- Email: [email protected]
RNA biosensor for the rapid detection of viable Escherichia coli in drinking water
Biosensors and Bioelectronics 18 (2003) 405 /413 www.elsevier.com/locate/bios
RNA biosensor for the rapid detection of viable Escherichia coli in drinking water Antje J. Baeumner *, Richard N. Cohen, Vonya Miksic, Junhong Min Department of Biological and Environmental Engineering, Cornell University, Ithaca, NY 14853, USA Received 13 December 2001; accepted 31 July 2002
Abstract A highly sensitive and specific RNA biosensor was developed for the rapid detection of viable Escherichia coli as an indicator organism in water. The biosensor is coupled with protocols developed earlier for the extraction and amplification of mRNA molecules from E. coli [Anal. Biochem. 303 (2002) 186]. However, in contrast to earlier detection methods, the biosensor allows the rapid detection and quantification of E. coli mRNA in only 15 /20 min. In addition, the biosensor is portable, inexpensive and very easy to use, which makes it an ideal detection system for field applications. Viable E. coli are identified and quantified via a 200 ntlong target sequence from mRNA (clpB ) coding for a heat shock protein. For sample preparation, a heat shock is applied to the cells prior to disruption. Then, mRNA is extracted, purified and finally amplified using the isothermal amplification technique Nucleic acid sequence-based amplification (NASBA). The amplified RNA is then quantified with the biosensor. The biosensor is a membrane-based DNA/RNA hybridization system using liposome amplification. The various biosensor components such as DNA probe sequences and concentration, buffers, incubation times have been optimized, and using a synthetic target sequence, a detection limit of 5 fmol per sample was determined. An excellent correlation to a much more elaborate and expensive laboratory based detection system was demonstrated, which can detect as few as 40 E. coli cfu/ml. Finally, the assay was tested regarding its specificity; no false positive signals were obtained from other microorganisms or from nonviable E. coli cells. # 2002 Elsevier Science B.V. All rights reserved. Keywords: E. coli ; RNA biosensor; NASBA; Rapid detection; Optical detection; Specific detection; Sensitive detection
1. Introduction The detection of pathogenic organisms in water and foods remains a challenging and important problem, since the safety of water and food supplies has to be ensured. Worldwide, infectious diseases account for nearly 40% of the estimated total 50 million annual deaths (Ivnitski et al., 1999). It is assumed that food borne diseases cause approximately 76 million illnesses, 325 000 hospitalizations, and 5000 deaths in the US each year (Mead et al., 1999). Conventional methods for the detection of bacterial pathogens typically involve culturing the organism in selective media and identifying isolates according to their morphological, biochemical, and/or immunological characteristics. While these meth-
* Corresponding author. Tel.: /1-607-255-5433; fax: /1-607-2554080 E-mail address: [email protected] (A.J. Baeumner).
ods have low limits of detection and can be used in complex food and sample matrices, they typically require days from initiation to readout. In addition, the interpretation of results is often very difficult (Sheridan et al., 1998). A number of new immunological and nucleic acidbased detection systems have been suggested which ameliorate the interpretation problems associated with biochemical assays. Nucleic acid based sensors, also often referred to as genosensors or gene probes, play an increasingly important role in the detection of pathogenic organisms for health care, environment monitoring and food safety (Ivnitski et al., 1999). Most of the gene-probe systems focus on the detection of specific DNA or rRNA sequences and utilize PCR or Reverse transcription-PCR (RT-PCR) for signal amplification (Hu et al., 1999; Spierings et al., 1993; Green et al., 1991; Rice et al., 1995; Wang et al., 1997). While some researchers have suggested the detection of only viable
0956-5663/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 5 6 - 5 6 6 3 ( 0 2 ) 0 0 1 6 2 - 8
406
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
cells via their ribosomal RNA (rRNA) (Jeffrey et al., 1994; Sheridan et al., 1998; van der Vliet et al., 1994), neither DNA nor rRNA are appropriate indicators of viability (Uyttendaele et al., 1997). However, viability is an important aspect when measuring for pathogenic organisms. In contrast to DNA and rRNA, mRNA molecules are degraded rapidly in living bacterial cells. Most mRNA species have a half-life of only a few minutes, because of their rapid degradation by enzymes (RNase), which are very stable even in harsh environments (Sheridan et al., 1998; Sela et al., 1957). Thus, if only viable pathogenic organisms are to be detected, specific mRNA sequences represent an excellent target molecule for the detection of viable pathogens. RT-PCR has been used to measure mRNA for confirming an organism’s viability (Bej et al., 1991; Patel et al., 1993). However, despite potential advantages, RT-PCR is difficult to exploit, because of the complexity of the method to measure low levels of intact mRNA from a few viable cells and small numbers of DNA molecules can produce false positive signals. nucleic acid sequence-based amplification (NASBA) can be used to achieve mRNA-based technology as an alternative method. NASBA has several advantages over RT-PCR. It exclusively amplifies RNA rapidly and isothermally. In addition, DNA cannot be used as template in NASBA. The presence of DNA samples, therefore, does not represent a potential for false positive signals. Lastly, since NASBA is isothermal, a heating block or water bath can be used instead of specialized expensive equipment (Baeumner et al., 2001; Simpkins et al., 2000; Johnson et al., 1995), which makes NASBA the better amplification tool for our application. We have chosen Escherichia coli as our model organism for the development of RNA biosensors, since it is an important analyte for food and water safety. E. coli is a natural inhabitant of the intestinal tract of
humans and warm-blooded animals. As it is found regularly in all animals capable of harboring Samonella and Shigella , the presence of this bacterium in foods or water indicates that fecal contamination may have occurred and consumers might, therefore, be exposed to enteric pathogens. Thus, E. coli is often used as an indicator organism in water sample analyses, because it is specific and most reliably reflects fecal origin (Gauthier and Archibald, 2001). We describe in this paper the development of a field-usable RNA biosensor for the specific, sensitive and rapid detection of viable E. coli in water. Highly specific DNA probes hybridize with an E. coli mRNA sequence that was amplified using the isothermal NASBA technique. The biosensor assay has been optimized, including DNA probe sequences, probe concentrations, and hybridization conditions. The biosensor’s limit of detection and specificity has been determined and the resulting signals correlated with a laboratory-based detection system for E. coli developed earlier in our lab. The biosensor assay principle is shown in Fig. 1.
2. Materials and methods All strains used in this study were provided by Dr Randy Worobo, Cornell University, Geneva, NY, which were routinely grown on Tryptic Soy broth (TSB) or agar at 37 8C as appropriate. Cells in growth phase were chosen as samples. RNA release, extraction, NASBA amplification were carried out using the Rneasy† Mini kit and RNampliFireTM (Qiagen, Valencia, CA). Two hundred proof ethanol, molecular biological grade water, and buffer reagents were obtained from Sigma Company. Lipids were purchased from Avanti Polar Lipids, Alabaster, AL. Sulforhodamine B, streptavidin etc. were purchased from Molecular Probes Company, Eugene, OR. Membranes were
Fig. 1. Biosensor assay principle. A DNA capture probe is immobilized on a polyethersulfone membrane. A DNA reporter probe is coupled to the surface of a liposome. When a specific E . coli mRNA is present (Figure A), a sandwich is formed between capture probe, mRNA and reporter probe, thus, liposomes are captured in the capture/detection zone. The number of liposomes is directly proportional to the amount of E . coli mRNA present. In Figure B it is shown that liposomes are not captured in the detection zone, when a nonspecific RNA molecule is present. Thus, no signal will be reported in the detection zone.
