Quantitation of Transgenic Bt Event-176 Maize Using Double

Mar 2, 2004 - In this work, a new procedure useful to quantitatively analyze genetically modified organisms (GMOs) in foods is described and applied t...
19 downloads 8 Views 149KB Size
Download PDF
Anal. Chem. 2004, 76, 2306-2313

Quantitation of Transgenic Bt Event-176 Maize Using Double Quantitative Competitive Polymerase Chain Reaction and Capillary Gel Electrophorsesis Laser-Induced Fluorescence Virginia Garcı´a-Can˜as, Alejandro Cifuentes,* and Ramo´n Gonza´lez*

Institute of Industrial Fermentations (CSIC), Juan de la Cierva 3, 28006 Madrid, Spain

In this work, a new procedure useful to quantitatively analyze genetically modified organisms (GMOs) in foods is described and applied to analyze transgenic Bt Event176 maize. The method developed consists of coamplifications of specific DNA maize sequences with internal standards using quantitative competitive PCR (QC-PCR). The QC-PCR products are quantitatively analyzed using a capillary gel electrophoresis (CGE) with laser-induced fluorescence detection (LIF) method developed at our laboratory that utilizes a physically adsorbed coating. The CGE-LIF procedure allows the use of internal standards differing by only 10 bp from the original target fragments, to our knowledge, the smallest size difference that can be found in the bibliography for QC-PCR of GMOs. A spectrofluorometric procedure using ROX reference dye is proposed to solve calibration problems of input DNA concentration. It is demonstrated that the use of ROX drastically enhances the accuracy of the quantitative analysis by QC-PCR. Reproducibility of analysis times and corrected peak areas (measured as target/competitor PCR products ratio) for the CGE-LIF separations are determined to be better than 0.91 and 1.93% (RSD, n ) 15) respectively, for three different days. It is shown that CGELIF provides better resolution and a signal/noise ratio improvement of ∼700-fold compared to slab gel electrophoresis. The good possibilities in terms of quantitative analysis of GMOs provided by this new method are confirmed by determining the Bt Event-176 maize content in certified reference maize powder and food samples of known composition. This procedure opens the possibility for accurate quantitation of multiple GMOs in a single run. Placing in the market of food or feed consisting or containing genetically modified organisms (GMOs) in the European Union (EU) is subject to compulsory labeling.1,2 Because an adventitious * Corresponding authors. E-mail: [email protected] and acifuentes@ ifi.csic.es. Fax: 34-91-5644853. (1) Regulation (EC) 258/97 of the European Parliament and of the Council of 27 January 1997 concerning novel foods and novel food ingredients. OJ No L043, 14.2.1997; pp 1-7. (2) Council Regulation (EC) 1139/98 of 26 May 1998 concerning the compulsory indication of the labeling of certain foodstuffs produced from genetically modified organisms of particulars other than those provided for in Directive 79/112/EEC, Oj No L159, 3.6.1998; pp 4-7.

2306 Analytical Chemistry, Vol. 76, No. 8, April 15, 2004

contamination of GM material in a non-GM background is difficult to avoid, and labeling as “GMO-containing” could severely affect the marketing of food products, EU regulations have fixed a 1% threshold for adventitious contamination that is not subject to labeling requirement. This has created a demand for analytical methods that can quantify the amount of GMO in foods. Moreover, it is expected that this regulation will be very soon modified to an even more demanding threshold of 0.9%.3 Most of the analytical methods proposed to determine the GMO content in foods are based on polymerase chain reaction (PCR) due to its sensitivity, specificity, and applicability to the analysis of complex food matrixes. The first methods developed for the quantitative analysis of GMOs in foods were based on competitive quantitative PCR (QC-PCR),4-7 which was initially developed in medicinal research to measure viral loads.8-9 QCPCR makes use of an internal standard (competitor DNA), added in a known amount to the amplification reaction, which competes with the target DNA for the same primers. Competitors used for QC-PCR should contain a differential property in order to be distinguished from the original target and are usually constructed by performing small insertions or deletions in the original target. However, because a good quantitation can only be obtained when the target and the internal standard sequences are amplified with the same efficiency, the difference in size between the target and the internal standard should be kept to a minimum.10,11 Besides, reliability of the quantitation by QC-PCR depends, to a great extent, on the technique used to detect the amplification products.12 The usual procedure consists (3) Commission Regulation (EC) 49/2000 of 10 January 2000 on the labeling of foodstuffs and food ingredients containing additives and flavorings that have been genetically modified or have been produced from genetically modified organisms, OJ No L006, 11.1.2000; pp 15-17. (4) Studer, E.; Rhyner, C.; Lu ¨ thy, J.; Hu ¨ bner, P. Z. Lebensm. Unters. Forsch., A 1998, 207, 207-213. (5) Hu ¨ bner, P.; Studer, E.; Lu ¨ thy, J. Food Control 1999, 10, 353-358. (6) Hardegger, M.; Brodmann, P.; Herrmann, A. Eur. Food Res. Technol. 1999, 209, 83-87. (7) Hupfer, C.; Hotzel, H.; Sachse, K.; Moreano, F.; Engel, K. H. Eur. Food Res. Technol. 2000, 212, 95-99. (8) Gilliland, G.; Perrin, S.; Blanchard, K.; Bunn, H. F. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 2725-2729. (9) Piatak, M.; Luk, K.; Williams, B.; Lifson, J. D. BioTechniques 1993, 14, 70-80. (10) Zimmermann, K.; Mannhalter, J. W. BioTechniques 1996, 21, 268-279. (11) McCulloch, R. K.; Choong, C. S.; Hurley, D. M. PCR Methods Appl. 1995, 4, 219-226. 10.1021/ac035481u CCC: $27.50

