Biosensors and Bioelectronics 20 (2005) 1744–1749
An integrated mini biosensor system for continuous water toxicity monitoring Jin Hyung Lee, Man Bock Gu∗ Department of Environmental Science and Engineering, Gwangju Institute of Science and Technology (GIST), National Research Laboratory on Environmental Biotechnology, 1 Oryong-dong, Puk-gu, Gwangju 500-712, South Korea Received 23 April 2004; received in revised form 19 June 2004; accepted 29 June 2004 Available online 28 July 2004
Abstract An integrated water toxicity monitoring system that uses recombinant bioluminescent bacteria was successfully developed for the continuous monitoring and classification of toxicities present in water. This system consists of four channels arranged horizontally inside of a cylinder, with each channel having two small bioreactors that are vertically connected to each other to maintain a separation of the culture reactor from test reactor. This system is easily handled and installed, making its application in the field a potential reality. As well, it performed stably and continuously due to the vertical separation of the culture reactor from the test reactor and a long term operation was also performed because of its small working volume, i.e., only 1 ml for the 1st bioreactor and 2 ml for the 2nd. During an operation with four strains, i.e., EBHJ2, DP1, DK1 and DPD2794, which are responsive to superoxide damage (EBHJ2 and DP1), hydrogen peroxide (DK1), and DNA damage (DPD2794), the O.D. and bioluminescence of the bacterial cultures inside the system were constant when no chemical was injected. However, with the addition of paraquat, hydrogen peroxide or mitomycin C, the bioluminescent responses of the strains were found to be dose-dependent to different concentrations of these chemicals. © 2004 Elsevier B.V. All rights reserved. Keywords: Bioluminescence; Recombinant bacteria; DNA damage; Oxidative damage; Continuous monitoring system
1. Introduction Biomonitoring using bacterial bioluminescence has received increasing attention and been extensively used in environmental monitoring due to its advantages, such as a minimization of the detection time, enhancement in the sensitivity, visible responses and other benefits. However, the utilization of living bacteria would be difficult to implement in field monitoring because the need for preparation of the test organisms, temperature control and culture media preparation. Several methods, including freeze-dried (Bulich and Isenberg, 1981; Choi and Gu, 2002), immobilization (Kim and Gu, 2003; Jerome et al., 2001) and a miniature bioreactor (Gu et al., 1996) have been evaluated to satisfy these ∗
Corresponding author. Tel.: +82 62 970 2440; fax: +82 62 970 2434. E-mail address:
[email protected] (M.B. Gu).
0956-5663/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2004.06.036
requirements and to successfully utilize bacteria to detect environmental pollutants in environmental monitoring. Over the last few decades, the importance of continuous biomonitoring of environmental pollutants in effluents has been raised for protecting natural environments and public health, and for managing waste treatment plants. Effluents from cities, industrial plants and agricultural sites that eventually end up in rivers, streams or lakes may have many dissolved chemicals. Therefore, a continuous monitoring system is needed near discharge sites to prevent and observe the release of contaminates. The continuous monitoring using immobilized Photobacterium phosphoreum has been performed through utilization of a flow-through system (Uck et al., 1996) and fluidized-bed reactor (Se et al., 2003) where immobilized cells are grown and leaking cells are continuously released into the effluent. When the immobilized P. phosphoreum within the system was exposed to
J.H. Lee, M.B. Gu / Biosensors and Bioelectronics 20 (2005) 1744–1749
pollutants, bioluminescence declined in response to toxicity of pollutants then recovered to a new stable rate of decline. Recently genetically engineered bacterial strains have been developed for use as biosensors (Van Dyk et al., 1994; Vollmer et al., 1997) and a two-stage mini-bioreactor system was made to use these strains in continuous monitoring (Gu et al., 1999). This system consists of two mini-bioreactors (their volumes are 10 ml and 20 ml for 1st reactor and 2nd reactor, respectively) arranged in tandem. The 1st reactor is for cell growth and 2nd is for toxicity testing. The physical separation of the cell cultures allows for a stable and reliable operation because the cell growth rate and concentration can be constantly maintained in the 1st reactor and then supplied into the 2nd reactor, thereby replenishing the cells that have died or been washed out. This system can be also operated continuously since the responsiveness to toxic chemicals is maintained through the use of the unexposed daughter cells. One of the unique advantages of using recombinant bacteria is the classification of toxicity when using various stress promoters (Choi and Gu, 2003; Min et al., 2003; Kim et al., 2000; Lee et al., 2003). To do classification of toxicity, thus, a multi-channel continuous monitoring system was developed based upon the two-stage mini-bioreactor system (Gu and Gil, 2001). This system is composed of a series of two-stage mini-bioreactor systems arranged in parallel channels. Each channel carries a different recombinant bacterial culture that can be induced by a specific stress. Therefore, classification of the toxicity is achieved in the multi-channel continuous monitoring system through the simultaneous changes seen in the bioluminescence of each channel. This method provides valuable information for the treatment of wastewater or in the investigation of accidental discharge sources. In this study, the development of an integrated water toxicity monitoring system that allows for the continuous monitoring and classification of toxicants present in water, and makes the application of continuous monitoring systems in the field a potential reality due to its simple structure, longterm operation and sensitivity, is described.
