A Recombinant Escherichia coli Sensor Strain for the Detection of

from bacterial luciferase operon of Photorhabdus lumi- nescens inserted under the control of tetracycline- responsive elements of the transposon Tn10,...
0 downloads 0 Views 94KB Size
Download PDF
AC Research

Anal. Chem. 1998, 70, 4457-4462

Accelerated Articles

A Recombinant Escherichia coli Sensor Strain for the Detection of Tetracyclines Matti T. Korpela, Jussi S. Kurittu, Jarkko T. Karvinen, and Matti T. Karp*

Department of Biotechnology, University of Turku, Tykisto¨ katu 6, 6th Floor, FIN-20520 Turku, Finland

A bioluminescent Escherichia coli K-12 strain for the specific detection of the tetracycline group of antibiotics is described. A sensor plasmid, containing five genes from bacterial luciferase operon of Photorhabdus luminescens inserted under the control of tetracyclineresponsive elements of the transposon Tn10, was constructed. Usage of the full-length luciferase operon in the sensor resulted in tetracycline-dependent light production without additions, i.e., self-luminescent phenotype, since all the substrates were intrinsically produced by the recombinant organism. The time needed for optimal induction of light emission was 90 min. Maximal induction of ∼100-fold over uninduced levels by using 20 ng of tetracycline, and picomole sensitivities for the seven different tetracyclines tested, were obtained without added Mg2+ ions. The higher the pH and the magnesium ion concentration in the assay medium the higher was the amount of membrane-impermeable tetracycline-Mg2+ chelate complex. In consequence, by adjusting the pH and the Mg2+ ion concentration, the sensitivity of the assay can be modified for different analytical purposes. Different non-tetracycline antibiotics did not cause induction of light emission.

Over the last two decades, there has been an ever-increasing usage and availability of various antimicrobial compounds for therapeutic veterinary purposes to maintain the health of farm animals and to enhance the productivity of the farming industry. Among these are the tetracycline antibiotics. The first naturally occurring antibiotics chlortetracycline and oxytetracycline were discovered in the late 1940s. Today, nearly 1000 tetracycline derivatives exist, but only 7 have been in extensive clinical and/ or veterinary use.1 * Corresponding author: (tel) +358-2-3338085; (fax) +358-2-3338050; (e-mail) [email protected]. 10.1021/ac980740e CCC: $15.00 Published on Web 10/03/1998

© 1998 American Chemical Society

The tetracyclines have all a four-ring carbocyclic structure as a basic skeleton, and they differ from each other by substituent variations at carbons 5, 6, and 7.2 They prevent bacterial growth by inhibiting protein synthesis. They bind to the bacterial 30S ribosomal subunit and prevent the attachment of aminoacyl-tRNA to the ribosomal receptor site in a reversible fashion. Tetracyclines are bacteriostatic agents which inhibit rather than kill bacteria, and they exhibit activity against a wide range of Gramnegative and Gram-positive bacteria. However, tetracycline resistance is widespread in Gram-negative and Gram-positive bacterial species. The resistance mechanism genes are situated in extrachromosomal elements, which often occur in multiple copies. These plasmids can be efficiently transferred from one strain to another, which has ensured the wide spreading of resistance over microbial communities. To decrease the spread of resistance to new strains, unnecessary usage of tetracyclines should be minimized. One way to decrease consumption is the use of monitoring methods for tetracycline residues in various samples. Conventional screening tests for antimicrobial agents are divided into microbial inhibition assays, enzymatic tests, and immunological tests. The growth inhibition-based assays are nonselective and time-consuming whereas the last two are characterized by their high selectivity and good sensitivity as well as a short test time. Certain enzymatic and immunological assays are also used as confirmatory tests, such as HPLC methods, for a high price but with possibilities for the automation of several tests. A novel approach to monitoring drug residue levels is to use a microbial whole cell sensor which is constructed by recombinantDNA technologies. In such strain, a plasmid carries a strictly regulated promoter connected to a sensitive reporter gene. Typical of regulatory systems is that the gene expression is turned on (i.e., induced) when necessary by using efficient inbuilt (1) Col, N. F.; O’Connor, R. W. Rev. Infect. Dis. 1987, 9 (Suppl. 3), S232S243. (2) Chopra, I.; Hawkey, P. M.; Hinton, M. J. Antimicrob. Chemother. 1992, 29, 245-277.

