Analytical and Bioanalytical Chemistry

, Volume 396, Issue 6, pp 1991–2002

New approaches in GMO detection

Authors

    • Molecular Biology and Genomics Unit, Institute for Health and Consumer Protection (IHCP)European Commission—Joint Research Centre (JRC)
  • Marc Van den Bulcke
    • Scientific Institute of Public Health (IPH)
  • Jana Žel
    • Department of Biotechnology and Systems BiologyNational Institute of Biology (NIB)
  • Guy Van den Eede
    • Molecular Biology and Genomics Unit, Institute for Health and Consumer Protection (IHCP)European Commission—Joint Research Centre (JRC)
  • Hermann Broll
    • Food Safety DepartmentFederal Institute for Risk Assessment (BfR)
Review

DOI: 10.1007/s00216-009-3237-3

Cite this article as:
Querci, M., Van den Bulcke, M., Žel, J. et al. Anal Bioanal Chem (2010) 396: 1991. doi:10.1007/s00216-009-3237-3

Abstract

The steady rate of development and diffusion of genetically modified plants and their increasing diversification of characteristics, genes and genetic control elements poses a challenge in analysis of genetically modified organisms (GMOs). It is expected that in the near future the picture will be even more complex. Traditional approaches, mostly based on the sequential detection of one target at a time, or on a limited multiplexing, allowing only a few targets to be analysed at once, no longer meet the testing requirements. Along with new analytical technologies, new approaches for the detection of GMOs authorized for commercial purposes in various countries have been developed that rely on (1) a smart and accurate strategy for target selection, (2) the use of high-throughput systems or platforms for the detection of multiple targets and (3) algorithms that allow the conversion of analytical results into an indication of the presence of individual GMOs potentially present in an unknown sample. This paper reviews the latest progress made in GMO analysis, taking examples from the most recently developed strategies and tools, and addresses some of the critical aspects related to these approaches.

Keywords

Genetically modified organismGenetically modified organism detectionDetection strategiesHigh-throughput systemsMatrix approachDecision support systems

Introduction

The detection of genetically modified (GM) organisms (GMOs) in food and feed samples has become a very complex matter which necessitates the integration of technical, legal and commercial information [1]. In the context of this paper, only GM plants (GMPs) will be considered; the genetic modification of micro-organisms or animals, or GMOs for therapeutic purposes is beyond the scope of this paper.

The development, adoption and deployment of genetically modified crops are constantly growing, and GMOs are increasing both in terms of acreage of cultivated land and in terms of event/trait diversification. According to the latest ISAAA report [2], 144 GM events, representing 24 crops, have so far received worldwide regulatory approval, and this number is expected to rise [3]. In 2008, 25 countries had planted GM crops, whereas an additional 30 countries had granted regulatory approvals for their import for food and feed use and for release into the environment [2]. The approval and introduction onto the market of GM crops and of the derived food/feed products are regulated in different countries by respective legal frameworks. Such legal frameworks range from legislation based on differences in the end products, including only a voluntary labelling procedure, such as in the USA [4], to legislation such as in the European Union (EU) which foresees stringent approval, labelling and traceability requirements [57]. Other countries (e.g. Japan, Canada, Australia and New Zealand) have implemented legal systems including features of both systems [4], whereas others have no regulations in place, neither for marketing nor for labelling [8]. Consequently GM crops authorized in one country do not necessarily have the same approval status in another country; such a so-called asynchronous approval may have a considerable impact, particularly on trade.

Key technical elements required for the implementation of the legislative requirements including monitoring of GMOs and labelling provisions as part of this authorization are the availability of appropriate sampling protocols, reference materials and analytical methods that allow an accurate determination of GMO content [9]. Among the different analytical approaches that can be used for this purpose, the most direct and widely applied approach targets the genetic modification itself, i.e. the modified DNA, using the polymerase chain reaction (PCR) [1012]. Indeed, PCR has been proven to be the most accurate technique applicable from seeds to final products without being affected by the level of processing of the food/feed samples. Moreover, so far no other analytical technique has reached the same level of specificity as that provided by PCR [10, 13].

The commonly applied testing strategy consists of several steps: the presence of GM material in a sample is first verified by methods targeting the most common genetic elements present in GM constructs (screening); in the case of a positive response, the second step is the identification of the GM event(s), eventually followed by its/their quantification [9]. This strategy, however, is no longer appropriate for the increasing number and complexity of the overall set of GMOs at a global level.

