Electrochemical Phagemid Assay for the Specific Detection of Bacteria

We describe a reporter phagemid system for the specific amperometric detection of bacteria. We constructed a phagemid a bacteriophage containing a bac...
1 downloads 0 Views 262KB Size
Download PDF
Anal. Chem. 2005, 77, 652-657

Electrochemical Phagemid Assay for the Specific Detection of Bacteria Using Escherichia coli TG-1 and the M13KO7 Phagemid in a Model System Tova Neufeld, Adrian Schwartz Mittelman, Virginia Buchner, and Judith Rishpon*

Department of Molecular Microbiology and Biotechnology, Tel-Aviv University, Ramat-Aviv, Israel 69978

We describe a reporter phagemid system for the specific amperometric detection of bacteria. We constructed a phagemid a bacteriophage containing a bacterial plasmid using the M13KO7 helper phage and a commercial plasmid, pFLAG-ATS-BAP, which contains a gene encoding for a reporter enzyme, alkaline phosphatase. In the bacteria, the enzyme reacts with the substrate, p-aminophenyl phosphate, in the periplamic space that separates the outer plasma membrane from the cell wall. Thus, the activity of the reporter enzyme can be measured directly in an electrochemical cell without further treatment. The product of the enzymatic activity, p-aminophenol, diffuses out and is oxidized at the working electrode with an applied potential of 220 mV vs the reference electrode Ag/AgCl. The lower detection limit was 1 cfu/ mL E. coli TG1 in less than 3 h in a very specific manner. The use of plasmid alkaline phosphatase as the reporter increased the sensitivity by 10-fold over our earlier electrochemical lytic phage method. Such a system can be used for the rapid detection of any strain of bacteria using the appropriate bacteriophage and reporter gene. The immediate detection of pathogenic bacteria in polluted water or food is essential for minimizing the risk of human exposure to harmful pathogens. Rapid and sensitive detection methods are essential for sanitary supervision. Conventional detection methods such as cultivation, immunoassays, or even the recently used PCR analyses are usually time-consuming, laborious, and expensive.1-3 Because monitoring for all potential bacterial pathogens in environmental samples is impossible, Escherichia coli is generally used to detect fecal pollution expressed as total coliforms, from human or animal waste in surface waters and food. Bacteriophage-based methods of environmental monitoring have been shown to be suitable indicators of fecal pollution in natural waters. A bacteriophage (phage) is a virus that exclusively infects bacteria in a very selective manner by attaching to specific * Corresponding author: Phone: +972-3-6409366. Fax: +972-3-6409407. E-mail: [email protected]. (1) Alvarez, J.; Sota, M.; Vivanco, A. B.; Perales, I.; Cisterna, R.; Rementeria, A.; Garaizar, J. J. Clin. Microbiol. 2004, 42, 1734-1738. (2) Chen, T. R.; Chiou, C. S.; Tsen, H. Y. Int. J. Food Microbiol. 2004, 92, 189197. (3) Li, J. W.; Shi, X. Q.; Chao, F. H.; Wang, X. W.; Zheng, J. L.; Song, N. Biomed. Environ. Sci. 2004, 17, 109-120.