Analytical Chemistry, Vol. 70, No. 21, November 1, 1998 4457

regulatory components. Luciferase genes are widely used reporter genes in prokaryotic as well as in eukaryotic systems, since they provide a sensitive and simple detection of gene expression and regulation.3 The quantification of light emission, i.e., bioluminescence, is one of the most sensitive means of detection and it can be measured with a liquid scintillation counter, a luminometer, or even with an X-ray film. The most commonly used reporter genes are the firefly luciferase (i.e., luc gene of Photinus pyralis) and the bacterial luciferase, lux genes of Photorhabdus luminescens and Vibrio fischeri. The bacterial luciferases (reviewed in ref 4) catalyze a reaction that involves the oxidation of a long-chain fatty aldehyde and FMNH2, whereas the firefly luciferase catalyses the oxidation of the substrate D-luciferin in the presence of ATP.5 The common features of different luminescence systems are the requirement of oxygen and the ability to emit visible light of different wavelengths. We and others have previously constructed specific detectors of heavy metals by incorporating a heavy metalresponsive element to control firefly or bacterial luciferase gene expression. These sensors were capable of detecting femtomolar concentrations of mercury ions6 or arsenic/antimonite ions7 and nanomolar concentrations of lead or cadmium ions.8 In this report, we present a construction of genetically engineered luminescent bacterial strain for the measurement of the tetracycline group of antibiotics. The sensor plasmid contains bacterial luciferase genes from P. luminescens and the regulation unit of tetracycline resistance factor from transposon Tn10 9 to control the expression of the lux genes. The regulation unit consists of the tetA promoter and the repressor protein TetR, which efficiently regulate the expression of the full-length luciferase operon from P. luminescens. In the absence of tetracyclines, the expression of the lux genes is repressed. The lux genes are induced by various tetracyclines, which inactivate the repressor encoded by the tetR gene. This assay system is tetracycline specific and antimicrobial agents other than tetracyclines do not cause the induction of tetA.

EXPERIMENTAL SECTION Materials. Tryptone, yeast extract, and agar were from Difco. The tetracyclines and other antibiotics used were from Sigma. All the molecular biology enzymes were obtained from Pharmacia (Uppsala, Sweden) or from New England Biolabs (Beverly, MA). Construction of the Sensor Plasmid pTetLux1. The plasmid pTetLux1 is based on a vector pASK75, which is constructed for the purposes of synthesis of foreign proteins in Escherichia coli under transcriptional control of a tetA promoter/ operator.10 The luciferase operon containing luxC,D,A,B,E genes of P. luminescens (without the regulatory genes) was transferred from plasmid pCGLS-11 11 as an EcoRI fragment into a unique (3) Pazzagli, M.; Devine, J. H.; Peterson, D. O.; Baldwin, T. O. Anal. Biochem 1992, 204, 315-323. (4) Meighen, E. A.; Dunlap, P. V. Adv. Microb. Physiol. 1993, 34, 1-67. (5) McElroy, W. D.; DeLuca, M. Chemi- and bioluminescence; Burr, J. G., Ed.; Marcel Dekker Inc.: New York, 1985. pp 387-399. (6) Virta, M.; Lampinen, J.; Karp, M. Anal. Chem. 1995, 67, 667-669. (7) Ramanathan, S.; Shi, W.; Rosen, B. P.; Daunert, S Anal. Chem. 1997, 69, 3380-3384. (8) Tauriainen, S.; Karp, M.; Chang, W.; Virta M. Biosens. Bioelectron. 1998, 13, 931-938. (9) Hillen, W.; Berens, C. Annu Rev. Microbiol. 1994, 48, 345-369. (10) Skerra, A. Gene 1994, 151, 131-135.

