Introduction

Agricultural biotechnology currently focuses mainly on the development of genetically modified plants with herbicide tolerance and the expression of resistance to insect pests [1, 2]. Insect-resistant cotton in India and insect-resistant corn and soybean in the United States are already grown at large scale [3].

In order to establish insect resistance in plants, genes from Bacillus thuringiensis (Bt) coding for insecticidal crystal (Cry) proteins are of particular importance. These genes encode proteins, which are toxic to the larvae of Lepidoptera, Coleoptera, and Diptera. More than 100 different Bt genes have been cloned and sequenced [4]. The groups of Cry proteins differ significantly in their amino acid sequences and in their different toxic effects on the target organisms.

Rice is the staple food in many Asian countries, and therefore, the development of insect-resistant rice is of high priority. Several genetically modified rice lines resistant to harmful insects are close to commercialization in these countries [5, 6]. The yellow stem borer (YSB; Scirpophaga incertulas Walker) and the rice leaf folder (RLF; Cnaphalocrocus medinalis Guenée) can lead to considerable crop loss and enormous costs for the use of chemical insecticides.

According to national biosafety regulations, genetically modified crops have to pass field and pre-production trials prior to approval for commercial use. Corresponding environmental safeguard measures mainly depend on the national regulation of the country where these trials take place. Confinement measures and isolation distances differ between the countries where the field trials are carried out and may not always be very effective. Admixtures of genetically modified rice with conventional rice happened in China and were detected in imported products in the EU and Japan since 2006 [7–9]. To date no events of genetically modified rice are authorized as food or feed in the EU. In China, safety certificates were issued for two Bt rice events, Huahui no. 1 and Bt Shanyou 63 (“Bt63”), both expressing a cry1Ab/Ac fusion gene, but commercial production of GM rice is still not permitted in China. For five other GM rice events, the pre-production field trials are completed, but safety certificates have not been issued up to now. The rice events close to authorization in China are Kemingdao (KMD), T1c-9, T2A-1, and Kefeng (KF) 6 and 8, and the latter two are expressing cry1Ac and cowpea-trypsin-inhibitor (cpti) genes [10].

In the EU, authorization of genetically modified crops intended for food and feed use is stipulated in Regulation (EC) No. 1829/2003 [11]. Since 2006, trace amounts of genetically modified rice in products placed on the EU market were notified to the rapid alert system for food and feed (RASFF). All products involved were recalled or withdrawn from the market. As a consequence of continuous findings, the EU Commission ruled emergency measures regarding all imports of rice products originating in or consigned from China [12, 13]. The event Bt63 developed in China was initially detected by PCR and DNA sequencing [7]. In 2009, Chinese rice products with DNA constructs differing from those of Bt63 were discovered and finally led to the detection of the event KF6 [9].

Based on these findings, several construct-specific real-time PCR methods for the detection of genetic modifications in Asian rice products were developed. The application of a cascade of detection methods allows the effective inspection of rice products for the presence of genetically modified material [9]. Using the cascade approach, genetic modifications were detected for the first time in Basmati rice by the authorities of Luxembourg in autumn 2011 and notified to RASFF [14]. In the following months, similar findings of genetic modifications in Basmati rice products were detected in France and Germany [15–18]. The molecular characterization of the genetic modifications and constructs found in a sample of the rice product corresponding to RASFF notification 2011.1646, and the resulting detection strategies are described in this paper.

Methods

Sample materials

Isolation of DNA

The DNA was extracted using a modified CTAB protocol as previously described [7, 19]. Briefly, 1 g of finely ground test material, 5 mL of CTAB extraction buffer, and 10 μL proteinase K (20 mg/mL) were incubated overnight at 65 °C. After centrifugation at 13,000×g, 1 mL of the clear supernatant was transferred into a new reaction tube. 750 μL of chloroform were added, vigorously shaken, and then centrifuged at 13,000×g for 5 min. The upper phase was transferred into a new vial; its volume was determined and mixed with two volumes of CTAB precipitation buffer. After incubation for 60 min at room temperature without agitation, the samples were centrifuged for 15 min at 13,000×g, the supernatant was discarded, and the pellet was suspended in 350 μL of a 1.2 mmol/L NaCl solution. 350 μL of chloroform were added; the samples were mixed and centrifuged for 10 min at 13,000×g. The upper phase was combined with 0.6 volumes of isopropanol for nucleic acid precipitation, and after 20 min incubation at room temperature, the samples were centrifuged for 10 min at 13,000×g. The supernatant was discarded; 500 μL of 70 % ethanol solution was added to the DNA pellet and vigorously shaken. Finally, the pellet was centrifuged 10 min at 13,000×g; air-dried overnight and resolved in 100 μL 0.1 × TE buffer. The DNA concentration was determined photometrically (Scan-Drop, Analytik Jena)