2. Experimental 2.1. Recombinant bioluminescent strains and chemicals The recombinant bioluminescent Escherichia coli strains—EBHJ2 (Lee and Gu, 2003), DP1 (Mitchell et al., 2004), DK1 (Mitchell and Gu, 2004) and DPD2794 (Vollmer et al., 1997)—were used in this study. All strains were stored at –70 ◦ C in 15% glycerol and were streaked on agar plates containing 50 g/ml ampicillin (Sigma, USA) to obtain isolated colonies. One colony, grown overnight, was inoculated into a flask containing 50 ml of Luria-Bertani (LB) medium containing 50 g/ml ampicillin (Sigma, USA) and cultured for seed at 30 ◦ C for EBHJ2 and DPD2794 and 37 ◦ C for DK1 and DP1, overnight with shaking at 250 rpm. All strains were cultured to an optical density of 0.8 at a wavelength
1745
of 600 nm and were grown with ampicillin as the selection marker to ensure a true comparison between the lux fusions. From these cultures, 100 l of the seed cultures was inoculated into the first bioreactor of each channel using a microliter syringe (Hamilton Co., USA) to begin culturing in the integrated water toxicity monitoring system. Mitomycin C and paraquat (Sigma, USA) were dissolved in distilled water at a concentration of 50 mg/l and 100,000 mg/l, respectively, to prepare the stock solutions. The 30% hydrogen peroxide solution was purchased from the Merck Company, and used directly. The injection volume was 100 l in all cases and was pulse-injected into the second bioreactor using a microliter syringe (Hamilton Co., USA). 2.2. Measurement system and data analysis The integrated water toxicity monitoring system consists of four channels horizontally arranged inside of a cylinder (Fig. 1). Each channel has two small bioreactors, with 1 ml and 2 ml working volumes, respectively, that are vertically connected to each other to maintain a separation of the culture reactor and test reactor. Each channel has a bottom port with a glass window to connect it with a fiber optic light probe, which itself is connected to a highly sensitive luminometer (Model 20C Turner Design, CA) to measure the on-line bioluminescence level (BL). The bioluminescent data measured was transferred from the luminometer to a PC-based automatic data acquisition program through a RS232 cable for real-time bioluminescence data display and for data storage. Oxygen was supplied into the second reactors at 20 cm3 /min through a head port with a sparge pipe using an air pump (Model DH5000S, Dae Haw Co.). Low-density of medium (0.125× LB medium) was supplied into first reactors to keep low-density of cell (O.D. 0.4–0.5) inside the reactors. The dilution rate of both reactors was 0.8 h−1 . The dilution rate of the second mini-bioreactor was naturally controlled by summing the outflow from the first mini-bioreactor and addition of 0.1× LB medium through a second port, in this case the same amount as the out-flow of the first bioreactor. The temperature of the system was maintained using a water bath (Fisher Scientific Co., USA). Each experiment was performed at least twice with same concentration of chemical and the results were similar in both tests. Shown are the typical responses seen.