The most common elements in GMO constructs, the 35 S promoter from Cauliflower mosaic virus (p35S) and the terminator from the nopaline synthetase gene of Agrobacterium tumefaciens (tNOS), have been typical targets of the screening approaches for the presence GMOs [14]. However, the variety of commercial GM crops to date invokes a broader approach. A high number of different GMOs contain the p35S and/or tNOS element in their inserted recombinant DNA; on the other hand, GMOs are emerging that do not contain these elements (http://www.agbios.com; http://www.bats.ch/gmo-watch/), meaning that they would escape from a “simple” p35S/tNOS screening approach. In addition, different variants of the individual elements are being used which might give rise to false negatives [15, 16]. This complexity results in an increasing number of methods being required and in increasing difficulty to have a comprehensive view of all the elements potentially present in GMOs.

So far, hundreds of GMO detection methods have been developed and details of them published in scientific journals as well as recorded in Web-based databases (http://mbg.jrc.ec.europa.eu/home/ict/methodsdatabase.htm; http://gmdd.shgmo.org/). Along with sequence information on primers (and probes, if applicable), method performance characteristics are described, allowing the user to select the most appropriate methods according to the intended application and their fitness for purpose [17]. Several reviews have already been published [1820] describing, evaluating and comparing the broad range of different analytical approaches applicable to GMO testing and the newest methods being adapted for these purposes.

The availability of individual, validated detection methods as such is not sufficient to cope with the complexity mentioned above. For an efficient detection strategy, comprehensive in terms of the GMOs and applicable in routine testing, the methods need to be carefully selected and combined, and the outcome needs to be interpreted in the correct manner. Some approaches following the above-mentioned strategies have recently been developed and are presented and discussed in the frame of this review.

GMO detection: a stepwise approach

The traditional strategy used in GMO analysis [21], based on the sequential detection of a few common targets, is being substituted by new approaches allowing a more comprehensive GMO analysis with the following common features: (1) definition of the frame—set of GMOs—as targets for the development of the approach, (2) dissection of the different GMOs into their recombinant or transformation-related constituents and (3) selection of an optimal set of detection methods to search for and to identify the GMO(s) present in the sample. Three key concepts applied in the development of a GMO analysis approach are the definition of the GMO universe, the description of GMOs in terms of their genetic elements/constituents (related to the transformation) and the use of a mathematical model describing the GMO universe and algorithms for identifying the GMO(s) present in a sample. All three concepts have been described in detail by Van den Bulcke et al. [22] and are only summarized here in light of the comparison of the practical development of these concepts into different GMO analysis approaches.

Three key concepts in GMO analysis

Description of the “GMO universe”

When establishing a comprehensive approach to answer complex questions, the definition of the scope of the field of application (and its limits) has to be defined. In such a formal description of the field of interest, the use of the mathematical concept “universe” defined as “the class that contains (as elements) all the entities one wishes to consider in a given situation” is considered useful. Applied to the field of GMOs, the most relevant universe in the EU would cover all the genetically modified plants authorized either under Directive 2001/18/EC [5] or under Regulation (EC) no. 1829/2003 [6] in the EU (EU-GMP). The “universe” can thus be differently defined according to the specific context and, in countries outside Europe, according to the applicable legislation.

A polynomial view on the EU-GMP universe

Each EU-GMP can be described as a combination of different genetic/recombinant elements inserted at a unique site in the host genome. Whereas the unique insertion site allows a precise identification of a particular GMO (so-called event-specific target), any inserted recombinant element could be used as a screening tag for the presence of that GM material in a product. Typical targets would be common regulatory sequences (e.g. p35S, tNOS), genetic traits (e.g. cry gene, epsps gene) and construct-specific elements (e.g. p35S-bar) [21]. The linkage of these transformation-related elements to particular ingredients (as defined by taxon-specific targets such as lectin for soybean and adh for maize) allows the GMP as a polynomial function: \( {\text{EU - GM}}{{\text{P}}_x} = {\text{Endo}}{X_0} + {\text{p35S}}{X_{\text{1}}} + \ldots + {\text{CP4 - EPSPS}}{X_5} + \ldots + {\text{Ev - GM}}{{\text{P}}_x}{X_N}, \) where EU-GMPx is EU-authorized GMP X, Endo is an endogenous species marker, XiΝ , CP4-EPSPS is the coding region of the 5-enolpyruvylshikimate 3-phosphate synthase, Roundup Ready herbicide-resistance gene and Ev is an event-specific marker.

Note that the representation of each GMP can also be seen as a particular set of elements, an array element or a vectorial equation (e.g. for Roundup Ready soybean, GTS 40-3-2 these representations could be, respectively, {p35S, tNOS, CP4-EPSPS, GMP GTS 40-3-2}, [p35S tNOS CP4-EPSPS GMP GTS 40-3-2] and (p35S, tNOS, CP4-EPSPS, GMP GTS 40-3-2) × (x1, x2, x3, x4).

