Anal. Chem. 2003, 75, 580-585
Combined Phage Typing and Amperometric Detection of Released Enzymatic Activity for the Specific Identification and Quantification of Bacteria T. Neufeld, A. Schwartz-Mittelmann, D. Biran, E. Z. Ron, and J. Rishpon*
Department of Molecular Microbiology and Biotechnology, Tel-Aviv University, Ramat-Aviv 69978, Israel
Here, we describe a novel electrochemical method for the rapid identification and quantification of pathogenic and polluting bacteria. The design incorporates a bacteriophage, a virus that recognizes, infects, and lyses only one bacterial species among mixed populations, thereby releasing intracellular enzymes that can be monitored by the amperometic measurement of enzymatic activity. As a model system, we used virulent phage typing and cellmarker enzyme activity (β-D-galactosidase), a combination that is specific for the bacterial strain Escherichia coli (K-12, MG1655). Filtration and preincubation before infecting the bacteria with the phage enabled amperometric detection at a wide range of concentrations, reaching as low as 1 colony-forming unit/100 mL within 6-8 h. In principle, this electrochemical method can be applied to any type of bacterium-phage combination by measuring the enzymatic marker released by the lytic cycle of a specific phage. A bacteriophage (phage) is an obligate intracellular parasite of bacteria that multiplies by using some or all of the host biosynthetic machinery (i.e., a virus that infects and lyses certain bacteria). The first step in the infection process is the specific adsorption of the phage to the bacterial cell, in which the phages attach to specific receptors on the bacterial cell surface. The host specificity of the phage (i.e., the range of bacteria that it can infect) is usually determined by the outer proteins of the bacteria. The bacterial receptors on the outer membrane include cell-surface proteins, lipopolysaccharides (LPS), pili, and lipoprotein. For example, bacteriophage lambda (λ) adsorbs specifically to Escherichia coli receptors, which are encoded by the bacterial lamB gene. Following the initial step of specific attachment, lytic or virulent phages multiply in the bacteria and are usually released by lysis at the end of the life cycle. The number of phage particles released per infected bacterial cell can be as high as 1000.1 Bacteriophages have been used in such areas as epidemiological phage-typing,2,3 phage therapy, overcoming bacterial antibiotic * Corresponding author. Phone: +972-3-6409366. Fax: +972-3-640-9407. E-mail:
[email protected]. (1) Brock, T. D., Madigan, M. T. Biology of Microorganisms, 6th ed.; Prentice Hall: Engelwood Cliffs, NJ, 1991. (2) Sechter, I.; Mestre, F.; Hansen, D. S. Clinical Microbiol. Infect. 2000, 6, 233-238.
580 Analytical Chemistry, Vol. 75, No. 3, February 1, 2003
resistance,4,5 and constructing single-chain antibodies.6-8 The specific detection of bacteria by phage typing is based on the principle that bacteriophages infect bacteria in a very selective manner,9,10 thereby eliminating the need for time-consuming conventional microbiological methods based on cultivation. The usual incubation period (24-48 h) of bacterial cultures is too long when immediate remedial or treatment measures must be taken.11 The phage-mediated detection also eliminates the need for selective media, because the selectivity of bacteria by phages relies on the natural, specific attachment of the phage to its specific receptor. The combination of phage-specific identification and the related release of an intrinsic enzymatic cell marker following lysis of the cell provides a powerful means for the highly specific detection of a given bacterial strain. Blasco et al.12 and Wu et al.13 used the phage-mediated release of the enzyme adenylate kinase (AK) as a cell marker for E. coli and Salmonella. The enzymatic conversion of ADP to ATP is followed by the appearance of bioluminescence using the external firefly luciferase. A method reported by Goodridge al.14,15 involves a fluorescent bacteriophage assay for detecting E. coli O157:H7 in ground beef and raw milk. This method, based on staining the bacteriophages with a fluorescent dye, enables the detection of 100 colony forming units (cfu)/mL (3) Capita, R.; Alonso-Calleja, C.; Mereghetti, L.; Moreno, B.; Garcia-Fernandez, M. D. J. Appl. Microbiol. 