Introduction

In 2010, 148 million hectares GM crops were cultivated globally with a total of more than 1 billion hectares of transgenic crops cultivated over the last 15 years [1]. Herein, transgenic herbicide tolerant (HT) and insect-resistant crops are the most important commercial applications of genetic modification in plants. These so-called first-generation HT glyphosate and glufosinate traits and the CryIAb insect-resistance B. thuringiensis (Bt) toxins are today still the most frequently engineered biotechnology traits (for a comprehensive overview, see http://www.gmo-compass.org).

The major commercial transgenic HT crops to date incorporate two different classes of genes: the bacterial phosphinotricin-N-acetyltransferases from Streptomyces viridochromogenes (pat) and from Streptomyces hygroscopicus (bar) [2] and the 5-enolpyruvylshikimate-3-phosphate synthase (epsps) from Agrobacterium tumefasciens strain CP4 or from plant origin (in casu petunia) [3, 4]. Both classes of HT genes have been engineered into commercially grown GM crops [5]. The CryIAb protein of B. thuringiensis represents still the key Bt toxin in GM-based insect control strategies [6]. Different variants of the cryIAb gene (e.g., the cryIAb/Ac) have been introduced into maize, cotton, and rice aiming at controlling different insect pests [5]. Various new HT and Bt-based insect control systems are being developed: new herbicide tolerances such as acetolactate synthase inhibiting herbicides and novel glyphosate tolerances [5] and new insect-resistance traits such as novel Bt toxins like CryF in TC 1507 maize or Cry3 toxins in MON 863 maize [6]. The first-generation traits will however remain prominent in commercial crops, and all their derived food and feed products for a very long period.

As the total amount and number of commercialized GM crops increases, the development of sensitive, reliable but also cost-effective, and flexible screening strategies for GMO presence in products will become more and more important. A number of novel approaches have been developed applying the “Polymerase Chain Reaction” (PCR) technology [7]. Several multiplex qPCR formats are available reducing the number of analyses and facilitating high throughput but requiring multi-channel detection devices and often including costly detection probes for at least some of the targets [8–10]. In other cases, the applied chemistries are quite complex or are to be combined with other technologies that are less appropriate for routine applications [11–16]. Several of these approaches involve the use of PCR with multiple targets and consecutive detection and identification of the amplification products using microarrays [11–13]. Apart from requiring additional costly array analysis equipment, these approaches are often prone to variable quality of the array chips making them less suitable for routine or enforcement purposes [17]. Raymond et al. [14] and Nadal et al. [15] combined PCR with capillary electrophoresis in a very effective way, but requiring costly investment in additional equipment. Finally, an effective novel approach based on so-called padlock-probes [18] was developed, which could in the future represent an elegant solution to cover the broad diversity of GMO [16]. However, this technology is less sensitive and per se requires time-consuming optimization [19]. All these approaches require still broader testing by other laboratories to support their common applicability and robustness in GMO detection and may in se impose some difficulties in establishing appropriate validation schemes of these methods according to ISO or Codex Alimentarius standards.

Contrary to the above “high-tech” strategies, a number of PCR-based approaches have also been developed essentially applying the so-called combinatory or matrix-approach [20]. In such GMO screening approach, a limited set of simplex or multiplex PCR methods targeting various types of elements (endogene markers, construct-specific markers, GM traits…) is selected in such a way that multiple GMO can be detected within a single analytical run. Careful selection of the markers allows to develop an approach wherein not only most GMO are detected but also discriminated [9, 20, 21]. Combined with an informatics decision support tool, such GMO screening represents a very useful approach in managing the experimental analysis of samples for regulatory or enforcement purposes [20, 21].

