Rapid detection of viruses using electrical biochips and anti-virion sera


Dr Grzegorz Wę grzyn, Department of Molecular Biology, University of Gdańsk, Kładki 24, 80–822 Gdańsk, Poland
(e-mail: wegrzyn@biotech.univ.gda.pl).


Aims:  Rapid detection and quantification of viruses is crucial in clinical practice, veterinary medicine, agriculture, basic research as well as in biotechnological factories. However, although various techniques were described and are currently used, development of more rapid, more sensitive and quantitative methods seems to be still important.

Methods and Results:  Here we describe a method for rapid detection of viruses (using bacteriophages as model viruses), based on electrical biochip array technology with the use of antibodies against capsid proteins.

Conclusions:  Using the procedure developed in this work, we were able to detect 2 × 104 virions on the chip. The whole assay procedure takes c. 50 min and the assay is quantitative.

Significance and Impact of the Study:  This procedure may be useful in various approaches, including detection of bacteriophage contamination in bioreactors and possibly detection of toxin gene-bearing phages or other viruses in food samples.


Rapid detection and quantification of viruses is crucial in clinical practice (human infections), veterinary medicine (animal diseases), agriculture (plant viruses), basic research (infections of all laboratory organisms, from bacteria to animals) as well as in biotechnological factories (bacteriophage contamination in bioreactors). There are many methods dedicated to detection of viruses (for recent reviews see, for example, Debiasi and Tyler 2004; Kim et al. 2004; Niesters 2004). However, it appears that usefulness of various methods may be different for detection and quantification of different viruses. For example Mahabir et al. (2004) found that sensitivity of the mouse antibody production test was 10 times higher than that of the viral plaque assay and 10 000 times higher than that of the RT-PCR for detection of the MHV-A59 virus, whereas, for detection of the MMVp virus, the PCR assay was 106 times more sensitive than the viral plaque assay and the mouse antibody production test. Therefore, the most widely used PCR-based methods are not always optimal ones, and it appears that the most sensitive method needs to be determined independently for each virus. Moreover, development of more rapid, more sensitive and quantitative methods for virological analyses seems to be still important.

Electrical biochip technology has been shown to provide a suitable platform for detection of biologically relevant molecules (Albers et al. 2003). This technology is based on miniaturized amperometric biosensor devices that enable evaluation of biomolecular interactions by measuring the redox recycling of enzymatic reaction products. Electrical biochips were reported to be useful for detection of RNA molecules with sensitivity of 1012 and 1010 molecules within 25 min and 4 h respectively (Gabig-Cimińska et al. 2004a). It was also proposed that this technology can be applied to detect bacteriophage DNA with sensitivity of 107–108 molecules (Gabig-Cimińska et al. 2004b).

Despite developing systems, which employ detection of nucleic acids, immunological tests are still the basis for most laboratories working on detection of viruses and other microbes (Mortimer 1996; Iqbal et al. 2000). An interesting system has been developed, which utilizes solution phase binding of antibodies to analyte, and the immunocomplex formed is subsequently captured on biotin-coated nitrocellulose membrane. A signal generator, which is an anti-fluorescein urease conjugate, is bound to the immunocomplex, whose detection is made with a sensor as a result of changes in pH as a result of the conversion of urea to carbon dioxide and ammonia (Panfili et al. 1994). High sensitivity in detection of bacterial cells was achieved using this system (Dill et al. 1999). Therefore, we aimed to combine the electrical biochip technology with procedures based on immunological reactions to develop a suitable system for detection of viral particles.

Here, we demonstrate detection of viruses (using bacteriophages as models) employing electrical biochip technology and using antibodies against capsid proteins as an alternative to nucleic acid probes. This modification allowed us to detect 2 × 104 virions on a chip, which corresponds to 3 × 107 virions per ml of sample, within 50 min.

Materials and methods


Bacteriophage λcIb2 (from our collection) and M13 KO7 helper phage (Pharmacia) were used.


Polyclonal rabbit anti-λ serum (Węgrzyn et al. 1995) was used. Immunoglobulins were purified by affinity chromatography using protein A-Agarose (Sigma), according to manufacturer's recommendations. Briefly, the serum was diluted five times in a running buffer (20 mmol l−1 sodium phosphate, 1 mol l−1 sodium chloride, pH 7·4) and applied to a protein A resin (total volume of 1 ml). For elution, 20 mmol l−1 sodium citrate (pH 3·0) was used. Fractions of 0·5 ml each were collected, neutralized (by addition of 0·1 volume of 1 mol l−1 Tris, pH 9·0) and pooled according to protein concentrations, using AKTA Purifier System (Amersham Biosciences). The samples were then dialysed against PBS pH 7·4. A part of the sample was mixed with a 10-fold molar excess of a solution (10 g l−1) of Sulfo-NHS-LC-Biotin (Pierce) in PBS (pH 7·4) and incubated for 90 min at room temperature. The labelled anti-λ anitbodies were dialysed against PBS. For detection of M13 virions, mouse monoclonal anti-M13 antibodies (Amersham Biosciences) were used as capture antibodies. Biotinylated rabbit polyclonal anti-fd antibodies (Sigma) were used as secondary antibodies.