2002, 92, 90-96. (4) Summers, W. C. Ann. Rev. Microbiol. 2001, 55, 437-451. (5) Sulakvelidze, A.; Alavidze, Z.; Morris, J. G. Antimicrob. Agents Chemother. 2001, 45, 649-659. (6) Hoogenboom, H. R.; de Bruine, A. P.; Hufton, S. E.; Hoet, R. M.; Arends, J. W.; Roovers, R. C. Immunotechnology 1998, 4, 1-20. (7) Huston, J.; George, A. Hum. Antibodies 2001, 10, 127-142. (8) Mertens, P.; Walgraffe, D.; Laurent, T.; Deschrevel, N.; Letesson, J.; De Bolle, X. Int. Rev. Immunol. 2001, 20, 181-199. (9) McNerney, R. Int. J. Tuberculosis Lung Dis. 1999, 3, 179-184. (10) Ronner, A. B.; Cliver, D. O. J. Food Prot. 1990, 53, 944-947. (11) Clesceri, L.; Greenberg, A.; Eaton, A. Standard Methods for the Examination of Water and Wastewater, 20th ed.; American Public Health Association, American Water Works Association, Water Environment Federation: Washington, DC, 1998. (12) Blasco, R.; Murphy, M. J.; Sanders, M. F.; Squirrell, D. J. J. Appl. Microbiol. 1998, 84, 661-666. (13) Wu, Y.; Brovko, L.; Griffiths, M. W. Lett. Appl. Microbiol. 2001, 33, 311315. (14) Goodridge, L.; Chen, J. R.; Griffiths, M. Int. J. Food Microbiol. 1999, 47, 43-50. (15) Goodridge, L.; Chen, J. R.; Griffiths, M. Appl. Environ. Microbiol. 1999, 65, 1397-1404. 10.1021/ac026083e CCC: $25.00
© 2003 American Chemical Society Published on Web 12/20/2002
of target bacteria in artificially contaminated milk after a 10-h enrichment step. Regarding other types of bacteria, the luciferase reporter phages (recombinant phages that contain the gene for the firefly luciferase enzyme) have been used for the detection and identification of Listeria monocytogenes16 and Mycobacterium tuberculosis.17 Thus, it is feasible that a merger between the measurable enzymatic activity released by the specific phagebacteria lytic interaction and the high sensitivity and relative simplicity offered by electrochemical techniques can provide the rapidity, ease of operation, and accuracy that are required for food and water quality control. The enzymes β-D-galactosidase and β-D-glucuronidase, respectively, are widely used for identifying total coliforms and E. coli in water and food samples. Many optical and flouregenic substrates exist for the specific detection of the activity of these enzymes, and various commercial tests are available.18 The enzymes are also used in the membrane-filter technique, in multiple-tube fermentation, and in the defined substrate technology known as colilert-18 (IDEXX, Atlanta. Ga.), widely used methods for detecting coliforms in water.19 Nelis and Van Poucke20 proposed the enzymatic detection of coliforms and E. coli within 4 h using a novel method for detecting a single cell with the aid of a laser ChemScan scanner. Because of expensive handling, however, such a method would be particularly useful only for the emergency monitoring of drinking water.21 Synthesis of the enzyme β-D-galactosidase is induced de nouveau in wild-type E. coli strains in response to the presence of lactose, the enzyme’s natural substrate, or β-D-galactopyranosides. The substrate specificity of the enzyme is such that any substrate having a β-D-galactopyranoside moiety can be hydrolyzed. Because of its high specificity and stability, β-galactosidase is used as a secondary detection reagent in enzyme-linked immunoassays. In a previous report, we described a sensitive electrochemical assay for the detection of β-galactosidase activity.22 In the present work, we have combined this method with bacterial detection in a model system involving virulent phage-typing (λ vir) and the release of intracellular enzyme activity (β-D-galactosidase) as a highly specific marker for the bacterial strain E. coli K-12, MG1655. The λ vir serves not only as a specific recognition element for E. coli but also as the releasing agent for β-Dgalactosidase activity, the product of which is measured amperometrically by monitoring its oxidation at the carbon anode. In principle, this electrochemical method can be applied to any type of bacterium-phage combination by measuring the enzymatic marker released by the lytic cycle of specific phage. EXPERIMENTAL SECTION Chemicals and Microorganisms. p-Aminophenyl-β-D-galactopyranoside (β-PAPG), isopropyl β-D-thio-galactopyranoside (IPTG), (16) Loessner, M. J.; Rudolf, M.; Scherer, S. Appl. Environ. Microbiol. 