CoSYPS, standing for “Combinatory SYBR®Green qPCR Screening” [20], was recently developed as a simplex qPCR GMO screening approach combining PCR methods detecting species/taxon markers, so-called generic recombinant markers as the 35S promoter of Cauliflower Mosaic Virus (CaMV) and the nopaline synthase terminator of A. tumefasciens [22, 23]. Here, we want to extend the system with GM trait markers such as the HT and Bt genes. All methods were developed to function under the same reaction conditions and apply SYBR®Green PCR allowing for post-PCR melting curve analysis to verify the nature of the amplification products [20]. Here, the performance and use of the SYBR®Green qPCR methods targeting the major herbicide tolerance GM traits (CP4-EPSPS for Roundup Ready® and PAT and BAR for LibertyLink®) and insect control GM trait (CryIAb) are presented. Their specificity, sensitivity, dynamic range, and PCR efficiencies are shown. The use of these GM trait PCR methods in combination with species-specific and generic recombinant SYBR®Green qPCR methods in CoSYPS is illustrated using samples from two proficiency tests and one from the Quality Control of a GM reference material. The advantage of including trait-specific methods into a GMO screening approach is discussed.

Materials and methods

Overview of test materials

Test materials were obtained from leaf tissue of in-house grown plants or were obtained from the Institute for Reference Materials and Measurements (IRMM), the American Oil Chemists’ Society (AOCS), the Biotech Companies (Bayer CropScience, Monsanto, Syngenta), or from FAPAS proficiency tests (Table 2).

Standard growth conditions for all plants were at 16 h/8 h day/night regime at 25 °C with chamber humidity at 80% in a Schnijder Scientific Plant growth chamber (S1084).

All Sybricon plasmids (see below) were constructed in-house, isolated using Qiagen mini/midi plasmid preparations and verified by DNA sequence analysis [23].

Bioinformatics analysis and primer design

Bioinformatics analysis of sequence information was performed applying the wEMBOSS software package [24, 25]. Relevant DNA sequence data were retrieved from public databases, patents, and scientific literature and from in-house DNA sequence analyses. A uniform primer design approach is applied in the development of primer pairs for the respective targets. The first step consists of identifying regions with high DNA sequence homology within the respective targets from the different GMO. Next, primer pairs (between 6 and 10 pairs) preferentially comprised in a common region are designed using the “Primer Express” program from Applied Biosystems (version 3.0). A bioinformatics specificity analysis for each primer is performed by probing each primer against several public and GMO DNA sequence databases. Any primer showing homology with non-relevant DNA sequences is discarded from further analysis. All primer pairs selected through this assessment were then evaluated further experimentally.

DNA target sequences and oligonucleotides

Primers applied in this study are listed in Table 1. Relevant physical parameters on the respective GM targets/amplicons including the primer sequences and the size of the expected PCR products are indicated.

Table 1 GM trait-specific targets, their corresponding primer sequences, and the amplicon size, the EMBL reference, and the Sybricon BCCM reference
Full size table

DNA extraction, DNA quantification, and PCR analysis

DNA extraction of all the different materials (seeds, powders, leaves) was performed as described in [23]. In short, 250 mg to 1 g homogenized material was extracted by a CTAB DNA extraction protocol (except for cotton materials for which a Qiagen kit was used). The final genomic DNA (gDNA) pellet was resuspended in 200 μL of DNase- and RNase-free water (ACROS) and stored at −20 °C. The extracted gDNA is quantified using the PicoGreen dsDNA quantitation kit (Invitrogen) on a VersaFluor™ Fluorimeter (Biorad). When considered necessary (e.g., unexpected low/high DNA yield, PCR inhibition…), the concerned DNA extracts were tested for purity either spectrophotometrically (260/280 nm ratio) or by agarose gel electrophoresis. Any non-conform extracts (degraded DNA, high protein content…) were either further purified (e.g., by column separation) or discarded.

Qualitative PCR was carried out on a Biorad iCycler using Amplitaq Gold (Applied Biosystems) and supplied buffers. Real-time qPCR was carried out on ABI 7300 PCR System (Applied Biosystems) using the SYBR®Green PCR mastermix (Diagenode, Liege-Belgium).

For both types of PCR, a standard 25 μl reaction volume was applied containing 5 μl of template (10 ng/μl gDNA or 40 copies plasmid DNA/μl), 1× SYBR®Green PCR Mastermix, and 250 nM of each primer). The thermal program consisted of a single cycle of DNA polymerase activation for 10 min at 95 °C, followed by 40 amplification cycles of 15 s at 95 °C (denaturing step) and 1 min at 60 °C (annealing-extension step). After completion of the run, a melting point analysis was performed by stepwise temperature increase (±1.75 °C/min) from 60 to 95 °C.