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
obtained from Pall/Gelman Company, Port Washington, NY. The BR-10 reflectometer was purchased from ESECO (l/560 nm). Cushing, OK. All oligonucleotides were purchased from the Biotechnology Center, Cornell University, Ithaca, NY. Heat shock procedure, nucleic acid extraction, and purification E. coli cells were grown to mid-exponential growth phase in TSB solution. For stimulation of heat shock mRNA production in E. coli , the cells were heated for 5 min at 41 8C in a heating block. The cells were disrupted by sonication (1 min at 20 W) and by incubating at 60 8C for 30 min in lysis buffer. RNA extractions were carried out using E. coli concentrations between 0 and 1 000 000 cfu/ml. The nucleic acid was purified using a Silica-guanidine thiocyanate protocol (Boom et al., 1990), provided as proprietary lysis buffer and isolation reagent modules in the RNeasyTM. The standard bacterial protocol recommended by the manufacturer was followed. Primers and probes DNA oligonucleotides for NASBA amplification (primers) and a reporter probe (for detection) had been optimized in earlier experiments (Min and Baeumner, 2002). Additionally, a variety of different oligonucleotides were available as capture probes in the biosensor to optimize sensitivity and specificity of the assay (all sequences are outlined in Table 1). The reporter probe was labeled with an amine group at the 3? end, and all capture probes were modified with a biotin moiety at the 5? end. All probes were obtained desalted and lyophilized. NASBA amplification of mRNA reactions were carried out using reagents and protocols included in the RNampliFireTM kit. A final volume of 20 ml was used containing 70 mM of KCl, 0.2 mM of each primer, 5 ml of sample and 5 ml of enzyme mix solution (RNase H, T7 polymerize and AMV-RT). Before addition of the enzyme mix, primers were annealed to the target at 65 8C for 5 min, and the mixture then cooled down to 41 8C. The final solution was incubated at 41 8C for 90 /120 min to allow exponential amplification of the target mRNA. A positive control was created that was used to determine the quality of the amplification
reaction. It consisted of a known positive E. coli NASBA amplification reaction that was diluted 1:200 in water and stored at /20 8C. Five microlitre of this solution was used as positive control in every set of NASBA reactions. For a negative control, water was added to the NASBA reaction instead of a cell extract. This control allowed the determination of possible sample cross contamination during a NASBA reaction. 2.1. Liposome preparation and nucleic acid coupling Liposomes were prepared following the protocol published in Siebert et al. (1993). In brief, liposomes were prepared using the reversed-phase evaporation method. One hundred to 150 mM Sulforhodamine B was dissolved in potassium phosphate buffer, pH 7.5 and entrapped in the liposomes. Subsequently, the liposomes were extruded at elevated temperatures through polycarbonate filters (3, and 0.4 mm, nine times each) for sizing using a mini extruder (Avanti Polar Lipids, Alabaster, AL) and purified away from unentrapped dye by gel filtration using Sephadex G50 columns and dialysis (MWCO 12/15 000) against potassium phosphate buffer with an osmolarity 50/100 mmol/kg higher than the osmolarity of the encapsulant solution. The osmolarity typically was adjusted using NaCl and sucrose. For the coupling of reporter probes to the liposome surface, a lipid tagged with a maleimide group (N -(4-(pmaleimidylmethyl)cyclohexane-1-carbonyl)-Diphosphatidyl palmitoyl ethanolamine [MMCC-DPPE]) was used during the liposome preparation and thus incorporated into the lipid bilayer. Typically, a 3 /4 mol% concentration of the overall amount of lipids was used. After liposomes were prepared and purified using gel filtration and dialysis, activated oligonucleotides were coupled to the maleimide groups available on the outer surface of the liposomes. First, the reporter probe (bearing an amine group at the 3? end) was derivatized with a sulfhydryl group. The probe was then dissolved in 0.05 M potassium phosphate buffer, pH 7.8 containing 1 mM EDTA to a concentration of 300 nmol/ml. About
Table 1 DNA sequences of oligonucleotides (NASBA primers and biosensor probes) Name
DNA sequence (5? /3? orientation)
Primer 1a Primer 2 Reporter probe Capture probe 1 (CP1) Capture probe 2 (CP2) Capture probe 3 (CP3) Capture probe long 1 (CPL1) Capture probe long 3 (CPL3)
AattctaatacgactcactatagggAAGGTTACTGGACGGCGACAA AAA TCC ACA TIT CTg ACg A GTC Tgg TgA ATT ggT TCC g CCg TTg gCA CAg CAA ATA gTC AAA CTg CTg AgC gAg AA CCT ggA AgT TAA TgA AgA CCg CgA AAA CCC gTT ggC ACA gCA AAT A TCg CCT ggA AgT TAA TCA AgA CCg gAT
a
407
T7 promoter sequence at 5? end (lower case) followed by a 4-nt GA-rich sequence.
408
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
30 nmol of probe were mixed with 90 nmol of N succinimidyl-S -acetylthioacetate (SATA) dissolved in DMSO. This mixture was incubated at room temperature for 1.5 h. Second, the thiol was deprotected in a deacetylation step. A fresh solution of 0.5 M hydroxylamine hydrochloride, 25 mM EDTA in 0.1 M potassium phosphate buffer with pH adjusted to 7.5 was prepared. The amount of solution added to the reporter probe mixture was 1/10 that of the probe solution volume. The mixture incubated for 2 h at room temperature. Third, the maleimide-derivatized liposomes were conjugated with SH-tagged reporter probes. In order to obtain the required surface density of reporter probe, the appropriate amount of liposomes was added to the probes (for example for a 0.4 mol% tag, 10.5% of liposomes was used). The pH of the reporter probe solution was adjusted to seven and the osmolarity adjusted to the appropriate value. Subsequently, the two solutions were mixed and then incubated at room temperature for 3 h and then overnight at 4 8C. In the final step, ethylmalcimide and Tris were added to quench the excess SH groups on the reporter probes and the unreacted maleimide groups on the liposomes, respectively. The liposomes were then purified from unreacted reporter probe on a Sepharose CL-4B column and finally dialyzed against the appropriate buffer overnight in the dark. The liposomes were stored at 4 8C in the dark. Different concentrations of reporter probes were tagged to the liposomes, ranging from 0.05 to 1 mol% tag. The liposome concentration of the liposome stock solution varied between the different preparations due to variations in the liposome yield after evaporation and differences in the gel filtration process. Thus, the amount of liposomes per assay was evaluated using a spectrophotometer (DU520Beckman, Fullerton, CA) at a wavelength of 570 nm and adjusted so that every assay contained the same amount of liposomes. 2.2. Membrane preparation The membranes were cut into strips of 4.5 /80 mm. Subsequently, the membranes were coated with a mixture of streptavidin and biotinylated capture probes. A mixture containing 15 pmol streptavidin and 45 pmol capture probe per ml in a sodium carbonate buffer (0.4 M NaHCO3/NA2CO3 with 5% methanol) was incubated for at least 15 min at room temperature. The streptavidin-capture probe mixture was immobilized on the membrane strips by pipetting 1 ml of the mixture directly onto the membrane, approximately 2.5 cm from the bottom. The membranes were dried for 5 min at room temperature and then for an additional 1.5 h in a vacuum oven (15 psi) at 52 /55 8C. Subsequently, the membranes were incubated in a blocking solution of
0.5% polyvinylpyrrolidone, 0.015% casein in Tris buffered saline (TBS: 20 mM Tris, 150 mM NaCl, 0.01% NaN3, pH 7 /7.5) for 30 min. The membranes were blotted dry and finally dried in the vacuum oven (15 psi) at 30 8C for 2 h. They were stored in vacuum-sealed bags at 4 8C until use.
2.3. Biosensor assay format A lateral flow assay was developed. First, liposomes, target sequence (amplicon or synthetic target sequence) and a hybridization solution (master mix) were incubated for 10 min at 41 8C. The mixture was pipetted onto the membrane strip about 8 mm from the center of the capture zone (toward the front of the strip). Immediately afterwards, 30/40 ml of running buffer were added to the front of the strip. After all of the running buffer traversed along the strips (/10 min), the capture zones were analyzed with the BR-10 reflectometer by placing the capture zone directly under the reflectometer opening or visually evaluated. The reflectometer measures the reflectance of light at a wavelength of 560 nm, which corresponds to the absorbance maximum of sulforhodamine B that is encapsulated in the liposomes. The reflectometer is calibrated about once a week using a white and magenta paint chip to a minimum value of zero arbitrary units and a maximum of 155 arbitrary units. Both solutions, running buffer and master mix contained formamide, SSC (1 /SSC contains 15 mM Sodium citrate and 150 mM NaCl, pH 7.0). Ficoll type 400, sucrose and Triton X-100. The concentrations of each compound were varied for optimization of the biosensor assay. Additionally, different hybridization conditions, as well as liposome and target sequence concentrations were investigated. Optimal conditions were: 2 ml of master mix (40% formamide, 9/SSC, 0.6 M sucrose, 0.6% Ficoll type 400) were mixed with 4 ml of liposomes and 4 ml of target sequence and incubated for 10 min at 41 8C. Thirty to 40 ml of running buffer (25% formamide, 6/SSC, 0.2 M sucrose, 0.2% Ficoll type 400, 0.01% Triton X-100) were applied to the membrane. For optimization of the reporter probe tag concentration on the liposome surface, a dipstick type assay was carried out (i.e. vertical assay). About 7 ml liposomes (volume adjusted depending on their concentration), 2 ml of synthetic target sequence and 51 ml of running buffer (25% formamide, 3 /SSC, 0.2 M sucrose, 0.2% Ficoll type 400) were mixed and incubated at 41 8C for 10 min in a glass tube. Subsequently, a membrane strip was inserted into the glass tube and the solution allowed to migrate up the strip. After approximately 8 min, the assay was complete (i.e. the solution reached the top of the membrane strip).
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
2.4. Synthetic target sequence A 37 nt long synthetic oligonucleotide was made representing a portion of the target sequence, the clpB mRNA of a heat shock protein, to which the E. coli specific detection probes hybridize. It consisted of the capture probe one sequence adjacent to the reporter probe sequence (Cgg AAC CAA TTC ACC AgA CTA TTT gCT gTg CCA ACg g). The lyophilized oligonucleotide was dissolved in water (DNase and RNase free) and stored frozen at /20 8C. Solutions were prepared with concentrations ranging from 0.1 to 10 000 fmol/ml. Comparison of biosensor with ECL detection Synthetic E. coli target sequence was used for the comparison of the biosensor signal with the previously established electrochemiluminescence (ECL) detection of E. coli in water samples (Min and Baeumner, 2002). In brief, 5 ml of target sequence or 1:10 in H2O diluted NASBA amplicon were mixed with 10 ml of magnetic bead labeled capture probe and 10 ml of ECL labeled reporter probe (Baron Consulting Co., Milford, CT). The mixture was incubated for 30 min at 41 8C with 5 s, vortexing every 10 min. Subsequently, 300 ml of hybridization buffer (supplied by Organon Teknika, Boxtel, NL) were added and the samples loaded into the autosampler of the NucliSense Reader for detection and quantification.
