Analytical and Bioanalytical Chemistry

, Volume 396, Issue 6, pp 2055–2064

Establishment of a system based on universal multiplex-PCR for screening genetically modified crops

Authors

  • I-Jen Lu
    • Institute of Microbiology and Biochemistry, College of Life ScienceNational Taiwan University
  • Chih-Hui Lin
    • Institute of Microbiology and Biochemistry, College of Life ScienceNational Taiwan University
    • Institute of Microbiology and Biochemistry, College of Life ScienceNational Taiwan University
Original Paper

DOI: 10.1007/s00216-009-3214-x

Cite this article as:
Lu, I., Lin, C. & Pan, T. Anal Bioanal Chem (2010) 396: 2055. doi:10.1007/s00216-009-3214-x

Abstract

The rapid development of many genetically modified (GM) crops in the past two decades makes it necessary to introduce an alternative strategy for routine screening and identification. In this study, we established a universal multiplex PCR detection system which will effectively reduce the number of reactions needed for sample identification. The PCR targets of this system include the six most frequently used transgenic elements: cauliflower mosaic virus (CaMV) 35S promoter, Agrobacterium tumefaciens nopaline synthase (nos) promoter, Agrobacterium tumefaciens nopaline synthase (nos) terminator, the neomycin phosphotransferase II (nptII) gene, the 5-enolpyruvylshikimate-3-phosphate synthase (CP4 epsps) gene of Agrobacterium tumefaciens strain CP4, and the phosphinothricin N-acetyltransferase (pat) gene. According to the AGBIOS database, the coverage of this detection system is 93% of commercial GM crops. This detection system could detect all certified reference materials (CRMs) at the 1.0% level. The correct combination of all the CRM amplicon patterns proved the specificity of this multiplex PCR system. Furthermore, the amplicon patterns of this multiplex PCR detection system could be used as an index of classification which will narrow the range of possible GM products. The simulation result of this multiplex PCR detection system on all commercialized 139 GM products in the AGBIOS database showed that the maximum number of PCR reactions needed to identify an unknown sample can be reduced to 13. In this study, we established a high-throughput multiplex PCR detection system with feasible sensitivity, specificity, and cost. By incorporating this detection system, the routine GM crop-detection process will meet the challenges resulting from a rapid increase in the number of GM crops in the future.

Keywords

GMOMultiplex PCRDetection systemTransgenic elementUniversal screening system

Abbreviations

UPS

Universal primer sequence

GM

Genetically modified

Introduction

In the past two decades worldwide development of genetically modified (GM) crops has been extensive. Following successful and rapid commercialization, regulation of GM crops has been one of the most important food issues. Many countries in the world have set up food-labeling laws that introduce labeling of products for which the GMO content is above definite threshold [1]. This requirement makes it necessary to develop qualitative and quantitative methods for detection of GMO. Most detection methods for GMO are based on DNA or heterologous protein approaches [2, 3]. The most frequently used GMO detection methods are based on DNA, because of its high stability even in highly processed food products [4]. PCR-based technology is currently implemented in GMO detection laboratories. PCR methods for detection of GMO can be classified into several types according to the sorts of primers, including: screening primer, gene-specific primer, construct specific primer, and product-specific primer [5]. PCRs introducing primers which target the promoter, terminator, or other transgene are suitable for screening purpose. Despite extensive research and development of GM crops, the diversity of transgenic elements used in commercialized GM crops is surprisingly low. The most frequently used transgenic element is CaMV 35S promoter. Eighty-two out of 139 records in the AGBIOS database carry this element, which means approximately 59% of commercialized GM crops could be detected if we simply target this sequence. GMO screening method based on the PCR detection of some frequently used transgenic elements such as cauliflower mosaic virus 35S promoter, Agrobacterium tumefaciens nopaline synthase (nos) terminator, and the nptII sequences has been reported [6, 7]. These methods could detect most GM products, but systematic coverage is still not feasible for full-scale screening purposes.