652 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

bacterial cell wall phage receptors, including sex pili, flagella, and cell wall lipoproteins, polysaccharides, and lipopolysaccharides. Following infection, the phage uses the host’s nucleic acid, enzymes, and other proteins to synthesize progeny, thereby suppressing the host’s gene expression and metabolic activities. The lytic cycle of the phage usually culminates in the release of viral progeny and bacterial cell content into the medium. Occasionally sensitivity is lost because of bacterial cell lysis in the early stages of infection or because of the decreased fluorescence or enzymatic activity caused by cell proteases that are released into the medium. Among the applications of phage commonly used for identifying specific bacteria are phage typing, phage therapy, and antibiotic resistance assays,4-7 which can use wild-type or genetically engineered phage. The genetic engineering of phages often requires complicated manipulation of the long or double-stranded DNA of the phage, e.g., 50 kb in length for phage λ DNA.8 One example of a commercial phagemid is λ Zap, from which a plasmid called pBluescript can be excised with a helper phage. pBluescript contains the LacZ sequence coding for β-galactosidase, which causes E. coli colonies to turn blue if plated on medium containing a substrate that turns blue when cleaved by the enzyme. Recently we described a phage-typing technique using electrochemical measurement for the specific detection of E. coli using phage λ, which lowered the detection limit to 1 cell/100 mL in less than 8 h.9 Direct amperometric monitoring of the reporter bacterial enzyme, β-D-galactosidase, provided high sensitivity for the detection of total coliforms.10 Alternatively, we suggest using a modified phage called a phagemid, which is a phage that contains a bacterial extrachromosomal, autonomously replicating circular DNA molecule called a plasmid. This phagemid has the ability to infect bacteria but (4) Park, D. J.; Drobniewski, F. A.; Meyer, A.; Wilson, S. M. J. Clin. Microbiol. 2003, 41, 680-688. (5) Westwater, C.; Kasman, L. M.; Schofield, D. A.; Werner, P. A.; Dolan, J. W.; Schmidt, M. G.; Norris, J. S. Antimicrob. Agents Chemothe. 2003, 47, 13011307. (6) McNerney, R.; Kambashi, B. S.; Kinkese, J.; Tembwe, R.; Godfrey-Faussett, P. J. Clin. Microbiol. 2004, 42, 2115-2120. (7) Hazbon, M. H.; Guarin, N.; Ferro, B. E.; Rodriguez, A. L.; Labrada, L. A.; Tovar, R.; Riska, P. F.; Jacobs, W. R. J. Clin. Microbiol. 2003, 41, 48654869. (8) Funatsu, T.; Taniyama, T.; Tajima, T.; Tadakuma, H.; Namiki, H. Microbiol. Immunol. 2002, 46, 365-369. (9) Neufeld, T.; Schwartz-Mittelmann, A.; Biran, D.; Ron, E. Z.; Rishpon, J. Anal. Chem. 2003, 75, 580-585. (10) Mittelmann, A. S.; Ron, E. Z.; Rishpon, J. Anal. Chem. 2002, 74, 903-907. 10.1021/ac0488053 CCC: $30.25