The basic equipment for detection of signals on biochips is a microprocessor-controlled multipotentiostat (eBiochip Systems GmbH), as described by Albers et al. (2003). From 16 array positions, to be connected in a chip adapter equipped with a flow cell, we used three. Each of the measurement positions on the chip is of 400 μm diameter and consists of a pair of interdigitated gold ultramicroelectrodes with 800 nm width and 800 nm gaps. The kind of antibodies, deposited on the particular position by spotting, determined whether it was an internal standard or a measurement position. All buffers and solutions were applied by a fluidic handling system, fully controlled by the device. The fluidic system consisted of a peristaltic pump, a six-way valve (Knauer) and a two-way magnetic valve.


Chips were spotted automatically, using a piezo driven micropipette system (Nano-Plotter, Modell 2·0, Software NPC 16; GeSiM GmbH). Each position on a chip was spotted using 50 droplets of a solution, which corresponds to 15 nl. When necessary, antibodies and sera used for spotting were diluted in PBS (pH 7·4). After spotting, chips were incubated in a humid chamber at room temperature for 60 min. Then, chips were washed, blocked with 10% instant milk in PBS (pH 7·4), and incubated for another 60 min.

Sample preparation

Bacteriophage lysates were prepared by infection of Escherichia coli cultures, according to standard procedures (Arber et al. 1983). A material from an uninfected culture, prepared in the same way as that containing phages, was used in control experiments.


Before the assay, a fluidic system of the instrument was washed, and after addition of 10% solution of milk in PBS (pH 7·4) the system was incubated for 10 min to block any unspecific interactions. Following connection of all tubes to appropriate buffers and sample containers, the programme was run, which consisted of washing with the washing buffer (0·05% Tween 20 in PBS, pH 7·4) for 30 s, sample loading (40 s), incubation of sample for 15 min, washing with the washing buffer for 2 min, loading of the solution of detection antibodies (40 s), incubation for 15 min, washing with the washing buffer for 2 min, loading of a solution of ExtrAvidin–Alkaline phosphatase (diluted 1 : 500 in PBS with 0·05% Tween 20, pH 7·4) (40 s), incubation for 8 min, washing for 3 min with the washing buffer, loading of the substrate (1 mg ml−1p-aminophenyl phosphate, pAPP, in TBS pH 8·1) (40 s), detection of the signal (1 min, see subsequent paragraph), and washing with the washing buffer for 2 min.

Detection of the signal on a chip

Transduction to the electrical signal was as a result of the enzymatic reaction of alkaline phosphatase, linked to ExtrAvidin (Sigma), which converted p-aminophenyl phosphate (pAPP) into a redox-active form of p-aminophenol (pAP). This form is oxidized at 350 mV at one of the finger pairs of electrodes and the formed quinoneimine is reduced at –50 mV on the other finger system. The close proximity of anodes and cathodes enable a 10-fold redox recycling. Amount of produced pAP was, therefore, proportional to enzymatic activity. The current slope was analysed for a period of 5 s after stop-flow of substrate solution. An external Ag/AgCl flowthrough reference electrode (type 16–702; Microelectrodes Inc.) was used. This reference electrode gave 0·0 V. The data were collected using OriginPro 7.0 software (OriginLab Corporation, Northampton, MA, USA), and slope was calculated using a script written in OriginC. The temperature in the reaction chamber was held at 37°C and controlled automatically.


To detect bacteriophage virions, we used the electrical biochip technology that was already tested for detection of nucleic acids (Albers et al. 2003; Gabig-Cimińska et al. 2004b; Gabig-Cimińska et al. 2004a). The fluidic system, driven by the main instrument, is presented schematically in Fig. 1. The detection chip (Fig. 2) was connected with this automatic system. All buffers and solutions were aspirated automatically. The only manual activity necessary was chip exchange and changing a test tube with the sample. Scheme of the sandwich that was formed on the chip is demonstrated in Fig. 3. Final conversion of pAPP to pAP, and redox recycling, which occurred on the chip, are depicted in Fig. 4. An example of the current time response in the amperometric analysis is shown in Fig. 5. Total assay time, including system washing after detection, was only 50 min.

Figure 1.

Scheme of the fluidic system used in this work

Figure 2.

Scheme of the chip used in this work. Three measurement electrodes and two counter electrodes are shown. Electrode section was magnified to show the structure of interdigitated microelectrodes

Figure 3.

Scheme of the sandwich that was formed on the chip

Figure 4.

A scheme for conversion of pAPP to pAP and redox recycling

Figure 5.