3. Results The system was operated continuously and three chemicals, paraquat, hydrogen peroxide and mitomycin C, were tested with the four strains, i.e., EBHJ2, DP1, DK1 and DPD2794. Each strain showed stable BL values and O.D.s during the operation (Data not shown). In this study, a pulse injection protocol, where a chemical is added all at once and the response of the strains is monitored, was employed. This was chosen since it simulated a situation where a water
1746
J.H. Lee, M.B. Gu / Biosensors and Bioelectronics 20 (2005) 1744–1749
Fig. 1. Schematics of the integrated water toxicity monitoring system (left) and sample flow in a channel (right). Items are labeled clock-wise from the top. (1) sample inlet, (2) medium inlet, (3) air outlet, (4) air inlet, (5) head plate, (6) main reactor body, (7) 2nd reactor, (8) 1st reactor, (9) channel consisting of both reactors, i.e., (7) and (8), (10) water jacket, (11) reactor housing, (12) medium outlet, (13) reactor base and stand, and (14) fiber optic probe and port.
sample is contaminated at various times during the system operation. When the chemicals were injected into the 2nd bioreactor, the bacterial strains showed a bioluminescent increase according to the promoter-lux fusion they harbored (Table 1). Although there are differences in the bioluminescent responses due to the properties of the promoters, the responses of the strains followed a similar trend, i.e., BL signal increase, attain a maximum value and then a BL decrease which levels out near the original uninduced value. This is seen for both strains DPD2794 (Fig. 2) and EBHJ2 (Fig. 3) when exposed to mitomycin C and paraquat, respectively. Mitomycin C (MMC) is a DDD (Direct DNA Damaging) agent that causes DNA damage by covalently binding to the minor groove of DNA, thus preventing the separation needed for DNA replication (Millard et al., 1998; Min et al., 1999; Li et al., 1995). Four different concentrations of mitomycin C were pulse injected into the 2nd bioreactor to a final concentration of 0.05 mg/l, 0.1 mg/l, 0.5 mg/l or 5 mg/l and strain DPD2794, which is responsive strain to DNA damage, was
induced (Fig. 2). This plot clearly shows the typical trend seen when these bacteria are induced by specific stresses, with the maximum response at a concentration of 5 mg/l and a minimum at 0.05 mg/l. During the tests with strain EBHJ2, shown in Fig. 3, seven different concentrations of paraquat were injected into the 2nd bioreactor and both the induction and recovery of the uninduced bioluminescence level were stable, proving that the responses of this system are unwavering even with numerous exposures, while the optical density within the second reactor, as measured by the outflow, was constant over the entire experiment. Therefore, since the basal level bioluminescence did not change significantly, when the values before the injection of paraquat and after stabilization were compared, and the optical density was also stable, this integrated system can be used in a continuous manner. Furthermore, strain EBHJ2 was highly sensitive to paraquat in this system and responded to ppb concentrations of paraquat. The minimum detectable concentration was found to be 50 ppb (0.05 mg/l).
Table 1 Responses of the strains to the different toxicants Strain
Chemical
Max RBLa
MRC (mg/l) b
MDC (mg/l) c
EBHJ2 DP1 DPD2794 DK1
Paraquat Paraquat Mitomycin C Hydrogen peroxide
250 2 12 3
50 0.5 5 100
0.05 0.1 0.05 10
a b c
Max RBL – maximum relative bioluminescence. MRC – concentration showing the maximum response. MDC – minimum detectable concentration (RBL of 1.5 or more).
J.H. Lee, M.B. Gu / Biosensors and Bioelectronics 20 (2005) 1744–1749
1747
Fig. 2. The response of DPD2794 to mitomycin C. The vertical arrows indicate when each concentration of chemical was injected for strain. Different concentrations of mitomycin C were injected into the 2nd bioreactor during the operation and a stable induction and recovery of bioluminescence was seen.