3. Results and discussion 3.1. Capture probe optimization
409
0.05, 0.1, 0.2, 0.4, 0.6, 0.8 and 1% of the total amount of lipids per liposome. Vertical assays were performed and each liposome preparation was analyzed in triplicate. Bound liposomes in the capture zones were quantified using the reflectometer (at l/560 nm). Fig. 2 shows the results of assays using 2 pmol of synthetic target sequence per assay. An optimum is observed at a reporter probe concentration of 0.4% tag, however, statistical analysis using a two-sample t-test shows that it is only 70% certain that the value at 0.4% tag is higher than the value at 0.6% tag. Thus, a concentration between 0.4 and 0.6% is found to be optimal for the analysis of target sequences in the range of 2 pmol per sample. Subsequently, a variety of different synthetic target sequence concentrations (20, 200, 2, and 20 pmol) were analyzed using liposomes with 0.05, 0.4, 0.6, 0.8 and 1% tag. The results are shown in Fig. 3. It is apparent that for lower target sequence concentrations, a lower percentage of tag on the liposomes is favorable (i.e. for 20 fmol the 0.05% tag results in the highest signals). In addition, for concentrations of 200 fmol per assay and above, a tag surface density of 0.4 /0.6% is optimal. The data point at 200 fmol target sequence and 0.6% tag liposomes is identified as an outlier and will be further investigated in the future with additional batches of liposomes. Future experiments described in this publication were carried out using 0.4% reporter probe tag on the liposome surface. However, when samples containing low concentrations of target sequence would be analyzed, liposomes with a 0.05% tag are suggested to be more appropriate.
In order to optimize the biosensor assay, six different capture probe sequences, immobilized on the polyethersulfone membrane, were investigated. E. coli clpB mRNA was amplified using NASBA and then applied to the six biosensor assays. Liposomes, master mix and running buffers remained constant in all experiments. All capture probes had been designed previously to have a similar degree of potential secondary structures (Min and Baeumner, 2002). It was found that under given conditions, capture probe one resulted in the strongest signals. Signals obtained with capture probe one was about twice as strong as those from capture probes 2 and 3 and about four times as strong as for any of the long capture probes (LCP1 and LCP3). While the longer capture probes (25 /27 nts) would have made a more specific detection assay, their slower hybridization kinetics resulted in the significantly lower signals. 3.2. Optimization of reporter probe tag concentration The concentration of reporter probe tags on the liposome surface was optimized. Seven different types of liposomes were prepared with tag concentrations of
Fig. 2. Optimization of reporter probe tag on liposomes. Synthetic target sequence (2 pmol per assay) was analyzed with the different biosensor assays in which the number of reporter probes on the liposome surface was varied. Reflectometer signals are plotted vs. the concentration of reporter probes. Background signals from negative controls (i.e. water instead of target sequence was used in the biosensor assay) are subtracted.
410
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
Fig. 3. Investigation of different liposome-reporter probe concentrations with different synthetic target sequence concentrations. Reflectometer signals are plotted vs. the concentration of reporter probes. Background signals from negative controls (i.e. water instead of target sequence was used in the biosensor assay) are subtracted.
3.3. Optimization of biosensor assay Polyethersulfonyl membranes were blocked against nonspecific liposome binding following a previously optimized protocol (data not published). Typically, only minimal nonspecific binding of liposomes to the membrane was observed which resulted in reflectance signals for negative samples between 5 and 18. The lateral flow biosensor assay hybridization buffers and incubation conditions were optimized. Presence of Ficoll type 400 in both the running buffer and master mix enhanced the hybridization of probes and target sequence. Concentrations higher than 0.6% in the master mix, however, yielded increased liposome aggregation. Sucrose was added for liposome stability. Concentrations lower than 0.6 M in the master mix and 0.2 M in the running buffer yielded lower signals. Also, it was found that the lateral flow assay would result in higher signals, if a pre-incubation of the liposomes and target molecules was carried out at 41 8C for 10 min. Formamide and SSC concentrations had the most significant effect on the hybridization assay. Concentrations lower than 9 / in the master mix and 6 /SSC in the running buffer were too stringent and caused less hybridization to occur. Also, higher concentrations resulted lower signals, which is probably due to secondary structure formation of probes and target sequence. In Fig. 4, the dependency of the biosensor signal on the formamide concentration in the master mix is shown. Two different assays were run, one using 5 ml of hybridization mixture (2 ml liposomes, 2 ml E. coli
Fig. 4. The effect of formamide concentration in the master mix. Reflectometer readings are plotted vs. the formamide concentration. Two different volumes of hybridization mixture are investigated 10 ml (4 ml liposomes, 4 ml E. coli amplicon and 2 ml master mix) and 5 ml (2 ml liposomes, 2 ml E. coli amplicon and 1 ml master mix).
amplicon (clpB mRNA amplified by NASBA) and 1 ml master mix) and one using twice the volume (i.e. 10 ml E. coli amplicon) is used in this experiment instead of synthetic target sequence. Higher formamide concentration increases the biosensor signal, thus the hybridization between DNA probes and E. coli target sequence. It is assumed that higher formamide concentration decreases the likelihood of secondary structures of all nucleic acid sequences involved and thus enhances hybridization. In the case of the 10 ml reaction volume, a slight decrease can be observed at a formamide concentration of 50%. Therefore, 40% formamide concentration in the master mix was chosen for all future experiments. In addition, the formamide concentration in the running buffer was investigated. While signals at higher formamide concentration (50%) were increased over those obtained using lower concentrations (10 and 25%), a significant fading of the signal was observed upon drying of the membrane strips (typically over night). At the high formamide concentrations liposomes probably lysed over night, released the entrapped dye, which then continued to migrate along the slightly wet membrane strip. This effect was much less obvious at 25% formamide concentration and not detectable at 10%. Thus, 25% formamide concentration was chosen as optimal for the running buffer since the signal was higher than with 10% and the fading was much less significant than with 50%. (While the membranes are typically investigated directly after completion of the assay, it is advantageous to have clear signals readable after the membranes are dried and stored for data collection.)
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
411
3.4. Determination of the detection limit The limit of detection was determined using the synthetic target sequence. A standard curve analyzing concentrations of 0.1 /5000 fmol per assay is shown in Fig. 5. The lateral flow biosensor assay was used with 0.4% reporter probe tag on the liposomes, capture probe 1 and optimized running buffer and master mix. An excellent detection limit of only 5 fmol was determined. A linear range of the standard curve is obtained between 5 and 1000 fmol of target sequence per assay. Above 1000 fmol a ‘hook effect’ is determined which is typical for an affinity assay with a sandwich format. Also, due to the good correlation of E. coli biosensor and ECL assay (see below), the detection limit of 40 cfu/ ml determined previously with the ECL assay (Min and Baeumner, 2002), can be applied as well to the biosensor assay. An ECL value of 87 860 was obtained for 40 cfu/ ml which would correlate to a biosensor value of about 40 which is well above the background signal of approximately 18 (see Fig. 5) and within the linear range of the biosensor’s dose response curve.
3.5. Correlation between ECL and biosensor values As described previously, a laboratory-based detection system for E. coli was developed which was based on an electrochemiluminescence detection system using the NucliSense Reader for quantification. It was shown in our laboratory that as few as 40 E. coli cells could be detected (Baeumner et al., 2001). Since the E. coli biosensor can be used in on-site applications, it was important to obtain a good correlation between the
Fig. 5. Determination of the limit of detection for the biosensor assay. Different concentrations of synthetic target sequence were analyzed under optimized conditions using 0.4% tag liposomes. Data are presented in a double log scale plot. Due to the log-scale presentation of the data, the negative control (water instead of target sequence used in the biosensor assay) is indicated as 0.01 fmol target sequence.