Example of measurement curves obtained during the experiment. Indications: bold line – M13 position; thin line –λ position; dotted line – control position (spotted with rabbit serum). The sample was spiked with M13 to final concentration of 109 PFU ml−1. Beginning of stop-flow is indicated by an arrow

In the first series of experiments, for detection of bacteriophage M13, primary unmodified mouse monoclonal anti-M13 antibodies were used, and secondary biotinylated rabbit anti-fd serum was employed. We have determined that sensitivity of the assay corresponds to detection of c. 105 virions, which is equal to c. 108 virus particles per ml (Fig. 6).

Figure 6.

Detection of virions of bacteriophage M13 using electrical biochip technology with biotinylated anti-fd serum and anti-M13 monoclonal antibodies. Indicated numbers of virions were used in reactions that were performed as described in Materials and methods. The results presented are average values from four to eight experiments with standard deviations indicated

Employing this method we aimed to detect bacteriophage λ virions using only polyclonal antibodies. The results shown in Fig. 7 indicate that this simplified procedure was even more effective. The detection limit was c. 2 × 104 virions on a chip, which corresponds to 3 × 107 virions per ml of the sample.

Figure 7.

Detection of virions of bacteriophage λ using electrical biochip technology with anti-λ serum. Indicated numbers of virions were used in reactions that were performed as described in Materials and methods. The results presented are average values from four to eight experiments with standard deviations indicated

We found that detection of bacteriophage virions was specific. No significant cross-reactions were observed when mixtures of anti-M13 (and the secondary serum) and anti-λ antibodies were used, and only one phage was present on the chip (Fig. 8).

Figure 8.

Detection of virions of bacteriophages λ (a) and M13 (b) using a mixture of antisera against λ and M13 virions. The reactions were performed analogously to those presented in Figs 1 and 2. Grey columns represent positions on chips corresponding to phage λ, open columns represent negative controls, and black columns represent positions on chips corresponding to phage M13. In experiments presented on panel A, phage M13 lysate was omitted and in experiments presented on panel B, phage λ lysate was omitted


The procedure of detection and quantification of viruses by means of electrical biochip technology, described in this report, is sensitive (c. 104– 105 virions on a chip can be detected, which corresponds to c. 107–108 viruses per 1 ml of the sample) and rapid (50 min). Therefore, we hope it may be useful in various aspects of virological studies, both basic and applied.

In our assays, we used bacteriophages as model viruses. Nevertheless, rapid and sensitive method for detection of bacteriophages may be not only of a very general meaning (i.e. bacteriophages as models in development of a new technique) but also may concern more particular aspects. Bacteriophage infections of bacterial cultures have serious and deleterious effects, especially in large-scale cultivations (Jones et al. 2000; Łośet al. 2004). There are procedures that lead to inhibition of bacteriophage propagation in infected bacterial cultures, however, to stop bacteriophage growth in such a culture, information about ongoing phage infection has to be obtained relatively quick (Czyżet al. 2001; Łośet al. 2004).

Bacteriophage development under conditions used in biofermentations is significantly slower than that observed under standard laboratory conditions in small-scale cultivations (Gabig et al. 1998; Łośet al. 2003). Moreover, because of significantly higher densities of bacteria in a bioreactor relative to flask cultures, higher number of bacteriophages per ml is required to lyse the culture. Therefore, it seems that the method presented in this report is quick and sensitive enough to assure infection control in biofermentation industries.

Bacteriophage contamination in biofermentation industries is often recurrent and is usually caused by virions of phages causing primary infection that survived subsequent sterilization procedures. Thus, the method described here may be useful for detection of recurring infection with the same bacteriophage. For this, it would be required to obtain a serum against phage which caused the first infection and to prepare it as presented in Materials and methods. Then, all further infections by the same phage could be detected using the electrical biochip technology. It is also possible to detect a few different targets at the same time. In this work chips with three positions were used, however, 8- and 16-position chips are being developed to work with electrical biochip instrumentation. This opens a broader spectrum of possible application for this method.

In this report, we discussed in detail only possibilities of using the method in detection and quantification of bacteriophages (as these viruses were used as models here), but we assume that it should work also with other viruses. Thus, we speculate that the method may potentially be applied in clinical practice, veterinary medicine, agriculture, biotechnology as well as in basic research. Nevertheless, even considering only bacteriophages, we assume that this method might be employed particularly in detection of bacteriophages bearing genes coding for toxins that cause serious medical problems. In fact, many bacterial toxins, including those produced by Vibrio cholerae, enteropathogenic (Shiga toxin-producing) Escherichia coli, Corynebacterium diphtheriae and Clostridium botulinum, are encoded in genomes of bacteriophages (Brussow et al. 2004). It is worth noting that in our experiments we used bacteriophage λ as a model, and a large number of carriers of toxin genes belong to lambdoid bacteriophages (Herold et al. 2004).


This work was supported by European Commission (grant no. LSHB-CT-2004-512009).