© 2004 American Chemical Society Published on Web 03/02/2004

of the separation of the PCR products by slab gel electrophoresis, staining the gels with ethidium bromide, recording the image by using a digital imaging device, and quantifying the DNA fragments with specialized software. However, these procedures are poorly reproducible and accurate, require large quantities of sample, and are time-consuming. In addition, the limited resolution of slab gel electrophoresis imposes restrictions to the design of internal standards. An alternative method, real-time PCR (RT-PCR) is gaining popularity over QC-PCR for the quantitation of GMOs in food samples.13-18 RT-PCR is based on the use of fluorescent markers for monitoring the PCR product formed during each cycle of the reaction. Quantitation is then performed in an early step of the reaction, when the efficiency is still constant and the PCR product concentration correlates well with the concentration of the initial target molecules.19 However, the interlaboratory reproducibility of the method is too low20,21 as has been demonstrated by the high RSD obtained. For example, Ho¨hne et al.17 reported RSD values equal to 40% for the quantitation of either 0.1 or 1% Bt Event176 maize by RT-PCR, and RSD values ranging from 17.3 to 26% are commonplace in the quantification of 1% GMO content by RTPCR,22,23,24 or 33.4% for 0.1% GMO content.24 In this work, we propose an alternative method for the quantitative analysis of GMOs in foods based on QC-PCR with internal standards differing by only 10 bp from the original target DNA, together with the use of capillary electrophoresis and laserinduced fluorescence detection (LIF) to analyze the amplification products. This technique involves a high degree of automatization, uses minimum quantities of sample and reagents, is able to produce separations of PCR products with great efficiency, and has proven to be a good alternative to obtain accurate, precise, and sensitive results in the quantitation of PCR-amplified DNA fragments.25-33 (12) Bouaboula, M.; Legoux, P.; Pesse´gue´, B.; Delpech, B.; Dumont, X.; Piechaczyk, M.; Casellas, P.; Shire, D. J. Biol. Chem. 1992, 267, 2183021838. (13) Vaı¨tilingom, M.; Pijnenburg, H.; Gendre, F.; Brignon, P. J. Agric. Food Chem. 1999, 47, 5261-5266. (14) Berdal, K. G.; Holst-Jensen, A. Eur. Food Res. Technol. 2001, 213, 432438. (15) Taverniers, I.; Windels, P.; Van Bockstaele, E.; De Loose, M. Eur. Food Res. Technol. 2001, 213, 417-424. (16) Terry; C. F.; Harris, N. Eur. Food Res. Technol. 2001, 213, 425-431. (17) Ho ¨hne, M.; Santisi, C. R.; Meyer, R. Eur. Food Res. Technol. 2002, 215, 59-64. (18) Holck, A.; Va, M.; Didierjean, L.; Rudi, K. Eur. Food Res. Technol. 2002, 214, 449-453. (19) Higuchi, R.; Fockler, C.; Dollinger, G.; Watson, R. BioTechnology 1993, 11, 1026-1030. (20) Pauli, U.; Liniger, M.; Schrott, M. Mitt. Lebensm. Hyg. 2001, 92, 145-158. (21) EU Tender Report, 2000. Development of qualitative as well as quantitative detection methods to identify a genetic modification in soybean and maize products. Report of the EU Tender XXIV/98/A3/001 or Internet http:// europa.eu.int/comm/food/fs/biotech/biotech05_en.pdf. (22) Brodmann, P. D.; Ilg, E. C.; Berthoud, H.; Herrmann, A. J. AOAC Int. 2002, 85, 646-653. (23) Hubner, P.; Waiblinger, H. U., Pietsch, K.; Brodmann, P. J. AOAC Int. 2001, 84, 1855-1864. (24) Pardigol, A.; Guillet, S.; Po ¨pping, B. Eur. Food Res. Technol. 2003, 216, 412-420. (25) Kuypers, A. W.; Meijerink, J. P.; Smetsers, R. F.; Linssen, P. C.; Mensink, E. J. J. Chromatogr., B 1994, 660, 271-277. (26) Fasco, M. J.; Treanor, C. P.; Spivak, S.; Finge, H. L.; Kaminsky, L. S. Anal. Biochem. 1995, 224, 140-147. (27) Borson, N. D.; Straus-bauch, M. A.; Wettstein, P. J.; Oda, R. P.; Johnston, S. L.; Lenders, J. P. Eur. Food Res. Technol. 1998, 213, 432-438.

