 607.8, and y = 24,875x + 243.8, respectively. Coefficients of determination (R 2 ) of these regression equations indicated statistically significant results (>0.98), but the significance was insufficient for trace analysis of Cr(VI). Moreover, two obvious linear intervals between the Cr(VI) concentration and the luminescence intensity were noted. Thus, two other sets of relationships between the Cr(VI) concentration and luminescence intensity at various concentration ranges were further evaluated. Figure 6B presents another set of regression equations for the T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors when the Cr(VI) concentration was ≤0.075 mg/L: y = 168,421x − 31.2, y = 115,788x + 23.9, and y = 38,893x + 96.1, respectively. R 2 values of these regression equations were high (>0.999), indicating their reliability even if applied in trace analysis of Cr(VI). Figure 6C indicated another set regression equations for the T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors while the Cr(VI) concentration ranged from 0.075 mg/L to 0.5 mg/L. These regression equations for these given data were y = 98,746x + 4019.1, y = 71,770x + 2161.6, and y = 23,480x + 717.9, respectively. R 2 values for these equations were quite high (>0.999), representing their high reliability. The concentration-dependent differences in these linear relationships may be due to differences in promoter characteristics [9,40]. In addition, the LOD values of Cr(VI) measurement by the T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors were calculated as 0.0005, 0.001, and 0.005 mg/L, respectively. Therefore, the T7-lux-E. coli biosensor was the most sensitive, and the SP6-lux-E. coli biosensor was the least sensitive. Table 2 [5,6,10,13,16,17,31]. Compared with the aforementioned biosystems, our luxAB-based recombinant E. coli biosensors have smaller LODs (0.0005-0.005 mg/L), indicating their high sensitivity. However, it was inefficient in indicating the level of genotoxicity of Cr(VI) at low concentrations [36]. The satisfactory LOD of our sensors relative to previous sensors was due to the high activity of the chrB gene from O. anthropi YC211 and the position-appropriate promoters placed before the reporter gene lux [9].
Consequently, by virtue of the aforementioned reliable calibration curves for the Cr(VI) concentration range of 0.075-0.5 mg/L for the all biosensors or the Cr(VI) concentration range of 0.0005-0.75 mg/L (for T7-lux-E. coli biosensor), 0.001-0.75 mg/L (for T3-lux-E. coli biosensor), and 0.005-0.75 mg/L (for SP6-lux-E. coli biosensor), the Cr(VI) concentration in the water samples could be accurately and rapidly determined. In sum, the broad-range T7-lux-E. coli biosensor is a practical device for Cr(VI) measurement.
To evaluate the reproducibility of the biosensors for detecting Cr(VI), T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli were tested under optimal conditions by using modified TMM with 0.5 mg/L Cr(VI). Relative standard deviation (RSD) for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli was 3.2, 3.6, and 3.3%, respectively (n = 10). The low values of RSD of our recombinant luminescent E. coli biosensors demonstrated operational stability.  [5,6,10,13,16,17,31]. Compared with the aforementioned biosystems, our luxAB-based recombinant E. coli biosensors have smaller LODs (0.0005-0.005 mg/L), indicating their high sensitivity. However, it was inefficient in indicating the level of genotoxicity of Cr(VI) at low concentrations [36]. The satisfactory LOD of our sensors relative to previous sensors was due to the high activity of the chrB gene from O. anthropi YC211 and the position-appropriate promoters placed before the reporter gene lux [9]. Consequently, by virtue of the aforementioned reliable calibration curves for the Cr(VI) concentration range of 0.075-0.5 mg/L for the all biosensors or the Cr(VI) concentration range of 0.0005-0.75 mg/L (for T7-lux-E. coli biosensor), 0.001-0.75 mg/L (for T3lux-E. coli biosensor), and 0.005-0.75 mg/L (for SP6-lux-E. coli biosensor), the Cr(VI) concentration in the water samples could be accurately and rapidly determined. In sum, the broad-range T7-lux-E. coli biosensor is a practical device for Cr(VI) measurement.