Fig. 6. Comparison of the biosensor signals with a laboratory-based electrochemiluminescence (ECL) detection system developed earlier. Different concentrations of synthetic target sequence (from 5 fmol to 5 pmol per assay) were analyzed in both detection systems. Biosensor results are correlated to the corresponding ECL signals. An r2 of 0.9361 was obtained.
biosensor and a laboratory-based ECL instrument. Synthetic target sequences were analyzed with the biosensor and with the ECL system (Fig. 6). Samples ranging from 5 fmol to 5 pmol were analyzed with both systems and a correlation coefficient (i.e. r2) of 0.9631 was obtained. The results demonstrate that the biosensor is as sensitive as the ECL system and that it has a similar dynamic range. 3.6. Analysis of assay specificity for viable E. coli The specificities of a nucleic acid amplification and hybridization system is determined by the specificities of primer and probe sets used for amplification and detection. While Gene Data Bank analysis using BLAST showed that theoretically no other organism can be detected with the biosensor assay system, a variety of microorganisms were investigated in our lab to ensure the specificity of this assay for E. coli. Microorganisms were chosen that is known to be closely related to E. coli or are potentially present in untreated domestic wastewater (i.e. Shigella sonnei , Salmonella typhimurium , Pseudomonas aeruginosa , Bacillus cereus ). Fresh cultures of all listed organisms were prepared by growing them to mid exponential phase (3 /3.5 h). The cells were heat shocked, the RNA extracted and subjected to amplification with NASBA. Subsequently, all samples were analyzed using agarose gel electrophoresis (2% agarose, in 1/TAE [0.04 Tris /acetate, 0.001 M EDTA (pH 8.0)], 30/60 min, 90 V, stained with SYBR Green H (1:10 000 diluted in water)) and the biosensor assay. Agarose gel electrophoresis (Fig. 7) was also used to confirm the successful RNA extraction from all micro-
412
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
Fig. 7. Investigation of the biosensor specificity. Four different microorganisms and E. coli cells were grown to mid exponential phase and then heat shocked. RNA was extracted from the cells and amplified using NASBA. The amplicons were investigated with a 2% agarose gel. The lanes were loaded with the following amplicons, lanes 2 and 3 with S. sonnei ; lane 4 and 5 P aeruginosa ; lane 6 and 7 with S . typhimurium ; lanes 8 and 9 with B. cereus ; lanes 10 and 11 with E . coli ; lane 12 with the positive control; lane 13 with the NASBA negative control; lanes 1 and 14 with molecular weight markers. The arrow points to an RNA marker of 200 nt.
organisms. Clear bands at about 200 nt length can be seen in lanes 9, 10 and 11. Lanes 9 and 10 are NASBA reactions using E. coli cells, lane 11 shows the NASBA positive control. No bands are observed from any of the other microorganisms. The same results were obtained from the biosensor analysis (Fig. 8) where E. coli samples and positive controls gave signals between 71 and 95, while all other organisms had values close to that of the negative control. Thus, no false positive signal was obtained when other microorganisms were analyzed using the biosensor.
Fig. 8. Investigation of the biosensor specificity. Four different microorganisms and E. coli cells were grown to mid exponential phase and then heat shocked. RNA was extracted from the cells and amplified using NASBA. The amplicons were subsequently analyzed in the biosensor assay in duplicates under optimized conditions. The numbers correspond to the following samples, #1 and 2, S. sonnei ; #3 and 4, P . aeruginosa ; #5 and 6, S . typhimurium ; #7 and 8, B. cereus ; #9 and 10, E . coli ; #11, positive control; #12, negative control (H2O).
Secondly, viable E. coli cells were investigated versus nonviable E. coli cells. Since the clpB mRNA can only be produced by viable cells, no false positive signal should be obtained from dead E. coli . This was proven when E. coli cells (10 000 per sample) were killed (by boiling at 98 8C for 15 min with vortexing every 5 min for 5 s) and subsequently analyzed with the biosensor (Fig. 9). No positive signal was obtained from dead cells, i.e. the biosensor and ECL signals were lower than
Fig. 9. Analysis of dead and viable E. coli cells. Each sample contained 10 000 cells. The cells were subjected to heat shock, RNA extraction and amplification and finally analyzed in the biosensor. For dead samples, cells were killed by heating to 98 8C for 15 min. Each condition was analyzed nine times with the biosensor (i.e. three samples, analyzed in triplicates) and in triplicates with the ECL method (i.e. three samples, each analyzed once in the ECL machine). Positive controls (E. coli mRNA) and negative controls (H2O) were analyzed in triplicates.
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
those obtained from samples containing water instead of cells (negative controls), but highly positive from viable cells (10 000 per sample). Thus, the biosensor is specific for viable E. coli.
4. Conclusion We have presented in this study the development of a highly specific and very sensitive biosensor for the rapid detection of viable E. coli in water. The biosensor can detect as few as 40 viable E. coli in water and is rapid and very simple to use. No false positive signals were obtained from other microorganisms or nonviable E. coli cells. An excellent correlation to a laboratory-based detection system was demonstrated. In the future, we will investigate several environmental samples from different sources in order to determine matrix effects on the biosensor. We will also investigate larger sample volumes and combine the detection assay with a filtration step in order to process volumes up to 1 l and thrive toward the optimization of the NASBA reaction in order to detect as few as one E. coli cell per 100 ml as the ultimate detection limit.
Acknowledgements The authors acknowledge financial support for this project from Innovative Biotechnologies International, Inc., and the Cornell University Hughes Scholar Program. The authors also want to thank Randy Worobo for helping with some of the microbiological aspects of the work and Richard Durst and Richard Montagna for their thoughts, which helped move the project forward.
References Baeumner, A.J., Humiston, M.C., Montagna, R.A., Durst, R.A., 2001. Detection of viable oocysts of Cryptosporidium parvum following nucleic acid sequence based amplification. Anal. Chem. 73, 1176 / 1180. Bej, A.K., Ng, W.Y., Mogan, S., Jones, D., Mabbubani, M.H., 1991. Detection of viable Vibrio cholerae by reverse-transcriptase polymerase chain reaction (RT-PCR). Mol. Biotechnol. 5, 1 /10. Boom, R., Sol, C.J.A., Salimans, M.M.M., Jansen, C.L., Wertheimvan-Dillen, P.M.E., van der Noordaa, J., 1990. Rapid and simple method for purification of nucleic acids. J. Clin. Microbiol. 28, 495 /503. Gauthier, F., Archibald, F., 2001. The ecology of ‘fecal indicator’ bacteria commonly found in pulp and paper mill water systems. Water Res. 35, 2207 /2218.
413
Green, D.H., Lewis, G.D., Rodtong, S., Loutil, M.W., 1991. Detection of fecal pollution in water by an Escherichia coli uidA -gene probe. J. Microbiol. Methods 13, 207 /214. Hu, Y., Zhang, Q., Meitzer, J.C., 1999. Rapid and sensitive detection of Eshcerichia coli O157:H7 in bovine faeces by a multiplex PCR. J. Appl. Microbiol. 87, 867 /876. Ivnitski, D., Abdel-Hamid, I., Atanasov, P., Wilkins, E., 1999. Biosensors for the detection of pathogenic bacteria. Biosens. Bioelectron. 14, 599 /624. Jeffrey, W.H., Nazaret, S., von Haven, R., 1994. Improved method for recovery of mRNA from aquatic samples and its application to detection of mer expression. Appl. Environ. Microbiol. 60, 1814 / 1820. Johnson, D., Pieniazek, N.J., Griffin, D.W., Misener, L., Rose, J.B., 1995. Development of a PCR protocol for sensitive detection of Cryptosporidium oocysts in water samples. Appl. Environ. Microbiol. 61, 3849 /3855. Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Shapiro, C., Griffin, P.M., Tauxe, R.V., 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5 (5), 607 /625. Min, J.H., Baeumner, A.J., 2002. Highly sensitive and specific detection of viable Escherichia coli in drinking water. Anal. Biochem. 303, 186 /193. Patel, B.K.R., Banjerjee, D.K., Butcher, P.D., 1993. Determination of Myeobacterium leprac viability by polymerase chain reaction amplification of 71-kDa heat shock protein mRNA. J. Infect. Dis. 168 (3), 799 /800. Rice, E.W., Covert, T.C., Johnson, S.A., Reasoner, D.J., 1995. Detection of Escherichia coli in water using a colormetric gene probe assay. J. Environ. Sci. Health A30, 1059 /1067. Sela, M., Anfinsen, C.B., Harrington, W.F., 1957. The correlation of ribonuclease activity with specific aspects of tertiary structure. Biochem. Biophys. Acta 26, 502. Sheridan, G.E.C., Masters, C.I., Shallcross, J.A., Mackey, B.M., 1998. Detection of mRNA by reverse transcription */RCR as an indicator of viability in Escherichia coli cells. Appl. Environ. Microbiol. 64, 1313 /1318. Siebert, S.T.A., Reeves, S.G., Durst, R.A., 1993. Liposome immunomigration field assay device for Alachlor determination. Anal. Chim. Acta 282, 297 /305. Simpkins, S.A., Chan, A.B., Hays, J., Popping, B., Cook, N., 2000. An RNA transcription-based amplification technique (NASBA) for the detection of viable Salmonella enterica . Lett. Appl. Microbiol. 30, 75 /79. Spierings, G., Ockhuijsen, C., Hofstra, H., Tommassen, J., 1993. Polymerase chain reaction for the specific detection of Escherichia coli /Shigella . Res. Microbiol. 144, 557 /564. Uyttendaele, M., Bastiaansen, A., Debevere, J., 1997. Evaluation of the NASBA nucleic acid amplification system for assessment of the viability of Campylobacter jejuni . Int. J. Food Microbiol. 37, 13 / 20. Van der Vliet, G.M.E., Schepers, P., Schukkink, R.A.F., Van Gemen, B., Klatser, P.R., 1994. Assessment of myeobacterial biability by RNA amplification. Antimicrob. Agents Chemother. 38, 1959 / 1965. Wang, J., Rivas, G., Parrado, C., Xiaohua, C., Flair, M., 1997. Electrochemical biosensor for detecting DNA sequences from the pathogenic protozoan Cryptosporidiun parvum . Talanta 44, 2003 / 2010.