4458 Analytical Chemistry, Vol. 70, No. 21, November 1, 1998

EcoRI site of pASK75. The 6.7-kb EcoRI fragment of pCGLS-11 was separated on a LGT-agarose gel and purified by using a Qiagen gel extraction kit (Santa Clarita, CA). This fragment was ligated by using T4-DNA ligase at 16 °C overnight to calf intestinal phosphatase-treated, EcoRI-digested vector pASK75, which was similarly purified as the fragment above. The resulting ligation mixture was transformed into E. coli K-12 strain M7212 (SmRlacZam, bio-uvr, ∆trpE42(λNam7-Nam53cI857∆HI), by electroporation.13 Twenty colonies were picked at random and cultivated in 1 mL of (0.5 µg/mL tetracycline for 4 h in L-broth after which they were measured for their bioluminescence using a LKB-Wallac (Turku, Finland) manual tube luminometer No. 1250 connected to a chart recorder (LKB-Bromma, Stockholm, Sweden). Two inducible, light-producing colonies were used for plasmid isolation in large scale. The correct construction of the plasmid was verified by proper restriction enzyme digestions. Cultivation of Bacteria. The sensor bacteria E. coli K-12/ pTetLux1 were cultivated in L-broth (10 g of tryptone, 5 g of yeast extract, and 5 g of NaCl per liter, pH 7.0) supplemented with 100 µg/mL ampicillin in a shaker at 30 °C overnight. An inoculum from the overnight culture was cultivated to OD600 nm of 1.5 at 37 °C after which the cells were diluted 1/100 into either L-broth or M9-medium.14 Luminescence Measurements. Bacterial dilution was pipetted into wells of a white microtitration plate (Labsystems Oy, Helsinki, Finland) in a volume of 200 µL over 10 µL of solution containing either L-broth (blank) or L-broth with the antibiotic dilutions. The plate was incubated for different time periods (optimally 90 min), and the in vivo bioluminescence was measured by using Luminoskan luminometer (Labsystems Oy) and an integral mode of measurement of 1 s. RESULTS The plasmid pTetLux1 shown in Figure 1 was constructed by transferring five lux genes from the luciferase operon of P. luminescens11 as an EcoRI fragment to the vector pASK7510 under the control of the tetA promoter from transposon Tn10. This vector was originally dedicated for the controlled expression of recombinant Fab fragments and it is unique for extremely low background expression. The tight repression is due to the expression of tetR gene introduced elsewhere in the parent plasmid pASK75 uncoupled from the tet control region. E. coli cells containing pTetLux1 are able to produce light “without additions”; i.e., the artificial lux operon contains all the components necessary for in vivo light production. The luxA and luxB genes code for the heterodimeric luciferase enzyme, and the luxC, luxD, and luxE code for the enzymes that synthesize the substrates needed for the generation of light emission by the luciferase. The E. coli K-12 cells containing the vector pTetLux1 were cultivated in L-broth in shake flasks in order to determine the optimal growth phase for induction of the artificial luciferase (11) Frackman, S.; Anhalt, M.; Nealson, K. H. J. Bacteriol. 1990, 172, 57675773. (12) Bernard, H.-U.; Remaut, E.; Hersfield, M. V.; Das, H. K.; Helinski, D. R.; Yanofsky, C.; Franklin, N. Gene 1979, 5, 59-76. (13) Dower, W. J.; Miller, J. F.; Ragsdale. C. W. Nucleic Acids Res. 1988, 16, 6126-6144. (14) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning: A Laboratory Manual, 2nd ed.; Cold Spring Harbor Laboratory: Cold Spring Harbor, NY, 1989.