1997, 63, 2961-2965. (17) Banaiee, N.; Bodadilla-del-Valle, M.; Bardarov, S.; Riska, P. F.; Small, P. M.; Ponce-De-Leon, A.; Jacobs, W. R.; Hatfull, G. F.; Sifuentes-Osornio, J. J. Clin. Microbiol. 2001, 39, 3883-3888. (18) Geissler, K.; Manafi, M.; Amoros, I.; Alonso, J. L. J. Appl. Microbiol. 2000, 88, 280-285. (19) Rompre, A.; Servais, P.; Baudart, J.; de-Roubin, M. R.; Laurent, P. J. Microbiol. Methods 2002, 49, 31-54. (20) Nelis, H.; Van Poucke, S. Water, Air, Soil Pollut. 2000, 123, 43-52. (21) Van Poucke, S. O.; Nelis, H. J. J. Microbiol. Methods 2000, 42, 233-244. (22) Schwartz-Mittelmann, A.; Ron, E. Z.; Rishpon, J. Anal. Chem. 2002, 74, 903-907.
and maltose were purchased from Sigma Chemicals (Israel). Bacterial media, Luria Bertani broth (LB), and agar were purchased from Difco, Becton Dickinson (U.S.A.). E. coli (MG1655), Klebsiella pneumoniae, and phage λ vir were obtained from our laboratory stocks. Screen print disposable electrodes were purchased from Gwent Electronics (England). Titration of Phage λ vir. Aliquots (100 µL each) of an overnight bacterial culture were infected with 100 µL of the phage at increasing concentrations (105-1010) in LB containing 10 mM MgSO4, 0.2% maltose, and 0.5 mM IPTG. Following an incubation period of 15 min without shaking, 3 mL of hot (50 °C) top agar (LB containing 0.6% agar) was added to the culture and immediately poured onto a plate containing 1.6% solid agar at room temperature. The plates were incubated overnight. Lysed bacterial cells, appearing as clear plaques in an opaque layer of bacteria on the agar surface, were counted to determine phage concentration as plaque forming units (pfu) per milliliter. Preparation of Initial Stock Cultures of E. coli (MG1655) and K. pneumoniae. Bacterial cultures were grown overnight to stationary phase in LB medium with aeration in a shaker. Total cell concentrations were determined by measuring the optical density at 600 nm. The overnight cultures were diluted in LB to concentrations ranging from 1 to 109 cfu/mL. The stationary-phase cultures were used as the stock cultures for the experiments. A concentration of 1 cfu/100 mL was prepared by diluting the bacteria into 2 L of LB or sterile water. The Electrochemical System. The screen-printed electrodes consisted of carbon working and counter electrodes and a silver/ silver chloride reference electrode, all printed on a ceramic support, as described in ref 22. The electrochemical cells were made of polystyrene tubes (volume of 0.3 mL) attached to the disposable screen print electrodes. The electrochemical measuring system consisted of an eight-channeled, highly sensitive potentiostat constructed by Prof. Yarnitzky, Technion, Israel Institute of Technology. This multipotentiostat allows the simultaneous measurement of eight electrochemical cells. The current resulting from the activity of β-D-galactosidase could be simultaneously visualized in real time on a computer screen. The potentiostat was interfaced to a PC via an A/D converter employing a visual basic software. We used a specific homemade apparatus for the electrical contacts of the screen print electrodes combined with suction-expulsion-based efficient stirring. β-D-Galactosidase Activity. Enzymatic activity was measured amperometrically using p-aminophenyl-β-D-galactopyranoside (βPAPG) as substrate (0.8 mg/mL).21 The product of the reaction, p-aminophenol (PAP) is oxidized at the carbon anode at 220 mV vs the reference electrode. Aliquots of 300 µL from the lysed and filtered