For each concentration, after reaching a maximum BL value, the BL signal decreased as the bacteria and chemical mixture was diluted and pumped out of the system. A previous study that the E. coli pqi-5 gene is induced by paraquat (Koh and Roe, 1995). Therefore, the BL from strain DP1, which carries a fusion of the pqi-5 promoter to the lux genes, was also induced when this strain was exposed to different concentrations of paraquat (Table 1). Although the induction level is not as high as those seen with strain EBHJ2,
clear responses to different concentrations of paraquat were seen. Hydrogen peroxide causes oxidative damaging to organism through the formation of hydroxyl radicals. Thus, strain DK1, having the katG promoter fused to the lux genes, showed an increase in its relative bioluminescence level (RBL = BLinduced /BLuninduced ) very shortly after hydrogen peroxide was injected (data not shown). The bioluminescence increased sharply after hydrogen peroxide injection and took
Fig. 3. The response of EBHJ2 to paraquat. Seven concentrations of paraquat were injected into the 2nd bioreactor during the operation and a stable induction and recovery of bioluminescence was seen. Also shown is the optical density, as triangles, over time, demonstrating that this system is stable.
1748
J.H. Lee, M.B. Gu / Biosensors and Bioelectronics 20 (2005) 1744–1749
only a short time to reach a maximum relative bioluminescence since hydrogen peroxide does not have to be converted to activate the OxyR protein, while other chemicals, such as paraquat, should first be converted to their ionized forms and then form radicals. The maximum bioluminescent level seen from each of the strains increased in a dose-dependent manner as the concentrations of the chemical increased until the chemical became overtly toxic, above which the maximum bioluminescence dropped because of metabolic inhibition. However, due to the separation of the two reactors, this integrated system was able to operate continuously and provide stable responses since a fresh supply of un-exposed cells was constantly available.
maintain a separation of the culture reactor from test reactor. During experiments with the four recombinant bioluminescent bacteria, the culture O.D. inside the system and the bioluminescence were constant when the cells were not exposed to chemicals. Furthermore, the bioluminescent responses of the strains were dose-dependent when they were exposed to different concentrations of specific chemicals, which were selected based upon the promoter-lux fusion present within the strain. Finally, this system can be used for the continuous monitoring and classification of toxicities through the use of other bioluminescent bacteria while its stable operation and ease of handling and installation make the application of such systems a potential in situ technology.
4. Discussion
Acknowledgements
In the application of recombinant bioluminescent bacteria to the continuous monitoring of river waters or wastewater treatment plants, it is essential to reduce amount of medium required to maintain the bacteria since the amount of medium determines both the operating time and the amount of spent media that should be treated later before disposal. It was thought, therefore, that a miniaturized system would satisfy both of these conditions for a continuous biomonitoring and, thus, an integrated water toxicity monitoring system was developed in this study. This system has one-tenth the working volume of previous system (Gu and Gil, 2001) and an integrated arrangement of four channels having two mini-bioreactors vertically, while the previous system consists of a series of four twostage mini-bioreactors, with 10 ml working volume for the 1st bioreactor and 20 ml for the 2nd bioreactor, arranged in parallel. These properties of the system lead to a smaller consumption of the medium and a lower production of waste, i.e., the bacteria-medium mixture. Therefore, the system can be operated for a longer period without management than other systems having same dilution rate after being set-up and, ultimately, it makes a remote control system more feasible. As well, the integration of the four channels into a single cylinder provides several advantages, such as ease in system handling and installation and, due to its integrated design, a reduced assembly and greater safety, both of which minimize the chance of the recombinant bacteria being accidentally released into the environment. These advantages make the application of recombinant bacteria in biomonitoring easier, as well as more practical.
This research was supported by Korea Science and Engineering Foundation (KOSEF) through Advanced Environmental Monitoring Research Center (ADEMRC) in Gwangju Institute of Science and Technology (GIST). Authors are grateful for the support.