Figure 1. Sensor plasmid construct pTetLux1 and the mode of action of the sensor on molecular level. The plasmid shown in (A) was constructed as described in the methods section. Schematical representation in (B) shows that in the repressed state the tetR protein produced constitutively by the tetR gene of the plasmid shuts down the expression of the luciferase operon consisting of five genes luxC, luxD, luxA, luxB, and luxE. If tetracycline molecules are present (1) they bind to tet repressor protein, causing a conformational change and thereby dissociation from the tetA promoter/operator region (2). This causes an induction in luciferase operon and luxAB reporter gene synthesis (3) which in turn causes the sensor cells to produce light. Other abbreviations: ptetA, for TN10 TcR promoter/operator region; f1-IG, intergenic region of filamentous phage f1; ORI, origin of DNA replication (ColE1); β-la, gene coding for the plasmid selection pressure marker β-lactamase.

operon. It was found that the optimal induction with tetracycline was achieved by using cells from the late logarithmic phase, and in subsequent cultivations, we always used cells from OD600 nm of 1.5 (data not shown). The induction of luciferase production is time dependent. The time course with use of different amounts of tetracycline as an inducer was investigated by detecting in vivo bioluminescence. A 90-min induction period was chosen for all subsequent work since it yielded the highest light emission especially in the lower concentration range of various tetracyclines (data not shown). The bacterial luciferase enzyme is conventionally considered rather temperature labile, which has limited the wide-spread usage of this sensitive reporter of gene expression. In this study, we used the luciferase from a terrestrial bacterium P. luminescens, which has a higher temperature stability profile than the luciferases from the marine bacterial organisms, such as V. fischeri and Vibrio harveyi.15 The optimum temperature with respect to induction with tetracycline of the sensor bacteria described in this study was 42 °C (data not shown). It is a well-known fact that the antimicrobial efficiency of tetracyclines depends on both the pH and the magnesium ion concentration. Yamaguchi et al.16 reported a ∆pH-dependent accumulation of tetracycline in E. coli. They showed that tetracycline accumulation was stimulated by decreasing the pH of the medium and inhibited by the addition of magnesium ions. The higher the pH and the magnesium ion concentration, the higher was the concentration of the membrane-impermeable chelate complex. Therefore, tetracycline accumulates in a compartment that has a higher pH and a higher magnesium ion concentration as shown schematically in Figure 2. We studied the effect of Mg2+ and pH on the ability of the sensor cells to respond to variations in tetracycline concentrations (15) Colepicolo, P.; Cho, K.-W.; Poinar, G. O.; Hastings, J. W. Appl. Environ. Microbiol. 1989, 55, 2601-2606. (16) Yamaguchi, A.; Ohmori, H.; Kaneko-Ohdera, M.; Nomura, T.; Sawai, T. Antimicrob. Agents Chemother. 1991, 35, 53-56.

Figure 2. Schematic mechanism for tetracycline accumulation in bacterial cells (adapted from Yamaguchi et al.16. Both in the cytoplasm and in the medium, tetracycline exists in equilibrium between a neutral form (TH2) and a magnesium ion-tetracycline chelate complex (THMg+) in the presence of magnesium ions. Only a neutral tetracycline can diffuse through the phospholipid bilayer membrane. The equilibrium between a neutral tetracycline and a chelate complex depends on both the pH and the magnesium ion concentration.

by the induction of bioluminescence. Figure 3 shows that an increase in the amount of Mg2+ ions in the medium shifts of the standard curve of tetracycline to the right. Light emission increased until the concentration of tetracycline reached the level where induction and inhibition are in balance (i.e., a peak is seen in the curve). Higher concentrations of tetracycline have inhibitory effects on sensor cells which can be seen as a decrease in the signal emitted. When no added Mg2+ is present, the peak in light emission occurs at 20 ng of tetracycline and is at ∼80 ng when there is 10 mM Mg2+ in the medium. The effect of pH change from pH 6.0 to 8.0 alone was significantly lower on the induction capacity of luciferase production than the combination of high pH and high concentration of magnesium ions (see Figure 4). As a consequence, it is possible Analytical Chemistry, Vol. 70, No. 21, November 1, 1998