EXPERIMENTAL SECTION Chemicals. All chemicals were of analytical reagent grade and used as received. Tris(hydroxymethyl)aminomethano (TRIS), sodium dodecyl sulfate (SDS), guanidine hydrochloride, ethidium bromide, ethylenediaminetetraacetic acid (EDTA), and λ DNA HindIII digest were from Sigma (St. Louis, MO). and 2-hydroxyethylcellulose (HEC) (MW 90 000) and poly(vinyl alcohol) (PVA; MW 50 000) were from Aldrich (Milwaukee, WI). MS8 agarose was from Pronadisa (Madrid, Spain). DNA molecular weight marker XIII, ScaI restriction enzyme, proteinase K, and RNAse A were from Roche (Barcelona, Spain), and chloroform from Scharlau (Barcelona, Spain), isoamylic alcohol and N-cetyl-N,N,Ntrimethylammonium bromide from Merck (Darmstadt, Germany), and phenol from LabClinics (Madrid, Spain) were used. YOPRO1 and SYBRGreen-I were from Molecular Probes (Leiden, The Netherlands), and ROX Reference dye was from Invitrogen (Barcelona, Spain). Buffers were stored at 4 °C and warmed at room temperature before use. Oligonucleotides were synthesized at Centro de Investigaciones Biolo´gicas (CSIC, Madrid, Spain). AmpliTaq Gold polymerase, including reaction buffer and MgCl2, was from Perkin-Elmer (Madrid, Spain). Deoxynucleotides and PstI restriction enzyme were from Amersham Pharmacia Biotech Europe GmbH (Barcelona, Spain). Uracil DNA glycosylase was purchased from New England Biolabs (Berverly, MA), and distilled water was deionized by using a Milli-Q system (Millipore, Bedford, MA). Samples. Certified reference maize powder MZ0 (conventional, i.e., containing 0% insect-resistant Bt Event-176 transgenic maize) and MZ0.1 and MZ2 411 series (containing 0.1 and 2% insect-resistant Bt Event-176 transgenic maize) produced by the Institute of Reference Materials and Measurements (IRMM) were purchased from Fluka Chemie GmbH (Buchs, Switzerland). Grains of conventional maize and Bt Event-176 transgenic maize were a gift of Syngenta Seeds S.A. (Zaragoza, Spain). To prepare bread samples, maize grains (transgenic and conventional) were separately milled to a fine powder using two grinders and then mixed at 1% transgenic maize. For preparation of doughs, 25 g of 1% transgenic maize flour was mixed with 2 g of olive oil, 0.2 g of salt, 1 g of baker’s yeast, and 12 mL of water. After mixing for 5 min and a subsequent resting period of 30 min at room temperature, the dough was baked. Three different baking temperatures were assayed (220, 180, and 100 °C) during 30 min. DNA Extraction. DNA purification was carried out by the SDS/proteinase K method. The procedure was modified from ref 34. Homogenized samples (1.5 g) were incubated at 37 °C overnight in 10 mL of extraction buffer (1% SDS, 100 µg/mL proteinase K, 50 mM TRIS-HCl (pH 8), 20 mM EDTA). The (28) Van Eekelen, J. A.; Shammas, F. V.; Wee, L.; Heikkila, R.; Osland, A. Clin. Biochem. 2000, 33, 457-464. (29) Lim, E.; Tomita, A. V.; Thilly, W. G.; Polz, M. F. Appl. Environ. Microbiol. 2001, 67, 3897-3903. (30) Garcı´a-Can ˜as, V.; Gonza´lez, R.; Cifuentes, A. J. Agric. Food Chem. 2002, 50, 1016-1021. (31) Garcı´a-Can ˜as, V.; Gonza´lez, R.; Cifuentes, A. J. Sep. Sci. 2002, 25, 1-6. (32) Garcı´a-Can ˜as, V.; Gonza´lez, R.; Cifuentes, A. Annals of the MCFA; Office for Official Publications of the European Communities: Luxembourg, 2003; Vol 2, pp 161-165. (33) Garcı´a-Can ˜as, V.; Gonza´lez, R.; Cifuentes, A. J. Agric. Food Chem. 2002, 50, 4497-4502. (34) Sambroock, J.; Fristsch, E. F.; Maniatis, T. Molecular Cloning. A Laboratory Manual; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 1989; Chapter 9, p 9.16.