A fluorescent reporter signal was measured against an internal reference dye (ROX) signal to normalize for non-PCR-related fluorescence fluctuations between samples. The threshold cycle (C t ) for each sample was calculated in automatic mode according to the manufacturer’s specifications. If considered necessary (e.g., due to highly diverging baseline values close to the exponential phase), the threshold and baseline were adjusted manually.

Cloning of the SYBR®Green amplification products obtained with the qPCR methods for CP4-EPSPS, CryIAb, PAT, and BAR

In order to confirm the correct nature of the amplification products obtained with each of the qPCR methods, DNA sequence analysis was performed on the amplicons obtained with representative GM events. For this, all amplicons were cloned into a pUC18 vector and the insert DNA sequence was determined applying dideoxy sequence analysis. These plasmids were designated as “Sybricons” (= SYBRGreen amplicons). Each of the obtained sequences was verified for homology by Blast sequence comparison against public DNA sequence databases (NBCI, EMBL). Only DNA sequences matching the expected traits at stringent homology could be retrieved (data not shown). All constructed plasmids have been registered under “Safe Deposit” at the “Belgian Coordinated Collections of Micro-organisms” in the “Plasmid Collection” (BCCM™/LMBP, Gent-Belgium) (Table 1) and were certified for their authenticity. The amplicon sequences were deposited at the EMBL sequence database (Accession number listed in Table 1).

Specificity of the SYBR®Green qPCR methods

The criteria for specificity set by Barbau-Piednoir et al. [23] for SYBR®Green qPCR methods were applied in this study. A recorded signal upon SYBR®Green qPCR analysis is considered positive when the four criteria are met: (1°) an (exponential) amplification at a similar level to the endogenous control marker (ΔCt (Endogene–GM marker) < 3) is obtained with 100% GM template DNA comprising the target sequence(s), while negative controls (the so-called No Template Controls (NTC) and the gDNA from wild-type crop plants) do not yield such amplification; (2°) with all target-containing template DNA, the obtained PCR product(s) represents a single peak upon melting analysis with a unique T m value corresponding to the nominal T m value obtained with the respective Sybricon as template DNA (with an acceptable SD ± 1 °C), while no specific peaks are detectable in the negative controls, and (3°) a single band on agarose gel analysis with (4°) an amplicon length of the predicted size (SD ± 10 bp).

Primer specificity was assessed by testing amplifications of positive and negative controls (50 ng of DNA template). “No Template” controls (NTC) were included in the assay to assess possible interference by primer dimer formation or due to a-specific background fluorescence.

Sensitivity of the SYBR®Green qPCR methods

To test the sensitivity of the real-time PCR methods in terms of (approximate) target copy number, serial dilutions of a well-characterized positive control were analyzed [23]. The LOD was set as outlined in the former AFNOR standard XP V 03-020-2 (2003) [26]. The LOD represents the required target copy number in a PCR reaction well to obtain a reproducible (= six-times repeatable) amplification. From these analyses, also the PCR efficiency (E) for each of the methods can be calculated according to [27]:

$$ E = \left( {10^{{ - 1/{\text{slope}}}} } \right) - 1 $$

The PCR efficiency (E) could be expressed in percentage:

$$ E = (10^{{ - 1/{\text{slope}}}} - 1) \times 100 $$

Agarose gel analysis

Agarose gel electrophoresis was performed on 3% ReadyAgarose 96 Plus Gel (Biorad) supplied with Ethidium Bromide. Standard electrophoresis running conditions were applied (as running buffer 89 mM TRIS–borate—2 mM EDTA was used; current was set at 100 v).

DNA sequencing

DNA sequencing analysis was carried out on a CEQ8000 Genetic Analysis System (Beckman Coulter) with the Genomelab Dye Terminator Cycle Sequencing (DTCS) kit (Beckman Coulter).

CoSYPS screening analysis

In short, two subsamples (250 mg) of a homogenized powder were extracted using the formerly described CTAB protocol [23]. From each subsample, a single analysis of 50 ng of extracted DNA, as measured by Picogreen, is performed for each target. The CoSYPS analysis was performed according to [20] but documenting explicitly the use of the four methods described in this study. The analysis for the presence of all markers in a sample was performed simultaneously on the same run. The obtained T m and C t values obtained from each target in each run are then independently scored applying a priori established acceptance ranges for both the T m and the C t value of each method as described in [22, 23] and in this study. The interpretation of the results is done according to the decision criteria described in [20].