cultures were analyzed. Bacterial Cultures. Bacteria were grown in the presence of IPTG to induce the synthesis of β-D-galactosidase. All incubations were carried out at 37 °C. In all experiments, we used stationaryphase bacterial cultures, which is typical for microorganisms in environmental samples. In general, stationary-phase bacteria require a long incubation period before they can be identified. At the end of the final 3-h incubation period, the total volume of each culture was passed through a 0.22-µm filter to separate the medium from intact cells and cell debris. The filtration ensures the measurement of only the β-galactosidase released to the Analytical Chemistry, Vol. 75, No. 3, February 1, 2003
581
medium following the specific lysis of E. coli by λvir phage, thereby excluding the activity of the enzyme in the remaining intact E. coli cells or in other β-galactosidase-producing bacteria.22-24 Only the cell-free filtrate was used for electrochemical measurements. High Bacterial Concentrations (105-109 cfu/mL). Bacteria were infected by introducing 100-µL aliquots into 15 mL sterile test tubes containing 2 µL of 1010 pfu/mL of λ vir phage in the presence of 10 mM MgSO4, 0.2% maltose, and 0.5 mM IPTG. The mixtures were incubated for 15 min without shaking. Control mixtures did not contain the phage λ. After 15 min, LB was added to a total volume of 4 mL, and the mixture was incubated for an additional 3 h with vigorous shaking. Medium-Low Bacterial Concentrations (102-105 cfu/ mL). Bacteria were infected by introducing aliquots into 4 mL of LB containing 0.2% maltose and 0.5 mM IPTG and preincubated for 2, 3, or 4 h with vigorous shaking. Samples (1 mL) of the phage solutions, containing from 104 to 105 pfu/mL or 1 mL LB alone, and 10 mM MgSO4 were added to the bacterial mixture. The mixtures were further incubated for 2-3 h with vigorous shaking until lysis occurred. Very Low Bacterial Concentrations (1-102 cfu/mL). Increasing volumes of 10, 50, or 100 mL of stock cultures containing 100, 10, or 1 cfu/mL were passed through a 0.22 µm filter. The filters were then transferred to 4 mL of LB medium containing 0.2% maltose and 0.5 mM IPTG. Following a preincubation period of 4 h with vigorous shaking, 1 mL of the phage solution containing 104 pfu/mL and 10 mM MgSO4 was added to the mixture and incubated for an additional 3 h with vigorous shaking until lysis occurred. Optimizing the Multiplicity of Infection (MOI). To determine the optimal ratio of the number of phage particles to the number of cells at risk for our experiments, 100-µL samples from a stationary-phase stock culture of E. coli containing 106 cfu/mL were inoculated with 2 µL of different concentrations of phage λ vir, ranging from 104 to 108 pfu/mL in LB containing 10 mM MgSO4, 0.2% maltose, and 0.5 mM IPTG. The mixtures were first incubated for 15 min without shaking and then for another 3 h at the same temperature with vigorous shaking. The MOI value was calculated as the total phage particles taken divided by the number of E. coli cells present in the initial experimental mixture (before incubation). Bacterial Cell Counting. Bacteria from serial dilutions were plated on LB-agar (1.6%) plates to determine viable cell concentration as colony forming units per milliliter. RESULTS Detection of High Bacterial Concentrations (105-109 cfu/ mL). In the first stage, we determined the β-galactosidase activity released from different concentrations of E. coli cells infected by a constant number of plaque forming units of λ vir phage. The electrochemical response to the PAP product is shown in Figure 1A,B. The results indicate that the system is clearly sensitive to a wide concentration range of bacteria (5 orders of magnitude). The results confirm that the action of the phage is specific for E. coli, because the samples containing K. pneumoniae (109 cfu/mL), (23) George, I.; Petit, M.; Servais, P. J. Appl. Microbiol. 2000, 88, 404-413. (24) Leclerc, H.; Mossel, D. A. A.; Edberg, S. C.; Struijk, C. B. Annu. Rev. Microbiol. 2001, 55, 201-234.
