A novel colorimetric method for the detection of Escherichia coli using cytochrome c peroxidase-encoding bacteriophage

Authors

  • Hoang A. Hoang,

    1. Department of International Development Engineering, Tokyo Institute of Technology, Meguro-ku, Tokyo, Japan
    Search for more papers by this author
  • Michiharu Abe,

    1. Department of International Development Engineering, Tokyo Institute of Technology, Meguro-ku, Tokyo, Japan
    Search for more papers by this author
  • Kiyohiko Nakasaki

    Corresponding author
    1. Department of International Development Engineering, Tokyo Institute of Technology, Meguro-ku, Tokyo, Japan
    • Correspondence: Kiyohiko Nakasaki, Department of International Development Engineering, 2-12-1-I4-2 Ookayama, Meguro-ku, Tokyo 152-8550, Japan. Tel./fax: +81 3 5734 3169; e-mail: nakasaki@ide.titech.ac.jp

    Search for more papers by this author

Abstract

A new rapid and simple method was developed for the detection of Escherichia coli by constructing a recombinant T4 phage carrying the cytochrome c peroxidase gene derived from Saccharomyces cerevisiae (T4ccp) using which, the colorimetric detection of E. coli K12 was examined. The oxidation activity toward the chromogenic substrate cytochrome c was demonstrated by the cytochrome c peroxidase (CCP) produced from the T4ccp genome. The color change caused by the oxidation of the substrate could be visually perceived. The possibility of interference in the detection by the coexistence of other bacteria was assessed using Pseudomonas aeruginosa as a nontarget bacterium, and it was confirmed that the coexistence of P. aeruginosa caused no interference in the detection of E. coli K12.

Introduction

Escherichia coli are commonly used as an indicator of fecal contamination in water and food samples. To selectively detect E. coli, media containing β-d-glucuronide substrates have been developed as β-d-glucuronidase activity is a typical characteristic of E. coli (Rice et al., 1990; Brenner et al., 1993). However, the β-d-glucuronidase-based agar assay is time-consuming because at least 1 day is required for colony formation on the agar plate. One of the approaches considered for shortening the detection time for E. coli is the use of polymerase chain reaction (PCR) for the amplification of the uidA gene encoding β-d-glucuronidase (Cleuziat & Robert-Baudouy, 1990; Bej et al., 1991). Although the method is rapid, it is inadequate for distinguishing the living cells from dead cells.

For the rapid detection of E. coli, methods involving immunomagnetic separation (IMS) have been investigated (Goodridge et al., 1999a; Willford & Goodridge, 2008). The IMS is performed with magnetic beads coated with antibodies that bind to antigens on the surface of the target bacterium. The specificity of the antibody together with the magnetic properties of the beads enables separation of the target bacterium from the nontargeted bacteria. The IMS exhibited excellent performance with respect to the rapidity and selectivity of separation; however, the requirement of expensive immunomagnetic beads renders the method costly. Furthermore, additional assays for determining the binding of the bacteria to the beads are essential for completing the detection.

The application of bacteriophages for the detection of specific bacteria is advantageous owing to the high specificity of the bacteriophages in host recognition. Until now, fluorescence- or bioluminescence-based detection methods utilizing bacteriophages have been investigated (Tanji et al., 2004; Brigati et al., 2007). The fluorescence-based detection method involved introducing the green fluorescent protein gene (gfp) into the lysozyme-inactivated T4 phage genome (Tanji et al., 2004). Production of the green fluorescent protein and consequently its accumulation in infected E. coli cells enabled selective observation under an epifluorescence microscope. On the other hand, bacterial luciferase genes (luxCDABE) were employed in a bioluminescence-based detection method (Brigati et al., 2007), for which a recombinant phage carrying the luxI gene that encodes LuxI protein was constructed. Infection of the recombinant phage to the target E. coli cells led to the production of the LuxI protein and initiated the expression of an autoinducer, which subsequently activated the expression of the luxCDABE genes leading to the emission of bioluminescence. The resulting bioluminescence could be detected using a luminescence counter. Although both fluorescence- and bioluminescence-based detection methods allow selective detection of E. coli in just 1 day, for the detection by those methods, the special apparatus is required to evaluate the results.