PCR and real-time PCR

PCR was done in a volume of 25 μL containing 2.5 μl of the Qiagen 10 × PCR buffer containing 15 mmol/L MgCl2, 0.5 μmol/L of each primer, 0.625 U Taq polymerase (HotStar, Qiagen), and 5 μL of template DNA corresponding up to 500 ng DNA. For thermal cycling, an initial denaturation step for 15 min at 95 °C was followed by 45 cycles of 30 s at 95 °C, 45 s at 60 °C, and 45 s at 72 °C with a final elongation step of 7 min at 72 °C.

Real-time PCR was performed in an ABI PRISM 7500 (Applied Biosystems). The 25-μL reaction mixtures for the construct-specific detection systems contained 12.5 μL Quantitect Multiplex PCR Mix (Qiagen), 0.4 μM of each primer, 0.1 μM probe (Table 1), and 5 μL of template DNA corresponding up to 500 ng DNA. The reaction conditions were as follows: Initiation step of 15 min at 95 °C followed by 45 cycles of 1 min at 94 °C and 1 min at 60 °C.

Table 1 Primers and probes used in this study
Full size table

Primer and probes

Primers and probes (Table 1) were designed using the Primer Express 3.0 software (Applied Biosystems). Primer and probe sequences for detection of taxon-specific reference genes were taken from published real-time PCR assays [20–23].

DNA sequencing

PCR products were purified with the QIAquick PCR purification kit (Qiagen). Purified PCR products were directly sequenced with the BigDye Terminator V 1.1 cycle sequencing kit (Applied Biosystems) in an ABI PRISM 310 device (Applied Biosystems). Nucleic acid sequence data were analyzed by searches in the GenBank database using the computer algorithm BLAST 2 [26].

Results

Identification of a P-ubi-cry1Ab/Ac-T-nos gene cassette similar to KF6

As previously described, the genetic construct TT51-1 present, for example, in the event Bt63 is composed of a fused cry1Ab and cry1Ac gene and regulated by the act1 promoter from rice and the NOS terminator (T-nos) from Agrobacterium tumefaciens, respectively [7]. In 2009 and 2010, the CaMV 35S promoter (P-35S) was detected in addition to the genetic elements of construct TT51-1 in several samples of rice noodle products from China. This finding indicated that another genetically modified rice variety was present in these samples [8]. By further analysis, event KF6 could be identified in these samples [9]. As a result of this study, three different construct-specific real-time PCR methods were developed for improved detection and identification of genetic modifications in rice products. By using these specific methods, several characteristic DNA sequences of genetic constructs were verified. One of these overlapping sequences covers the junction of the maize ubiquitin promoter (P-ubi) and the 5′-part of the cry1Ab gene. In KF6, a fused cry1Ab/cry1Ac gene is linked to a DNA sequence coding for T-nos. The expression of the insect resistance in KF6 is enhanced by a second genetic construct consisting of the cpti gene combined with the rice actin promoter and T-nos. Additionally, KF6 contains P-35S, a hygromycin-phosphotransferase gene (hpt) and a terminator derived from the cauliflower mosaic virus (T-35S) as well.

In autumn 2011, a rapid alert was notified in the European Union reporting about a Basmati rice product that was positively tested for the genetic elements P-35S, T-nos, and the construct P-ubi-cry [14]. The two other construct characteristics for KF6 and the Bt63 construct TT51-1 were not detected. A detailed DNA sequence and PCR analysis confirmed that the identical construct as in KF6 is present in this product, consisting of P-ubi-cry1Ab/Ac-T-nos. For further verification, the construct-specific real-time PCR method targeting the construct cry1Ac-T-nos as described by Akiyama et al. [8] was applied and tested positive. Using the event-specific assay for KMD1 [27], the DNA extracted from the sample reacted negative. At that point of the analysis, it was suspected that the Basmati rice product corresponding to RASFF 2011.1646 contains material of an unknown GM rice event.

Detection of a P-35S-hpt construct

The following analyses focused on the characterization of the unknown genetic construct containing the P-35S element. It was assumed that sample RASFF 2011.1646 contained a P-35S-hpt construct which could not be detected by the construct-specific real-time PCR assay described in [9]. This assay was developed for specific detection of KF6 and did not amplify the P-35S-hpt construct present in KMD1, because the junction sequence in KMD1 is different.