Analytical Chemistry, Vol. 76, No. 8, April 15, 2004

2307

Table 1. Primer Sequences Used for Construction of Competitor Molecules and To Perform Competitive PCR Reactionsa sequence (5′-3′)

CryIA(b)-V5 CryIA(b)-V6 CryM1

GATCGGCAACTACACCGACCAC TTGGTGTAAATCTCGCGGGTCAG GGGTCAGCTCGAAGTACTGAGCGGCG AACTGGTTGTACC GTTCCGCCGCTCAGTACTTCGAGCTGA CCTGACCGTGCTG TCAACATCCGTGGATTGCATC TTCAGGGAAATCATCAGTTAATTGC TGACGAAGAAGGGTTGATTATAGTACT AGCTAGTTCATGTGAAGAAGGC

CryM2 MSS-S MSS-A Dull10I a

accession number

primer

position

I41419 I41419 I41419

597-618 787-809 683-702

I41419

703-723

AF023159 AF023159 AF023159

933-954 1142-1166 1062-1082

Boldface type, inserted sequence.

suspension was centrifuged for 10 min at 5000 rpm. The supernatant was extracted with 1 volume of phenol and subsequently with 1 volume of chloroform/isoamyl alcohol (24:1). The aqueous phase was transferred to a new tube and 300 µg of RNAse A was added; the resultant mixture was incubated at 37 °C for 15 min. The extract was mixed with 0.1 volume of 3 M sodium acetate at pH 4.8. Then, the mix was overlaid with 2.5 volumes of 100% ethanol. The two phases were mixed carefully by gentle agitation and then centrifuged for 10 min at 12000g. The supernatant was discarded and the pellet washed with 70% ethanol. Finally, DNA was dissolved in 500 µL of TE buffer (10 mM TRIS-HCl (pH 8.0); 1 mM EDTA). Total dsDNA was quantified in a RF1501 spectrofluorometer (Schimadzu, Spain) at λexc ) 480 nm and λemi ) 520 nm, using a 1:2000 SYBRGreen-I dilution from commercial stock solution and a calibration curve made with λ DNA HindIII digest. Recombinant DNA Techniques. Plasmid DNA from Escherichia coli cells was purified by using the Wizard Plus Minipreps DNA Purification System (Promega). Restriction enzymes were used as suggested by the supplier. Sequencing was performed at the Sequencing service of Centro de Investigaciones Biolo´gicas (Spanish Council for Scientific Research, Madrid, Spain). The internal standard for the cryIA(b) system was constructed by the overlap extension method.35 Briefly, two separate amplifications were performed by using primer pairs CryIA(b)-V5/CryM1 and CryIA(b)-V6/CryM2 (see Table 1 and Figure 1A) and Bt Event-176 transgenic maize DNA as template. The two amplification products, which contain a single overlapping region of 30 bp (Figure 1B), were purified from a 2% agarose gel using QIAquick gel extraction kit (Quiagen) and used as template for a third PCR reaction with primers CryIA(b)-V5 and CryIA(b)-V6 (Figure 1C). The following thermal parameters were used for all the amplification steps: first denaturation, 10 min at 95 °C, 30 cycles of 1 min at 95 °C, 30 s at 66 °C, 30 s at 72 °C, and a terminal elongation step of 10 min at 72 °C. Reaction mixtures contained 1× AmpliTaq Gold reaction buffer, 2 mM MgCl2, 0.25 mM dATP, 0.25 mM dCTP, 0.25 mM dGTP, 0.5 mM dTTP, 3 µM each primer, 50 ng of template DNA, and 1.25 units of AmpliTaq Gold polymerase. The final 223-bp PCR product was ligated into the pGEM-T vector (Promega) and used for the transformation of E. coli DH5R (supE44 ∆lacU169(φ80 lacZ∆M15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1) resulting in plasmid pBTIS.

For the construction of the internal standard for the dull1 system, a PCR reaction was performed with primers MSS-S and MSS-A (see Table 1), and conventional maize DNA as template, under the conditions described above. The 234-bp dull1 PCR product was cloned in pGEM-T and the resulting plasmid used as template for mutagenesis using the QuikChange site-directed mutagenesis kit (Stratagene) and a single mutagenic primer, dull10I (Table 1), as described.36 The following thermal parameters were used for the mutagenic reaction: 18 cycles of 30 s at 95 °C, 1 min at 55 °C, and 7 min at 68 °C. Reaction mixtures contained 1× Pfu Turbo reaction buffer, 0.25 mM dATP, 0.25 mM dCTP, 0.25 mM dGTP, 0.2 mM dTTP, 125 ng of primer, 10 ng of plasmid, and 1.25 units of Pfu Turbo DNA polymerase. The reaction products were digested with DpnI and used to transform supercompetent Epicurian Coli XL1-Blue cells (recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac [F proAB lac1qZ∆M15 Tn10 (Tetr)]c), to obtain plasmid pDULLIS.

(35) Higuchi, R.; Krummel, B.; Saiki, R. K. Nucleic Acids Res. 1988, 16, 73517367.

(36) Makarova, O.; Kamberov, E.; Margolis, B. BioTechniques 2000, 29, 970972.

2308

Analytical Chemistry, Vol. 76, No. 8, April 15, 2004

Figure 1. Scheme for the construction of the internal standard for the cryIA(b) QC-PCR system.