Results and discussion

Bioinformatics analysis and development of PCR primers for Roundup Ready (CP4-EPSPS), LibertyLink® (PAT and BAR), and the Bt CryIAb traits

The LibertyLink® (LL) traits (PAT and BAR phosphinothricin herbicide tolerance), the Roundup Ready® (RR) traits (CP4-EPSPS-based glyphosate herbicide tolerance), and the CryIAb Bt traits are present in a large number of GMO [7] (Table 2). Aiming at developing primers to highly conserved regions, a bioinformatics analysis (Align program) was performed on a number of available DNA sequences for the respective traits.

Table 2 Specificity assessment of the four SYBR®Green GM trait qPCR methods: “CP4-EPSPS” and “CryIAb,” “PAT,” and “BAR.” Specificity assessment of the four GM trait SYBR®Green qPCR methods using gDNA from relevant GM events and wild-type gDNA as template, using “Sybricon” plasmids as reference for the T m value
Full size table

For the LL traits, no regions of homology could be defined and specific primers were developed for the PAT and the BAR gene separately (Genebank AY346130.1; DQ156557.1). For the CP4-EPSPS genes existing to date, two types of DNA sequences can be distinguished: a sequence-type I EPSPS found in soybean, maize, and cotton GM events (indicated CP4-EPSPS I) and a sequence-type II EPSPS found in GM rapeseed and sugar beet (indicated CP4-EPSPS II) [28, 29]. A duplex qPCR method was developed comprising a small highly homologous region close to the NH2-terminus of the EPSPS mature protein (the forward primer matching both sequence types of CP4-EPSPS) and two reverse primers specific to each of the respective chloroplast transit peptide sequences of the EPSPS I and II (Genebank AY125353.1; own unpublished data). For the cryIAb gene (Genebank AF465640; EF094884.1), a single primer pair was developed allowing the detection of the major CryIAb events on the market (Bt11 and MON 810 GM maize, MON531 GM cotton). For both the CP4-EPSPS and the CryIAb targets, the respective different types have been specified in Table 2. The DNA sequence of the selected primers are listed in Table 1.

Development of Sybricon control plasmids for CP4-EPSPS, CryIAb, PAT, and BAR targets

To minimize bias due to the genetic background in determining the nominal value of the melting temperature for each target, so-called Sybricon plasmids containing the respective amplification products of the respective CP4-EPSPS, CryIAb, PAT, and BAR amplicons were constructed. All constructed plasmids have been registered under “Safe Deposit” at the “Belgian Coordinated Collections of Micro-organisms” in the “Plasmid Collection” (BCCM™/LMBP, Gent-Belgium) (Table 1) and were certified for their authenticity. All DNA sequences of the respective Sybricon inserts have been deposited at EMBL (for references, see Table 1). T m values for the different amplicons are all distinct from each other: T m values at 85 and 80.5 °C for the CP4-EPSPS I and II, respectively, at 77.2 and 78.9 °C for CryIAb from Bt11 and MON 810 maize, respectively, at 75.3 °C for PAT and at 78.4 °C for BAR. It is generally accepted that the T m obtained with SYBR®Green qPCR can vary between 0.5 and 1 °C for the same amplicon [30, 31]. Therefore, to cover slight deviations in the T m value between reference materials and samples due to analyte impurities, a standard deviation (SD) of ± 1 °C on the nominal T m value will be applied, as the acceptance range, in further analysis. In the case of the CryIAb targets, the T m values of the amplicons are overlapping, and in this case, the acceptance range of a positive signal is set at a T m = 78.1 ± 2 °C.