Product-specific PCR has been regarded as the most comprehensive approach for GM crop detection and quantification. However, the highly specific nature of this method makes it less suitable for screening purposes. For this reason, multiplex PCR detection systems with combinations of several product-specific PCR have been introduced to improve the efficiency of the screening process [8, 9]. Onishi et al. [10] reported a multiplex PCR system for simultaneous detection of eight GM crops. Xu et al. [11] reported microarray-based GMO detection method for seven GM crops. Because of increasing taxonomic (diverse taxon host plants) and biotechnological (diverse genetic constructs) diversity of GMOs, product-specific detection strategies are not facilitated by the long testing time and large associated cost, even when using a multiplex PCR approach [12]. To overcome the challenge raised by the rapid increase in the number of GM crops, we have proposed a strategy using multiplex PCR as a preliminary screening and classification method. This method should be easily carried out, with high throughput, low cost, high specificity, and high sensitivity, and is expandable. In this study, we target six most frequently used transgenic elements to accomplish a feasible universal multiplex PCR detection system for routine screening of GM crops.

Materials and methods

Plant material

Seventeen certified reference materials (CRMs), NK603, 3272, MON863, MON810, GA21, MIR604, TC1507, DAS59122, MON863 × MON810, Bt11 and Bt176 maize; Roundup Ready (GTS40-3-2), 304523 and 356043 soybean; 281-24-236 × 3006-210-23 cotton; H7-1 sugar beet, and EH92-527-1 potato, were purchased from the EU Joint Research Centre, IRMM (Institute for Reference Materials and Measurements, Geel, Belgium). Genuine GM crop seeds of MON810, NK603, and GA21 maize were kindly provided by Monsanto (St Louis, MO, USA). A genetically modified CMV-resistant tomato line R8 developed by AVRDC (Asian Vegetable Research and Development Center, Shanhua, Tainan), was also used to test the capability of this system. The transgenic elements carried in the transgenic tomato R8 were one copy of nos promoter, nptII, 35S promoter, and two copies of nos terminator.

Extraction of genomic DNA

Genomic DNA was isolated from 50–70 mg of each sample using the DP02-150 plant genomic DNA purification kit (Genemark, Taipei, Taiwan). DNA concentration was quantified by use of a DNA quantitation kit, fluorescence assay (Sigma, St Louis. MO, USA). Fluorescence emission at 460 nm due to the fluorescent dye binding to the double-stranded DNA was detected by use of a microplate fluorescence reader (FLx800; BioTek Instruments, Winooski, VT, USA). DNA was diluted with deionized water to a concentration of 10–50 ng μL−1 before PCR analysis.

Simulation of a multiplex PCR system

To evaluate the feasibility of our multiplex PCR system, simulated classification of all 139 GM records in the AGBIOS database (www.agbios.com; retrieved 19 March 2009) was carried out by SPSS 12.0 software (SPSS, Chicago, IL, USA) using patterns of expected PCR product. The presence of transgenic elements in GM products was converted into a binary data matrix (1 = present, 0 = absent) according to the information provided in the database. Hierarchical cluster analysis using the UPGMA (unweighted pair group method with arithmetic mean) method, which creates a phylogenic tree [13], was carried out to simulate the result of our multiplex PCR system.

Oligonucleotide primers

Oligonucleotide primers used in this study were designed using Vector NTI Advance 9 software (Invitrogen, Carlsbad, CA, USA). The following criteria were applied to design primers:
  1. 1.

    high efficiency and robustness (each amplification spans 70–350 bp); and

     
  2. 2.

    annealing temperature between 55 and 65 °C.