Such a representation allows one to establish a matrix model wherein a relationship between targets and GMPs can be established in a hermeneutic way (presence represented by a cross symbol and absence represented by an empty field, as in Table 1), in a logical expression (“true” or “false”) or by numbers, e.g. binary (0 or 1) [23] or prime numbers [22].
Table 1

Matrix description of the universe of EU-authorized genetically modified plants (2009), comprising all genetically modified organisms (GMOs) authorized under Regulation (EC) no. 1829/2003 and the unauthorized GMOs Bt10 maize, LLRICE601 and Bt63 rice

A

B

C

D

E

F

G

H

I

J

K

L

M

N

O

P

Q

R

S

Species

Event GMO

p35S

tNOS

pFGW

t35S

pNOS

rice actin

tOCS

nptII

CP4 EPSPS

mEPSPS

PAT/pat

PAT/bar

barnase

Cry1Ab

CryIAc

Cry1F

Cry3Bb1

soybean

GTS 40/3/2

X

X

      

X

        

soybean

A 2704-12

X

  

X

      

X

      

soybean

A 5547-127

X

  

X

      

X

      

soybean

MON 89788

                 

maize

Bt 11

X

X

        

X

  

X

   

maize

Bt 176

X

  

X

       

X

 

X

   

maize

MON 810

X

            

X

   

maize

GA 21

 

X

   

X

   

X

       

maize

T25

X

  

X

      

X

      

maize

NK 603

X

X

   

X

 

X

X

        

maize

MON 863

X

X

     

X

        

X

maize

TC1507

X

  

X

      

X

    

X

 

maize

DAS59122

X

  

X

      

X

      

maize

Bt10

X

X

        

X

  

X

   

canola

GT73

  

X

     

X

        

canola

MS1/RF2/

 

X

  

X

 

X

X

   

X

X

    

MS1xRF2

canola

MS1/RF1/

 

X

  

X

 

X

X

   

X

X

    

MS1xRF1

canola

MS8/RF3/

 

X

         

X

X

    

MS8xRF3

canola

TOPAS 19/2

X

  

X

X

 

X

X

  

X

      

canola

T45

X

  

X

      

X

      

canola

Liberator

X

  

X

      

X

      

L62

canola

Falcon

X

  

X

      

X

      

pHoe6/AC

canola

Oxy235

                 

cotton

MON 1445

X

X

X

    

X

X

        

cotton

MON 531

X

X

     

X

      

X

  

cotton

LLCOTTON25

X

                

rice

LLRICE601

X

          

X

     

rice

Bt63

X

    

X

       

X

   

sugar beet

RUR H7-1

  

X

     

X

        

Crosses indicate the corresponding target is present in the GMO; p35S and t35S are the promoter and terminator of the 35 S gene from Cauliflower mosaic virus; pFGW is the promoter of the 35 S gene of Figwort mosaic virus; pNOS and tNOS are the promoter and terminator of the nopaline synthetase gene of Agrobacterium tumefaciens; rice actin is the promoter of the rice actin gene; nptII is the coding region of the neomycin phosphotransferase gene; CP4-EPSPS is the coding region of the 5-enolpyruvylshikimate 3-phosphate synthase, Roundup Ready herbicide-resistance gene mEPSPS is the mutated coding region of the 5-enolpyruvylshikimate 3-phosphate synthase gene of maize; PAT/pat and PAT/bar are coding regions of the phosphinotricin acetyltransferase, Liberty-link herbicide-resistance gene; CryIAb, CryAc, CryIF and Cry3Bb1 are coding regions of Bacillus thuringiensis insect-resistance genes

The use of the matrix approach in GMO analysis

The easiest representation of the polynomial EU-GMP universe and of the GMP present is a table (Table 1). Each row represents a particular EU-GMP, whereas the respective targets are listed in the columns.

Such a matrix table can then be used firstly as a tool to define the (minimal) set of targets allowing the identification of the presence of an EU-GMP, e.g. by compiling frequency tables (Y. Bertheau, personal communication), CoSYPS (for “combinatory SYBR® Green real-time PCR screening”) [22] and GMOtrack [23], and, secondly, as a decision support system (DSS) in identifying which GMP may be present in a particular sample, e.g. the CoSYPS DSS [22].

Such a matrix representation forms the basis of decision tools to conclude which GMP is present in a sample.

As indicated above, the elements most commonly present in GMOs are the 35 S promoter and the NOS terminator, but the current variety of commercial GM crops requires a broader set of targets to be used to cover the whole range of GMOs.

For this, other targets, such as the GM constructs in which the primer set (or eventually the probe) spans two neighbouring elements (e.g. the pNOS-nptII element in Topas 19/2 rapeseed) can be included in the screening approaches. In addition, trait-specific elements targeting, e.g., the Roundup Ready cp4 epsps gene or the Bt cryIAb gene can be included, resulting in the selective detection of classes of herbicide-tolerant insect-resistant GMPs [22, 24].