© 2005 American Chemical Society Published on Web 12/16/2004

cannot replicate as an intact phage. Under most conditions, the phagemid propagates autonomously like a normal plasmid and replicates along with the bacterial cell cycle, but it can be released from the bacteria (rescue) only by coinfection with a second helper phage to lyse the bacteria.11 In this work, we used E. coli TG-1 as a model and the M13KO7 phagemid as the reporter. We inserted the commercial genetically engineered plasmid pFLAG-ATS-BAP into phage particles with the aid of a helper phage, M13KO7, a mutated form of the M13 filamentous phage. The plasmid pFLAG-ATS-BAP carries a “reporter” gene coding for the detectable bacterial enzyme, alkaline phosphatase, as well as a genetically selective marker, Ampr, which confers resistance to ampicillin. Filamentous phages such as M13 infect hosts by injecting their DNA through a protein fiber structure of the bacteria called the sex pilus. To initiate infection, the phage binds to the a minor coat protein, pIII, present only on the tip of the F pilus on male bacteria, E. coli F+.12 The binding induces retraction of the pilus, resulting in the injection of the single-stranded DNA of the phage into the bacterial cell. A specific peptide in the phagemid then genetically directs the reporter enzyme, bacterial alkaline phosphatase, to the periplasmic space between the outer plasma membrane and the cell wall of the bacteria. The cell wall is porous, allowing the substrate, p-aminophenyl phosphate, to enter easily into the perplasmic space. Hence, the activity of the reporter enzyme can be directly measured in the electrochemical cell without further treatment of the sample. The product of the reaction, p-aminophenol (p-AP), diffuses out and oxidizes at the working electrode with an applied potential of 220 mV.13-16 The improved assay can detect 1 colonyforming unit (cfu), an individual bacterial cell that grows into a visible colony of identical cells, per milliliter of bacteria in less than 3 h. EXPERIMENTAL SECTION Chemicals. Ampicillin, kanamycin, polymixin B, isopropylthiogalactoside (IPTG) poly(ethylene glycol) 6000 (PEG), and p-nitrophenyl phosphate (p-NPP) were purchased from Sigma Chemicals. D-Glucose was purchased from BDH, CaCl2 and MgCl2 were from Baker, and NaCl was from Merck. The enzyme substrate, p-aminophenyl phosphate (p-APP), was synthesized from p-NPP by Prof. Carmeli at the Department of Organic Chemistry, Tel-Aviv University (Israel). The Gen Elute Plasmid Miniprep Kit was purchased from Sigma Chemicals. Bacterial Media. Lennox Broth (LB), granulated agar, yeast extract, and Bacto trypton were purchased from Difco, Becton Dickinson. TYE growth medium (1 L) comprises 10 g of trypton, 5 g of yeast extract, 8 g of NaCl, and 15 g of granulated agar. 2YT growth medium (1 L) comprises 16 g of trypton, 10 g of yeast extract, and 5 g of NaCl. (11) Kramer, R. A.; Cox, F.; van der Horst, M.; van den Oudenrijn, S.; Res, P. C. M.; Bia, J.; Logtenberg, T.; de Kruif, J. Nucleic Acids Res/ 2003, 31. (12) Brock, T. D.; Madigan, M. T. Biology of Microorganisms, 6th ed.; Prentice Hall: Englewood Cliffs, NJ, 1991; pp 261-263. (13) Hadas, E.; Soussan, L.; Rosenmargalit, I.; Farkash, A.; Rishpon, J. J. Immunoassay 1992, 13, 231-252. (14) Rishpon, J.; Gezundhajt, Y.; Soussan, L.; Rosenmargalit, I.; Hadas, E. ACS Symp. Ser. 1992, 511, 59-72. (15) Rishpon, J.; Gezundhajt, Y.; Soussan, L.; Rosen, I.; Hadas, E. Abstr. Pap. Am. Chem. Soc. 1991, 201, 38-BIOT. (16) Rosen, I.; Rishpon, J. J. Electroanal. Chem. 1989, 258, 27-39.

Microorganisms and Plasmids. E. coli TG1 and Salmonella typhimurium were a generous gift from Prof. E. Z. Ron, Faculty of Life Sciences, Tel-Aviv University. The M13KO7 helper phage was a gift from Dr. Benhar, Faculty of Life Sciences, Tel-Aviv University. Bacillus cereus 7064 and Staphylococcus aureus are maintained in our laboratory. All bacterial strains were P1 safety level (allowing their use in standard chemical laboratories). A commercial plasmid, pFLAG-ATS-BAP, carrying the reporter gene coding for alkaline phosphatase and the genetic marker Ampr conferring ampicillin resistance, was purchased from Sigma Chemicals. Electrodes. Disposable screen-printed electrodes were purchased from Gwent Electronics. Preparation of Competent E. coli TG1 Cells. An aliquot (100 mL) of a logarithmic E. coli TG1 culture was centrifuged for 5 min at 7000 rpm. The pellet was diluted with 10 mL of 100 mM MgCl2 solution and incubated for 30 min on ice. The cells were centrifuged for 5 min at 7000 rpm, diluted with 10 mL of 100 mM CaCl2 solution, and placed for 30 min on ice. The cells were then centrifuged for 5 min at 7000 rpm, and the pellet was diluted with 1 mL of 10 mM CaCl2 solution containing 10% glycerol. Aliquots of 200 µL were kept at -70 °C. Transformation of E. coli TG1 Cells. Competent TG1 bacterial cells (200 µL, as prepared above) were allowed to thaw slowly for 30 min on ice. The plasmid pFLAG-ATS-BAP (100 ng) was added to the tube and incubated for 30 min on ice. The E. coli tube was moved to heat shock for 3 min at 42 °C and cooled for 5 min on ice. LB medium (1 mL) was added to the E. coli tube and incubated for 1 h at 37 °C. An aliquot of 100 µL of the transformed cells was plated onto an LB-ampicillin (100 µg/mL) agar plate. pFLAG-ATS-BAP Plasmid Production. An overnight culture of transformed E. coli cells, 3 mL, was centrifuged for 1 min at 13 000 rpm. The plasmid was purified using a Gen Elute Miniprep Kit, according to the manufacturer’s protocol, and stored at -20 °C. Rescue of M13KO7 Phagemid Containing pFLAG-ATSBAP. Transformed E. coli cells bearing the plasmid pFLAG-ATSBAP in LB liquid medium containing ampicillin were grown overnight with shaking at 20 °C. An aliquot (100 µL) of the overnight E. coli culture was inoculated into 3 mL of 2YT medium containing 100 µg/mL ampicillin and 1% glucose and grown until reaching the logarithmic phase. For infection, 15 µL of 4 × 1010 pfu/mL M13KO7 helper phage was added and incubated for 30 min at 37 °C without shaking. A plaque-forming unit (pfu) arises from a single infectious phage. The culture was further incubated for 30 min at the same temperature with shaking. The cells were centrifuged for 5 min at 13 000 rpm, and the pellet was diluted with 100 mL of 2YT medium containing ampicillin (100 µg/mL) and kanamycin (50 µg/mL) for overnight incubation with mild shaking at 30 °C. After the overnight incubation, the culture was centrifuged for 20 min at 5000 rpm at 4 °C and the supernatant collected. A sterile solution of 20% PEG containing 2.5 M NaCl was added gently to the supernatant at a ratio related to one-fifth (1/5) of the supernatant volume. The solution was incubated for at least 2.5 h on ice, without stirring. The solution was then centrifuged for 30 min at 3000 rpm, 4 °C, and the pellet was diluted with 1 mL of sterile PBS buffer. After recentrifugation for 5 min Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