     
Primers used in this study are listed in Table 1. To enhance the specificity and efficiency of multiplex-PCR, each primer is “tagged” on the 5′ end with an unrelated 20-nucleotide sequence (universal primer sequence) 5′-GCGGTCCCAAAAGGGTCAGT-3′ [14]. The primers were synthesized by MDbio (Taipei, Taiwan).
Table 1

The primers used in this study

Primer sets

Sequence 5′-3′

Target element

Amplicon size*

Source

nosFZMP1

GAATCCTGTTGCCGGTCTTG

nos terminator

125

[15]

nosFZMP2

GCGGGACTCTAATCATAAAAACC

 

nos-proF

TGAGACTCTAATTGGATACCGAGGG

nos promoter

201

This study (GeneBank AB294471)

nos-proR

TTTGGAACTGACAGAACCGCAAC

35S3F

GCCATCATTGCGATAAAGGAAAGG

35S promoter

173

This study (GeneBank CS437393)

35S6R

TTGTGCGTCATCCCTTACGTCAGTG

NPTII-3

GAGGC TATTC GGCTA TGACT

nptII

271

[15]

NPTII-4R

AAGGT GAGAT GACAG GAGAT

 

epsps-11F

YGTGTTGAACCCRCTKCGCGAAAT

epsps

101

This study

epsps-12R

TGATYGGCGTYGGMGTCTTYGGY

PAT-4F

CACTCTTGTGGTGTTTGTGGCTCTG

pat

102

This study

PAT-4R

AGCTACAGCAGCTGATATGGCCG

*The amplicon sizes are results of the amplifications by use of a specific primer (without the utilization of UPS-tag)

PCR conditions

Uniplex PCR was performed with Super-Therm PCR core reagents (Bertec Enterprise, Taipei, Taiwan) in 25 μL reaction volume including 1× reaction buffer, 50 μmol L−1 dNTP, 0.3 μmol L−1 of each primer, 0.375 units DNA polymerase, and DNA template. Fifty nanogram template DNA was used for each reaction. PCR reactions were performed in a PCRSprint thermal cycler (Model HBSP02110; HBPXE05110; Thermo Fisher Scientific, Waltham, MA, USA). The touch-down thermal cycling program included initial denaturation for 5 min at 95 °C; and then 10 cycles of 30 s at 95 °C, 30 s at 58 °C, and 30 s at 72 °C; then 25 cycles of 30 s at 95 °C, 30 s at 60 °C, and 30 s at 72 °C, followed by a final extension for 10 min at 72 °C. The concentration of primers, dNTP, and DNA polymerase were tested. The annealing temperature was tested from 58 to 68 °C using gradient PCR. The optimized multiplex PCR conditions were: 25 μL PCR mixture containing 100 μmol L−1 dNTP, 1.6 μmol L−1 of each primer, and 0.5 units DNA polymerase. The thermal cycling program was: initial 10 min denaturation at 95 °C; 35 cycles with 30 s at 95 °C, 30 s at 62 °C, and 30 s at 72 °C; and a final extension for 10 min at 72 °C.

Agarose gel electrophoresis

Electrophoresis of PCR products were carried out in 3% agarose (UltraPure Agarose, Invitrogen) gels for 40 min at 100 V in 1 × TAE buffer. The gels were stained with EtBr and photographed by DigiGel (DGIS-8 Digital Gel Image System; Topbio, Taipei, Taiwan).

Results

Simulation of multiplex PCR system

To maximize the coverage of our system, we selected the most frequently used transgenic elements as PCR target based on information from the AGBIOS biotech crop database. According to the database, the top six of most frequently used transgenic elements among 139 products were: cauliflower mosaic virus (CaMV) 35S promoter (present in 82 products), Agrobacterium tumefaciens nopaline synthase (nos) terminator (present in 59 products), neomycin phosphotransferase II (nptII) gene (present in 41 products), phosphinothricin N-acetyltransferase (pat) gene (present in 34 products), 5-enolpyruvylshikimate-3-phosphate synthase (CP4 epsps) gene of Agrobacterium tumefaciens strain CP4 (present in 27 products) and Agrobacterium tumefaciens nopaline synthase (nos) promoter (present in 20 products). These were selected and incorporated into our system. To evaluate the feasibility of the multiplex PCR system with these six transgenic elements as targets, we used SPSS 12.0 software (SPSS Inc., Chicago, IL, USA) to simulate the results of this system on all GM products in the AGBIOS database. The presence of these six transgenic elements was treated as measured characteristics (square Euclidean distance) of hierarchical cluster analysis. As a result, GM products with the same combination of these transgenic elements were classified within the same clusters. The result of simulation (Table 2) revealed that all of 139 GM products will be classified into 25 (24 positive, 1 negative) groups according to the PCR product patterns generated by our multiplex PCR system. The largest positive group (group 9) contains 12 products. The reasonable distribution of GM products among 24 groups makes the result of this multiplex system applicable as an index of classification. Preliminary classification will drastically reduce the list of possible products in an unknown sample. According to the result of simulation, to identify the product in an unknown positive sample it will be necessary to carry out a maximum of 1 (multiplex screening) +12 (product-specific) PCR reactions. The result of simulation showed that a multiplex PCR detection system targeting the six most frequently used transgenic elements is applicable for GM crops screening.
Table 2