RNA biosensor for the rapid detection of viable Escherichia coli in drinking water Antje J. Baeumner *, Richard N. Cohen, Vonya Miksic, Junhong Min Department of Biological and Environmental Engineering, Cornell University, Ithaca, NY 14853, USA Received 13 December 2001; accepted 31 July 2002
Abstract A highly sensitive and specific RNA biosensor was developed for the rapid detection of viable Escherichia coli as an indicator organism in water. The biosensor is coupled with protocols developed earlier for the extraction and amplification of mRNA molecules from E. coli [Anal. Biochem. 303 (2002) 186]. However, in contrast to earlier detection methods, the biosensor allows the rapid detection and quantification of E. coli mRNA in only 15 /20 min. In addition, the biosensor is portable, inexpensive and very easy to use, which makes it an ideal detection system for field applications. Viable E. coli are identified and quantified via a 200 ntlong target sequence from mRNA (clpB ) coding for a heat shock protein. For sample preparation, a heat shock is applied to the cells prior to disruption. Then, mRNA is extracted, purified and finally amplified using the isothermal amplification technique Nucleic acid sequence-based amplification (NASBA). The amplified RNA is then quantified with the biosensor. The biosensor is a membrane-based DNA/RNA hybridization system using liposome amplification. The various biosensor components such as DNA probe sequences and concentration, buffers, incubation times have been optimized, and using a synthetic target sequence, a detection limit of 5 fmol per sample was determined. An excellent correlation to a much more elaborate and expensive laboratory based detection system was demonstrated, which can detect as few as 40 E. coli cfu/ml. Finally, the assay was tested regarding its specificity; no false positive signals were obtained from other microorganisms or from nonviable E. coli cells. # 2002 Elsevier Science B.V. All rights reserved. Keywords: E. coli ; RNA biosensor; NASBA; Rapid detection; Optical detection; Specific detection; Sensitive detection
1. Introduction The detection of pathogenic organisms in water and foods remains a challenging and important problem, since the safety of water and food supplies has to be ensured. Worldwide, infectious diseases account for nearly 40% of the estimated total 50 million annual deaths (Ivnitski et al., 1999). It is assumed that food borne diseases cause approximately 76 million illnesses, 325 000 hospitalizations, and 5000 deaths in the US each year (Mead et al., 1999). Conventional methods for the detection of bacterial pathogens typically involve culturing the organism in selective media and identifying isolates according to their morphological, biochemical, and/or immunological characteristics. While these meth-
* Corresponding author. Tel.: /1-607-255-5433; fax: /1-607-2554080 E-mail address: [email protected] (A.J. Baeumner).
ods have low limits of detection and can be used in complex food and sample matrices, they typically require days from initiation to readout. In addition, the interpretation of results is often very difficult (Sheridan et al., 1998). A number of new immunological and nucleic acidbased detection systems have been suggested which ameliorate the interpretation problems associated with biochemical assays. Nucleic acid based sensors, also often referred to as genosensors or gene probes, play an increasingly important role in the detection of pathogenic organisms for health care, environment monitoring and food safety (Ivnitski et al., 1999). Most of the gene-probe systems focus on the detection of specific DNA or rRNA sequences and utilize PCR or Reverse transcription-PCR (RT-PCR) for signal amplification (Hu et al., 1999; Spierings et al., 1993; Green et al., 1991; Rice et al., 1995; Wang et al., 1997). While some researchers have suggested the detection of only viable
0956-5663/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 5 6 - 5 6 6 3 ( 0 2 ) 0 0 1 6 2 - 8
406
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
cells via their ribosomal RNA (rRNA) (Jeffrey et al., 1994; Sheridan et al., 1998; van der Vliet et al., 1994), neither DNA nor rRNA are appropriate indicators of viability (Uyttendaele et al., 1997). However, viability is an important aspect when measuring for pathogenic organisms. In contrast to DNA and rRNA, mRNA molecules are degraded rapidly in living bacterial cells. Most mRNA species have a half-life of only a few minutes, because of their rapid degradation by enzymes (RNase), which are very stable even in harsh environments (Sheridan et al., 1998; Sela et al., 1957). Thus, if only viable pathogenic organisms are to be detected, specific mRNA sequences represent an excellent target molecule for the detection of viable pathogens. RT-PCR has been used to measure mRNA for confirming an organism’s viability (Bej et al., 1991; Patel et al., 1993). However, despite potential advantages, RT-PCR is difficult to exploit, because of the complexity of the method to measure low levels of intact mRNA from a few viable cells and small numbers of DNA molecules can produce false positive signals. nucleic acid sequence-based amplification (NASBA) can be used to achieve mRNA-based technology as an alternative method. NASBA has several advantages over RT-PCR. It exclusively amplifies RNA rapidly and isothermally. In addition, DNA cannot be used as template in NASBA. The presence of DNA samples, therefore, does not represent a potential for false positive signals. Lastly, since NASBA is isothermal, a heating block or water bath can be used instead of specialized expensive equipment (Baeumner et al., 2001; Simpkins et al., 2000; Johnson et al., 1995), which makes NASBA the better amplification tool for our application. We have chosen Escherichia coli as our model organism for the development of RNA biosensors, since it is an important analyte for food and water safety. E. coli is a natural inhabitant of the intestinal tract of
humans and warm-blooded animals. As it is found regularly in all animals capable of harboring Samonella and Shigella , the presence of this bacterium in foods or water indicates that fecal contamination may have occurred and consumers might, therefore, be exposed to enteric pathogens. Thus, E. coli is often used as an indicator organism in water sample analyses, because it is specific and most reliably reflects fecal origin (Gauthier and Archibald, 2001). We describe in this paper the development of a field-usable RNA biosensor for the specific, sensitive and rapid detection of viable E. coli in water. Highly specific DNA probes hybridize with an E. coli mRNA sequence that was amplified using the isothermal NASBA technique. The biosensor assay has been optimized, including DNA probe sequences, probe concentrations, and hybridization conditions. The biosensor’s limit of detection and specificity has been determined and the resulting signals correlated with a laboratory-based detection system for E. coli developed earlier in our lab. The biosensor assay principle is shown in Fig. 1.
2. Materials and methods All strains used in this study were provided by Dr Randy Worobo, Cornell University, Geneva, NY, which were routinely grown on Tryptic Soy broth (TSB) or agar at 37 8C as appropriate. Cells in growth phase were chosen as samples. RNA release, extraction, NASBA amplification were carried out using the Rneasy† Mini kit and RNampliFireTM (Qiagen, Valencia, CA). Two hundred proof ethanol, molecular biological grade water, and buffer reagents were obtained from Sigma Company. Lipids were purchased from Avanti Polar Lipids, Alabaster, AL. Sulforhodamine B, streptavidin etc. were purchased from Molecular Probes Company, Eugene, OR. Membranes were
Fig. 1. Biosensor assay principle. A DNA capture probe is immobilized on a polyethersulfone membrane. A DNA reporter probe is coupled to the surface of a liposome. When a specific E . coli mRNA is present (Figure A), a sandwich is formed between capture probe, mRNA and reporter probe, thus, liposomes are captured in the capture/detection zone. The number of liposomes is directly proportional to the amount of E . coli mRNA present. In Figure B it is shown that liposomes are not captured in the detection zone, when a nonspecific RNA molecule is present. Thus, no signal will be reported in the detection zone.