To evaluate the reproducibility of the biosensors for detecting Cr(VI), T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli were tested under optimal conditions by using modified TMM with 0.5 mg/L Cr(VI). Relative standard deviation (RSD) for T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli was 3.2, 3.6, and 3.3%, respectively (n = 10). The low values of RSD of our recombinant luminescent E. coli biosensors demonstrated operational stability.   To date, some biosensors have been developed for detecting Cr(VI), but few biosensors have been applied for detecting the concentration of Cr(VI) in real water samples, particularly for determining compliance with the MAL in water bodies. Table 3 summarizes the measured Cr(VI) concentrations in three industrial effluent, three domestic effluent, and three surface water samples using the standard colorimetric method and our three recombinant luminescent E. coli biosensors. The results demonstrated that the Cr(VI) concentration determined using our biosensors and the colorimetric DPC method were highly correlated (R 2 > 0.999), excluding those for the B water sample of surface water. The B water sample of surface water was difficult to measure using the colorimetric DPC method because the Cr(VI) concentration was lower than its LOD (0.01 mg/L). Moreover, the deviation between the Cr(VI) concentrations measured using the colorimetric DPC method and those measured using the T7-lux-E. coli, T3-lux-E. coli, and SP6-lux-E. coli biosensors was −5.0-7.9%, −11.0-14.6%, and −2.6-18.4%, respectively. Some water samples required dilution because the Cr(VI) concentration exceeded the linear measurement range (>0.5 mg/L). These samples (industrial effluents B and C and domestic effluents B and C) caused positive deviation when the T7-lux-E. coli biosensor was used, but it caused negative deviation when the T3-lux-E. coli and SP6-lux-E. coli biosensors were used. The SP6-lux-E. coli biosensor had high environmental adaptability, as mentioned previously; thus, it was accurate (−2.6-1.3%) when it was applied for Cr(VI) detection in complex matrices (e.g., industrial and domestic effluents). However, its performance was unfavorable when applied to detect low Cr(VI) concentrations (e.g., <0.03 mg/L), with high positive deviation (11.0-18.4%). By contrast, the performance of the T7-lux-E. coli biosensor was favorable when applied to detect low Cr(VI) concentrations (e.g., <0.03 mg/L), with low deviation (2.6-4.2%). Considering the measurement range and accuracy, the T7-lux-E. coli biosensor provided the most accurate and reliable Cr(VI) measurement in these aqueous matrices. However, for Cr(VI) detection in industrial and domestic effluents, the SP6-lux-E. coli biosensor was the optimal biosensor in terms of accuracy.
The measurement deviation of the T7-lux-E. coli biosensor for Cr(VI) detection in real water samples was much lower than that (−13-6.2%) of an O. anthropi YC152 MFCbased biosensor, and the measurement deviation of the T3-lux-E. coli and SP6-lux-E. coli biosensors was comparable to that of an O. anthropi YC152 MFC-based biosensor for Cr(VI) detection [6]. Taken together, these results clearly indicate that the developed recombinant luminescent bacterial biosensors can determine low Cr(VI) concentrations in different water bodies.

Conclusions
In this study, three novel recombinant E. coli biosensors containing the chrB gene (from O. anthropi YC211); the T3, T7, or SP6 promoter; and the reporter gene lux were constructed for the rapid and accurate measurement of low Cr(VI) concentrations in industrial effluent, domestic effluent, and surface water. Of these biosensors, the T7-lux-E. coli biosensor exhibited the highest illuminance intensity, the most sensitivity to Cr(VI) (LOD: 0.0005 mg/L), and the widest measurement range (0.0005-0.5 mg/L) for Cr(VI) concentrations. Moreover, the SP6-lux-E. coli biosensor had the shortest detection time (0.5 h) and the highest environmental adaptability but low LOD (0.005-0.5 mg/L). The three recombinant E. coli biosensors exhibited advantages over previously reported biosystems in terms of optimal LOD, wide measurement range, and low measurement deviation. To the best of our knowledge, this is the first report on the use of recombinant biosensors to monitor the Cr(VI) concentration in real wastewater, especially from the perspective of whether they meet water quality standards (e.g., Cr(VI) ≤ 0.05 mg/L). In sum, our biosensors, particularly the T7-lux-E. coli biosensor, are sensitive, reliable, specific, and stable systems for preliminary in-field detection of Cr(VI) in water samples.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.