Grouping of events using simulated results of multiplex PCR as index

Group no.

Transgenic elements

Number of events

nptII gene

nos promoter

CaMV 35S promoter

nos terminator

pat gene

CP4 epsps gene

1

+

8

2a

28

3

+

1

4

+

7

5

+

+

1

6

+

+

1

7

+

+

12

8

+

3

9b

+

12

10

+

+

4

11

+

+

+

+

3

12

+

+

+

+

+

1

13

+

+

+

+

6

14

+

+

+

11

15

+

+

+

8

16

+

+

1

17

+

+

+

5

18

+

+

7

19

+

+

6

20

+

+

+

+

2

21

+

+

+

3

22

+

+

+

+

1

23

+

+

+

1

24

+

+

+

+

4

25

+

+

+

+

+

3

aNegative group

bThe largest positive group

Design of primers

The specific primers used in this study are listed in Table 1. Primer pairs NPTII-3/4R and nosFZMP1/2 were adopted from a previous report [15] for nptII gene and nos terminator detection. Primer pair 35S-3F/6R and nosproF/R were designed on the basis of CaMV 35S promoter sequence (GeneBank accession no. CS437393) and nos promoter sequences (GeneBank accession nos AB294471 and AB294451), respectively. Because of to the limited access and unexpected variation among pat gene sequences in GM crops, neither designed primers nor adopted primers based on published reports could successfully detect all pat genes. Therefore, the sequence of pat gene amplified from the TC1507 maize using published primers [16] was used to design new primer pair pat-4F/4R. Primer pair pat-4F/4R was proved to amplify pat gene sequence of the products including TC1507, DAS59122, Bt11 maize, and 281-24-236 × 3006-210-23 cotton. A similar problem was also encountered when we were acquiring primers to detect the CP4 epsps gene. The primers we designed on the basis of the sequence of Glycine max CP4 epsps gene (GeneBank accession nos AB209952 and AF464188) and CP4 epsps gene of synthetic construct from Monsanto Technology (GeneBank accession no. CS362717) could not consistently amplify the target sequence from product H7-1. On the contrary, primers from Xu’s report [17] could only amplify the CP4 epsps gene in H7-1. According to the sequence of the amplicon fragment, we found that the CP4 epsps gene in H7-1 was a variant compared with the same gene in other products. We redesigned a pair of new CP4 epsps primer on the basis of the conserved region of amplicon from H7-1 and GeneBank sequence AF464188. The primer pair epsps-11F/12R was proved to effectively amplify all the CP4 epsps target region of NK603, Roundup Ready soybean, and H7-1.

UPS-tagged primers for multiplex PCR

The major considerations of a multiplex PCR are efficiency and robustness. Although the annealing temperature and primer concentrations may be calculated to some degree, conditions generally must be defined empirically in multiplex reactions [14]. As primer sets were added to the multiplex PCR reactions, the problem of efficiency bias may arise from complex primer–primer interaction or artifacts resulting from nonspecific amplification. In the early stage of multiplex PCR system development, we tried to optimize the multiplex system by adjusting PCR conditions such as annealing temperature, components of reaction mixture, and the ratio among primer sets. After optimization, the bias of efficiency was reduced to an acceptable level. Nevertheless, the PCR conditions were too critical to achieve feasible reproducibility when applied to CRMs. Therefore, we adopted the chimeric-specific primer strategy in which a “universal primer sequence (UPS)” was tagged to the 5′ of primers [14].