5. Conclusions An integrated water toxicity monitoring system was successfully developed with four channels horizontally arranged inside of a cylinder, with each channel having two small bioreactors that are vertically connected to each other to
References Bulich, A.A., Isenberg, D.L., 1981. Use of the luminescent bacterial system for the rapid assessment of aquatic toxicity. ISA Trans. 20, 29–33. Choi, S.H., Gu, M.B., 2002. A portable toxicity biosensor using freezedried recombinant bioluminescent bacteria. Biosens. Bioelectron. 17, 433–440. Choi, S.H., Gu, M.B., 2003. Toxicity biomonitoring of degradation byproducts using freeze-dried recombinant bioluminescent bacteria. Anal. Chim. Acta 481, 229–238. Gu, M.B., Gil, G.C., 2001. A multi-channel continuous toxicity monitoring system using recombinant bioluminescent bacteria for classification of toxicity. Biosens. Bioelectron. 16, 661–666. Gu, M.B., Gil, G.C., Kim, J.H., 1999. A two-stage minibioreactor system for continuous toxicity monitoring. Biosens. Bioelectron. 14, 355–361. Gu, M.B., Prasad, S.D., Tina, K.V.D., LaRossa, R.A., 1996. A miniature bioreactor for sensing toxicity using recombinant bioluminescent Escherichia coli cells. Biotechnol. Prog. 12, 393–397. Jerome, R.P., Ovadia, L., Rachel, R., Shirmshon, B., 2001. Encapsulation of luminous recombinant E.coli in sol-gel silicate films. Adv. Mater. 13 (23), 1773–1775. Kim, B.C., Gu, M.B., 2003. A bioluminescent sensor for high throughput toxicity classification. Biosens. Bioelectron. 18, 1015–1021. Kim, S.W, Choi, S.H, Min, J.H., Gu, M.B, 2000. Toxicity monitoring of endocrine disrupting chemicals (EDCs) using freeze-dried recombinant bioluminescent bacteria. Biotech. Bioproc. Eng. 5, 395–406. Koh, Y.S., Roe, J.H., 1995. Isolation of a novel paraquat-inducible (pqi) gene regulated by the soxRS locus in Escherichia coli. J. Bacteriol. 177, 2673–2678. Lee, H.J., Gu, M.B., 2003. Construction of a sodA::luxCDABE fusion Escherichia coli: comparison with a katG fusion strain through their responses to oxidative stresses. Appl. Microbiol. Biotechnol. 60, 577–580. Lee, H.Y., Choi, S.H., Gu, M.B., 2003. Response of bioluminescent bacteria to sixteen azo dyes. Biotech. Bioproc. Eng. 8 (2), 101–105. Li, V.S., Choi, D., Tang, M., Kohn, H., 1995. Structural requirements for mitomycin C DNA bonding. Biochemistry 34, 7120–7126.
J.H. Lee, M.B. Gu / Biosensors and Bioelectronics 20 (2005) 1744–1749 Millard, J.T., Spencer, R.J., Hopkins, P.B., 1998. Effect of nucleosome structure on DNA interstrand cross-linking reactions. Biochemistry 37, 5211–5219. Min, J., Pham, C.H., Gu, M.B., 2003. Specific responses of bacterial cells to dioxins. Environ. Toxicol. Chem. 22, 233–238. Min, J.H., Kim, E.J., LaRossa, R.A., Gu, M.B., 1999. Distinct responses of a recA::luxCDABE Escherichia coli strain to direct and indirect DNA damaging agents. Mutat. Res. 442, 61–68. Mitchell, R.J., Gu, M.B., 2004. An Escherichia coli biosensor capable of detecting both genotoxic and oxidative damage. Appl. Microbiol. Biotechnol. 64 (1), 46–52. Mitchell, R.J., Ahn, J.M., Gu, M.B., 2004. The use of X. luminescens and V. fischeri lux fusions to study gene expression patterns: a comparative study. J. Microbiol. Biotechnol. In press.
1749
Se, K.K, Baek, S.L., Jeong, G.L, Hyung, J.S., Eun, K.K., 2003. Continuous water toxicity monitoring using immobilized Photobacterium phosphoreum. Biotech. Bioproc. Eng. 8, 147–150. Uck, H.C., Nina, S., Yaping, C., Margaret, L.B., 1996. Continuous pollution monitoring using Photobacterium phosphoreum. Resour. Conserv. Recycl. 18, 25–40. Van Dyk, T.K., Majarian, W.R., Konstantinov, K.B., Young, R.M., Dhurijati, P.S., LaRossa, R.A., 1994. Rapid and sensitive pollutant detection of heat shock gene-bioluminescence gene fusion. Appl. Environ. Microbiol. 60, 1414–1420. Vollmer, A.C., Belkin, S., Smulski, D.R., Van Dyk, T.K., LaRossa, R.A., 1997. Detection of DNA damage by use of Escherichia coli carrying recA::lux, uvrA::lux, or alkA::lux reporter plasmids. Appl. Environ. Microbial., 2566–2571.