Specificity assessment of the SYBR®Green qPCR Methods for CP4-EPSPS, CryIAb, PAT, and BAR presence in GM events

The specificity of the respective qPCR methods was demonstrated on both wild type and genetically modified control samples from different species (including Zea mays, Glycine max, Gossypium hirsutum var Coker, Brassica napus, Oryza sativa var japonica, Beta vulgaris, and Solanum tuberosum) (Table 2). The results of these analyses demonstrated that each method recognized the correct targets in the respective GMO and that wild-type crop material did not give any aberrant signals (summarized in Table 2). Melting curve analysis of the produced amplicons yielded clearly defined peaks for all qPCR products in the positive controls and no primer dimer formation could be detected (data not shown). The amplification products obtained by qPCR using the respective trait-specific methods were analyzed by 3% agarose gel electrophoresis. All qPCR methods produce a single DNA product of expected size in the positive controls (data not shown). Using genomic DNA, the T m values obtained for the respective amplicons were consistently slightly higher than the T mvalues obtained with pure plasmid DNA due to the lower initial target copies in the plasmid samples (copy number at about 200 in the case of plasmids vs. >1,000 in case of gDNA).

Identification of GMO by melting curve analysis

Melting curve analysis yielded clearly defined qPCR products in all tested GMO. In case of the BAR and PAT positive controls, a clearly distinct T m value was observed for both traits (T m (BAR) at 78.4 °C; T m (PAT) at 75.3 °C) (see Table 2). These results confirm the low homology at DNA level between both targets as deduced from DNA sequence data. With the CP4-EPSPS SYBR®Green qPCR method, a distinct T m value was observed with different GMO. In the case of GTS 40-3-2 soybean and NK 603 and MON88017 maize, melting analysis of the amplicons yielded a T m value = 85 °C. In case of MON 89099 soybean, GT73 oilseed rape, H7 sugar beet, and MON 1445 cotton, a T m value = 80.5 °C was observed. Again, these results are in line with reported DNA sequences and confirm the very distinct open reading frames applied in the different classes of Roundup Ready® GMOs. The plant EPSPS as present in GA21 maize was not recognized by this method.

For the CryIAb SYBR®Green qPCR method, two distinct amplicon types were generated: one type with a T m value at 77.2 °C corresponding to the Bt11 type CryIAb and a second type with T m value at 78.9 °C corresponding to the MON 810 type CryIAb. The CryIAb amplicons obtained from maize GMO Bt176 and Bt10 had a similar T m value as GMO Bt11. The cotton GMO MON 531 and MON 15,985 generated CryIAb amplicons with a similar T m value as GM Bt11 maize.

Efficient detection of CP4-EPSPS, CryIAb, PAT, and BAR targets in GMO at low copy numbers

The LOD for the four SYBR®Green qPCR methods was estimated by serial dilution using leaf DNA from GTS 40-3-2 soybean, MS8 rapeseed, or Bt11 maize as representative materials that contained each of the traits at least as a single copy (Table 3). All methods performed adequately within a range from 1 to 400 copies [32], and the LOD was estimated lower than 5 copies of the respective targets (two, four, two, and a single theoretical copy for the PAT, BAR, CP4-EPSPS, and CryIAb targets, respectively). It should be stressed that these values represent only an estimated copy number. In the dilution series of the CP4-EPSPS analysis, at the 0.2 copy dilution, no positive signals could be observed anymore (data not shown). The PCR efficiencies of these 4 SYBR®Green qPCR methods are estimated at about 91, 92, 96, and 109% for the PAT, BAR, CP4-EPSPS, and CryIAb qPCR methods, respectively (Table 3). All methods thus meet the acceptance criteria set by the European Network of GMO Laboratories (ENGL) (a LOD below 20 target copies with a PCR efficiency between 89.6 and 110.2%) [27].

Table 3 Sensitivity assessment of the four SYBR®Green RT-qPCR methods: “PAT,” BAR,” “CP4-EPSPS,” and “CryIAb”
Full size table

Application of the CP4-EPSPS, CryIAb, PAT, and BAR SYBR®Green qPCR methods in CoSYPS GMO analysis

This set of trait-specific SYBR®Green qPCR methods completes the GMO analysis methods as described and applied in combinatory SYBR®Green qPCR screening (CoSYPS) [20]. In short, CoSYPS for GMO detection applies 6 SYBR®Green qPCR methods detecting the four major GM traits, the CryIAb, the CP4-EPSPS, the PAT, and the BAR (this paper), two generic recombinant markers, the CaMV P-35S promoter and the Agrobacterium T-Nos terminator targets, as the two most frequently present targets in GMO to date [23]. These GM markers are combined with the RBCl plant marker and a lectin, an alcohol dehydrogenase, and a cruciferine species marker for the detection of materials derived from soybean, maize, and rapeseed, respectively [22]. CoSYPS allows to detect the presence of most EU-authorized GMO [7].