653

at 13 000 rpm, the supernatant was purified by filtering through a 0.45-µm sterile filter membrane and stored at -20 °C. Titration of Phagemid Concentration. A mid-logarithmic phase E. coli TG1 culture and a decimal dilution of the purified phagemid solution were used for titration. An aliquot (10 µL) of each dilution was added to 90 µL of bacteria. All samples were incubated for 1 h at 37 °C without shaking. A 10-µL drop of each sample was plated on a TYE-Amp plate and dried for 10 min in a sterile hood. The plate was incubated overnight at 37 °C and the titer determined according to the number of pfu. The titer used for this work was 4 × 1010 pfu/mL. Phagemid Infection of 105-108 Cells/mL E. coli TG1 Samples. Liquid bacterial cultures were grown in 2YT medium at 37 °C until O.D600 ) 0.5 was reached. Decimal serial bacterial dilutions, ranging from 105 to 108 cells/mL, were used for the experiments. An aliquot (1 mL) of each bacterial dilution was infected with 20 µL of M13KO7 phagemid (4 × 1010 pfu/mL). Control mixtures did not contain phagemid. The mixtures were incubated for 1 h at 37 °C without shaking to allow the phagemid to attach to its receptor. Ampicillin and IPTG (for bacterial alkaline phosphatase induction under tac promoter) were then added to the bacterial mixtures at the respective concentrations of 100 µg/ mL and 0.5 mM. The samples were incubated at 37 °C with shaking for different times, ranging from 30 min to 3-4 h. Phagemid Infection of 1-104 Cells/mL E. coli TG1 Samples. Bacterial cultures were grown on 2YT medium at 37 °C until O.D600 ) 0.5 was reached. Fifty milliliters of each dilution of bacteria, ranging from 1 to 104 cells/mL, was filtered and the cells were collected on a 0.22-µm sterile filter. The filters were transferred to 2 mL of 2YT medium; the bacteria were released by shaking for 10 min at 37 °C. The content of each tube was divided and placed into two separate tubes: one tube was infected with 20 µL of phagemid (4 × 1010 pfu/mL), and the other, into which the phagemid was not introduced, served as a control for the collection filter. All samples were incubated for 30 min at 37 °C without shaking for phagemid attachment. Ampicillin and IPTG were added to the bacterial mixtures at the respective concentrations of 100 µg/mL and 0.5 mM and incubated for 90 min at 37 °C. Lysis of Phagemid-Infected E. coli TG1 Samples. Bacterial samples, infected with phagemid as described in the former section, were incubated for 1 h with polymixin B and lysozyme at the respective concentrations of 10 and 25 µg/mL at 37 °C to lyse the bacteria. Phagemid Infection of Other Bacterial Strains. Different bacterial strains, S. typhimurium, B. cereus, S. aureus, and E. coli TG1 cultures, were used for this experiment. Infection was performed as described above for the high-concentration samples. Phagemid Infection in the Presence of Other Bacterial Strains. Different bacterial strains, B. cereus, Klebsiella pneumoniae, and E. coli TG1 cultures were grown on 2YT medium at 37 °C until O.D600 ) 0.5 was reached. Dilutions of 1-104 E. coli TG1 cells/mL were used for these experiments. Different dilutions of E. coli were introduced to a constant high concentration (105 cells/ mL) of other strains of bacteria (B. cereus, K. pneumoniae) in 2YT medium. Infection was performed as described above. Electrochemical System. The electrochemical system consisted of an eight-channeled multipotentiostat built by Prof. 654 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