Generally, colorimetric examinations are easy and convenient for result evaluation simply because of the ease in detection by eyes without the need for specific apparatus, and for quantifying the color change using a spectrophotometer that is relatively popular and more easily available compared to an epifluorescence microscope and a luminescence counter. However, to the best of our knowledge, a phage-based colorimetric detection of bacteria has not been reported before. Convenience of the colorimetric assay together with the rapidity and selectivity of the phage-based detection may allow the development of a useful detection method for a specific bacterium. In this study, we developed the colorimetric detection of E. coli K12 using a recombinant T4 phage bearing the cytochrome c peroxidase gene (ccp) derived from Saccharomyces cerevisiae, in which the cytochrome c peroxidase (CCP) catalyzes the oxidation of cytochrome c that leads to the decolorization of the assay solution.

Materials and methods

Bacterial strains and bacteriophage

Escherichia coli K12 W3110 was used as the host for the T4 phage. The general protocol for cloning involved the use of E. coli DH5α as host. Pseudomonas aeruginosa NBRC 12689 was employed to investigate the possibility of its interference in the detection of E. coli K12 by a recombinant T4 phage constructed in this study. The source of the ccp gene was S. cerevisiae IAM 4178.

Construction of a recombinant T4 phage carrying the ccp gene

Oligonucleotide primers used in the PCR amplification are listed in Table 1. The DNA fragment (T4soc-mrh2) corresponding to T4 soc gene and N-terminal sequence of mrh2 gene was amplified using T4socF/mrh2R primer set that was designed based on the primers used in the research by Tanji et al. (2004), where the sequences for the restriction digestion were changed, and wild-type T4 phage (T4wt) lysate as the template. Following the digestion of the amplified fragment by EcoRI/ApaI, the PCR fragment was cloned into the EcoRI/ApaI site of pCR2.1-TOPO (Invitrogen, CA) to produce the plasmid vector pCRmrh2. The genome of S. cerevisiae IAM 4178 was extracted using the ISOIL for Beads Beating Kit (Nippon Gene, Toyama, Japan). The ccp gene was amplified from the extracted genome using the ccpF/ccpR primer set that was designed based on the primers used in the research by Iffland et al. (2000) where the sequences for the restriction digestion were changed. The SacI/EcoRI-digested PCR fragment was inserted into the SacI/EcoRI site of the pCRmrh2 to produce the plasmid vector pCRccpmrh2. The DNA fragment (T4g69) amplified from the T4wt genome using the g69F/T4socNR primer set that was designed based on the primers used in the research by Tanji et al. (2004) where the sequences for the restriction digestion were changed, were digested with KpnI/SacI, and inserted into the pCRccpmrh2 to produce the vector pCRT4ccp (Fig. 1).

Table 1. Oligonucleotide primers and a probe used for PCR amplification and plaque hybridization
Primers, probeSequence (5′→3′)References
T4socF (EcoRI)TTCCGGAATTCCATGGCTAGTACTCGCGGTanji et al. (2004)
mrh2R (ApaI)TTAGGGCCCGGTTTAATCCAACGATTTAACATTanji et al. (2004)
g69F (KpnI)CGGGGTACCAGAAAAATCATATGAAGTTGATanji et al. (2004)
T4socNR (SacI)TTGCGAGCTCCTCCTTTATTTAAATTACATGTanji et al. (2004)
ccpF (SacI)TTGCGAGCTCATGACACCGCTCGTTCATGTIffland et al. (2000)
ccpR (EcoRI)TTCCGGAATTCCTATAAACCTTGTTCCTCIffland et al. (2000)
ccp probeGGGACTCTAAGAGCGGCTACATGAThis study
Figure 1.

The constructed vector pCRT4ccp.

