You are here: Agriculture > Genfood and Food > Food from GMO > Methods for identifying > PCR-based methods > PCR diagnostics

3.1.4 PCR diagnostics - problems and possible solutions in application

This section will briefly review some general considerations important to the design and execution of PCR. It will also discuss certain problems that may arise when applying PCR for the analysis of food stuffs with special regard to problems due to the nature of the food matrix and the applicability to processed food.

3.1.4.1 Choice of primers and general methodological parameters

The main criterion determining the specificity of a PCR assay is the choice of the primers. To ensure uniqueness of a sequence the primer should be at least 20 bases in length for statistical reasons (Berry and Peter, 1984). On the other hand, the primers should not exceed this size too much, since this would result in unspecific annealing of the primers during the extension phase generally performed at 72 °C. Given a 50 % A/T-content of a primer, this would limit the primer length to approximately 24 bases or to 29 bases assuming an A/T-content of 76 %. The average length of the primers listed in Table 14 is 21 bases. The primary choice and optimisation of primer sequences for PCR can be facilitated by the use of software programmes (Meyer, 1995b). Although a single mismatch between primer and target-DNA can influence the hybridisation under certain experimental conditions (Ikuta et al., 1987), the efficiency of PCR under conditions used in routine diagnostics is only expected to be affected if the mismatch is located in one of the extreme 3' nucleotide positions of a primer. Exchanges on 4 positions in a primer sequence, however, can be specifically detected and even be used for monitoring purposes of genetically altered microorganisms (Hertel et al., 1992; Ludwig et al., 1995). Other important methodological parameters for the development of a PCR test are the optimisation of the amplification reaction and the choice of a positive control (Wolcott, 1992; Mullis et al., 1994; Karch et al., 1995). Furthermore, the use of standardised reagents and protocols is essential for the reproducibility of such tests (Mahony et al., 1994).

3.1.4.2 Avoiding false-positive and false-negative results

Avoiding false-positive as well as false-negative results is very important for the reliability of a PCR test. False-positive results can arise from carry-over contamination (in particular, from previous PCR-assays). Several techniques have been published for avoiding carry-over effects (for a review, see: Carrino and Lee, 1995; Wolcott, 1992). The most frequently used method for 'preamplification sterilisation' employs the enzyme uracil-DNA-glycosylase, which removes all uracil bases from the DNA sugar-phosphate backbone (Longo et al., 1990, Müller et al., 1996). Unspecific primers or insufficiently restrictive conditions during the amplification reaction should also be avoided to prevent false-positive results.

False-negative results can be assessed by the parallel processing of a second PCR designed as positive control, such as a PCR-specific eucaryotic DNA (Allmann et al., 1993; Meyer, 1995a) or plant-specific sequences (Pietsch et al., 1997). These sequences are generally present in many copies within a single cell. When the target sequence is expected to be present only in a very low concentration, indicating that sensitivity will be an issue, it may be advantageous to include also a positive control targeted to a sequence which will be present in similarly low concentrations. If necessary, the positive control can be processed together with the target sequence by using multiplex-PCR (Feldmann et al., 1996; Cha and Tilly, 1993).

One very important and effective means of optimising the specificity of a PCR assay for the detection of GMOs is to choose the primers in such a way that they are located on different genetic elements (e.g. promoter, structural gene, terminator, vector-sequence). The primers should be specific for target sequences which do not occur naturally (at least not in that specific combination) in the respective crops (e.g. when the genetic elements originate from different phyla) (Meyer, 1995a), thus ensuring a high specificity of the test. In order to develop a truly specific method for a given GMO product, it is highly effective to choose a unique combination of elements (eventually by including the criteria of the length of the amplicon) that occurs neither in conventional products nor in other genetically engineered organisms that have been generated or approved. The interface between inserted DNA (T-DNA) and host-DNA may offer another unique nucleotide sequence providing an ideal target sequence for a highly specific PCR test. Such nucleotide sequences from interfaces between host DNA and transforming DNA have been described for several approved products in their respective petition documents.

3.1.4.3 Sensitivity

The sensitivity of a PCR test can be significantly improved by increasing the number of cycles (Candrian, 1994; Meyer et al., 1994). The application of 'magnetic capture-hybridisation-technique' has also been shown to augment the sensitivity of an assay by two orders of magnitude (Kirchhof et al., 1996; Jacobsen, 1995). Using 'hemi-nested PCR' or 'nested PCR' (Brockmann et al., 1996; Meyer, 1995b; Lunel et al., 1995) instead of conventional PCR represents another way of increasing assay sensitivity. Sensitivity may be assessed through a positive control which targets a sequence of similar length expected to be present in similar quantity as the actual target sequence.

3.1.4.4 DNA quality

The 'quality' of the DNA present in the samples is of particular significance in food diagnostics. The average length of DNA fragments present in the test sample is the main determinant of DNA 'quality'. It is essential that the average size of the DNA fragments in the probe not be significantly smaller than the target sequence (amplicon length) in the assay. Damage within the DNA fragments caused by chemical, physical or enzymatic processes (e.g. depurination, UV-damage) is also relevant.

Various factors may contribute to the degradation of DNA in food stuffs: (i) hydrolysis of the DNA due to prolonged heat treatment (Average DNA fragment length in processed food stuffs); (ii) enzymatic degradation by nucleases; and (iii) increased depurination and hydrolysis of DNA at low pH.