In 2008, Waiblinger et al. [25] established an Excel spreadsheet describing the application of five different screening real-time PCR methods: P35S [26]; T-nos [27]; CTP2-EPSPS [28]; bar [29]; P35S-pat [26]. All methods described have been validated in German national collaborative studies [2629].

With the exemption of only three events (maize LY038, soybean 305423 and cotton 281-24-236 × 3006-210-23), all authorized and non-authorized GMPs in the EU and listed in publicly available databases can be identified by one or more of the above-mentioned screening methods. The table also indicates if the methods have been already tested on the event described or if the indication is based only on in silico analysis. The table is available online and will be constantly updated according to new events approved and recorded in public databases (http://www.gdch.de/strukturen/fg/lm/ag/bioanal/screening.htm); although designed as a table, it could be seen as the first available matrix approach. The outcome of laboratory analyses can be compared with the indications given in the table, resulting in a proposal for the GM event(s) present in the sample under investigation.

Since it is just an Excel spreadsheet, its maintenance and updating seem to be simple and most probably much cheaper than for any other database approach.

Although this approach may serve as a universal screening, more complete and complex combinations are required, and DSSs have to be developed, for the conversion of analytical results into the identification of the individual GMO(s) potentially present in an unknown sample. The matrix approach has also been used by Hamels et al. [30] in combination with a multiplex PCR-microarray method developed to screen and identify the 24 EU-authorized GM events.

New methods will have to be developed to fill up any gaps in the data to cover the complete matrix. Among current developments in this field, the recently started project GMOseek [Development of screening methods for GMOs. Project under the Sixth Framework Programme ERA-net SAFEFOODERA coordination action (CA-515726). Topic 5.] is focusing on the development of such new screening methods which would then feed the matrix approach with available relevant methods.

Analytical approaches allowing high throughput—multitarget detection

A number of analytical approaches allowing the simultaneous detection of several targets have been developed recently. Some of them also include an interpretation of the results. The ones reported later are the methods that appear to be most promising to cope with the future challenges of an increased number of GMO authorizations around the world and that, in most cases, can be integrated within the already existing routine analysis without the need to establish totally new techniques with respective expensive instrumentation.

When the increasing complexity is considered, it is important to distinguish two aspects: on the one hand, an increase in diversity of targets and crops and, on the other hand, the increased possibility of accidental releases of similar GMOs. To assess the impact of both types of evolutions, an analysis of the current status of the commercial GMP and an estimation of the near-future situation is to be made constantly in an efficient way.

From an evaluation of several recent sources of information [2, 3], it is estimated that the targets currently applied in the approaches described later will still be able to cover the field. As such, a matrix-based approach with relatively few targets is still acceptable, but in the future may have to be extended to a broader number of targets. For this, the different platforms were compared with one another with regard to the following aspects: performance (validation status, reported performance in proficiency testing and on real-life samples), flexibility (especially with respect to the introduction of new targets), applicability (especially evaluation of compatibility with currently applied strategies) and cost (essentially reagents and equipment).

Real-time PCR-based approaches

In addition to the system developed by Waiblinger et al. [25] and described earlier, another simplex real-time PCR-based approach combining analyses and interpretation of results is the GMO screening platform CoSYPS. It was recently developed by the Scientific Institute of Public Health (IPH; Brussels, Belgium) and is based on the SYBR® Green real-time PCR method [22].

The CoSYPS approach combines the detection of the presence of major commodity crops (such as soybean, maize, oilseed rape, rice and cotton) with the detection of common genetic recombinant elements (such as the p35S/tNOS elements) and GM-specific elements (such as herbicide-resistance genes and insect-resistance genes). A limited set of 11 targets covers the current EU-GMP universe for commercial releases, including not only the EU-authorized GMOs, but also most of the GMOs authorized in non-EU countries (except LY038 maize) [22]. The GMO screening set-up as developed by the IPH has been validated according to ISO 5725 [31].

The CoSYPS analysis relies on the combination of two physical parameters of each real-time PCR method: the Tm value(s) of the amplicon [as a decision criterion for the presence of a particular target (authentication)] and the Ct value of the amplification [as an estimate of the target concentration (quantification)]. As a post-PCR authentication criterion, the Tm value represents a major advantage over most PCR-based methods, as it allows verification of the results obtained in the validation dossiers with the results obtained in a sample (in particular, the amplification of the correct target and the absence of unspecific by-products in the sample). In addition, owing to the strict primer set choice, any post-PCR characterization by, e.g., DNA sequencing is greatly facilitated as only one (or a maximum of two in the case of some trait methods) major PCR product is produced. Such strict recognition is essential when the Ct value is being applied as a criterion in deciding on the presence (limit of detection) and/or amount (limit of quantitation) of a particular target in a sample [22].