To test for the presence of a P-35S-hpt construct, a conventional PCR with the primers 35S-HPH-1F and 35S-HPH-3R [8] amplifying a larger fragment (Table 1) was used. This PCR test using DNA of sample RASFF 2011.1646 resulted in a PCR product of 350 bp. Based on the sequence determined for the amplified fragment, a real-time PCR system for broader screening of P-35S-hpt constructs was designed. Primers located entirely within the sequence of the 35S promoter (307-F) and in the 5′ region of the hpt coding sequence (652-R) were chosen. The probe (307-T) anneals completely to the sequence of P-35S. This P-35S-hpt screening PCR assay should allow the detection of the P-35S-hpt construct in GM events that contain a combination of these genetic elements. Using DNA of sample RASFF 2011.1646 a PCR product of 134 bp is obtained (Fig. 1).

Fig. 1
figure 1

Consensus and differing sequence of the P-35S-hpt construct present in KF6; KMD1 and basmati rice sample RASFF 2011.1646. The construct-specific interjacent DNA-sequences between P-35S and the hpt-gene result in different sizes of the amplified fragments. The location of the forward and reverse primers are underlined and given in bold italics, the probe sequence is shown in bold and underlined letters

Full size image

Attempts to analyze the hpt gene cassette

To completely characterize the hpt gene cassette present in sample RASFF 2011.1646, attempts were made to amplify its 3′-part and thereby determine the transcription termination element.

Real-time PCR-based screening for a hpt-T-35S construct

Event KF6 contains a P-35S-hpt construct combined to T-35S [28]. Based on the sequence of this construct in KF6, a PCR system was designed with primers and probe annealing completely within the 3′-end of the hpt gene (663-F2) and within the T-35S sequence (663-Tm and 663-R), respectively. It should allow the detection of a hpt−T-35S construct independently of the junction sequence. The specificity of this construct-specific PCR assay was confirmed using KF6 DNA. However, using this assay, no amplification product could be detected with the DNA of sample RASFF 2011.1646 (Fig. 2).

Fig. 2
figure 2

DNA sequence and location of the hpt−T-35S construct in KF6. The PCR product is 133-bp long. Sequences corresponding to the primers and the probe sequences are underlined. The hpt sequence is given in italics, the interjacent sequence is shaded, and the CAMV 35S terminator sequence is shown in bold

Full size image

Real-time PCR-based screening for a hpt-T-nos construct

In event KMD1, the hpt gene cassette contains T-nos as the terminator [27]. To test for the presence of a hpt-T-nos construct, a reverse primer (Tnos-uni-R) and probe (Tnos-uni-Tm) located in the T-nos sequence [9] and hpt-forward primer 663-F2 (see Table 1) were used (Fig. 3) which should allow the detection of a hpt-T-nos construct independently of the junction sequence. The specificity of this construct-specific PCR assay was confirmed using KMD1 DNA.

Fig. 3
figure 3

DNA sequence of the hpt-T-nos construct in rice event KMD1. The PCR product is 156-bp long. The sequences of the primers and the probe are underlined. The hpt sequence is given in italics, the interjacent sequence is shaded, the T-nos sequence is shown in bold

Full size image

Again, no PCR product was obtained with sample DNA RASFF 2011.1646, and therefore, the hpt gene cassette could not be completely characterized. It might be that neither T-35S nor T-nos was used for the termination of the hpt gene, or the terminator element is truncated like in maize event Mon810 or rice event LL601. At least we cannot exclude the possibility that primers and/or probes do not perfectly match to the terminator sequences used in the rice sample RASFF 2011.1646. Other strategies than PCR, for example, DNA sequencing and genome walking could not be applied due to the low percentage of the genetically modified DNA in the Basmati rice sample.

Identification of a P-35S-cry2A-T-35S gene cassette

PCR tests using the P-35S/T-nos duplex real-time PCR [29] resulted in lower ct-values for P-35S than for T-nos with DNA of sample RASFF 2011.1646. Obviously, higher copy numbers of the P-35S target sequence are present in the sample compared with the T-nos copy number. This finding could not be explained by the analytical results at this stage, unless the P-35S-hpt gene cassette would be present in multiple copies. This assumption was not confirmed by the results obtained in the P-35S-hpt PCR tests. Thus, it was suspected that additional genetic constructs are present in the genome of the unknown genetically modified rice.