As a result, the UPS added to the 5′ of primers not only raised the efficiency but also increased the specificity of primers (Fig. 1a). The larger PCR products of UPS–primers resulted from the extra length of universal primer sequence (20 bp). Each UPS–primer set was tested individually before addition to the multiplex PCR reaction (Fig. 1b).
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-009-3214-x/MediaObjects/216_2009_3214_Fig1_HTML.gif
Fig. 1

The effect of UPS–primer. (a) The effect of UPS tag on the efficiency and specificity of PAT-4F/4R primer set. (lanes 1 and 1′: TC1507; lanes 2 and 2′: DAS59122; lanes 3 and 3′: Bt11; lanes 4 and 4′: 281-24-236 × 3006-210-23). Lanes 1, 2, 3, and 4: primer set PAT-4F/4R; lanes 1′, 2′, 3′, and 4′: primer set UPS-PAT-4F/4R. Lane M: 100 bp DNA ladder. (b) Uniplex PCR test for individual transgenic elements. Lanes: 1, nptII gene; 2, nos promoter; 3, 35S promoter; 4, nos terminator; 5, pat gene; 6, CP4 epsps gene; M. 100-bp DNA ladder

Optimization of the multiplex PCR conditions

The PCR conditions were optimized with UPS–primers. To optimize the annealing temperature, a gradient PCR with annealing temperature from 58 to 68 °C was performed. Because there was no single product containing all six transgenic elements, three products including R8, DAS59122, and NK603 were used. The components of the PCR reaction mixture including UPS–primers, dNTP, and DNA polymerase concentration were also evaluated. The results showed that the optimum annealing temperature was 62 °C. The optimized PCR reaction mixture contains 1.6 μmol L−1 of each UPS–primer, 100 μmol L−1 dNTP, and 0.5 units DNA polymerase (data not shown) in the total volume 25 μL.

Application of multiplex PCR to GM crop samples

The performance of each UPS–primer set on CRM or GM tomato (R8) samples was tested before addition to the multiplex PCR. Primer sets targeting CaMV 35S promoter, nos terminator, nptII gene, nos promoter, pat gene, and CP4 epsps gene were tested independently. The CaMV 35S promoters are present in NK603, MON863, MON863 × MON810, Bt11, R8, RRS, MON810, TC1507, 59122, and Bt176. The nos terminators are present in NK603, MON863, MON863 × MON810, Bt11, R8, RRS, 3272, GA21, MIR604, and EH92-527-1. The nptII genes are present in MON863, MON863 × MON810, R8, and EH92-527-1. The nos promoters are present in R8 and EH92-527-1. The pat genes are present in TC1507, DAS59122, Bt11, and 281-24-236 × 3006-210-23. The CP4 epsps genes are present in NK603, H7-1, and RRS. All PCR reactions were performed in triplicate. The only unexpected PCR product (approx. 420 bps) observed was amplification of the 35S promoter sequence using primer set UPS-35S-3F/6R on NK603 (Fig. 2). The BLAST result of this sequence showed that the unexpected PCR product resulted from the enhanced 35S promoter sequence (GeneBank AY739898.1). The enhanced 35S promoter contains a partial tandem repeat of the original 35S promoter sequence, which leads to an extra PCR product amplified by the same primer set UPS-35S-3F/6R. The specificity and efficiency of UPS–primers were proved in uniplex PCR test.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-009-3214-x/MediaObjects/216_2009_3214_Fig2_HTML.gif
Fig. 2

Specificity of UPS–primers targeting for 35S promoter: lane 1, no template control; lanes 2–11, NK603, MON863, MON863 × MON810, Bt11, R8, RRS, MON810, TC1507, DAS59122, and Bt176, in which the CaMV 35S promoters are present; lanes, 12–19, 3272, GA21, MIR604, EH92-527-1, 281-24-236 × 3006-210-23, H7-1, 356043, and 305423, in which the CaMV 35S promoters are absent; lane M, 100-bp DNA ladder