4459

Figure 3. Effect of the presence of Mg2+ ions on the doseresponse reaction of the tetracycline sensor strain. The sensor strain E. coli K-12/pTetLux was incubated in the presence of varying concentrations of MgCl2 in L-broth buffered with MES, pH 6.0, in the wells of a white microtitration plate. After 90 min, the light emission was measured using a microtiter luminometer. Symbols used: (9) blank experiment without added Mg2+, (b) 0.625 mM Mg2+ in the well, (2) 2.5 mM Mg2+, and (1) 10 mM Mg2+ in the well. Each measurement point represents the mean of three parallel measurements. The coefficient of variation is smaller than the size of the symbol in each measurement point, i.e., between 1 and 2%.

to change the “assay window” from a low to a higher concentration range of tetracyclines. The capacity of different non-tetracycline antibiotics, rifampicin, kanamycin, nalidixic acid, chloramphenicol, streptomycin, and erythromycin, to induce the luciferase production in the E. coli K-12/pTetLux1 cells was also studied. The non-tetracycline antibiotics did not induce bioluminescence even at high concentrations over a wide range (data not shown). The dose-response reactions for seven different tetracyclines (chlortetracycline, demethylchlortetracycline, doxycycline, methacycline, minocycline, oxytetracycline, tetracycline hydrochloride) are presented in Table 1. It is surprising that all tetracyclines tested in this study behave in a very similar manner and that the induction capacity of various tetracyclines is roughly in the same concentration area. This makes the sensor even more attractive for analytical use of the assay of the tetracycline family of antibiotics. DISCUSSION We have described the construction of a new type of bacterial sensor strain for an extremely fast and sensitive detection of tetracycline antibiotics. The sensor is based on genetically engineered bacteria which start to emit visible blue light in the presence of tetracyclines. The sensor is self-luminescent; i.e., light is emitted without additions of substrates or cofactors due to the expression of full-length bacterial luciferase operon inserted under the control of tetracycline-responsive repressor tetR and the tetA promoter.17 The bacterial luciferase we used in this study was obtained from a terrestrial bacterium P. luminescens. This (17) Chopra, I.; Hacker, K.; Misulovin, Z.; Rothstein, D. M. Antimicrob. Agents Chemother. 1990, 34, 111-116.