Another straightforward approach for the rapid detection of multiple GM events in a single experiment is the “real-time PCR based ready-to-use multitarget analytical system” recently developed by Querci et al. [32]. The system, conceived as a ready-to-use product, is based on TaqMan® real-time PCR technology and it consists of 96-well plates containing lyophilized primers and probes for the individual detection and the simultaneous identification of 39 GM events in seven plant species.

The published formulation is based on event-specific detection: the methods included are the ones that, according to EU legislation [6], have been submitted by applicants to the Community Reference Laboratory for GM Food and Feed (CRL-GMFF; http://gmo-crl.jrc.ec.europa.eu/) for validation, as an integral part of the approval process, and it includes the taxon-specific methods for maize, cotton, rice, oilseed rape, soybean, sugar beet and potato. The system has been proven to retain the individual methods’ specificity and to meet performance requirements for method limit of detection (http://gmo-crl.jrc.ec.europa.eu/guidancedocs.htm). The ready-to-use format allows the immediate implementation by enforcement laboratories. As illustrated in Fig. 1, the complete analysis of a food or feed sample requires only a few simple steps to be performed: DNA is extracted from the sample and added to the PCR master mix, the reaction mix is subsequently loaded on the plate and the thermal cycling programme is started. Upon completion of the programme, the results are directly extrapolated from the ad hoc instrument software. The approach as such is flexible and adaptable to meet different detection needs: the multicrops formulation reported by Querci et al. [32] and described above was conceived for the analysis of food and feed samples, whereas a crop-specific formulation, applicable to the analysis of crop commodities, has recently been developed (data not shown) for the exclusive detection of maize and soybean events. A multitarget approach based on the use 96-well plates containing different primer/probe combinations preloaded into the plates’ wells has also been reported by Mano et al. [33] for the detection of GM events in maize, soybean, rice and canola. This approach is based on a combination of targets with different specificity levels (screening elements to event-specific targets) and, even though it has been tested on only 16 GM events, it constitutes a first attempt towards the possibility to detect unapproved GM crops as it is combined with a spreadsheet application (Unapproved GMO Checker version 2.01) developed to facilitate interpretation of the results and to pinpoint a possible unapproved GM crop contamination in the sample.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-009-3237-3/MediaObjects/216_2009_3237_Fig1_HTML.gif
Fig. 1

The four-step workflow for genetically modified organism (GMO) analysis using the ready-to-use multitarget analytical system [32]: 1 DNA is extracted from the sample to be analysed, 2 extracted DNA is added to the polymerase chain reaction (PCR) master mix and the reaction mix is loaded on the plate, 3 real-time PCR (RTi-PCR) amplification, and 4 visualization of the results using the ad hoc instrument software. In the example, a composite sample (test material GeM MU01 from FAPAS® GEMMA Proficiency Scheme; http://www.fapas.com/) containing less than 0.72% GTS 40-3-2 soybean, 1.29% MON810 maize and 1.33% NK603 maize was tested

In spite of the fact that PCR technology is so far the method of choice for GMO detection, its application in the detection of several targets is limited. Multiplex PCR, allowing the detection of several GMOs, has been developed [3436], but this approach is limited by the poor multiplexing capabilities of PCR. Alternative PCR-based strategies or combinations of PCR with hybridization or capillary electrophoresis have been explored and have resulted in promising alternatives capable of overcoming the drawbacks linked to PCR technology.

Microarray-technology-based approaches

Several approaches have been developed that rely on microarray technology; microarrays, also called “DNA chips”, consist of glass supports containing specific oligonucleotide capture probes immobilized on their surface. Detection of the selected targets is achieved by hybridization, usually after preamplification of the target(s) of interest by one or more multiplex PCRs. The formulations published so far vary in relation to the specific characteristics of the capture probes, the specificity of the probe/primer combinations and the detection modes. Different DNA microarray-multiplex PCR approaches have been reported for the detection of GM events in maize and soybean [37, 38] using fluorescent probes for target detection. A low-density DNA chip for the identification of nine GM events (Bt176, Bt11, GA21, MON810, CBH351, T25 maize, Topas 19/2, T45 rapeseed and Roundup Ready soybean) has been developed [39] and validated [40] using biotinylated nucleotides, which allow target detection by colorimetric reaction, therefore avoiding the high cost associated with the use of fluorescent probes and the drawbacks derived from their photosensitivity, requiring great care in their handling. The same approach has been recently used by Hamels et al. [30] for the development of a multiplex PCR-microarray method for the screening of GMOs.

The first description of the use of padlock probes for GMO detection was provided by Prins et al. [41]. Linear padlock probes targeting GM events, GM elements and GMP species have been established that can hybridize to their genomic DNA targets and are visualized using microarray hybridization. Prins et al. [41] demonstrated the applicability using ten different padlock probes in a single assay with a detection limit down to at least 1%.