The result of the multiplex test on 18 GM samples is shown in Fig. 3a. Few unexpected PCR products appeared when we applied multiplex PCR to specific products (Fig. 3b). The results for MON863 and MON863 × MON810 (lanes 3 and 4, respectively) showed that three of unexpected PCR products in two sizes (420 and 700 bp, approximately) were consistently amplified. These unexpected bands amplified constantly under a wide range of PCR conditions. Sequence analysis revealed that in these three PCR products was the sequence between the 35S promoter-nptII gene and the nos terminator-35S promoter (4-AS1) of the MON 863 transgene construct. After cloning and sequencing, the band with size approximately 420 bp actually contains two PCR products with very similar size. One sequence of shorter PCR products (approx. 420 bp) was the same as the longer one (approx. 700 bp), but missing a part of sequence in the 35S promoter and nptII gene junction (Fig. 4). Another sequence of shorter PCR products (approx. 420 bp) resulted from amplification of nos terminator-35S promoter (4-AS1). The sequence of the unexpected PCR product (approx. 510 bp) amplified from GTS-40-3-2 (RRS) resulted from the enhanced 35S promoter as in NK603. The result of multiplex PCR on Bt11 showed that three of unexpected products were present. These three PCR products resulted from amplification of pat gene to nos terminator (718 bp) , 35S promoter to nos terminator (991 bp), and nos terminator to another nos terminator (very faint, 1202 bp). TC1507 and DAS59122 resulted in amplification of an unexpected PCR product of size approximately 414 bp. Because of the same orientation of the transgene construct of these two products, the unexpected PCR products were actually the same sequence resulting from the 35S promoter-pat gene sequence. All the sequences of these unexpected PCR products have revealed that they are the co-products between primer sets rather than unspecific PCR products [18].
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-009-3214-x/MediaObjects/216_2009_3214_Fig3_HTML.gif
Fig. 3

Test of universal multiplex PCR on GM products: (a) pattern of PCR products; (b) expected pattern of PCR products. Lane 1, no template control; lanes 2–19, NK603, MON863, MON863 × MON810, Bt11, R8, RRS, MON810, TC1507, DAS59122, Bt176, 3272, GA21, MIR604, EH92-527-1, 281-24-236 × 3006-210-23, H7-1, 356043, and 305423; lane M, 100-bp DNA ladder

https://static-content.springer.com/image/art%3A10.1007%2Fs00216-009-3214-x/MediaObjects/216_2009_3214_Fig4_HTML.gif
Fig. 4

Schematic diagram of the unexpected PCR co-product amplified from MON 863 and MON 863 × MON 810. p35S, CaMV 35S promoter; nptII, neomycin phosphotransferase II

The PCR co-products of MON863 and MON863 × MON810 resulted from the primers 35S-6R/nosFZMP1 and 35S-3F/NPTII-4R. The PCR co-products of Bt11 resulted from primers of PAT-4F/nosFZMP2, 35S-3F/nosFZMP2, and nosFZMP1/nosFZMP2. The PCR co-products of TC1507 and DAS59122 resulted from primers of PAT-4F/35S-6R. Product 305423, which has no targeted transgenic element in this system, results in a negative pattern. It could represent the products belonging to the negative group of this multiplex PCR detection system. Product 356043 contains a constitutive synthetic promoter SCP1, which contains a portion of the CaMV 35S promoter. However, we could not obtain information about this new product as a public commercialized GM product in the GMO database at the time we designed primers for the CaMV 35S promoter. Consequently, the primer pair of the CaMV 35S promoter have not been made to adapt the 35S promoter partial region of the synthetic promoter. We confirmed relativity between the sequence of the SCP1 promoter and the 35S promoter region we target, and corroborated that the reverse primer binding site of the 35S promoter is not involved in the SCP1 synthetic promoter.