The ccp gene was integrated into the T4wt genome by homologous recombination. The E. coli K12 was transformed with the vector pCRT4ccp by electroporation using the Gene Pulser II (Bio-Rad, CA) with the following settings: 1.8 kV, 25 μF, and 200 Ω. The transformant E. coli K12 cells were cultivated in Luria–Bertani (LB) medium supplemented with 50 mg L−1 ampicillin. At an OD600 nm of 0.1, T4wt was added at a multiplicity of infection (MOI) value of 0.01. Following incubation at 37 °C, 120 r.p.m. for 4 h, the culture was filtered through a 0.45-μm membrane filter (Advantec, Tokyo, Japan) to obtain the phage lysate. The phage lysate was diluted with the SM buffer [100 mM NaCl, 10 mM MgSO4, 0.01% gelatin, and 50 mM Tris-HCl (pH 7.5)]. The diluted phage lysate was mixed with E. coli K12 in 0.7% agar and overlaid onto LB plates. The recombinant T4 phage bearing the ccp gene (T4ccp) was isolated by plaque hybridization with a digoxigenin (DIG)-labeled probe. The plaque hybridization was carried out according to the DIG Application Manual for Filter Hybridization (Roche, Upper Bavaria, Germany). The hybridization with the DIG-labeled ccp probe was conducted at 37 °C for 6 h. Probe–target hybrids were detected using a chromogenic assay involving NBT/BCIP Stock Solution (Roche, Upper Bavaria, Germany).

Evaluation of the activity of CCP produced by the T4ccp

Escherichia coli K12 was cultivated at 37 °C until an OD600 nm of 0.5 (approximately 1 × 108 CFU mL−1) was attained. Then, the culture was divided into three aliquots, of which two aliquots were mixed with either T4ccp or T4wt phage lysate at MOI of 5.0. One aliquot was left blank without phage addition. The aliquots were incubated at 37 °C for 1 h, and then, they were passed through a 0.45-μm membrane filter to obtain filtrates. In addition to the filtrates, the LB medium was also used for the assay. The cytochrome c from equine heart (Sigma-Aldrich, St. Louis, MO) was used as a substrate for the enzymatic assay, and cytochrome c was reduced prior to the assay in accordance with the protocol described by Spinazzi et al. (2012), with minor modifications. The concentration of the reduced cytochrome c was determined by measuring the A550 nm [Extinction coefficients (ε) of oxidized and reduced forms of cytochrome c are 9.5 and 28 mM−1 cm−1, respectively. Therefore, Δε = 18.5 mM−1 cm−1 (Gelder & Slater, 1962; Iffland et al., 2001)]. The filtrates or the LB medium was mixed with phosphate buffer (50 mM KH2PO4, pH 6.0), cytochrome c, and H2O2 to obtain a 10-fold dilution. The final concentrations of cytochrome c and H2O2 were 0.7 and 360 μM, respectively. The mixture was incubated at 30 °C, and the A550 nm of the reaction solution was measured every minute using a spectrophotometer (UV mini-1240; Shimadzu, Kyoto, Japan). All the enzyme assays were conducted in triplicate.

Interference of P. aeruginosa in the colorimetric assay

Pseudomonas aeruginosa NBRC 12689 was cultivated in LB medium at 37 °C until it attained an OD600 nm of 0.5. Then, the culture was divided into two aliquots. T4ccp was added to one aliquot with a 5 : 1 concentration ratio of the phage to P. aeruginosa NBRC 12689. The second aliquot of the culture was mixed with an equal volume of E. coli K12 suspension in LB medium that was prepared by concentrating two times by the centrifugation of the culture incubated until the OD600 nm reached 0.5 and pipetting out of the half volume of supernatant. The mixture was then infected with T4ccp at MOI of 5.0. The twofold concentrated E. coli K12 culture was mixed with an equal volume of LB medium, and then, the culture was also infected with T4ccp at MOI of 5.0. The LB medium mixed with T4ccp was also prepared. The aliquots, the E. coli K12 culture, and the LB medium mixed with T4ccp were then incubated for 1 h with shaking. The preparation of the filtrates and running the enzymatic assays with substrate cytochrome c were according to the procedure mentioned above.