Figure 1. Amperometric response of bacterial alkaline phosphatase expressed in high E. coli concentrations samples tested after infection with the M13K03 phagemid: (1) 108 cfu/mL with phagemid, (2) 107 cfu/mL with phagemid, (3) 106 cfu/mL with phagemid, (4) 105 cfu/mL with phagemid, (5) 108 cfu/mL without phagemid, (6) 107 cfu/mL without phagemid, (8) 106 cfu/mL with phagemid, and (8) 105 cfu/mL without phagemid.

Yarnitzky, from the Technion, Israel Institute of Technology. The multipotentiostat allows simultaneous measurement of eight electrochemical cells. The potentiostat was interfaced to a PC via an A/D converter controlled by LabVIEW software. The electrochemical cells were made from disposable screen-printed electrodes attached to a polystyrene well with a total volume of 0.3 mL. The screen-printed electrodes consisted of a carbon working electrode (3.14 mm2), a carbon counter electrode, and a silver/ silver chloride reference electrode printed on a ceramic support. Specific laboratory-made apparatus was used for electrical contacts of the screen-printed electrodes, combined with suction-expulsion stirring. Amperometric Measurements. A volume of 270 µL of bacterial sample was placed in the electrochemical cell. The substrate, p-APP, in a volume of 30 µL was added to the sample to make a final concentration of 1 mg/mL. The product of the bacterial alkaline phosphatase reaction, p-AP, is oxidized at the carbon working electrode at 220 mV potential versus the reference electrode Ag/AgCl. The amperometric signal generated by the enzymatic activity was measured, and the current versus time plot was recorded on line. RESULTS Detection of high E. coli Concentrations by Infection with the M13KO7 Phagemid. Bacterial cultures of E. coli TG1 were grown to mid-log stage at 37 °C in order to induce pilus formation and diluted to samples ranging from 109 to 105 cfu/mL. Each sample was infected with the phagemid at high and constant concentration of 4 × 1010 pfu/mL following a 1-h incubation period. The samples were measured immediately without any additional treatment such as filtration, chemical permeabilization, or lysis. The electrochemical measurements of the alkaline phosphatase activity expressed in the periplasmic space are shown in Figure 1. Increasing the incubation time and decreasing the bacterial concentration abolished the basal activity of the bacterial enzyme detected at high bacterial concentrations. Bacterial Strain Specificity. By introducing the M13KO7 phagemid into other strains of bacteria, we examined whether

Figure 3. Effect of incubation time on the amperometric signal generated by different E. coli TG1 samples ranging between 107 and 104 cfu/mL infected with the phagemid.