Preparation of DNA Working Dilutions. All DNA solutions were prepared in TE buffer. To determine the equivalence points, pBTIS and pDULLIS competitor DNA 1:2 dilution series were performed, starting from stock solutions of known concentration, previously determined spectrofluorometrically using SYBRGreenI, and 10 ng/µL λ DNA HindIII digest was added to each tube in order to stabilize DNA.37 Genomic DNA dilutions assayed to obtain standard curves in quantitation experiments were prepared using conventional maize DNA (0% Bt Event-176) and transgenic maize DNA (100% Bt Event-176) of known concentration. Working dilutions used for constructing the cryIA(b) standard curve were prepared by mixing DNA maize solutions to obtain 5% DNA transgenic maize in a background of conventional maize DNA and preparing 1:2 serial remaining dilutions in DNA conventional maize. 1:2 dilution series used to obtain dull1 standard curves were prepared using DNA conventional maize. ROX reference dye was used in some experiments in order to prepare competitor and genomic DNA working dilutions for calibration of the QC-PCR system. To this end, a calibration curve was obtained by preparing ROX reference dye solutions ranging from 7 to 0.43 µM in TE buffer and measuring fluorescence in a RF1501 spectrofluorometer (Schimadzu) (λexc ) 580 nm; λem ) 605 nm). A known quantity of ROX was added to a first DNA working dilution and the actual DNA content of a series of dilution tubes calculated by interpolating the ROX fluorescence value in the standard curve. PCR Conditions. Unless otherwise specified, reaction mixtures contained 1× AmpliTaq Gold reaction buffer, 2 mM MgCl2, 0.25 mM dATP, 0.25 mM dCTP, 0.25 mM dGTP, 0.5 mM dUTP, 50 ng of template DNA, 3 µM each primer, 0.25 unit of UracilDNA Glycosilase (UDG), and 1.25 units of AmpliTaq Gold polymerase. Amplification reactions were run in a Mastercycler gradient thermocycler (Eppendorf). Both amplifications were preceded by an incubation period of 10 min at 37 °C for UDG activity. The thermal parameters used for the cryIA(b) system were as follows: 10 min at 95 °C; followed by 40 cycles of 1 min at 95 °C, 30 s at 66 °C, 30 s at 72 °C; and a final elongation step of 10 min at 72 °C. For the dull1 system, thermal parameters were as follows: 10 min at 95 °C, followed by 40 cycles of 1 min at 95 °C, 30 s at 58 °C, 30 s at 72 °C; and a final elongation step of 10 min at 72 °C. Capillary Gel Electrophoresis. The analyses were carried out in a PACE-MDQ (Beckman Instruments, Fullerton, CA) equipped with an Ar+ laser working at 488 nm (excitation wavelength) and 520 nm (emission wavelength). Bare fused-silica capillaries with 75-µm i.d. were purchased from Composite Metal Services (Worcester, England). Injections were made at the cathodic end using N2 pressure of 1 psi for 12 s (1 psi ) 6894.76 Pa). Data acquisition and integration were performed with 32 Karat Software (Beckman Instruments). Before first use, any uncoated capillary was preconditioned by rinsing with 0.1 M HCl for 30 min. Between injections, capillaries were physically coated using 0.1 M HCl for 4 min, 1% PVA for 2 min, and separation buffer for 4 min. At the end of the day, the capillary was rinsed with deionized water for 5 min and stored overnight with water inside. (37) Ko ¨hler, T.; Rost, A. K.; Remke, H. BioTechniques 1997, 23, 722-726.

The following conditions were used for PCR product separation: separation buffer (20 mM Tris, 10 mM phosphoric acid, 2 mM EDTA, 1.5 M urea, 500 nM YOPRO1, and 4.5% HEC at pH 7.3); temperature of separation 45 °C; running electric field -217 V/cm. Slab Gel Electrophoresis. Fifteen microliters of each PCR mixture was electrophoresed in 4% TAE buffer (40 mM TRISacetate (pH 8), 1 mM EDTA) MS8 agarose gels. Gels were run at 1.8 V/cm during 18 h, stained with 0.5 µg/mL ethidium bromide, and photographed using a Kodak DC265 digital camera (Kodak). Quantitative analysis was carried out using Adobe Photodeluxe software (Adobe Systems Inc.) and Quantity One software (BioRad). RESULTS AND DISCUSSION Construction of the DNA Competitors for QC-PCR. Internal standards based on previously described conventional PCR systems for the amplification of cryIA(b) and dull1 sequences,33 were constructed by two different site-directed mutagenesis strategies as described in the Experimental Section. In both cases, the final plasmid construct contains the particular amplicon sequence carrying a 10-bp insertion, including the target sequence for the restriction enzyme ScaI, cloned into the pGEM-T vector. A scheme showing the construction of the internal standard for the cryIA(b) system is shown in Figure 1. The overlap extension method used for this construction involves the cloning of just the internal standard molecule, while the strategy based in QuikChange site-directed mutagenesis, used for the construction of pDULLIS, involves the cloning of both the original amplicon and the internal standard. This is an important consideration because contamination of PCR reactions by these cloned DNA fragments cannot be avoided by using dUTP and Uracil-DNA Glycosilase, which we routinely use as an effective way to avoid carryover contamination by the product of previous PCR reactions. In both cases, the presence of the 10-bp insertion was verified by restriction with ScaI and confirmed by sequencing the complete insert, including primer recognition regions (Figure 2). Both recombinant plasmids were linearized with the restriction enzyme PstI prior to the preparation of dilutions for QC-PCR. Calibration of QC-PCR Systems. A 50-ng sample of genomic DNA from 1% Bt maize was coamplified with different dilutions of pBTIS and pDULLIS internal standards. The equivalence point (i.e., the concentration at which the relative areas of target and standard products are equal) was calculated by linear regression between the logarithm of the ratios of the amounts of amplification products and the logarithm of the initial amount of competitor. To compensate for the differences in molecular weight of target and standard (namely, 213 bp/223 bp and 234 bp/244 bp), corrected areas of competitor amplicons were normalized using a factor of 0.96. QC-PCR analyses were performed in two steps. In a first step, using varying amounts of pBTIS or pDULLIS and 50 ng of 1% Bt Event-176 DNA, the equivalence point was confirmed. In a second step with a constant amount of competitor, determined in the previous step, and varying percentages of transgenic DNA (for the cryIA(b) system) or varying amounts of total DNA (for the dull1 system), the slopes of the calibration curve were calculated (theoretically the value of these slopes should be equal to 1, if the same amplification efficiency applies for target and standard Analytical Chemistry, Vol. 76, No. 8, April 15, 2004