In several scientific publications, the development of Bt rice by transformation with gene constructs using cry2A in addition to the widely used cry1Ac gene is described [24, 30, 31]. In these reports, Basmati rice lines were transformed using the cry2A gene in combination with P-35S and T-nos in addition to a gene cassette consisting of the cry1Ac with P-ubi as promoter and T-nos as terminator. A PCR assay described in one of these publications and designed for the detection of integrated sequences [24] was used to test for the presence of a cry2A gene. A fragment of 632 bp could be amplified with DNA of the Basmati rice samples RASFF 2011.1646.

DNA sequencing of the PCR product and comparison with sequences in GenBank showed high identity to a synthetic cry2A gene sequence [entry EU109565; 32]. No other significant homology to a cry2A DNA sequence was found by the BLAST search in GenBank.

Primer 340-RR-R4 was designed which was fitting to the 5′ region of the published sequence of the cry2A-gene to amplify and to determine the sequence spanning the region between the assumed P-35S and the cry2A gene. In combination with primer 307-F, the overlapping sequence was amplified and analyzed. The DNA sequence of this amplified fragment was determined, and an optimized assay for the detection of the P-35S-cry2A construct could be developed (Fig. 4).

Fig. 4
figure 4

DNA sequence of the P-35S-cry2A construct in sample RASFF 2011.1646. The PCR product is 113-bp long. The location of the primers and the probe is underlined. The P-35S sequence is shown in bold, the sequence in between is shaded, and the cry2A sequence part is given in italics

Full size image

Using a similar approach, the terminator of the construct was identified. Based on the synthetic cry2A DNA sequence of GenBank entry EU109565, the PCR primer 340-RR-F4 annealing at the 3′-end of the cry2A gene was designed. This primer was combined with a probe and primer annealing at the 5′-end of T-nos and the T-35S sequence, respectively, (Tnos-uni-Tm, Tnos-uni-R) and T-35S (663-R, 663-TM). Only the combination of primer 340-RR-F4 and the T-35S probe resulted in a PCR product, indicating that the full gene cassette consists of the elements P-35S, cry2A, and T-35S. Based on the DNA sequence of the amplified fragment, an optimized real-time PCR assay for the detection of this cry2A-T-35S construct was designed (Fig. 5).

Fig. 5
figure 5

DNA sequence of the cry2A-T-35S construct in sample RASFF 2011.1646. The PCR product is 87-bp long. The sequences of the primers and the probe are underlined. The cry2A sequence part is given in italics, the sequence in between is shaded, and the T-35S sequence is shown in bold

Full size image

The BLASTN analysis of the 1,902-bp-long DNA sequence of the cry2A-gene present in sample RASFF 2011.1646 resulted in 98 % identity to the sequence of GenBank entry EU109565 coding for a synthetic cry2A gene [32] (data not shown).

Discussion

Genetic modifications using different cry genes

A major problem discussed in the context of using Bt genes for genetic modification of crops is the likelihood of emerging resistances of the target organisms to the specific variants of the Cry proteins [33]. To cope with this problem, one strategy applied is the simultaneous integration of several different Bt genes. The scientific basis for this gene stacking is the idea that the probability of more than one random mutation in the genome of the target insects resulting in resistance to Cry proteins is extremely low. As a conclusion, different constructs could be present in genetically modified rice lines which should allow distinguishing them from each other.

A research group from Pakistan reported the simultaneous integration of two different Bt genes (cry1Ac and cry2A) into the genome of Basmati rice varieties 370 and M7 [34, 35]. These lines were co-transformed with a construct of a gene coding for the snowdrop lectin gene (gna). However, tests for the presence of gna in sample RASFF 2011.1646 were negative (results not shown). Another report described conventional crossing of two different Bt rice lines [36]. A similar strategy was used for breeding of genetically modified cotton carrying two different Bt genes [37, 38]. The authors reported that the combination of cry1Ac and cry2A genes appeared to be promising to circumvent insect resistance manifestation, since the DNA sequence identity of the two genes is <45 % [31].

Identification and origin of the genetically modified rice event

The molecular analysis of the DNA extracted from the Basmati rice sample RASFF 2011.1646 shows that in this product at least three different genetic constructs can be identified. However, the identification of the transgenic event(s) and the complete molecular characterization of the constructs remained difficult.