Figure 2. Bacterial strain specificity tested at high concentrations of B. cereus, S. typhimurium, and S. aureus and compared with E. coli TG1 (A) Amperometric response of infected bacterial cultures. (1) E. coli TG1 infected with the M13K07 phagemid; (2-8) other bacterial responses to the phagemid infection. (B) Comparison of infected E. coli TG1 shown in (A) with the other cultures tested. Results are represented as ∆current/∆time (with error bars (n ) 6)).

the phagemid infection is specific for E. coli TG1. The strains tested were B. cereus, S. typhimurium, and S. aureus. Figure 2 shows the results of infection at very high bacterial concentrations of 108 and 109 cells/mL. The infection was specific because only the E. coli TG1 culture showed alkaline phosphatase activity. A minor signal, 15-20-fold less than that for E. coli TG1, detected in B. cereus at 109 cells/mL, was probably due to the innate alkaline phosphatase enzyme activity typical for Bacillus species.17,18 Such activity, which disappeared almost completely in B. cereus, is not considered a problem because polluted water samples contain about 4-5 orders of magnitude lower bacterial concentrations than those tested in this experiment. S. typhimurium and S. aureus, showed no activity at all. Effect of Incubation Time on the Amperometric Signal. At low E. coli concentrations, prolonged incubation with the phagemid was required to obtain a signal. To test the effect of the incubation time on the amperometric signal, we withdrew samples at 30-min intervals for electrochemical measurement. As low as 104 cfu/mL were detectable after 4-h incubation without any prefiltering, preincubation, or further lysis-aided treatment, all of which are required in other methods.19 The time dependence of enzymatic activity is shown in Figure 3. Low bacterial concentrations required additional treatment such as prefiltration and a preincubation period before phage infection. Detection of Low E. coli Concentrations by the M13KO7 Phagemid. To measure very low concentrations of E. coli, we (17) Bursik, M.; Nemec, M. Folia Microbiol. 1999, 44, 90-92. (18) Vinter, V.; Smid, F. Folia Microbiol. 1984, 29, 389-389. (19) Van Poucke, S. O.; Nelis, H. J. J. Microbiol. Methods 2000, 42, 233-244.

Figure 4. Amperometric response to alkaline phosphatase activity expressed at low concentrations of E. coli TG1 infected with the phagemid at (1) 680, (2) 68, (3) 10, and (4) 1 cfu/mL and (5-8) in noninfected bacterial samples . The results in the inset are represented as ∆current/∆time (n ) 6).

collected on a filter 50 mL of bacteria, ranging from 680 to 1 cfu/ mL, which were shaken for 10 min to release the bacteria into the liquid medium. The samples were divided into two tubes for the 30-min infection step and the 90-min incubation period in the presence or absence of the phagemid. The duration of the whole procedure was 130 min. The results of the electrochemical measurements are shown in Figure 4. We found a good correlation between the sample concentration and the electrochemical signal. No signal was obtained from control experiments in the absence of the phagemid because the low concentration of the natural bacterial alkaline phosphatase activity is under the detection limit of the system, and noninfected bacteria were excluded by ampicillin selection. This procedure can be manipulated for measuring even lower concentrations of bacteria (1 cfu/100 mL) by filtering between 2 and 5 L of the test water sample. Optional Further Cell Lysis. To emphasize the efficiency of this method, we further permeabilized the bacterial cells detected in the former section by subjection to polymixin B (10 µg/mL) Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

655

Figure 5. Comparison of cultures subjected to further permeabilization with polymixin B/lysozyme and cultures without permeabilization. Bacterial concentrations ranged from 1 to 100 cfu/mL E. coli TG1. Results are represented as ∆current/∆time (n ) 5).