2309

Figure 2. Sequence alignments of the original target sequences, cryIA(b) and dull1, with their cognate internal standards, pBTIS and pDULLIS, respectively. The annealing sites for mutagenic or PCR primers are shown in boldface type. ScaI recongnition sites are underlined.

fragments). However, the plots obtained in all these experiments showed slope values ranging from 1.35 to 1.57 for the dull1 system and from 1.29 to 1.60 for the cryIA(b) (R2 values were always above 0.991). These slopes would suggest completely different amplification efficiencies for the original targets and the cognate internal standards, precluding any accurate quantitation. The generation of calibration curves is recognized as a bottleneck for quantitative PCR analysis, mainly derived from handling minute volumes of highly diluted DNA samples for the preparation of working dilutions and PCR mixtures. These problems have to be addressed by accurate determination of DNA concentration in the actual working dilutions.37 To investigate whether errors introduced during the preparation of working DNA dilutions were responsible for the high slopes obtained in the previous experiments, we performed a QC-PCR assay using an amount of pDULLIS roughly equivalent to 50 ng of genomic DNA (after a previous experiment) and a series of maize genomic DNA dilutions in which the actual amount of DNA was determined by spectrofluorometry in the presence of SYBRGreen-I, as described in the Experimental Section. The slopes obtained by using the 2310 Analytical Chemistry, Vol. 76, No. 8, April 15, 2004

theoretical or the corrected data were strikingly different: 1.53 and 1.07, respectively. These results confirm the hypothesis that the preparation of serial DNA dilutions could be one of the main sources of error for the quantitation based on QC-PCR. Unfortunately, due to the minute amounts of standard or genomic DNA used, which fall below the detection limits of the SYBRGreen I-based quantitation method, this approach was not feasible for the cryIA(b) system or for the first step in the dull1 QC-PCR analysis. To solve this problem, we designed a new strategy based on the use of ROX reference dye. It has been broadly used as passive reference dye in real-time PCR to compensate for variations in fluorescence signal between reactions originated by pipetting errors or instrument variability.38 It takes advantage of both the high quantum yield of the fluorescent dye and the large linear dynamic range that it provides. Our approach consists of the addition of a known amount of ROX reference dye to the DNA stock solution, to calculate, using spectrofluorometry, the actual (38) User Bulletin 2, ABI PRISM 7700 Sequence DetectionSystem, Perkin-Elmer Applied Biosystems, 1997.

Table 2. Reproducibility of Migration Times (tm), Corrected Areas, and Ratio (213 bp/223 bp Amplicons) Target/Competitor Corrected Areas for the Same Day and Three Different Daysa same day (n ) 5)

a

Figure 3. Electrophoregrams of a series of QC-PCR reactions with 50 ng of template DNA mixture containing 1% Bt Event-176 maize DNA and an increasing amounts of pBTIS internal standard: (A) 0, (B) 0.03, (C) 0.08, (D) 0.22, (E) 0.47, and (F) 1.45 fg; (G) electrophoregram of PCR reaction using 1.45 fg of pBTIS as template in the absence maize genomic DNA. Separation conditions: uncoated fused-silica capillary with 60 cm of total length, 50 cm of effective length, and 75-µm i.d. Separation voltage: -13 kV, 45 °C running temperature. Running buffer: 20 mM Tris, 10 mM orthophosphoric acid, 2 mM EDTA, 1.5 M urea, 500 nM YOPRO1, and 4.5% HEC at pH 7.3. Injection for 12 s using N2 pressure (1 psi).