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
obtained from Pall/Gelman Company, Port Washington, NY. The BR-10 reflectometer was purchased from ESECO (l/560 nm). Cushing, OK. All oligonucleotides were purchased from the Biotechnology Center, Cornell University, Ithaca, NY. Heat shock procedure, nucleic acid extraction, and purification E. coli cells were grown to mid-exponential growth phase in TSB solution. For stimulation of heat shock mRNA production in E. coli , the cells were heated for 5 min at 41 8C in a heating block. The cells were disrupted by sonication (1 min at 20 W) and by incubating at 60 8C for 30 min in lysis buffer. RNA extractions were carried out using E. coli concentrations between 0 and 1 000 000 cfu/ml. The nucleic acid was purified using a Silica-guanidine thiocyanate protocol (Boom et al., 1990), provided as proprietary lysis buffer and isolation reagent modules in the RNeasyTM. The standard bacterial protocol recommended by the manufacturer was followed. Primers and probes DNA oligonucleotides for NASBA amplification (primers) and a reporter probe (for detection) had been optimized in earlier experiments (Min and Baeumner, 2002). Additionally, a variety of different oligonucleotides were available as capture probes in the biosensor to optimize sensitivity and specificity of the assay (all sequences are outlined in Table 1). The reporter probe was labeled with an amine group at the 3? end, and all capture probes were modified with a biotin moiety at the 5? end. All probes were obtained desalted and lyophilized. NASBA amplification of mRNA reactions were carried out using reagents and protocols included in the RNampliFireTM kit. A final volume of 20 ml was used containing 70 mM of KCl, 0.2 mM of each primer, 5 ml of sample and 5 ml of enzyme mix solution (RNase H, T7 polymerize and AMV-RT). Before addition of the enzyme mix, primers were annealed to the target at 65 8C for 5 min, and the mixture then cooled down to 41 8C. The final solution was incubated at 41 8C for 90 /120 min to allow exponential amplification of the target mRNA. A positive control was created that was used to determine the quality of the amplification
reaction. It consisted of a known positive E. coli NASBA amplification reaction that was diluted 1:200 in water and stored at /20 8C. Five microlitre of this solution was used as positive control in every set of NASBA reactions. For a negative control, water was added to the NASBA reaction instead of a cell extract. This control allowed the determination of possible sample cross contamination during a NASBA reaction. 2.1. Liposome preparation and nucleic acid coupling Liposomes were prepared following the protocol published in Siebert et al. (1993). In brief, liposomes were prepared using the reversed-phase evaporation method. One hundred to 150 mM Sulforhodamine B was dissolved in potassium phosphate buffer, pH 7.5 and entrapped in the liposomes. Subsequently, the liposomes were extruded at elevated temperatures through polycarbonate filters (3, and 0.4 mm, nine times each) for sizing using a mini extruder (Avanti Polar Lipids, Alabaster, AL) and purified away from unentrapped dye by gel filtration using Sephadex G50 columns and dialysis (MWCO 12/15 000) against potassium phosphate buffer with an osmolarity 50/100 mmol/kg higher than the osmolarity of the encapsulant solution. The osmolarity typically was adjusted using NaCl and sucrose. For the coupling of reporter probes to the liposome surface, a lipid tagged with a maleimide group (N -(4-(pmaleimidylmethyl)cyclohexane-1-carbonyl)-Diphosphatidyl palmitoyl ethanolamine [MMCC-DPPE]) was used during the liposome preparation and thus incorporated into the lipid bilayer. Typically, a 3 /4 mol% concentration of the overall amount of lipids was used. After liposomes were prepared and purified using gel filtration and dialysis, activated oligonucleotides were coupled to the maleimide groups available on the outer surface of the liposomes. First, the reporter probe (bearing an amine group at the 3? end) was derivatized with a sulfhydryl group. The probe was then dissolved in 0.05 M potassium phosphate buffer, pH 7.8 containing 1 mM EDTA to a concentration of 300 nmol/ml. About
Table 1 DNA sequences of oligonucleotides (NASBA primers and biosensor probes) Name
DNA sequence (5? /3? orientation)
Primer 1a Primer 2 Reporter probe Capture probe 1 (CP1) Capture probe 2 (CP2) Capture probe 3 (CP3) Capture probe long 1 (CPL1) Capture probe long 3 (CPL3)
AattctaatacgactcactatagggAAGGTTACTGGACGGCGACAA AAA TCC ACA TIT CTg ACg A GTC Tgg TgA ATT ggT TCC g CCg TTg gCA CAg CAA ATA gTC AAA CTg CTg AgC gAg AA CCT ggA AgT TAA TgA AgA CCg CgA AAA CCC gTT ggC ACA gCA AAT A TCg CCT ggA AgT TAA TCA AgA CCg gAT
a
407
T7 promoter sequence at 5? end (lower case) followed by a 4-nt GA-rich sequence.
408
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
30 nmol of probe were mixed with 90 nmol of N succinimidyl-S -acetylthioacetate (SATA) dissolved in DMSO. This mixture was incubated at room temperature for 1.5 h. Second, the thiol was deprotected in a deacetylation step. A fresh solution of 0.5 M hydroxylamine hydrochloride, 25 mM EDTA in 0.1 M potassium phosphate buffer with pH adjusted to 7.5 was prepared. The amount of solution added to the reporter probe mixture was 1/10 that of the probe solution volume. The mixture incubated for 2 h at room temperature. Third, the maleimide-derivatized liposomes were conjugated with SH-tagged reporter probes. In order to obtain the required surface density of reporter probe, the appropriate amount of liposomes was added to the probes (for example for a 0.4 mol% tag, 10.5% of liposomes was used). The pH of the reporter probe solution was adjusted to seven and the osmolarity adjusted to the appropriate value. Subsequently, the two solutions were mixed and then incubated at room temperature for 3 h and then overnight at 4 8C. In the final step, ethylmalcimide and Tris were added to quench the excess SH groups on the reporter probes and the unreacted maleimide groups on the liposomes, respectively. The liposomes were then purified from unreacted reporter probe on a Sepharose CL-4B column and finally dialyzed against the appropriate buffer overnight in the dark. The liposomes were stored at 4 8C in the dark. Different concentrations of reporter probes were tagged to the liposomes, ranging from 0.05 to 1 mol% tag. The liposome concentration of the liposome stock solution varied between the different preparations due to variations in the liposome yield after evaporation and differences in the gel filtration process. Thus, the amount of liposomes per assay was evaluated using a spectrophotometer (DU520Beckman, Fullerton, CA) at a wavelength of 570 nm and adjusted so that every assay contained the same amount of liposomes. 2.2. Membrane preparation The membranes were cut into strips of 4.5 /80 mm. Subsequently, the membranes were coated with a mixture of streptavidin and biotinylated capture probes. A mixture containing 15 pmol streptavidin and 45 pmol capture probe per ml in a sodium carbonate buffer (0.4 M NaHCO3/NA2CO3 with 5% methanol) was incubated for at least 15 min at room temperature. The streptavidin-capture probe mixture was immobilized on the membrane strips by pipetting 1 ml of the mixture directly onto the membrane, approximately 2.5 cm from the bottom. The membranes were dried for 5 min at room temperature and then for an additional 1.5 h in a vacuum oven (15 psi) at 52 /55 8C. Subsequently, the membranes were incubated in a blocking solution of
0.5% polyvinylpyrrolidone, 0.015% casein in Tris buffered saline (TBS: 20 mM Tris, 150 mM NaCl, 0.01% NaN3, pH 7 /7.5) for 30 min. The membranes were blotted dry and finally dried in the vacuum oven (15 psi) at 30 8C for 2 h. They were stored in vacuum-sealed bags at 4 8C until use.
2.3. Biosensor assay format A lateral flow assay was developed. First, liposomes, target sequence (amplicon or synthetic target sequence) and a hybridization solution (master mix) were incubated for 10 min at 41 8C. The mixture was pipetted onto the membrane strip about 8 mm from the center of the capture zone (toward the front of the strip). Immediately afterwards, 30/40 ml of running buffer were added to the front of the strip. After all of the running buffer traversed along the strips (/10 min), the capture zones were analyzed with the BR-10 reflectometer by placing the capture zone directly under the reflectometer opening or visually evaluated. The reflectometer measures the reflectance of light at a wavelength of 560 nm, which corresponds to the absorbance maximum of sulforhodamine B that is encapsulated in the liposomes. The reflectometer is calibrated about once a week using a white and magenta paint chip to a minimum value of zero arbitrary units and a maximum of 155 arbitrary units. Both solutions, running buffer and master mix contained formamide, SSC (1 /SSC contains 15 mM Sodium citrate and 150 mM NaCl, pH 7.0). Ficoll type 400, sucrose and Triton X-100. The concentrations of each compound were varied for optimization of the biosensor assay. Additionally, different hybridization conditions, as well as liposome and target sequence concentrations were investigated. Optimal conditions were: 2 ml of master mix (40% formamide, 9/SSC, 0.6 M sucrose, 0.6% Ficoll type 400) were mixed with 4 ml of liposomes and 4 ml of target sequence and incubated for 10 min at 41 8C. Thirty to 40 ml of running buffer (25% formamide, 6/SSC, 0.2 M sucrose, 0.2% Ficoll type 400, 0.01% Triton X-100) were applied to the membrane. For optimization of the reporter probe tag concentration on the liposome surface, a dipstick type assay was carried out (i.e. vertical assay). About 7 ml liposomes (volume adjusted depending on their concentration), 2 ml of synthetic target sequence and 51 ml of running buffer (25% formamide, 3 /SSC, 0.2 M sucrose, 0.2% Ficoll type 400) were mixed and incubated at 41 8C for 10 min in a glass tube. Subsequently, a membrane strip was inserted into the glass tube and the solution allowed to migrate up the strip. After approximately 8 min, the assay was complete (i.e. the solution reached the top of the membrane strip).
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
2.4. Synthetic target sequence A 37 nt long synthetic oligonucleotide was made representing a portion of the target sequence, the clpB mRNA of a heat shock protein, to which the E. coli specific detection probes hybridize. It consisted of the capture probe one sequence adjacent to the reporter probe sequence (Cgg AAC CAA TTC ACC AgA CTA TTT gCT gTg CCA ACg g). The lyophilized oligonucleotide was dissolved in water (DNase and RNase free) and stored frozen at /20 8C. Solutions were prepared with concentrations ranging from 0.1 to 10 000 fmol/ml. Comparison of biosensor with ECL detection Synthetic E. coli target sequence was used for the comparison of the biosensor signal with the previously established electrochemiluminescence (ECL) detection of E. coli in water samples (Min and Baeumner, 2002). In brief, 5 ml of target sequence or 1:10 in H2O diluted NASBA amplicon were mixed with 10 ml of magnetic bead labeled capture probe and 10 ml of ECL labeled reporter probe (Baron Consulting Co., Milford, CT). The mixture was incubated for 30 min at 41 8C with 5 s, vortexing every 10 min. Subsequently, 300 ml of hybridization buffer (supplied by Organon Teknika, Boxtel, NL) were added and the samples loaded into the autosampler of the NucliSense Reader for detection and quantification.
