Introduction

On 31 August 2006, public laboratories found at Rotterdam in a cargo ship containing US long grain rice, traces of transgenic rice. Further analyses revealed that this rice corresponds with the not so far in US officially approved transgenic rice line LL601 from Bayer crop science. At that moment, the closely related line transgenic rice LL62 has been commercialised in the US and Canada already. The LL601 rice as well as the later in rice noodles detected transgenic Bt-63-rice (Shanyou63) from China was solely used for experimental field studies and not for human consumption. All of the transgenic rice lines are neither approved in the European community nor in Switzerland. Switzerland imports rice mainly from the European countries and from USA. But rice and rice products are also imported from eastern countries in high quantities. As a reaction, the European and Swiss authorities decided to implement a severe control of rice and rice products from the US and China in order to protect consumers from illegal transgenic rice and its products. As only an efficient diagnostic tool would enable laboratories to perform these investigations in a reasonable time and to reasonable costs, a novel pentaplex quantitative polymerase chain reaction (PCR) system called AllRice, to detect and quantify these three transgenic rice lines was developed.

Materials and methods

Plant material and DNA samples

Plant material from LL62-rice was obtained from American Oil chemists’ Society DNA samples. Plant material and DNA samples from LL601-transgene rice lines were kindly provided by the Federal Office of Public Health. Due to lack of standard material (Bt-63-rice) or insufficient amounts of reference material (LL601), an artificial standard was created.

Artificial standards

For the transgenic rice Bt63, no transgenic plant material was available. Therefore, the only solution for this lack was to produce an artificial reference material as reported in the past already [8–15, 17]. As the amplicon for the transgenic rice Bt63 is only 84 bp long, it was possible to synthesise two overlapping oligonucleotides and to amplify this fragment using the published primers (for details see Fig. 1).

Fig. 1
figure 1

Construction of the artificial standard for the analysis of the transgene Bt-63-rice Rice phospholipase

Full size image

This artificial DNA standard was calibrated against non-transgenic rice DNA using its additionally cloned part containing the gene for phospholipase enclosed in all rice genomes. Parallel amplification lead to the factor for the dilution of the artificial standard to assign the appropriate percentage. Dilutions of this calibrated artificial standard solutions were used for the quantification of the samples. It is not known how many insertions of the transgenic cassette, the transgenic line Bt63 comprises [1, 2]. Therefore, if the transgenic line Bt63 comprises more than one introduction site, the quantitative results presented here may overestimate the real percentage of the transgenic rice in the sample.

The equal procedure for the construction was used to create an artificial standard enclosing the target sequence for LL601-rice and the target sequence of the phospholipase-PCR system. The amount of reference material received for LL601 was sufficient to calibrate the artificial DNA standard. This was done by amplification of dilution rows of the genomic DNA of LL601 (100%) and the artificial standard for LL601. The factor gained was used to dilute the artificial standard to the appropriate percentage used for the validation and the routine diagnostic.

For routine diagnostic, these artificial standards were applied to generate quantitative results. The amounts obtained after construction is immense, posing also a high risk for contamination. To reduce this, risk-only dilutions of at least 1:1,000 were handled.

DNA extraction

DNA extraction from all sample matrices was performed using a Wizard Plus Minipreps DNA purification system (Promega, Madison, USA). Usually 200 mg of grinded sample material was extracted and DNA was eluted in 50 μl elution buffer according to the producers manual. The concentration was determined photometrically and adjusted by dilution to 20 ng/μl.

Primers and probes

The primers and probes were taken from published single PCR systems (details see Table 1). All primers and probes were synthesised by Microsynth AG, Balgach, Switzerland. Labelling of probes with fluorescent markers was done according to the recommendations of the Rotorgene 6000 manual and are listed in Table 1.

Table 1 Multiplex qPCR-system AllRice for the simultaneously determination of rice-phospholipase, 35S:Bar, LL62-rice, LL601-rice and Bt-63-rice
Full size table

Real-time PCR procedure

5 μl DNA extracts were added to 20 μl of reaction mix containing QuantiTect multiplex PCR no ROX Master Mix (Qiagen AG, Germany), and all primers and probes listed in Table 1. For final concentrations see Table 1. PCR was performed on a Rotorgene 6000 real-time system (Corbett, Australia) according to the following cycling protocol: Initial step of 15 min at 95 °C; followed by 45 cycles of 10 s at 95 °C and 1 min at 60 °C (according to the recommendations for the QuantiTect multiplex PCR no ROX MasterMix).

Results

Design of the multiplex real-time PCR Systems

For all applied PCR systems specificity is prerequisite. After checking primers for unspecific primer sites in the NCBI-databases by the Beacon-designer 5.1 software (Premier Biosoft, Palo Alto, CA, USA), they were first tested without probes in the single SYBR-green format. This approach is useful to detect amplification of unspecific products. After this, a first check of the specificity was performed. Therefore, DNA of several related plant species was isolated and used as template for the PCR reaction. In addition, a dilution row of each target revealed the amplification efficiency, which is crucial to reach maximal sensitivity for each target. If these first results were showed no cross-reactivity and amplification efficiencies close to 1, TaqMan-probes were ordered and labelled according to the channels of the Rotorgene 6000. We choose FAM, JOE, ROX, Cy5 and DY681 labelled TaqMan-probes (for details see Table 1).