and lysozyme (25 µg/mL). This combination usually results in almost 10-fold higher signals than in nonpermeabilized bacteria because of leakage of the cellular material.10 The results shown in Figure 5, however, revealed no considerable difference in the enzyme activity because the signal obtained in the treated cultures was only 2-fold higher than in untreated cultures and the errors were larger. Detection in the Presence of Other Bacterial Strains. As water samples usually contain more than one species of bacterial pollutant, we tested the ability of our phagemid assay to identify E. coli TG1 in a mixture containing two other strains of bacteria. E. coli TG1 cultures in the range of 1-103 cfu/mL were mixed with 105 cells/mL B. cereus and 105 cells/mL K. pneumoniae. A 50-mL sample of each mixture collected on a filter was released into a tube containing 2 mL of fresh medium and incubated for 90 min in the presence or absence of the phagemid. Figure 6 shows that even at the lowest concentration of 1 cfu/mL a measurable signal was detected in the phagemid-infected E. coli samples. DISCUSSION The results presented here demonstrate the high efficiency of our model M13KO7 phagemid system for the specific identification of E. coli TG1, either alone or in a mixture. Enzymatic activity at a lower detection limit of 1 cfu/mL was recorded in the extremely short time of ∼2 h. The system described here is not only highly specific but also offers several advantages over other phage-based detection methods. In our method, the M13KO7 phagemid serves as both the specific recognition agent and the transfecting factor of the reporter enzyme, alkaline phosphatase, a product of the plasmid DNA. The ability to measure the enzymatic reaction directly in the electrochemical cell eliminates the need for additional treatment, such as filtration, chemical permeabilization, or lysis. Because the phagemid cannot replicate as an intact phage, the multiplicity of infection ratio (MOI, the ratio of virus particles to bacterial cells) is less critical than when using lytic phages that can cause the loss of bacteria through lysis during the initial stages of the assay. Phagemids are predominantly used as vectors to produce a collection of phages displaying on their surfaces a population of 656 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

Figure 6. Detection of low concentrations of E. coli TG1 in the presence of mixtures of high (105 cfu/mL) concentration of other bacterial strains: B. cereus and K. pneumoniae. (A) E. coli TG1 infected with the phagemid at (1) 1000, (2) 100, (3) 10, and (4) 1 cfu/mL; (5-8) noninfected bacterial samples at the same concentrations, respectively. (B) Results are represented as ∆current/∆time.

related but diverse bacterial polypeptides (phage-display libraries). Phagemids are also useful for gene amplification and screening.20-22 The mainstream of work published until now deals with the use of natural or genetically engineered phages for phage typing, a technique that differentiates between bacteria or strains of bacteria by their susceptibility to one or more phages. Recombinant luciferase reporter phages, which produce quantifiable light measured by bioluminescence, serve as good agents for identifying bacteria.23-26 We developed a method that overcomes such disadvantages by using the nonlytic, nonreplicating phage, M13, to construct the phagemid M13K07. The phagemid retains the advantages of both vectorssthe specific recognition contributed by bacteriophage and the effortless genetic manipulation of a plasmid. The (20) Benhar, I.; Eshkenazi, I.; Neufeld, T.; Opatowsky, J.; Shaky, S.; Rishpon, J. Talanta 2001, 55, 899-907. (21) Benhar, I. Biotechnol. Adv. 2001, 19, 1-33. (22) Larocca, D.; Jensen-Pergakes, K.; Burg, M. A.; Baird, A. Mol. Ther. 2001, 3, 476-484. (23) 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. (24) Loessner, M. J.; Rudolf, M.; Scherer, S. Appl. Environ. Microbiol. 1997, 63, 2961-2965. (25) Blasco, R.; Murphy, M. J.; Sanders, M. F.; Squirrell, D. J. J. Appl. Microbiol. 1998, 84, 661-666. (26) Wu, Y.; Brovko, L.; Griffiths, M. W. Lett. Appl. Microbiol. 2001, 33, 311315.