3. Results and discussion 3.1. Capture probe optimization
409
0.05, 0.1, 0.2, 0.4, 0.6, 0.8 and 1% of the total amount of lipids per liposome. Vertical assays were performed and each liposome preparation was analyzed in triplicate. Bound liposomes in the capture zones were quantified using the reflectometer (at l/560 nm). Fig. 2 shows the results of assays using 2 pmol of synthetic target sequence per assay. An optimum is observed at a reporter probe concentration of 0.4% tag, however, statistical analysis using a two-sample t-test shows that it is only 70% certain that the value at 0.4% tag is higher than the value at 0.6% tag. Thus, a concentration between 0.4 and 0.6% is found to be optimal for the analysis of target sequences in the range of 2 pmol per sample. Subsequently, a variety of different synthetic target sequence concentrations (20, 200, 2, and 20 pmol) were analyzed using liposomes with 0.05, 0.4, 0.6, 0.8 and 1% tag. The results are shown in Fig. 3. It is apparent that for lower target sequence concentrations, a lower percentage of tag on the liposomes is favorable (i.e. for 20 fmol the 0.05% tag results in the highest signals). In addition, for concentrations of 200 fmol per assay and above, a tag surface density of 0.4 /0.6% is optimal. The data point at 200 fmol target sequence and 0.6% tag liposomes is identified as an outlier and will be further investigated in the future with additional batches of liposomes. Future experiments described in this publication were carried out using 0.4% reporter probe tag on the liposome surface. However, when samples containing low concentrations of target sequence would be analyzed, liposomes with a 0.05% tag are suggested to be more appropriate.
In order to optimize the biosensor assay, six different capture probe sequences, immobilized on the polyethersulfone membrane, were investigated. E. coli clpB mRNA was amplified using NASBA and then applied to the six biosensor assays. Liposomes, master mix and running buffers remained constant in all experiments. All capture probes had been designed previously to have a similar degree of potential secondary structures (Min and Baeumner, 2002). It was found that under given conditions, capture probe one resulted in the strongest signals. Signals obtained with capture probe one was about twice as strong as those from capture probes 2 and 3 and about four times as strong as for any of the long capture probes (LCP1 and LCP3). While the longer capture probes (25 /27 nts) would have made a more specific detection assay, their slower hybridization kinetics resulted in the significantly lower signals. 3.2. Optimization of reporter probe tag concentration The concentration of reporter probe tags on the liposome surface was optimized. Seven different types of liposomes were prepared with tag concentrations of
Fig. 2. Optimization of reporter probe tag on liposomes. Synthetic target sequence (2 pmol per assay) was analyzed with the different biosensor assays in which the number of reporter probes on the liposome surface was varied. Reflectometer signals are plotted vs. the concentration of reporter probes. Background signals from negative controls (i.e. water instead of target sequence was used in the biosensor assay) are subtracted.
410
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
Fig. 3. Investigation of different liposome-reporter probe concentrations with different synthetic target sequence concentrations. Reflectometer signals are plotted vs. the concentration of reporter probes. Background signals from negative controls (i.e. water instead of target sequence was used in the biosensor assay) are subtracted.
3.3. Optimization of biosensor assay Polyethersulfonyl membranes were blocked against nonspecific liposome binding following a previously optimized protocol (data not published). Typically, only minimal nonspecific binding of liposomes to the membrane was observed which resulted in reflectance signals for negative samples between 5 and 18. The lateral flow biosensor assay hybridization buffers and incubation conditions were optimized. Presence of Ficoll type 400 in both the running buffer and master mix enhanced the hybridization of probes and target sequence. Concentrations higher than 0.6% in the master mix, however, yielded increased liposome aggregation. Sucrose was added for liposome stability. Concentrations lower than 0.6 M in the master mix and 0.2 M in the running buffer yielded lower signals. Also, it was found that the lateral flow assay would result in higher signals, if a pre-incubation of the liposomes and target molecules was carried out at 41 8C for 10 min. Formamide and SSC concentrations had the most significant effect on the hybridization assay. Concentrations lower than 9 / in the master mix and 6 /SSC in the running buffer were too stringent and caused less hybridization to occur. Also, higher concentrations resulted lower signals, which is probably due to secondary structure formation of probes and target sequence. In Fig. 4, the dependency of the biosensor signal on the formamide concentration in the master mix is shown. Two different assays were run, one using 5 ml of hybridization mixture (2 ml liposomes, 2 ml E. coli
Fig. 4. The effect of formamide concentration in the master mix. Reflectometer readings are plotted vs. the formamide concentration. Two different volumes of hybridization mixture are investigated 10 ml (4 ml liposomes, 4 ml E. coli amplicon and 2 ml master mix) and 5 ml (2 ml liposomes, 2 ml E. coli amplicon and 1 ml master mix).
amplicon (clpB mRNA amplified by NASBA) and 1 ml master mix) and one using twice the volume (i.e. 10 ml E. coli amplicon) is used in this experiment instead of synthetic target sequence. Higher formamide concentration increases the biosensor signal, thus the hybridization between DNA probes and E. coli target sequence. It is assumed that higher formamide concentration decreases the likelihood of secondary structures of all nucleic acid sequences involved and thus enhances hybridization. In the case of the 10 ml reaction volume, a slight decrease can be observed at a formamide concentration of 50%. Therefore, 40% formamide concentration in the master mix was chosen for all future experiments. In addition, the formamide concentration in the running buffer was investigated. While signals at higher formamide concentration (50%) were increased over those obtained using lower concentrations (10 and 25%), a significant fading of the signal was observed upon drying of the membrane strips (typically over night). At the high formamide concentrations liposomes probably lysed over night, released the entrapped dye, which then continued to migrate along the slightly wet membrane strip. This effect was much less obvious at 25% formamide concentration and not detectable at 10%. Thus, 25% formamide concentration was chosen as optimal for the running buffer since the signal was higher than with 10% and the fading was much less significant than with 50%. (While the membranes are typically investigated directly after completion of the assay, it is advantageous to have clear signals readable after the membranes are dried and stored for data collection.)
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
411
3.4. Determination of the detection limit The limit of detection was determined using the synthetic target sequence. A standard curve analyzing concentrations of 0.1 /5000 fmol per assay is shown in Fig. 5. The lateral flow biosensor assay was used with 0.4% reporter probe tag on the liposomes, capture probe 1 and optimized running buffer and master mix. An excellent detection limit of only 5 fmol was determined. A linear range of the standard curve is obtained between 5 and 1000 fmol of target sequence per assay. Above 1000 fmol a ‘hook effect’ is determined which is typical for an affinity assay with a sandwich format. Also, due to the good correlation of E. coli biosensor and ECL assay (see below), the detection limit of 40 cfu/ ml determined previously with the ECL assay (Min and Baeumner, 2002), can be applied as well to the biosensor assay. An ECL value of 87 860 was obtained for 40 cfu/ ml which would correlate to a biosensor value of about 40 which is well above the background signal of approximately 18 (see Fig. 5) and within the linear range of the biosensor’s dose response curve.
3.5. Correlation between ECL and biosensor values As described previously, a laboratory-based detection system for E. coli was developed which was based on an electrochemiluminescence detection system using the NucliSense Reader for quantification. It was shown in our laboratory that as few as 40 E. coli cells could be detected (Baeumner et al., 2001). Since the E. coli biosensor can be used in on-site applications, it was important to obtain a good correlation between the
Fig. 5. Determination of the limit of detection for the biosensor assay. Different concentrations of synthetic target sequence were analyzed under optimized conditions using 0.4% tag liposomes. Data are presented in a double log scale plot. Due to the log-scale presentation of the data, the negative control (water instead of target sequence used in the biosensor assay) is indicated as 0.01 fmol target sequence.