Detection limit of the universal multiplex-PCR detection system

The limit of detection (LOD) of this screening system was tested using Certified Reference Materials with different GM content (0.1%, 0.5%, 1.0%, 5.0% and 10%, if available). All 13 products (NK603, MON863, MON863 × MON810, Bt11, RRS, MON810, TC1507, DAS59122, Bt176, 3272, GA21, MIR604 and 281-24-236 × 3006-210-23) available from IRMM were used to determine the lowest detectable relative percentage of GM materials. Each test was performed in triplicate in three independent experiments. Results showed that the general detection limit of the universal multiplex PCR system is 1.0%. Some products could be detected at levels of 0.5% to 0.1% (Table 3).
Table 3

Results from detection limit test

Product

GM content level (%)

0.1

0.5

1

5

10

NK603

+

+

+

NA

MON863

NA

+

NA

+

MON863 × MON810

NA

+

NA

+

Bt11

+

+

+

+

NA

GTS 40-3-2

+

NA

NA

NA

MON810

+

+

+

NA

TC1507

NA

+

NA

+

DAS59122

+

+

+

Bt176

+

+

+

NA

3272

NA

+

NA

+

GA21

+

+

+

+

NA

MIR604

+

+

+

281-24-236 × 3006-210-23

NA

NA

+

NA

+

+, positive, −, negative; NA, not available

Discussion

Product-specific PCR has been regarded as the most comprehensive approach for GM crop detection and quantification. Multiplex PCR detection systems with combinations of several product-specific PCR were introduced to improve the efficiency of screening processes [811]. Because of the increasing taxonomic (diverse taxon host plants) and biotechnological (diverse genetic constructs) diversity of GMOs, product-specific detection strategies are not facilitated in terms of large testing time and huge associated cost even using a multiplex PCR approach [12].

To overcome the challenge arising from the rapid increase in the number of GM crop products, we proposed a strategy using universal multiplex PCR for preliminary screening and classification. According to information from the AGBIOS biotech crop database, 139 GM crops have been commercially approved throughout the world. Products originating from natural mutant, chemical mutagenesis, and somaclonal variants selection were excluded from our target list for detection, because the point-mutated genes carried in these products are not detectable using conventional PCR. Therefore, these 20 products were excluded, bringing the number of target products to 119. Most frequently used transgenic elements were selected as PCR targets according to information in the AGBIOS database. Hierarchical cluster analysis was carried out to simulate the result of multiplex PCR on all the GM products. Various combinations of PCR target were evaluated using a data matrix converted from the AGBIOS database to maximize the coverage and feasibility of multiplex PCR. Because of the marginal effect on system coverage (only approx. 1%) and the difficulty of incorporating the seventh PCR target, the number of multiplex PCR targets was set to 6. The best combination of PCR targets includes CaMV 35S promoter, nos terminator, nptII gene, pat gene, CP4 epsps gene, and nos promoter. This combination of PCR targets will classify all GM products into 25 groups according to the expected multiplex PCR product pattern. The coverage of this combination was approximately 93% (111 out of 119) and the largest positive group contains only 12 products.

Besides the high coverage, the relatively even distribution of products among groups also effectively reduced the range of possible products in an unknown sample. Therefore, to identify the product in an unknown sample will need 1 universal multiplex PCR and 12 product-specific PCR at maximum. Compared with full-scale product-specific identification, the maximum number of PCR reactions needed for one sample was greatly reduced from 119 to 13. Cluster analysis makes it easy to evaluate detection of GM crops. During the development of this method, about 10 products were approved and added to the AGBIOS database. The performance of universal multiplex PCR method could be easily recalculated and updated by cluster analysis.

Design of primers is the major challenge of the universal multiplex PCR system. Only few sequences of transgenic constructs are accessible to public. In effect, the sequences of transgenic elements in GM crops usually differ from the published records in the database because of the transformation process or the modification done by developers. Primers designed solely on the basis of the published sequence usually do not work well. There is no record in databases of variation of CP4 epsps and pat gene sequences in GM crops. The limited access to detailed information about GM crops means some primers in published reports designed for these two genes only work for some specific products. In this study, some of primers had to be redesigned according to the consensus of sequences in GM crops. Eight products could not be detected by use of this multiplex PCR system. The reasonable approach to cover these products is to develop an product-specific multiplex PCR method. Unfortunately, limited access to information and to genuine GM crop material also makes this difficult. For these reasons, we believe that fully accessible information and genuine commercialized GM crop material is essential for reliable detection methods. Regulation of GM crops will also benefit from public access to detailed information and genuine commercialized GM crop material as a result of reliable and efficient detection methods.