Specificity

As one of the first steps establishing these PCR systems, all primers and probes were checked for relevant homologies by BLAST nr search within GenBank databases. To test the specificity extensively, DNA from a wide range of animals and other food ingredients was isolated. DNA of the following organisms were isolated and tested by this pentaplex PCR system: beef, trout, white beans, lentils, kidney beans, mung beans, runner beans, chickpeas, peas, beans, wheat, tomato, potato, rice, plum, apricots, peanut, hazelnut, almonds walnut, lupinus albus, lupinus augus, onion, garlic, carrot celery, chive, nutmeg, white pepper, cinnamon, aniseed, coconut and paprika. The cross-reactivity was measured as signal in % which a non-target DNA simulates (100% equal to 100 ng). Cross-reactivity above 1% emerged surprisingly for tomato. The cross-reactivity in the phospholipase-gen-system (rice) was 94% for tomato. However, it is unlikely to have tomato in raw rice but for products composed by many different ingredients this has to be considered. Additionally, several other transgene plants were tested for cross-reactivity: Corn (Bt176, Bt11, Mon810, GA21, T25, NK603, CBH-351, Mon863, TC1507, Mon810x683) and Rondup-Ready soy. Only CBH-351-corn lead to a 3.7% signal in the 35S:Bar-system. This must be considered when analysing mixed products of rice and maize.

Amplification characteristics

The amplification characteristics like amplification efficiency and correlation to the concentration of the target DNA was investigated using the multiplex row of standards (Table 2). This standard row simulates disproportional target concentrations leading to the numbers presented in Table 3. The system for LL62 shows an underperformance for the amplification efficiency which was also found in the single PCR modus (data not shown). We consider this to suboptimal primer sites. However, this system is an official method validated by the JRC [5] and therefore we decided to apply it.

Table 2 Multiplex row of standards for AllRice in the range from 0.0032 to 1% for transgene rice markers and 100–0.32% for the quantification of the amplifiable rice DNA using the phospholipase-PCR system
Full size table
Table 3 Amplification characteristics of the AllRice-system
Full size table

Sensitivity, precision and uncertainty

To evaluate the sensitivity, DNA extracts of each target species (DNA concentration 20 ng/μl) were diluted in half-logarithmic steps. Herring sperm DNA solution (20 ng/μl) was used as diluent. These single analyte dilution rows were mixed according to Table 3 to produce a multiplex standard, simulating calibration samples with signals for all five analytes. Each data-point was analysed six times (N = 6) during four runs. Data were collected and the relative standard deviation for the estimation of the precision and for the accuracy the mean deviation from the true value was taken (see compilation Table 4). The PCR systems detecting the transgene signals 35S:Bar, LL62 and Bt-63-rice 0.01% of target DNA was reliable detected, for LL601 even 0.0032% target DNA was detected reliable. None of the six negative controls showed a positive signal.

Table 4 Characterization of the performance of the AllRice-system using RSD (precision) and the difference to the true value (accuracy)
Full size table

The quantification performance and range are shown in Table 4.

Due to the fact that the number of insertions is unknown, results for Bt-63-rice have to be interpreted carefully. However, artificial standards once created are stable and available in high amounts. They enable laboratories to solve principal problems due to the lack of reference material with reasonable input in a short time. Unfortunately, only one ringtrial was organised from Fapas (GeMD16) for LL62-rice including z scores. The assigned level was 1.17%. Applying the AllRice-system, we measured 1.5% leading to a satisfactory z score of 0.5.

Analysis of commercial samples

Fifteen rice-products were analysed. Three of them gave reproducible signals for traces of LL601-rice. Also signals for Bt-rice were detected, but due to the low concentration of target sequence, none of these signals were reproducible. In addition, processed products like rice-noodles were analysed for the presence of transgene rice. Of 15 samples most signals were below the quantification range for LL601. Signals for Bt-63-rice were detected but were not reproducible and/or below the quantification range. In total, only one product (Table 5: Parboiled Rice Liberty Gold, USA) gave reproducible signals for the LL601 and, as a confirmation, the 35S:Bar-construct was detected in this sample in addition. Therefore, this sample was considered to be GMO-positive according to the regulations [16]. The phospholipase-gene system that is enclosed in this pentaplex PCR system showed for all samples a positive signal. In addition, we diluted every sample to check if the results are proportional to the input-DNA. The column-based DNA cleanup in conjunction with these two tests confirm that no sample was inhibiting and therefore leading to false-negative results. In summary, this multiplex PCR system produced clear results without laborious repeated pipetting. Inconsistent results were produced in the range below the detection limit, which has to be expected. The accuracy of the here presented multiples PCR system has to be investigated further when certified reference material becomes available and in future ringtrials.

Table 5 Results of rice-products for the GMO-markers 35S:Bar, LL62, LL601 and Bt-63-rice
Full size table

Conclusion

We state, that the pentaplex PCR system presented here, gives quantitative results with high sensitivity showing that multiplex quantitative results can be generated for five target sequences at once. Additionally, it was demonstrated that artificial standards as amplicons can be created rapidly and make analysis possible in cases where reference material is limited or not available at all. This multiplex quantitative real-time PCR system enables laboratories to process many sample in a reliable and cost-effective manner.