use of plasmid alkaline phosphatase as the reporter increased the sensitivity of the assay by 10-fold compared with the prior method developed in our laboratory using lytic phage. The lower level detection limit (without a preconcentration step) in that assay was 105 cfu/mL.9 The antibiotic selection gene, Ampr, in the plasmid facilitates the specific selection of the bacteria of interest because E. coli is resistant to ampicillin (due to plasmid presence), whereas other strains of bacteria are susceptible. The increased sensitivity to a lower detection limit of 1 cfu/mL, obtained by concentrating a contaminated 50-mL water sample, decreased the detection time to ∼2-3 h. Further permeabilization of the cells before measuring was not necessary because the signal obtained in the treated samples was only 2-fold higher than in untreated samples, which is not significant for bacterial cultures. The ability of phagemid assay to detect E. coli TG1 in mixtures containing 2-4 orders of magnitude higher concentrations of B. cereus and K. pneumoniae confirmed the high specificity of the assay because the other bacterial strains tested have neither a pilus receptor nor antibiotic resistance that is independent of phagemid presence. The minor enzymatic activity detected in uninfected bacterial mixtures, probably originating from the innate enzyme activity of B. cereus, should not be a problem for environmental monitoring because such bacteria are susceptible to ampicillin. Additionally, such high concentrations of bacteria are not expected in drinking water. The phagemid method described here is fast and efficient. For comparison, Van Poucke and Nelis19 developed a sensitive chemiluminometric assay for the detection of 1 coliform/100 mL within 6-9 h using a permeabilization step followed by a 45-min assay. Other phage methods were developed primarily for bacterial identification in food. Goodridge et al.27,28 used fluorescencestained phage for detecting 100 cfu/mL E. coli O157:H7 in artificially contaminated milk in 10 h. A new approach recently published uses the phage PP01 small outer capsid as a platform to present genetically the marker protein, a green fluorescence protein, on the phage capsid for rapid detection of the pathogenic E. coli O157:H7.29 Again, this approach requires genetic manipulation with the whole phage genome, and not only the MOI ratio of this lytic phage but also the expensive epifluorescence microscope needed for observing the results must be taken into consideration. Such drawbacks emphasize the convenience of an electrochemical method that overcomes the problem of natural fluorescence of bacterial contaminants such as Pseudomonas.30 (27) Goodridge, L.; Chen, J. R.; Griffiths, M. Int. J. Food Microbiol. 1999, 47, 43-50. (28) Goodridge, L.; Chen, J. R.; Griffiths, M. Appl. Environ. Microbiol. 1999, 65, 1397-1404. (29) Oda, M.; Morita, M.; Unno, H.; Tanji, Y. Appl. Environ. Microbiol. 2004, 70, 527-534. (30) Ekholm, D. F.; Hirshfield, I. N. J. AOAC Int. 2001, 84, 407-415.

The extremely short assay time presented here is due partly to the overproduction of plasmid alkaline phosphatase in the infected cells. Moreover, the specific activity of alkaline phosphatase is much higher than that of the commonly used reporter enzyme β-galactosidase. A short detection time is the most important demand concerning assays of polluted water sources and even more critical when bioterrorism is suspected. The phagemid system suggested here is applicable in the food poisoning area as well because it allows measurements in opaque samples. A longer incubation period is recommended for determining food contamination. The phagemid system, generally designed to create DNA libraries, can also be used to construct similar large libraries for the specific identification of bacteria. Furthermore, the identification of a multitude of bacterial strains also can be achieved by using broad host range origin of replication phagemids. For example, the P1 phagemid can detect members of the family Enterobacteriaceae, which are Gram-negative bacteria like E. coli, Shigella flexneri, Shigella dysenteriae, K. pneumoniae, and Citrobacter freundii. CONCLUSIONS The phagemid strategy presented here has several advantages over other bacteriophage methods for the specific selection of bacteria. Phagemids retain the advantages of the specific recognition contributed by bacteriophage and the effortless genetic manipulation of a plasmid. The M13KO7 phagemid specifically recognizes and infects its host, E. coli TG-1 by attaching to its complementary receptors but cannot replicate as an intact phage. The directed secretion of the reporter enzyme into the periplasm eliminates the need to lyse the bacteria with lytic phages or to add chemical agents to increase the permeability of the bacterial cell wall. The use of phagemids for the detection and identification is a powerful tool that in principle can be used with any specific pair of bacteria and its selective phage, packaged in a plasmid containing a reporter enzyme directed to the periplasmic space. Choosing the appropriate reporter enzyme and its location of expression is the key to saving detection time and eliminating the need for further treatment to achieve increased sensitivity. This approach, in combination with amperometric detection, provides an assay that is rapid, highly sensitive, easy to operate, and economical. A concentration of 1 cfu/mL E. coli TG1, collected from only 50 mL of a contaminated water sample, was detected rapidly with our phagemid assay within a very short time of 2-3 h. The assay is specific and can be exploited for water- or foodborne bacterial contamination. Received for review August 12, 2004. Accepted November 3, 2004. AC0488053

Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

657