The junction sequence between the maize ubiquitin 1 promoter and the 5′ section of the cry1Ab gene was already known. A very similar construct was used in the KF6 with a 5′ region of the cry1Ab gene fused to the end of the cry1Ac gene and T-nos as terminator. The junction sequence from cry1Ac into T-nos identified in the sample RASFF 2011.1646 is exactly the same as the one detected in the event KF6. This was confirmed by applying the real-time PCR detection method described by Akiyama et al. [8]. Despite the fact that in the scientific literature KF6 is being described to comprise an integration of the cry1Ac gene [28], based on our sequencing data we presume that the two rice events Bt63 and KF6 do contain a fusion of cry1Ab- and the cry1Ac-gene.

Comparison of the DNA sequence of the P-35S-hpt construct identified in sample RASFF 2011.1646 shows significant differences to the sequences present in the events KF6 and KMD1, respectively. Presumably the P-35S-hpt construct identified in sample RASFF 2011.1646 was developed independently by a different research group. The identification of the terminator present in the hpt gene cassette failed.

The third genetic construct identified in sample RASFF 2011.1646 comprises the cry2A gene. To our knowledge, this is the first case of detection of a cry2A gene sequence in a rice product notified by the RASFF. Presumably the cry2A gene is a synthetic gene not published by the developers. It was possible to identify that the construct is composed of P-35S as promoter and T-35S as terminator. The molecular characterization of the constructs identified in sample RASFF 2011.1646 did not provide enough evidence to name the rice event. However, some parallels to genetically modified events based on Basmati rice variety B-370, which were developed and tested in field trials at the University of Punjab, Lahore in Pakistan obviously exist.

Basmati rice events transformed with plasmids pSM6 and pROB5 are described in detail [39]. PSM6 contains a P-ubi-regulated cry1Ac gene and a P-35S-driven cry2A gene. Our findings basically match with the structure of this construct. The plasmid pROB5 used for co-transformation contains the selectable marker gene cassette P-35S-hpt in combination with T-nos. This Basmati rice event was named NCB-313.

By using plasmid pSM6, another group at the University of Punjab in Pakistan developed a Basmati rice event named L-8-22 [40–42]. Event L-8-22 contains the constructs P-ubi-cry1Ac and P-35S-cry2A and also a hpt gene, because this event was selected by using this marker. Six generations of the homozygous line L-8-22 were tested in field trials in 2003–2005.

A detailed map or description of plasmid pSM6 is not publically available, but it is published that in pSM6, the cry2A element is used in conjunction with T-nos [42]. However, for the Basmati rice sample, RASFF 2011.1646 a cry2A-T-nos construct was not detected. Moreover, in the events L-8-22 and NSB-313, the cry1Ac is regulated by the maize ubiquitin 1 promoter, whereas in sample RASFF 2011.1646, we identified the presence of a fusion gene of cry1Ab and cry1Ac driven by this promoter element.

The exact assignment of the constructs identified in the Basmati rice sample RASFF 2011.1646 is currently not possible without precise knowledge of the DNA sequences of the plasmids pSM6 and pROB5 and without access to clearly characterized control materials. The exchange of information with developers of genetically modified Basmati rice should further help to clarify the situation.

The minimum specifications for labeling a rice product as Basmati are laid down in a so-called code of practice (CoP) on Basmati rice [43]. According to the CoP, the non-Basmati rice content must not exceed 7 %. The content of genetically modified rice in the analyzed sample RASFF 2011.1646 is estimated to be below 0.1 %; thus, the source of the genetically modified rice could be also a non-Basmati rice event. The identification of the event(s) present in sample RASFF 2011.1646 will provide clear evidence regarding this possibility.

Advanced detection strategy and screening table

The method cascade proposed in 2010 enabled the unambiguous detection of a genetically modified rice event [9]. Initial screening for the genetic elements P-35S and T-nos followed by analysis with construct-specific detection methods in case of a positive P-35S and/or T-nos result in unambiguous identification of genetic modifications in food products. The detailed molecular analysis of Basmati rice sample RASFF 2011.1646 facilitated the design of additional and improved construct-specific detection methods applicable for the identification of genetic modifications in rice products. The strategy was successfully applied to DNA obtained from Basmati rice sample RASFF 2012.0252 and 2012.0388. All genetic elements detected in RASFF 2011.1646 are present in these samples. It is proposed to extend the method cascade according to the findings described in the present work. A screening table is established to illustrate the current status of method verification data for known rice events (Fig. 6). The detection systems for the hpt constructs could be supplemented by an element-specific real-time PCR system for the hpt gene as an additional screening step [46].

Fig. 6
figure 6

Screening table to illustrate the current status of method verification data for known rice events. P positive, experimentally verified; N Negative, experimentally verified; − theoretically negative; + theoretically positive; n.d. not determined

Full size image