Eighteen GM crop products including all available 17 CRMs and GM tomato product R8 were used to prove the specificity and feasibility of the universal multiplex PCR system. The result of multiplex PCR on all 18 products proved that this system was feasible for GM crop-screening purposes. A laboratory-phase GM tomato product R8 was included to demonstrate the potential of this system to detect unproved products which may result from contamination or gene flow. PCR co-products which result from co-amplification of adjacent transgene elements were present in the amplification of some specific products. These PCR co-products did not interfere with the results of multiplex PCR because of their larger size. In effect, these distinct PCR co-products will help identify possible products in unknown samples. The sequences of all PCR co-products matched the description of published information except one of the PCR co-products from GM maize MON863 and MON863 × MON810 [18]. According to the description of MON 863 and MON 863 × MON 810 in published reports [19, 20], there is only a single intact copy of the 35S promoter–nptII construct in the maize genome. The PCR co-product presents at approximately 700 bp exactly matched the 35S promoter–nptII fragment of the MON 863 construct. Another shorter PCR co-product presents at 420 bp was the 35S promoter–nptII construct fragment which was missing part of its sequence at the junction. These shorter PCR co-products were consistently present in the multiplex amplification of maize MON 863 and MON 863 × MON 810 among all the CRM1 batches used in this study, because this PCR co-product is specific for MON 863 and MON 863 × MON 810 and there is no other such construct (missing a part of sequence at the 35S promoter–nptII junction) present in other GM crops. It is not possible to explain the existence of the shorter PCR co-product if only one copy of intact 35S promoter–nptII construct exists in the maize genome, so we think it plausible that an unknown, non-functional copy of the 35S promoter–nptII construct which is missing part of sequence at junction (380 bp: 202 bp at 5′ of nptII gene, 178 bp at 3′ of 35S promoter) is present in maize MON 863 and MON 863 × MON 810.

The detection limit of this system could be considered as 1% of the GM content according to tests on different CRMs. This detection limit will satisfy labeling requirements for most countries in the world. In some places, for example the European Union and Russia, a 0.9% threshold has been set up for GM content labeling. Results from detection limit testing showed that most of CRMs could be detected at the 0.5% level. Judging from this result, the universal multiplex PCR detection system is plausible for detection of GM crops at 0.9% level.

Agarose gel electrophoresis is an ideal method for routine GMO screening because of its popularity and low cost. For this reason, the universal multiplex PCR detection system was successfully designed to take advantage of agarose gel electrophoresis. The capability of the universal multiplex PCR system also could be further expand by incorporating other analytical methods, for example microarray or capillary gel electrophoresis (CGE) depending on the resources available.

Conclusion

In this study, we established a high-throughput multiplex PCR detection system with feasible sensitivity, specificity, and cost. This detection system will cover about 93% of conventional PCR detectable commercialized GM crops. Besides the high coverage, the relative even distribution of products among groups also effectively reduced the range of possible products in an unknown sample to 12. Therefore, to identify the product in an unknown sample will need 1 universal multiplex PCR and 12 product-specific PCR at maximum. The detection limit of this system could be regarded as 1% of GM content according to the test on different levels of CRMs. The result for a laboratory-phase GM tomato product R8 has demonstrated the potential of this system in detecting unproved products which may result from contamination or gene flow. The capability of the universal multiplex PCR system also could be further extended by incorporating other methods of analysis, for example microarray or capillary gel electrophoresis (CGE), depending on the resources available. In summary, this versatile detection system will meet the challenges of routine GM crop detection resulting from a rapid increase in the number of GM crops in the future.

Footnotes
1

The latest CRMs used in this study were ERM-BF417d (MON863 × MON810; sample no. 0046) and ERM-BF416d (MON863; sample no. 0133).

 

Copyright information

© Springer-Verlag 2009