4460 Analytical Chemistry, Vol. 70, No. 21, November 1, 1998

luciferase is stable with regard to the temperature.15 The choice fits well to the host bacteria of the sensor plasmid, E. coli, whose metabolism is optimal at 37 °C. This temperature is too high for most of the luciferases of bacterial origin. It was noticed that at 30 °C the bioluminescence signals were rather low presumably due to the lower rate of metabolism compared to 37 °C and that at 42 °C the cells might be stressed by high temperature. Therefore, 37 °C was found optimal with regard to high enough levels of light emission and sensitive detection of all tetracyclines tested. Our sensor strain was found to luminesce effectively in the presence of tetracyclines. The signal over background noninduced state (induction coefficient) was over 100 in most cases. All of the seven clinically most relevant tetracyclines were found to induce the light emission effectively in the sensor cells. At their best the detection limits were ∼2 ng/sample; i.e., picomole levels of tetracyclines were detected. The sensor strain is very fast, giving a measurable signal in few minutes at best. There is no need for the cultivation of microorganisms in order to get a measurable signal, which is the case in conventional assay methods for various antimicrobial agents. The conventional microbial methods rely on plate counting where either the amount of cells or the size of the inhibitory zone on a cultivation plate around the disk containing the sample is estimated. Also, various chromogenic dyes changing their color in cultivation medium are used together with spore-forming bacteria such as Bacillus stearothermophilus (for instance, the Delvo Test, supplied by Gist Brocades). The conventional microbial methods take ∼3 h or more to be completed and their sensitivities are 0.2 µg/mL or more, thereby being far less sensitive than the method described here. Furthermore, conventional microbial methods detect all antimicrobial compounds the test bacterium is sensitive to and hence the group-specific method described here provides a clear advantage. In addition to microbial methods, HPLC, MS, or their combination, TLC and various immunometric assays are currently used methods for analysis of tetracycline residues of different food matrixes.18 Microbial methods are simple, and they detect all antimicrobial compounds the used test bacterium is sensitive to, whereas immunometric assays are used to detect certain antibiotics or antibiotic groups. Both of these tests are widely used in application areas, which require simplicity and analysis of large amounts of samples. Unlike microbiological and immunometric methods, HPLC and MS necessitate the usage of elaborative extraction protocols from various matrixes. However, due to their high sensitivity, they are used to confirm the results of screening and group-specific assays and to identify antibiotics. The sensitivity of the method described here is very comparable to the sensitivities of microbial methods and group-specific immunometric assays as well as confirmatory methods, such as HPLC and MS. Even though tetracyclines are inhibitors of protein synthesis, we are able to measure their concentration by means of the system presented in this communication. The affinity of tetracyclines is ∼1000-fold higher to the repressor than to ribosomes, which are the actual targets of the antimicrobial action of tetracyclines.19 Therefore, small amounts of tetracyclines are able to activate luciferase synthesis visualized as light production before the (18) Barker, S. A.; Walker, C. C. J. Chromatogr. 1992, 624, 195-209. (19) Hinrichs, W.; Kisker, C.; Du ¨ vel, M.; Mu ¨ ller, A.; Tovar, K.; Hillen, W.; Saenger, W. Science 1994 264, 418-420.

Table 1. Response of the Tetracycline Reporter Strain to Different Types of Tetracyclinesa relative induction coefficients ng

tetracycline

oxytetracycline

chlortetracycline

methacycline

doxycycline

demeclocycline

minocycline

1.25 5 20 80

3.9 21.5 100.0 27.4

6.1 15.3 100.0 27.5

3.3 15.7 100.0 16.1

5.5 36.2 100.0 34.1

2.6 15.9 100.0 35.6

4.0 44.1 100.0 35.4

1.8 11.3 100.0 48.6

a The experiment was done as in Figure 3 in L-broth buffered to pH 6.0 with 50 mM MES. The light emission value at 20 ng/microtiter plate well is taken as 100% for each tetracycline analogue tested. Induction coefficient means light emission in relative light units obtained at various concentrations of tetracyclines over the background, noninduced levels.

Figure 4. Combined effect of pH and Mg2+ ions on the doseresponse reaction of the tetracycline sensor strain. The experiment was done essentially as described in Figure 3 except that L-broth was buffered either with 50 mM MES at pH 6.0 [(A) (9) blank without added Mg2+ and (b) 10 mM Mg2+] or with 50 mM TRIS-Cl at pH 8.0 [(B) (9) blank without added Mg2+ and (b) 10 mM Mg2+]. The figure shows induction coefficient, i.e., measured signal in relative light units over the signal of uninduced background light emission as a function of various tetracycline concentrations (in ng).