Fig. 6. Comparison of the biosensor signals with a laboratory-based electrochemiluminescence (ECL) detection system developed earlier. Different concentrations of synthetic target sequence (from 5 fmol to 5 pmol per assay) were analyzed in both detection systems. Biosensor results are correlated to the corresponding ECL signals. An r2 of 0.9361 was obtained.
biosensor and a laboratory-based ECL instrument. Synthetic target sequences were analyzed with the biosensor and with the ECL system (Fig. 6). Samples ranging from 5 fmol to 5 pmol were analyzed with both systems and a correlation coefficient (i.e. r2) of 0.9631 was obtained. The results demonstrate that the biosensor is as sensitive as the ECL system and that it has a similar dynamic range. 3.6. Analysis of assay specificity for viable E. coli The specificities of a nucleic acid amplification and hybridization system is determined by the specificities of primer and probe sets used for amplification and detection. While Gene Data Bank analysis using BLAST showed that theoretically no other organism can be detected with the biosensor assay system, a variety of microorganisms were investigated in our lab to ensure the specificity of this assay for E. coli. Microorganisms were chosen that is known to be closely related to E. coli or are potentially present in untreated domestic wastewater (i.e. Shigella sonnei , Salmonella typhimurium , Pseudomonas aeruginosa , Bacillus cereus ). Fresh cultures of all listed organisms were prepared by growing them to mid exponential phase (3 /3.5 h). The cells were heat shocked, the RNA extracted and subjected to amplification with NASBA. Subsequently, all samples were analyzed using agarose gel electrophoresis (2% agarose, in 1/TAE [0.04 Tris /acetate, 0.001 M EDTA (pH 8.0)], 30/60 min, 90 V, stained with SYBR Green H (1:10 000 diluted in water)) and the biosensor assay. Agarose gel electrophoresis (Fig. 7) was also used to confirm the successful RNA extraction from all micro-
412
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
Fig. 7. Investigation of the biosensor specificity. Four different microorganisms and E. coli cells were grown to mid exponential phase and then heat shocked. RNA was extracted from the cells and amplified using NASBA. The amplicons were investigated with a 2% agarose gel. The lanes were loaded with the following amplicons, lanes 2 and 3 with S. sonnei ; lane 4 and 5 P aeruginosa ; lane 6 and 7 with S . typhimurium ; lanes 8 and 9 with B. cereus ; lanes 10 and 11 with E . coli ; lane 12 with the positive control; lane 13 with the NASBA negative control; lanes 1 and 14 with molecular weight markers. The arrow points to an RNA marker of 200 nt.
organisms. Clear bands at about 200 nt length can be seen in lanes 9, 10 and 11. Lanes 9 and 10 are NASBA reactions using E. coli cells, lane 11 shows the NASBA positive control. No bands are observed from any of the other microorganisms. The same results were obtained from the biosensor analysis (Fig. 8) where E. coli samples and positive controls gave signals between 71 and 95, while all other organisms had values close to that of the negative control. Thus, no false positive signal was obtained when other microorganisms were analyzed using the biosensor.
Fig. 8. Investigation of the biosensor specificity. Four different microorganisms and E. coli cells were grown to mid exponential phase and then heat shocked. RNA was extracted from the cells and amplified using NASBA. The amplicons were subsequently analyzed in the biosensor assay in duplicates under optimized conditions. The numbers correspond to the following samples, #1 and 2, S. sonnei ; #3 and 4, P . aeruginosa ; #5 and 6, S . typhimurium ; #7 and 8, B. cereus ; #9 and 10, E . coli ; #11, positive control; #12, negative control (H2O).
Secondly, viable E. coli cells were investigated versus nonviable E. coli cells. Since the clpB mRNA can only be produced by viable cells, no false positive signal should be obtained from dead E. coli . This was proven when E. coli cells (10 000 per sample) were killed (by boiling at 98 8C for 15 min with vortexing every 5 min for 5 s) and subsequently analyzed with the biosensor (Fig. 9). No positive signal was obtained from dead cells, i.e. the biosensor and ECL signals were lower than
Fig. 9. Analysis of dead and viable E. coli cells. Each sample contained 10 000 cells. The cells were subjected to heat shock, RNA extraction and amplification and finally analyzed in the biosensor. For dead samples, cells were killed by heating to 98 8C for 15 min. Each condition was analyzed nine times with the biosensor (i.e. three samples, analyzed in triplicates) and in triplicates with the ECL method (i.e. three samples, each analyzed once in the ECL machine). Positive controls (E. coli mRNA) and negative controls (H2O) were analyzed in triplicates.
A.J. Baeumner et al. / Biosensors and Bioelectronics 18 (2003) 405 /413
those obtained from samples containing water instead of cells (negative controls), but highly positive from viable cells (10 000 per sample). Thus, the biosensor is specific for viable E. coli.
4. Conclusion We have presented in this study the development of a highly specific and very sensitive biosensor for the rapid detection of viable E. coli in water. The biosensor can detect as few as 40 viable E. coli in water and is rapid and very simple to use. No false positive signals were obtained from other microorganisms or nonviable E. coli cells. An excellent correlation to a laboratory-based detection system was demonstrated. In the future, we will investigate several environmental samples from different sources in order to determine matrix effects on the biosensor. We will also investigate larger sample volumes and combine the detection assay with a filtration step in order to process volumes up to 1 l and thrive toward the optimization of the NASBA reaction in order to detect as few as one E. coli cell per 100 ml as the ultimate detection limit.
Acknowledgements The authors acknowledge financial support for this project from Innovative Biotechnologies International, Inc., and the Cornell University Hughes Scholar Program. The authors also want to thank Randy Worobo for helping with some of the microbiological aspects of the work and Richard Durst and Richard Montagna for their thoughts, which helped move the project forward.
References Baeumner, A.J., Humiston, M.C., Montagna, R.A., Durst, R.A., 2001. Detection of viable oocysts of Cryptosporidium parvum following nucleic acid sequence based amplification. Anal. Chem. 73, 1176 / 1180. Bej, A.K., Ng, W.Y., Mogan, S., Jones, D., Mabbubani, M.H., 1991. Detection of viable Vibrio cholerae by reverse-transcriptase polymerase chain reaction (RT-PCR). Mol. Biotechnol. 5, 1 /10. Boom, R., Sol, C.J.A., Salimans, M.M.M., Jansen, C.L., Wertheimvan-Dillen, P.M.E., van der Noordaa, J., 1990. Rapid and simple method for purification of nucleic acids. J. Clin. Microbiol. 28, 495 /503. Gauthier, F., Archibald, F., 2001. The ecology of ‘fecal indicator’ bacteria commonly found in pulp and paper mill water systems. Water Res. 35, 2207 /2218.
413
Green, D.H., Lewis, G.D., Rodtong, S., Loutil, M.W., 1991. Detection of fecal pollution in water by an Escherichia coli uidA -gene probe. J. Microbiol. Methods 13, 207 /214. Hu, Y., Zhang, Q., Meitzer, J.C., 1999. Rapid and sensitive detection of Eshcerichia coli O157:H7 in bovine faeces by a multiplex PCR. J. Appl. Microbiol. 87, 867 /876. Ivnitski, D., Abdel-Hamid, I., Atanasov, P., Wilkins, E., 1999. Biosensors for the detection of pathogenic bacteria. Biosens. Bioelectron. 14, 599 /624. Jeffrey, W.H., Nazaret, S., von Haven, R., 1994. Improved method for recovery of mRNA from aquatic samples and its application to detection of mer expression. Appl. Environ. Microbiol. 60, 1814 / 1820. Johnson, D., Pieniazek, N.J., Griffin, D.W., Misener, L., Rose, J.B., 1995. Development of a PCR protocol for sensitive detection of Cryptosporidium oocysts in water samples. Appl. Environ. Microbiol. 61, 3849 /3855. Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Shapiro, C., Griffin, P.M., Tauxe, R.V., 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5 (5), 607 /625. Min, J.H., Baeumner, A.J., 2002. Highly sensitive and specific detection of viable Escherichia coli in drinking water. Anal. Biochem. 303, 186 /193. Patel, B.K.R., Banjerjee, D.K., Butcher, P.D., 1993. Determination of Myeobacterium leprac viability by polymerase chain reaction amplification of 71-kDa heat shock protein mRNA. J. Infect. Dis. 168 (3), 799 /800. Rice, E.W., Covert, T.C., Johnson, S.A., Reasoner, D.J., 1995. Detection of Escherichia coli in water using a colormetric gene probe assay. J. Environ. Sci. Health A30, 1059 /1067. Sela, M., Anfinsen, C.B., Harrington, W.F., 1957. The correlation of ribonuclease activity with specific aspects of tertiary structure. Biochem. Biophys. Acta 26, 502. Sheridan, G.E.C., Masters, C.I., Shallcross, J.A., Mackey, B.M., 1998. Detection of mRNA by reverse transcription */RCR as an indicator of viability in Escherichia coli cells. Appl. Environ. Microbiol. 64, 1313 /1318. Siebert, S.T.A., Reeves, S.G., Durst, R.A., 1993. Liposome immunomigration field assay device for Alachlor determination. Anal. Chim. Acta 282, 297 /305. Simpkins, S.A., Chan, A.B., Hays, J., Popping, B., Cook, N., 2000. An RNA transcription-based amplification technique (NASBA) for the detection of viable Salmonella enterica . Lett. Appl. Microbiol. 30, 75 /79. Spierings, G., Ockhuijsen, C., Hofstra, H., Tommassen, J., 1993. Polymerase chain reaction for the specific detection of Escherichia coli /Shigella . Res. Microbiol. 144, 557 /564. Uyttendaele, M., Bastiaansen, A., Debevere, J., 1997. Evaluation of the NASBA nucleic acid amplification system for assessment of the viability of Campylobacter jejuni . Int. J. Food Microbiol. 37, 13 / 20. Van der Vliet, G.M.E., Schepers, P., Schukkink, R.A.F., Van Gemen, B., Klatser, P.R., 1994. Assessment of myeobacterial biability by RNA amplification. Antimicrob. Agents Chemother. 38, 1959 / 1965. Wang, J., Rivas, G., Parrado, C., Xiaohua, C., Flair, M., 1997. Electrochemical biosensor for detecting DNA sequences from the pathogenic protozoan Cryptosporidiun parvum . Talanta 44, 2003 / 2010.