amount of the original DNA stock solution in each tube. The results of one experiment using this strategy are shown in Figure 3. A series of five amplifications were performed using 50 ng of 1% genomic DNA and varying amounts of cryIA(b) internal standard (Figure 3) demonstrating that the CGE procedure allows a good separation of the two DNA fragments (213 and 223 bp) without noticeable effect from ROX. The graphs obtained by plotting the logarithm of adjusted internal standard peak area/ target peak area ratio (log y) against the logarithm of the pBTIS internal standard theoretical (log xt) and corrected concentration (log xc) showed different slopes. Namely, log y ) -0.62 + 1.45 log xt (R2 ) 0.996) and log y ) -0.69 + 1.03 log xc (R2 ) 0.992), respectively (range from 0.03 to 1.45 fg of pBTIS/PCR reaction). The slope value of 1.03 obtained using the ROX procedure indicates nearly equal efficiency for the amplification of the original target and the internal standard, which is an essential requisite for a good quantitation by QC-PCR. An identical effect was observed for the dull1 system, where the slope changed from 1.40 (R2 ) 0.991) to 0.96 (R2 ) 0.997) when ROX was utilized. Therefore, ROX reference dye was used in all subsequent experiments to correct for the actual DNA concentration in the working DNA dilutions. The equivalence points were calculated to be 0.21 fg of pBTIS and 4.48 fg of pDULLIS for 50 ng of a template DNA mixture containing 1% Bt DNA. Quantitation of the Bt Event-176 Maize Content in Food Samples. Quantitative GMO analyses of maize raw material and processed samples were carried out using the transgenic-specific cryIA(b) and the plant-specific dull1 QC-PCR systems. For each analysis, a standard curve was constructed for absolute quantitation of unknown target DNA concentrations. CryIA(b) competitive reactions were performed starting with 0.21 fg of pBTIS (see the previous section) and 50 ng of template maize DNA mixture

three days (n ) 15)

tm (min)b % RSD

21.519 0.63

21.729 0.91

corrected areab % RSD

23082.83 3.1

22779.42 5.83

ratio target/competitor corrected areas % RSD

0.056

0.056

1.72

1.93

All conditions as in Figure 3. b For 213-pb DNA fragment.

containing different quantities of Bt maize ranging from 0.04 to 2.5 ng (5 points). The plant-specific competitive dull1 system was calibrated by amplifying 4.48 fg of pDULLIS (see the previous section) with varying amounts of conventional maize DNA ranging from 3.3 to 125 ng (5 points). The standard curves were created by plotting the logarithm ratio of target/competitor PCR-amplified DNA yields for both systems (log ydull and log yBt), calculated by integrated peak areas from electrophoregrams versus the logarithm of nanograms of genomic DNA/PCR reaction (log xdull and log xBt, respectively). Again the efficiency of amplification of the target and competitor DNAs was nearly equal, as demonstrated by the close agreement to a slope of 1 for each curve. Namely, the equation log ydull ) -1.39 + 0.98 log xdull (R2 ) 0.993) was obtained for the DULL system, and the equation log yBt ) -1.01 + 1.03 log xBt (R2 ) 0.999) was obtained for the BT system, for the concentration ranges and number of points mentioned above. A reproducibility study of the whole procedure was next performed. First, reproducibility of migration times and corrected peak areas obtained using CGE-LIF were tested by injecting the same PCR reaction products from a given cryIA(b) QC-PCR. Table 2 shows the results of this reproducibility study using the GCELIF method. Thus, RSD for migration times were lower than 0.91% for three different days (n ) 15), which demonstrates the usefulness of the physically adsorbed coating used in this work. Moreover, as can be seen, RSD values for corrected peak areas range from 1.72 to 1.93% for the same day (n ) 5) and three different days (n ) 15) respectively, which confirms the good possibilities of the CGE-LIF procedure for quantitative purposes. Next, five different cryIA(b) QC-PCR reactions containing different amounts of genomic Bt Event-176 DNA and 0.21 fg of pBTIS were prepared. Each QC-PCR reaction was repeated three times and each sample injected twice in the capillary electrophoresis instrument. In this case, the highest RSD value obtained for reproducibility of corrected peak areas (n ) 6; i.e., two injections of each triplicate, for a total of 15 QC-PCR reactions and 30 injections done) was 9.98%, which is well below typical RSD values obtained using RT-PCR,20,21 for which RSD values ranging from 17.3 to 26% have recently been reported for quantitation of 1% of transgenic maize or soya.17,22-24 To analyze the transgenic maize content in real samples, 50 ng of genomic DNA extracted from certified reference maize powder or maize bread samples was coamplified using two competitive systems with the fixed quantity of each cognate Analytical Chemistry, Vol. 76, No. 8, April 15, 2004

2311

Table 3. Results from Quantitation of Food Samples Using Double Competitive PCR Method Combined with CGE-LIF Analysis % Bt maize content sample

actual

calculated

Certified Reference Material MZ0 Certified Reference Material MZ0.1 Certified Reference Material MZ2 maize flour maize bread 1 (220 °C) maize bread 2 (180 °C) maize bread 3 (100 °C)

0 0.1 2.0 1.0 1.0 1.0 1.0

0 0.07 1.80 0.94 0. 44 0.43 1.00

Figure 5. Electrophoretic analysis of a series of QC-PCR reactions (cryIA(b) system) with 0.21 fg of pBTIS internal standard per reaction. The analysis was performed by (A) slab agarose gel or (B) CGELIF. Samples: M, DNA molecular weight marker XIII; (1) control in the absence of DNA template; (2) control in the absence of internal standard; (3-8) competitive reactions with increasing amounts of transgenic Bt Event-176 maize genomic DNA ((3) 2.5, (4) 0.77, (5) 0.26, (6) 0.12 , (7) 0.04, and (8) 0 ng per reaction). All the separation conditions as described under Experimental Section.