concentration of the drug becomes too high and the inhibition of protein synthesis begins. At sufficiently high concentrations, there does appear to exist a decrease in light emission, which is an indication of serious toxicity of tetracycline to the sensor cells protein synthesis machinery. This kind of low signal samples (i.e., wrong negative samples) can be avoided by dilution of the samples or by careful adjustment of the reaction conditions. The transport of tetracyclines to the cells is carried out through both passive and energy-requiring mechanisms.16,20,21 The passive transport is dependent on the pH and the external concentration of Mg2+ ions since tetracyclines diffuse through membranes in neutral forms and high external concentration of Mg2+ prevents the (20) Argast, M.; Beck, C. F. Antimicrob. Agents Chemother. 1984, 26, 263265. (21) Smith, M. C. M.; Chopra, I. Antimicrob. Agents Chemother. 1984, 25, 446449.

transport. Indeed, we have been able to make the detection less sensitive and hence move the occurrence of the “hook effect” to higher concentrations of the tetracycline tested by simple adjustment of the external Mg2+ ion concentration and further by moving the pH to the alkaline region as shown in Figures 3 and 4, respectively. For instance, in cow milk the concentrations of divalent cations Ca2+ and Mg2+ are rather constant from one milk sample to another at about 30 and 6 mM, respectively. The E. coli sensor strain was found to be specific for the tetracycline family of antimicrobial drugs only. The other inhibitors of protein synthesis and transcription or compounds affecting DNA synthesis were not inducers of the light emission. The different tetracyclines tested (Table 1) were able to activate light production with approximately equal efficiency, even though they have structural differences in the substituents of the carbon backbone. Furthermore, the affinities of the tetracycline-Mg2+ complex inactivating the repressor have rather large differences between different tetracyclines, in vitro.22 However, these facts do not necessarily have marked effects on this kind of whole cell sensor in vivo, since results similar to ours have been reported previously with a reporter construction utilizing the less sensitive β-galactosidase gene under the control of tetA promoter.17 However, Table 1 shows that the induction coefficient of minocycline was slightly lower at low amounts (1.25 and 5.0 ng) and higher at high amounts (80 ng) of antibiotic than the coefficients of other tetracyclines. The reason for this must be the poorer penetration of minocycline into the sensor cells, but the effect was expected to be higher according to the hydrophobicities and MIC values of tetracyclines.23 The sensor plasmid described in this article can be transferred not only to different E. coli strains but also to various other Gramnegative strains such as Salmonella. This broadens the applicability of the approach to the research and development of new antibiotics against pathogenic bacteria with a tetracycline backbone by using, for instance, combinatorial chemistry approaches. The use of the sensor allows the use at high-throughput screening on 96- or 384-well microtitration plates and automated measuring instruments. We have found that this kind of sensor cell can be freeze-dried without any loss of sensitivity or overall performance,24 which further simplifies the applicability of the assay system to, for instance, the search for new tetracycline lead compounds. (22) Lederer, T.; Kintrup, M.; Takahashi, M.; Sum, P.-E.; Ellestad, G. A.; Hillen, W. Biochemistry 1996, 35, 7439-7446. (23) Leive, L.; Telesetsky, S.; Coleman, W. G., Jr.; Carr, D. Antimicrob. Agents Chemother. 1984, 25, 539-544. (24) Tauriainen, S.; Karp, M.; Chang, W.; Virta, M. Appl. Environ. Microbiol. 1997, 63, 4456-4461.

Analytical Chemistry, Vol. 70, No. 21, November 1, 1998

4461

Another large area of applications of the sensor approach is sensitive and fast detection of tetracyclines from various biological materials such as different food matrixes or serum. The mastitis of cows is often cured by tetracyclines, which have a specific clearing period before the milk can be used for cheese manufacturing. Small batches of contaminated milk can spoil large volumes collected from several breeders, and a fast answer would be beneficial to prevent such losses preferably already during the transportation from a farm to a dairy.

4462 Analytical Chemistry, Vol. 70, No. 21, November 1, 1998

ACKNOWLEDGMENT Dr. Kenneth Nealson and Dr. Arne Skerra are gratefully thanked for their kind gifts of the plasmids pCGLS-11 and pASK75, respectively. Received for review July 8, 1998. Accepted September 8, 1998. AC980740E