582 Analytical Chemistry, Vol. 75, No. 3, February 1, 2003
Figure 1. Amperometric response of β-galactosidase released from 100 µL of stationary E. coli samples after infection and incubation for 3 h at 37 °C with 2 µL of 108 pfu/L λ vir at concentrations ranging between (A) 107 and 109 and (B) 105 and 107 cfu/mL and of K. pneumonia (109 cfu/mL).
which is β-D-galactosidase positive, did not show enzymatic activity (Figure 1B, lowest curve). Figure 2 summarizes the effect of 108 pfu/mL of λ vir on the reaction rate of the released enzyme at a 5 orders-of-magnitude concentration range. As shown in Figures 1 and 2, a good correlation was found between the initial concentrations of E. coli infected by the virulent phage λ and the electrochemical response generated by the activity of β-D-galactosidase. The enzymatic activity of fully lysed E. coli samples at high concentrations was at least 3 orders of magnitude higher than that of K. pneumonia samples, because the latter lacked the phage-mediated release of the enzyme to the medium. Our results correlate well with the data of Tryland et al.25 and Schwartz-Mittelmann et al.,22 indicating that the detection of nontarget β-D-galactosidase positive bacteria is low, because the enzymatic activity of K. pneumoniae cells was at least 2 log units below that of E. coli cells. Multiplicity of Infection (MOI). In the next step, we conducted experiments to determine the ratio of the number of (25) Tryland, I.; Fiksdal, L. Appl. Environ. Microbiol. 1998, 64, 1018-1023.
Figure 2. Reaction rates of β-D-galactosidase released from 100µL E. coli samples at different concentrations with and without 108 pfu/mL λ vir.
infectious virions to the number of susceptible cells in the tested culture, which is important for optimizing the mode of operation. The results shown in Figure 3, using E. coli cultures at a constant initial concentration of 2.5 × 104 cfu/mL and different concentrations of phage λ vir show that the optimal MOI was in the low range, between 0.05 and 0.005. Using such a low MOI value allows the bacterial culture to grow and produce enough β-D-galactosidase before full lysis occurs so that a large number of virulent phage progeny can be produced during a single 3-h incubation period. Detection of Medium-Low Bacterial Concentrations. Our next step was to increase the sensitivity of the method to measure lower concentrations. For this purpose, preincubation of the bacterial cultures (before adding the phage) was necessary. In this way, we achieved a higher enzyme activity in a relatively short time when compared with the mode described above. Furthermore, the preincubation allowed us to use a relatively high phage concentration rather than the low concentration calculated in the former section, meaning that higher than 0.05-0.005 MOI ratios must be used. Note that in the former mode of infection, the detection of a lower bacterial concentration would also require an extremely low concentration of phage λ, resulting in a very long incubation period. To verify this claim, we preincubated for 2 h an E.coli culture infected with two different concentrations of phage and then incubated the culture further until lysis occurred (2 h). Figure 4 shows that increasing the phage concentration from 2.5 × 104 to 2.5 × 105 pfu/mL led to an increase in enzymatic activity, and thus, in the same detection time (overall of 4 h), we could detect a 100 µL sample of 105 cfu/mL. Detection of lower concentrations of bacteria (