Padlock probes are based on the principle of ligation-detection probes. A single-stranded DNA molecule containing from left to right a 5′ target sequence, a reverse primer recognition site, a forward primer recognition site, a cZIP-code and a 3′ target site hybridizes to the complementary genomic target sequence. Upon hybridization, both ends ligate to form a circular molecule. Subsequently, the circular molecule is amplified by PCR with a universal forward and reverse primer labelled with a Cy3 fluorescent dye. Each probe contains a unique DNA sequence, which the authors call “cZIP-code”. Large amounts of linear single-stranded DNA with a Cy3-labelled cZIP-code are generated, which can be visualized on a microarray after hybridization. The method is based on the amplification with universal primers to generate unique cZIP-codes and is therefore suitable for multiplexing.

Morisset et al. [42, 43] have developed a novel multiplex quantitative DNA-based target amplification method suitable for use in combination with microarray detection named nucleic acid sequence based amplification implemented microarray (NAIMA). In the first step, the use of tailed primers allows the multiplex synthesis of template DNAs in a primer extension reaction. The second step of the procedure consists of transcription-based amplification using universal primers. The complementary RNA product is then directly ligated to 3DNA dendrimers labelled with fluorescent dyes, allowing signal amplification, and hybridized without further purification on an oligonucleotide-probe-based microarray for multiplex detection. Two triplex systems have been applied to test maize samples containing several transgenic lines. The great advantage and further potential of the NAIMA method is its possibility for quantification in a multiplex platform [44].

Luminex xMAP technology

An innovative high-throughput system allowing multitarget GM analysis using the Luminex xMAP technology has recently been reported by Fantozzi et al. [45]. The approach is based on the use of fluorescent beads which are commercially available in 100 distinctly coloured sets. Each coloured bead set is individually coupled to an oligonucleotide probe specific to a unique target DNA sequence. Once individually coupled and mixed, the bead complex can be used to analyse complex solutions containing multiple tagged sequences in a single step. The DNA sample is first subjected to few cycles of PCR or multiplex PCR, required to label the target DNA sequences, and is then incubated with the bead complex for the hybridization step. Finally, the detection is performed using the Luminex device (Luminex-100 or HTS, http://www.luminexcorp.com). Although Fantozzi et al. [45] have tested the approach only on a limited number of GM-specific targets (p35S and epsps gene), theoretically any additional primer/probe sets can be combined and the system can be easily adapted according to needs. The adoption of this approach however, although elegant and promising, is still hampered by the limited use of the Luminex device in GMO control laboratories.

Multiplex PCR—capillary gel electrophoresis

Recently, the simultaneous detection of six cotton and five maize targets by multiplex PCR-capillary gel electrophoresis with identification of amplified targets by size and colour was reported [46, 47]. Multiplex PCRs were performed by forward and reverse primers corresponding to primers of validated real-time PCR assays. Forward primers were fluorescently labelled with different fluorescent dyes to allow identification of each amplicon by capillary gel electrophoresis. The most-similar-sized amplicons were labelled with different dyes. 6-Carboxyfluorescein (6-FAM), tetrachloro-6-carboxyfluorescein (TET) and hexachloro-6-carboxyfluorescein (HEX) were used. PCR products were loaded onto a capillary in a sequencer device and spectra showing different peaks and colours determining different amplicon sizes and identification of amplicons were obtained. The specificity and repeatability of the method and the limits of detection are similar to those obtained from validated real-time PCR assays. This approach can be considered as a promising tool for GMO screening, since it is flexible in that different multiplex PCR products can be combined in a single capillary gel electrophoresis with identification of amplified targets by size and colour run. Also, such an approach may be particularly interesting in monitoring for the presence of unauthorized events by identifying unexpected additional amplification products (different size and Tm value).

Comparison of high-throughput multitarget technologies

From the previous discussion, it can be concluded that a DSS for GMO analysis requires close interactions at different levels in the procedure. For comparing the feasibility to fate of the platforms described for use in such an integrated approach, the following features were compared: performance (reported performance in proficiency testing and/or on real-life samples), flexibility (especially with respect to the introduction of new targets), applicability (especially evaluation of compatibility with currently applied strategies), cost (essential reagents and equipment) and, finally, the status of implementation conforming to ISO or other accepted standards.