Therefore, the quality of the DNA in processed foods, heat-treated in conditions of low pH, such as tomato ketchup or soy sauce (Meyer, 1995b), is much diminished and represents a particular challenge to performing nucleotide-based amplification and detection methods. Average DNA fragment length in processed food stuffs provides an overview of the average DNA length that can be expected from fresh and processed food stuffs.

3.1.4.5 PCR methods applied to processed foods

It can be concluded from data presented in Average DNA fragment length in processed food stuffs that PCR assays in routine diagnostics should certainly not target sequence stretches longer than 500 basepairs. Instead, it may be rather favourable to restrict amplicon length to below 300 basepairs, when the assay is to be used with processed foods.

It may, therefore, be appropriate to test the average length of DNA fragments present in a given probe. For this purpose, the probe may be tested for the detectability of a sequence that must be present in the actual (non-degraded) probe using an amplicon with a size similar to the actual target sequence. In a project evaluating the possibility of detecting honey containing genetically engineered pollen, DNA fragments from 226 up to more than 1,300 basepairs have been amplified out of honey (Du Prat et al., 1996) using plant universal primers (Chaw et al., 1993). These fragments could also be amplified in tomato puree from conventional and genetically modified tomatoes, tomato soup, passata, ketchup, sun-dried tomatoes, genetically engineered potatoes, mashed potatoes, potato salad, canned potatoes, frozen chips, and oil-fried chips produced from genetically engineered potatoes (Ballaron et al., 1996; Ford et al., 1996). In addition, the identification of a 150 basepair fragment by PCR analysis of lecithin probes has been reported (Personal communication D. Bobbink, Greenpeace e. V., Hamburg, and A. Wurz, Hydrotox GmbH, Freiburg).

Further aid for the development of PCR assays applicable to processed foods can be obtained from articles on authenticity testing in food diagnostics (for a review, see: Meyer and Candrian, 1996; Candrian, 1994).

3.1.4.6 Inhibition of PCR and DNA extraction procedures

PCR can be inhibited by various compounds present in food stuffs. Hemoglobin (Ruano et al., 1992), nitrite salts used in sausages (Hertel et al., 1995b) and diary products (Bickley et al., 1996) have been shown to be potent inhibitors of the PCR reaction. A long list of salts, carbohydrates and other compounds frequently used in buffer solutions also decrease the performance of PCR (Rossen et al., 1992; Hammes and Hertel, 1995). The choice and optimisation of the DNA extraction procedures which eliminate potential inhibitory components may thus be of pivotal importance for the success of a given PCR method (Du Prat et al., 1996; Ford et al., 1996; Meyer and Candrian, 1996). Overly high concentrations of DNA itself may also inhibit PCR (Candrian, 1994).

Apart from optimising the DNA extraction there are other ways of counteracting the inhibitory effects on PCR. If the DNA content is not limiting, the simplest and possibly most effective way to avoid inhibition of PCR is the dilution of the sample. Application of 'nested PCR' appears to be particularly advantageous for the analysis of highly processed tomato products (Personal communication, H. Parkes, Laboratory of the Government Chemist, Middlesex, UK). Repeated freeze-thawing (Stary et al., 1996) and the addition of single-strand DNA-binding proteins (Vahjen and Tebbe, 1994; Kreder, 1996) have also been reported as effective methods for minimising PCR inhibition effects. A detailed discussion on compounds that inhibit PCR and on methods for removing inhibitors has recently been published by Gasch et al. (1997).

3.1.4.7 Verification of PCR results

There are several methods for verifying PCR results; they vary in reliability, precision and cost. In almost all methods used, PCR products are separated using gel electrophoresis and checked for the expected size. Parallel to that separation or subsequent to it, there are various verification techniques applicable. Specific cleavage of the amplification product by the use of restriction enzyme(s) followed by an additional separation of the fragments by electrophoresis represents one method (Meyer, 1995a, Pietsch et al., 1997). More time-consuming but also somewhat more specific is the transfer of the separated amplification products onto membranes (Southern Blot) followed by hybridisation with a DNA probe specific for the target sequence (LMBG-Methodensammlung, 1996; Schulze et al., 1996). Also worth considering is the application of a special electrophoresis technique that separates DNA fragments not only by size but by the relative composition of bases (Wawer et al., 1995). Verification of PCR products may be done by direct sequencing (Kocher, 1992; Feldmann et al., 1996). Other elegant techniques which can be performed on microtitre plates analogous to an ELISA test are also available: one technique is using DNA double strand-specific antibodies (DNA-Hybridisation Immuno-Assay, DIA), as described by Müller et al. (1996); another method employs biotinylated and digoxigenin-labelled primers (Börchers et al., 1997).

3.1.4.8 Reviews on the application of PCR in other areas

Apart from authenticity testing, mentioned already earlier, PCR has been used for several years in other fields of diagnostic applications, such as the detection of pathogens in food (Olsen et al., 1995), in parasitology (Felleisen et al., 1996) and veterinary (Pfeffer et al., 1995) and clinical diagnostics (Karch et al., 1995; Ronai and Yakubovskaya, 1995). In addition, PCR has been employed for monitoring of genetically engineered microorganisms in the environment (Jansson, 1995).