To illustrate the CoSYPS approach, three examples were chosen: two of them represent the GMO screening results of GEMMA proficiency tests (PT), the third one is an in-house verification of a reference material (Table 4). Both GEMMA proficiency tests were flours containing, respectively, the GM soybean event GTS 40-3-2 or the GM maize event Bt11.

Table 4 CoSYPS application of the trait-specific SYBR®Green qPCR methods: detection of GTS40-3-2 soybean and Bt11 maize in two different proficiency tests and a Quality Check of RF3 rapeseed reference material
Full size table

In the first GEMMA PT, CoSYPS screening gave positive signals for the generic plant marker and for the following GM crop markers: lec (soy), P-35S, T-Nos, and the CP4-EPSPS trait. Then, the possible GMOs were determined by Prime Number Modulation analysis according to Van den Bulcke et al. [20]. These evaluation concluded that GTS 40-3-2 and MON 89788 soybean, Bt63 rice, MON 1445 and MON 531 and MON 15985 cotton and H7-1 sugar beet could be present in the sample (7/27 GMO; data not shown). The presence of these GMO was then verified applying the event-specific Taqman® qPCR methods validated by the EURL-GMFF (http://gmo-crl.jrc.ec.europa.eu/statusofdoss.htm). This identification analysis showed that only GM soybean GTS 40-3-2 was present (data not shown).

A similar analysis was performed for a second GEMMA PT containing GM maize Bt11. For this PT, CoSYPS gave positive signals for the generic plant marker, the lec (soy), the adh (maize), the P-35S, the T-Nos, the CryIAb-Bt11, and the PAT marker. Prime Number Modulation analysis identified the following GMO as possibly present in the sample: A2704-12 soybean, T25 and GA21 and Bt11 and Bt10 and MON 810 and TC 1507 and DAS 59122 and MON 863 maize and Bt63 rice (10/27 GMO; data not shown). A similar verification analysis as described above correctly showed that only Bt11 maize was present in the sample.

Finally, an example of the Quality Control verification at our Institute is shown testing a 100% GM rapeseed RF3 reference material. In this case, CoSYPS gave positive signals only for the generic plant marker, the cru (rapeseed), the T-Nos and the BAR marker. Prime Number Modulation analysis identified only the so-called SeedLink™ rapeseed family as possible candidates (MS1, RF1, RF2, MS8, RF3) (5/27 GMO; data not shown). A verification analysis as described above correctly showed that in the reference material sample only RF3 GM rapeseed was present.

In all three cases, CoSYPS indicated correctly the possible presence of the GM event present in the sample out of the GMO universe defined at June 2010. The CoSYPS (10 screening tests plus seven (PT1) or ten (PT2) or five (QC) identification tests, respectively) reduced the number of tests required compared to for example testing with all event-specific methods (27 tests in all three cases), applying a P-35S screening (one screening test plus 19 identification tests for both PTs; RF3 would not be detected in the QC testing) or even a P-35S/T-Nos screening (2 screening tests plus 19 identification tests for both PTs and 12 tests for the RF3 reference material). Considering that identification tests are costly due to the presence of the fluorescently labelled probes, CoSYPS also represents a reduction in cost. CoSYPS functioned appropriately when the expected targets were present in a complex matrix such as was the case for the Bt11 material.

Conclusion

In this study, four novel SYBR®Green qPCR methods that perform at equal efficiency were developed to allow the detection of the most common GM traits present in genetically modified crops to date. Successful use of these SYBR®Green qPCR methods together with species and generic recombinant targets (such as P-35S, T-Nos …) in a CoSYPS GMO screening approach is shown. The inclusion of such methods allows a very efficient screening for potential GMO presence and will therefore generally result in a limited number of further identification steps to be considered, as such reducing the cost of analysis. CoSYPS represents a powerful tool to determine GMO presence in food/feed products and is readily amendable into an efficient tool for enforcement purposes and GMO traceability along the food and feed chain. For this, the applicability, the robustness, and the reproducibility of the CoSYPS need, however, to be confirmed for example in an inter-laboratory trial.