PCR itself is already established in most of enforcement laboratories in the EU and also outside Europe. Most laboratories have had experience with PCR methods for many years. They are widely accepted as the methods of choice for GMO detection and a major advantage of this technology is the broad availability of suppliers of PCR reagents and equipment around the world at an acceptable price. Moreover, as one of the leading orgnaizations in the GMO detection world, the CRL-GMFF in the EU is so far validating exclusively real-time PCR methods for GMO quantification and is providing all the information necessary to carry out the analysis to the public without any restriction. On the basis of this experience, the European Commission—Joint Research Centre has recently introduced so-called prespotted plates, which clearly represent an interesting end-point analysis for the presence of a GMO. The system is still under assessment, but may in the future develop into an interesting complementation to other GMO analysis approaches. The other three simplex PCR-based strategies, CoSYPS [22], the TaqMan® PCR system developed by Waiblinger et al. [25] and the DualChip® array approach described by Hamels et al. [30] represent integrated screening approaches combining the decision process, the analytical part as well as the final interpretation of the laboratory results. All three of them have been tested on real samples and are to different extents currently applied in enforcement. CoSYPS and the TaqMan® PCR system can relatively easily be “updated” as they represent simplex PCR approaches. The DualChip® may, however, require more technical adaptations and will require expertise and specific know-how to be adapted. The DualChip® approach described by Hamels et al. [30] has been validated in an international collaborative study. The methods applied in the TaqMan® PCR system developed by Waiblinger et al. [25] have been assessed in interlaboratory trials within the German GMO network [2629] and the CoSYPS approach has been accredited at the Belgian national level and was recently tested in an interlaboratory trial (M. van den Bulcke, personal information). At the cost and investment level, CoSYPS is more attractive than the TaqMan® PCR system owing to the absence of the costly fluorescently labelled probes used in the TaqMan® PCR system; applying the DualChip® technology requires a considerable investment in new equipment and in training of the personnel.

Next to these simplex PCR technologies, multiplex systems have been developed [3436]. For real-time PCR it has become almost routine to develop new primer/probe systems and to combine these into a single reaction. A promising example is presented by the combination of multiplex PCR with capillary electrophoresis. Multiplex PCR with capillary electrophoresis needs conventional PCR and conventional sequencer devices.

However, any multiplexing PCR approach faces limits and becomes highly sophisticated and more prone to false positives if more than five to ten real-time PCR systems are to be combined. This is considered as one of the major disadvantages in the long-term development of the PCR method as a key technology in GMO screening analysis. Another clear prerequisite to carry out any type of PCR analysis is the expensive equipment and laboratory environment necessary. The logistic cost factor may become even more explicit if adaptation to the complex GM world would invoke switching from the common 96-well format to a larger format, which may demand major investment in new types of equipment (robots), adaptation of facilities and technical education of personnel.

With respect to the array-based detection methods that do not apply classic PCR amplification strategies, the recently developed NAIMA method uses multiplex isothermal amplification [44] and is considered to have good potential for multitarget detection. The method still needs further optimization before it can find its way in routine test laboratories, but it may provide an appropriate answer to the issue of sample complexity in the future. At the cost level, this approach does not need special equipment for the amplification steps, and it therefore can generate target and signal amplification at lower costs. The detection of an NAIMA amplicon requires investments in low-density array production and scanning equipment.

The padlock probe technology [41] and the Luminex xMAP technology [45] have so far not left the status of research; however, the potential for multiplexing is a significant advantage and might be even higher than for the PCR approach. Nevertheless, the need for specific constructs (padlock probes) or carrier bead sets (Luminex) makes these techniques more complicated than the classic PCR approach, which may have a considerable impact on the cost of development.

When all these aspects are into consideration, it is still the (simplex) PCR and its modification using real time that is considered to be the most applied and most suitable method for GMO analysis. Only if techniques are not only newly developed but also validated for the purpose of GMO detection will laboratories have the choice to select the most appropriate technique to enforce their national legal provisions. In general, the validation is one of the most crucial criteria for a laboratory to endorse any new method. Indeed, only validated methods should be used for control purposes as the criterion is required by the generally applicable ISO 17025 standard in GMO analysis. Therefore, it is not sufficient only to develop new methods but one must also decide on the best possible way to validate the new developed methods. In relation to multiplex approaches, it is too time-consuming and labour-intensive to validate all individual primer/probe systems on all different targets and with different concentrations of them. Therefore, new strategies need to be developed. It might be useful to think about the establishment of method criteria rather than validating individual methods. In the Codex Committee for Methods of Analysis and Sampling, a working group has been created to try to figure out those criteria. However, the work is ongoing and might need a few more years to be finished.