Results

Construction of T4ccp

The ccp fragment and two flanking regions were inserted into the vector pCR2.1-TOPO to produce the vector pCRT4ccp (Fig. 1). Then, T4ccp was produced by homologous recombination of the vector pCRT4ccp with the genome of T4wt. Plaque hybridization for the isolation of T4ccp involved the detection of about 10 positive plaques in a plate containing about 2 × 104 plaques. Therefore, the frequency of recombination seemed to be approximately 5 × 10−4. The positive plaques were picked and suspended in the SM buffer, and T4ccp was isolated from the suspension by repeated plaque hybridization. Integration of the ccp gene into the genome of the T4 phage was confirmed by PCR (Fig. 2). PCR using the primer set for the ccp gene yielded an amplified gene fragment with the expected size (882 bp) from the T4ccp genome, while no fragment was amplified from the T4wt genome (lanes 1 and 2 in Fig. 2). PCR using primers g69F and mrh2R provided an amplified fragment having a size corresponding to the combination of ccp and the flanking regions (1757 bp), whereas the size of the fragment amplified from T4wt corresponded only with that of the flanking regions (857 bp; lanes 3 and 4 in Fig. 2). The size of the PCR amplified fragment from T4ccp using the primer set g69F and ccpR corresponded to that of ccp along with the upstream flanking region (1386 bp), while no DNA fragment was amplified from the T4wt (lanes 5 and 6 in Fig. 2). These results ensured that the ccp gene was integrated into the region between g69 and soc genes of the T4wt genome in the correct direction. Moreover, the sequence of the ccp gene and the adjacent two regions in the T4ccp genome was confirmed by sequencing (data not shown). In terms of host specificity, T4ccp is expected to have no alteration in its host range as Tanji et al. (2004) demonstrated that insertion of a foreign gene into the region between g69 and soc genes in T4 phage genome did not alter the host range of T4 phage.

Figure 2.

Agarose gel electrophoresis of the amplified fragments from T4ccp or T4wt genome. M: marker. Lanes 1, 2: fragments amplified with ccpF/ccpR primers; lanes 3, 4: fragments amplified with g69F/mrh2R primers; lanes 5, 6: fragments amplified with g69F/ccpR primers. T4ccp lysate was used as template for lanes 1, 3, and 5, and T4wt lysate was used as template for lanes 2, 4, and 6.

Activity of CCP expressed from T4ccp genome

Cytochrome c oxidation activity of the CCP produced through the T4ccp infection of E. coli K12 was examined. The oxidation of cytochrome c results in the fading of the red color of the reaction solution. Preliminary, to investigate the relation between substrate concentrations and the extent in the color change between the enzyme assays with T4ccp and T4wt, some different concentrations of cytochrome c and H2O2 were examined. Among the H2O2 concentrations examined with a range from 120 to 720 μM, the largest difference in the A550 nm values was shown between the enzyme assays using T4wt and T4ccp at the H2O2 concentration of 360 μM. And among the cytochrome c concentrations examined with a range from 0.48 to 1.0 μM, the difference in the A550 nm value between the assays using T4ccp and T4wt could be observed even at the lowest cytochrome c concentration examined (0.48 μM), which suggested that the lower limit of necessary concentration of cytochrome c was lower than 0.48 μM. However, the difference in the color change between the assays using T4ccp and T4wt was more distinctly observed in the assay with 0.7 μM cytochrome c, while in the assay with the cytochrome c at the concentration higher than 0.7 μM, the difference in the A550 nm value was similar to that of the assays using 0.7 μM cytochrome c. Based on those preliminary examinations, we decided to conduct the enzyme assay using 360 μM H2O2 and 0.7 μM cytochrome c.

In the enzyme assay using the lysate obtained by the T4ccp infection of E. coli K12, the change in the color of the reaction solution could be visually perceived and the change in A550 nm during the assay with cytochrome c is shown in Fig. 3. As the results of the assay using the lysate obtained by the T4wt infection of E. coli K12 and the filtrate of the E. coli K12 culture without phage addition were almost identical to the result obtained using LB medium without any bacterial inoculation, it was confirmed that the presence of E. coli K12 or the lysis of E. coli K12 by the infection of T4wt did not affect the oxidation of cytochrome c. Previously, a direct oxidation of cytochrome c by H2O2 was reported by Beetlestone (1960), and the slight decreases in A550 nm in the filtrates of experiments with T4wt or with E. coli K12 culture without phage or with LB medium without any bacteria inoculation were considered to be caused by a direct oxidation of cytochrome c by H2O2. In addition, dissolved oxygen also may have contributed to the oxidation of cytochrome c in those samples. In contrast, using the lysate obtained by the T4ccp infection of E. coli K12 in the assay resulted in a significant decrease in the A550 nm, and it was suggested that the CCP expressed from the T4ccp genome contributed substantially to the oxidation of cytochrome c.