Figure 4. Electrophoregrams of a series of QC-PCR reactions with DNA extracted from 1% transgenic maize bread baked at different temperatures (I, 220 °C; II, 180 °C; III, 100 °C; and IV, untreated maize powder). (A) cryIA(b) system with 0.21 fg of pBTIS internal standard per reaction. (B) dull1 system with 4.48 fg of pDULLIS internal standard per reaction. All CGE-LIF conditions were as in Figure 3.

internal standard previously established for the equivalence point. Absolute quantities of Bt DNA and genomic maize DNA were calculated by interpolation of the ratio of target/competitor PCRamplified DNA in the appropriate standard curve. The Bt Event176 maize content, expressed as a percentage of total maize genomic DNA, was then calculated by dividing the measured amount of the transgenic sequence by the amount of the dull1 sequence and multiplying this number by 100. All the calculated Bt contents for Certified Reference Powders were very close to the theoretical values (see Table 3). For maize bread samples, the calculated transgenic maize content was also in accordance with the theoretical values with the exception of samples processed at the highest baking temperatures (i.e., 220 and 180 °C), which resulted in ∼50% underestimation (Table 3). Figure 4 shows the CGE-LIF electrophoregrams for the injections of cryIA(b) and dull1 amplification reactions from maize bread DNA extracts. Further studies would be necessary in order to understand or confirm this apparent differential susceptibility to heat treatment between cryIA(b) and dull1 sequences. Comparative Study of CGE-LIF and Slab Gel Electrophoresis. In this work, the internal standards pBTIS and pDULLIS have been constructed in order to minimize the differences in size with the original target, with the aim of approaching the amplification efficiencies between them. The use of the CGE-LIF method is an essential element of the quantitation protocol since it brings about the independent quantitation of the two QC-PCR products. However, we found it interesting to compare the results 2312 Analytical Chemistry, Vol. 76, No. 8, April 15, 2004

obtained with CGE-LIF with the most commonly used slab gel electrophoresis analysis. To this end, the same cryIA(b) standard curve was analyzed by CGE-LIF and slab gel electrophoresis as described in the Experimental Section. No sufficiently resolved separation of PCR products was achieved by slab gel electrophoresis (as can be seen in Figure 5); therefore, quantitation using the QuantityOne software (Bio-Rad) was not possible, and no comparison of quantitative data could be performed. Nevertheless, sensitivity could be compared between both methods, showing differences of signal/noise ratios up to 688-fold (2750 for CGELIF and 4 for slab gel electrophoresis) for the 213-bp fragment (electrophoregram and lane 3 in Figure 5). In conclusion, slab gel electrophoresis is not an appropriate analytical method for the QC-PCR reactions developed in this work. CONCLUSIONS A double QC-PCR system was developed for transgenic Bt Event-176 maize detection and quantitation. Accurate relative quantitation can be achieved by the combination of two absolute quantitation reactions: one for the transgenic-specific sequence (cryIA(b)) and a second for the plant-specific gene (dull1). With the assumption that Bt material has been submitted to the same treatment as the non-Bt material, the measurement can be expressed as a percentage weight/weight (w/w %). One particularity of the method developed is that the internal standards differ by only 10 bp from the original target fragments, and this has been possible because of the high-resolution power of the CGELIF method used for analyzing the QC-PCR products. In fact, this is the smallest size difference that can be found in the bibliography between the original target and the internal standard used for QCPCR analysis. In comparison to the results obtained with CGELIF, slab gel electrophoresis showed worse resolution and sensitivity for the separation of the QC-PCR products. In addition, we have shown the importance of an accurate quantitation of the actual input DNA concentration in order to obtain calibration curves appropriate for an accurate GMO content quantitation by

QC-PCR and have successfully used ROX reference dye in order to solve this technical limitation. This solution would also be useful for the analysis of other biological samples by any quantification method using serial dilutions for preparing standards or as a different part of the procedure. The double competitive quantitative PCR-CGE-LIF method described in this paper has been successfully used for the quantitation of the GMO content of IRMM materials and processed food samples. The use of this method to quantitatively analyze multiple GMOs in a single analysis is now being developed at our laboratory.

ACKNOWLEDGMENT V.G.-C. thanks Consejerı´a de Educacio´n y Cultura (Comunidad de Madrid) for a fellowship. The authors thank the Comunidad de Madrid for supporting 7B-0021-2002 project.

Received for review December 15, 2003. Accepted January 28, 2004. AC035481U

Analytical Chemistry, Vol. 76, No. 8, April 15, 2004

2313