DSS for target selection and data interpretation tools

A DSS is a computerized information system used to support decision-making activity. A DSS is a valuable tool to generate decisions during the whole GMO testing process, from method development, its validation, selection of methods used on specific sample to final interpretation of the results obtained in the laboratory (Fig. 2). Development of the DSS for comparison of important characteristics of methods, such as its performance characteristics, practicability, applicability and cost, was presented at the final COEXTRA conference (http://www.coextra.eu/conference/programme). The DSS for evaluation of method performance in validation was developed by Bellocchi et al. [48]. Systems using the matrix approach, such as PCR-based CoSYPS and the DualChip® (Eppendorf, Germany), use a DSS. It may become a valuable tool in assessing target selection/combination to be used in laboratories during routine testing of samples, as well as for supporting the interpretation of the results.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-009-3237-3/MediaObjects/216_2009_3237_Fig2_HTML.gif
Fig. 2

Decision support systems (DSS) in GMO testing from method development and evaluation to final interpretation of results in the laboratory. At DSS level 1 the information available on the GMO (indicated at GMO universe) is combined with validated methods into a matrix table. This information is used in combination with sample data to define the optimal screening and identification strategy at DSS level 2. Finally the results obtained in the experimental PCR analysis are integrated and interpreted (DSS level 3)

An integrated DSS covering both method selection and result analysis has been presented as a prototype called “GMOtrack”. The system was developed by the National Institute of Biology in cooperation with the Jozef Stefan Institute (Ljubljana, Slovenia) and aims at assisting GMO testing to choose the most cost-effective testing strategies for a given sample. The approach is based on a data matrix containing, on the one hand, the GMOs associated with their plant species and, on the other hand, the methods for detecting GMOs; it can be easily updated according to the evolving market situation. The core algorithm generates a proposal for GMO testing strategies and evaluates them according to their expected cost. GMOtrack also supports the interpretation of wet-laboratory results. After the screening has been performed on a sample, the screening results can be introduced into the GMOtrack system. On the basis of the laboratory results and the data matrix, it computes the set of GMOs that are possibly present in the sample and lists the event-specific assays that need to be performed in the event-specific phase. The core algorithm is freely available under the terms of the General Public License on the Web page http://kt.ijs.si/petra_kralj/GMOtrack/ [23].

Concluding remarks

As described herein, many different approaches and techniques have been developed to detect commercial GMOs, including those which have received authorization only in a limited number of countries (asynchronous authorization). The effective combination of DSSs with high-throughput analytical approaches shows the possibility for additional efficient solutions of GMO testing of the rapidly increasing number of GMOs on the market.

From the methodological point of view, the CoSYPS [22], Waiblinger et al. [25] and Hamels et al. [30] strategies are integrated approaches combining the decision process, the analytical part as well as the final interpretation of the laboratory results. All three have been tested on real samples and have passed validation to different extents (see earlier) From the decision support point of view, the strategy applied in the GMOtrack approach is the most advanced so far. As such, GMOtrack could represent the basis to create an open platform specifically giving support for deciding on which GMO elements should be looked for and finally allowing the interpretation of laboratory results for the selected target elements. With this approach the user can apply any analytical method he/she has established in the enforcement laboratory. Theoretically, all the analytical methods described herein can be combined with GMOtrack to obtain a complete and integrated approach.

The recently developed matrix-based approaches described in this paper have the possibility of being implemented within a reasonable time and being able to cope in a practical and suitable way with the increasing number of GMOs authorized around the world. However, some aspects need to be carefully considered when deciding on what kind of matrix will be applied in combination with which type of methods. A key element is the proper definition of the scope and of the GMOs under investigation (the universe), which will by far determine the most appropriate method selection criterion. Indeed, depending on the national legislation and the specific situation regarding the authorization of certain GMOs, different sets of methods might be more suitable and/or required for optimal GMO screening/identification analysis. Also, the different types of sample(s) to be analysed (seeds, raw material or processed food/feed products) can have a considerable impact on the choice of the methods to be combined and applied.

The accessibility to correct DNA sequence information, validated methods and reference material is a fundamental prerequisite for setting up an effective strategy. The most appropriate source of information on the inserted DNA sequences is the notification dossier for authorization/regulation of the use of GMOs (see GMO Compass, Aphis and other national legislation sources). Next, the development of validated methods detecting such sequences according to accepted standards (e.g. ISO/CEN) is essential. Accessibility to all necessary information on the performance of such methods is crucial. Regular verification of the method performance for GMOs entering the market as well as previously commercialized varieties of approved GMOs is important [16]. Also, in view of the rapid increasing number of GMOs and their diversity, constant updating of this information will be necessary. In this respect, a centralized official database with reliable data may represent the most appropriate means to provide this information to the largest group of stakeholders.

Next to the availability of a formal description of the scope of the GMO analysis, the availability of reference material and, in an ideal situation of Certified Reference Material, is an additional key factor. So far, only in Europe does GMO legislation require that applicants make available such reference materials and provide information on where such materials can be purchased [6].

Which of the new approaches described will be developed further and be used on a larger scale in routine analyses depends on many factors, such as the cost and adaptability of the laboratories, which are currently using mainly PCR systems. Maybe more of them will find their appropriate place in GMO testing or be further combined in even more efficient systems.

Acknowledgements

The authors would like to thank Kristina Gruden and Dany Morisset from the National Institute of Biology for their valuable inputs to the manuscript.

Copyright information

© Springer-Verlag 2009