Figure 3.

Change in the A550 nm during the enzyme assay and color change after 6 min with lysate obtained from T4ccp or T4wt infection. Error bars indicating 95% confidence intervals for the averaged values (= 3) are not graphically detectable as the intervals were too narrow.

Effect of the coexistence of the P. aeruginosa on the detection

Pseudomonas aeruginosa is a cytochrome oxidase and CCP-positive bacterium (Horio et al., 1961; Rönnberg et al., 1989), and it was therefore used to assess the possibility of the undesirable oxidation of the substrate by the presence of contaminant bacteria in the sample. The changes in A550 nm (Fig. 4) during the enzymatic assay using the filtrate of P. aeruginosa NBRC 12689 culture, the filtrate of the mixed culture of P. aeruginosa NBRC 12689 and E. coli K12, the filtrate of E. coli K12 culture and LB medium without inoculation of any bacterium were obtained. There was almost no difference between the changes in A550 nm for the filtrate of P. aeruginosa NBRC 12689 culture and that for the LB medium without inoculation of any bacterium, thereby confirming that the P. aeruginosa NBRC 12689 did not result in the oxidation of cytochrome c. The slight decreases in A550 nm in the filtrates of P. aeruginosa NBRC 12689 culture and the LB medium without inoculation of any bacterium were considered to be caused by a direct oxidation of cytochrome c by H2O2 and dissolved oxygen as mentioned above. In addition, A550 nm decrease in the assay using the filtrate of the mixed culture of P. aeruginosa and E. coli was almost the same as that obtained using the culture of E. coli. The results indicated that E. coli K12 can be detected using T4ccp without interference caused by P. aeruginosa NBRC 12689.

Figure 4.

Change in A550 nm during the enzyme assay for the evaluation of the possibility of interference of Pseudomonas aeruginosa culture. The experiment was conducted with addition of T4ccp. Error bars indicating 95% confidence intervals for the averaged values (= 3) are not graphically detectable at some points as the intervals were too narrow.

Discussion

Alkaline phosphatase (AP) and horseradish peroxidase (HRP) are most commonly used for colorimetric detection in enzyme immunoassays, and these enzymes were expected to be useful for the detection of E. coli using a recombinant phage. However, the size of the gene encoding the AP (5.4 kb) was considered too large to be introduced into the T4 phage genome. Moreover, HRP from the T4 phage genome cannot be used for production in the E. coli cells because the disulfide bonds essential for the folding process of HRP cannot be formed in the cytoplasm of E. coli, and E. coli does not possess the enzymes required for the glycosylation of HRP (Marston, 1986; Ortlepp et al., 1989). Therefore, neither AP nor HRP was considered to be applicable for the detection of E. coli using a recombinant phage. By contrast, CCP is known to fold correctly in E. coli cells (Goodin et al., 1991), and the reaction between CCP and its original substrate cytochrome c results in a color change. Therefore, CCP was examined for the detection of E. coli K12 using a recombinant phage in this study.

In addition to cytochrome c, 3,3′-diaminobenzidine (DAB), which is one of the chromogenic substrates popularly used for the enzyme immunoassay with HRP, was also examined as substrate of the CCP, and it was confirmed that CCP produced from the genome of T4ccp had the DAB-oxidizing activity (data not shown). The color change in the assay solution could be clearly perceived by eyes. However, on using DAB for the detection of E. coli in the culture containing the P. aeruginosa NBRC 12689, undesirable oxidation of DAB was caused by P. aeruginosa NBRC12689, implying that DAB cannot be used for the detection of E. coli in samples that contain other bacteria. Therefore, a combination of CCP and cytochrome c is considered to be appropriate for the colorimetric detection of E. coli K12 using a recombinant phage.

In this study, detection of the E. coli at concentrations of 1 × 107 and 1 × 106 CFU mL−1 was also examined, and the change in A550 nm could be detected at 1 × 107 CFU mL−1, while, in case of E. coli K12 at a concentration of 1 × 106 CFU mL−1, the color change in the assay could not be detected because of insufficient CCP activity (data not shown). We consider that not only how the detectable concentration of E. coli is low but how long the detection takes is also important. Namely, even though the low concentration of E. coli cannot be detected by a detection method, if the detection procedure itself is rapid enough, the method that requires the precultivation to complement the low sensitivity must be able to detect the E. coli as rapid as the methods that have the high sensitivity but take long time for the detection. It can be considered that fluorescent- and bioluminescent-based detection methods are more sensitive than colorimetric-based detection method. And, the bioluminescent phage-based and the fluorescent phage-based methods are supposed to be able to detect lower concentrations of E. coli. Therefore, these methods are considered to take shorter time of the precultivation (in case the precultivation is required) than the present method. While the detection limit of the present method is 10CFU mL−1, it was reported that in the bioluminescent-based detection method, an extremely low concentration of E. coli of 103 CFU mL−1 can be detected (Ripp et al., 2008). It means the bioluminescent-based detection method is much superior in the sensitivity; however, the time required for the detection itself was reported to be 5 h in the same reference. In addition, when the E. coli concentration is 101 CFU mL−1, after increasing the concentration of E. coli by the precultivation for 6 h (though the concentration of E. coli after the precultivation was not described, it can be deduced to be around 3 × 106 CFU mL−1 by assuming the mean generation time of E. coli as 20 min), time required for the detection itself was 2.5 h, and the E. coli can be detected in the total time of 8.5 h including the time required for the precultivation. If this time is compared with that of the present method, the time required for the precultivation to increase E. coli concentration from 101 to 107 CFU mL−1 is about 7 h, and the detection itself take around 1 h in the present method; therefore, the time required for the whole detection is about 8 h, which is comparable with the time required for the bioluminescent-based detection. On the other hand, as to the fluorescent phage-based method, it took 10 h for precultivation and 2 h for detection step for detection of E. coli concentration of 101 CFU mL−1 (Goodridge et al., 1999b). Therefore, it took 12 h for the whole detection of E. coli at the concentration of 10CFU mL−1 that was longer than that of the present method. The long time for the whole detection in the fluorescent phage-based method is due to the long precultivation time of 10 h, and the reason for the long precultivation time was not clear. However, similarly to the bioluminescent method, if the fluorescent phage-based method also requires about 3 × 106 CFU mL−1 of E. coli for the detection in 2 h, in other words, if the fluorescent phage-based method also requires 6 h of precultivation, the time for the whole detection would be around 8 h. Thus, by considering not only the sensitivity but also the time required for the detection itself, in the comparison of the whole detection, the present method is as enough applicable as the bioluminescent and fluorescent phage-based methods. In addition, the present method is simple as the detection can be accomplished by eyes without the need for specific apparatus. Therefore, the present method enables the detection of E. coli easily with the total time for the detection comparable with those of previous phage-based methods.

In this study, T4 phage was used to construct the recombinant phage having the ccp gene to prove the concept of the detection method as it is thoroughly understood. The host range of T4 phage is narrow as previously reported by Miyanaga et al. (2006) that T4 phage infected only about 8% of total E. coli strains in the sewage influent, and it might be insufficient to be used for the detection of fecal contaminants in the water and food samples. Therefore, practically, several recombinant phages having the ccp gene may need to be constructed, and a cocktail of those recombinant phages would enable the detection of fecal contaminant in the water and food samples. Namura et al. (2008) showed that a cocktail of two T4-related recombinant phages with the broad host rage could infect more than 50% of total E. coli strains in the sewage water. Therefore, usage of an appropriate cocktail of several recombinant phages with broad host range can be considered as promising for the detection of fecal contaminants in the water and food samples.

Ancillary