The basic goal of many fluorescence-based quantitative PCR procedures is to produce signal indicative of the presence of double-stranded DNA during PCR amplification. In this study, we have developed a new combined method for fluorescent signal generation, termed AUDP and illustrated in Figure 1. A complete AUDP PCR assay is composed of a target DNA-specific primer pair, with one primer containing an attached UT sequence, and a set of universal duplex probes containing both FP and QP (Figure 1A). During the first and second cycle of amplification, one new synthetic chimeric DNA fragment with UT sequence at the 5' end can be generated from each initial template DNA molecule by PCR amplification (Figure 1B and 1C). During the annealing phase of the third cycle, FP hybridizes with these newly synthesized chimeric DNA strands, separating the QP from the FP (Figure 1D). Meanwhile, FP bound to the newly synthesized chimeric DNA strand can also function as a forward primer for synthesis of a new DNA strand based on the mother strand (Figure 1E). During the annealing period of the fourth cycle, the QP hybridizes with the chimeric mother strand having a FP at its 5' end, thus it can be hydrolyzed by the 5' exonuclease activity of Taq DNA polymerase (Figure 1Fa) or displaced (Figure 1Fb) by new strand synthesis. Either displacement or hydrolysis of the duplex probes will result in increasing fluorescent signal. Throughout the following PCR cycles, more and more DNA strands (Figure 1C and 1E) are employed for amplification and fluorescent signal emission. Unlike the signal generation mechanisms of TaqMan (hydrolysis) and double-stranded probes (strand displacement) at the first or second cycle, the AUDP PCR method can generate fluorescent signal in the third cycle through a combination of hydrolysis and strand displacement 26 .

To assess the ability of generating signal depending upon strand displacement in AUDP PCR assay, Pfu DNA polymerase without 5'→3' exonuclease activity was used to detect different target sequences, including cp4EPSPS, Cry1A(b), Lectin, and the Invertase I gene. In the AUDP PCR assay for the cp4EPSPS gene, similar fluorescent curves were obtained with Pfu and Taq DNA polymerases, except that Pfu DNA polymerase resulted in reduction of fluorescent intensity (Figure 2). Similar results were also obtained in Cry1A(b), Lectin, and Invertase I assays, indicating that DNA polymerases without 5'→3' exonuclease activity can still produce florescent signals by strand displacement of probes, verifying that the fluorescent signal emission of this assay can be independent of the 5'→3' exonuclease activity. The decreased fluorescence intensity is likely to be due to the different mechanisms of fluorescent signal generation between Pfu (strand displacement) and Taq (hydrolysis and strand displacement) in AUDP PCR.

After optimizing the concentration of primers, AUDP probes and MgCl₂, the repeatability and reproducibility of CaMV35s promoter and NOS terminator AUDP PCR assays were tested using a series of dilutions of transgenic RRS genomic DNA solutions. The series were diluted to give a concentration range of 10–100,000 copies of soybean genomic DNA. Each serial dilution was assayed in triplicate per PCR run and runs were repeated three times. The repeatability standard deviation (SDr) and reproducibility standard deviation (SDR) of these two AUDP PCR assays were calculated. The SDr values of the CaMV35s and NOS PCR assays ranged from 0.09 to 0.17 and 0.08 to 0.19, respectively (Table 1); the SDR values ranged from 0.04 to 0.13 in the CaMV35s PCR assays, and from 0.05 to 0.14 in the NOS PCR assays (Table 1). This result indicates that the AUDP PCR assays were highly reproducible and could be used for further nucleic acids quantitative analysis.

To prove the applicability of this method in DNA quantification, the AUDP PCR method was employed to quantify the CaMV35s promoter and NOS terminator contents in genetically modified RRS by relative quantitative method. Both standard curves of the endogenous and exogenous PCR system were constructed by serial dilution of the target DNA in separate tubes. Samples diluted from 100 ng to 10 pg of RRS DNA were employed by simplex CaMV35s and NOS AUDP PCR assay, respectively. The PCR efficiencies of CaMV35s and NOS assay were 0.94 and 1.06, respectively. The square regression coefficients (R²) of these two exogenous PCR assays were both 0.9978. Comparable results were also obtained from the AUDP PCR assay of soybean endogenous Lectin gene with a square regression coefficient of 0.9935 (Figure 3). The limits of detection (LODs) of CaMV35s and NOS AUDP assays were 10 copies. The excellent linearity between DNA concentrations and fluorescence values (Ct) visualized in the calibration curves and low LOD values indicate that the AUDP assays established in this study are well suited for quantitative measurements. Furthermore, more simplex AUDP PCR assays for GMOs quantification were developed, such as CaMV35s promoter and NOS terminator AUDP PCR assays for GMOs screening detection, RRS Cp4-EPSPS and Bt176 Cry1A(b) for gene-specific detection, and event-specific analyses for Bt11, MON863, MON810, and GA21 maize respectively. Totally twenty-four GM soybean and maize samples were quantified by the relative quantification method according to the ratio of transgene to total DNA quantity 28 . The bias of quantitative analysis was calculated based on the formula of ABS (Measurement Value - True value)/True value. For screening assays, three GM soybean samples with different GM contents (5.0%, 3.0%, and 1.0%) were quantified, and the results showed that all the GM soybean samples with different GM contents could be quantified with low bias value compared with their virtual GM contents ranging from 3.92% to 15.12%. As to gene-specific assays of cp4EPSPS and Cry1A(b), the GM soybean and Bt176 maize samples with different GM contents (5.0%, 3.0%, and 1.0%) could be quantified with low bias (4.34%~19.22%) and standard deviation (0.04~0.20). For event-specific assays, the bias was between 3.53 and 21.13, and the SD between 0.05 and 0.40. All these indicate that the AUDP PCR assay can be successfully used in the simplex quantification of target DNAs. All the data is shown in Table 3.

Duplex PCR is less vulnerable to pipetting errors, less labor intensive, and less expensive than simplex PCR. In this study, duplex PCR systems for the quantitative analysis of RRS (cp4EPSPS and Lectin genes) and Bt176 maize (Cry1A(b) and Invertase 1 genes) were established, respectively. The sensitivity test for RRS indicated that the exogenous gene cp4EPSPS and endogenous Lectin gene could both be detected using as low as 10 pg, and which corresponded to 10 copies of haploid genomic DNA. The quantitative standard curve, which plots the log RRS concentration versus ΔCt(Ct_cp4EPSPS-Ct_Lectin), was constructed using 100 ng of total DNA varying in GM contents from 0.1 to10% (w/w) according to the described method (

24

As to the duplex PCR assay for Bt176 maize, the generated standard curve showed that the R² value was 0.9954. The tests of the three GM maize samples (5.0%, 3.0%, and 1.0%) showed that the mean GM content was of 4.69% to the 5.0% sample, 3.41% for the 3.0% sample, and 0.84% for the 1.0% sample. All above data indicated that the established duplex AUDP PCR assays are credible.

Gene expression analysis using AUDP PCR

To test the applicability of AUDP PCR in gene expression analysis, several Arabidopsis anther specific genes were selected, such as SPL, DYT1, TPD1, AMS, and EMS1, and their AUDP PCR probe and primers were designed. Gene expression in Arabidopsis flower bud was analyzed using AUDP real-time PCR according to the 2^-ΔΔCt methods 35 . In these optimized AUDP real-time PCR assays of these genes, the PCR amplified efficiencies were all above 0.90 and their amplified curves had good linearity. The results from AUDP real-time PCR assays indicated that these genes were all expressed in Arabidopsis flower bud, and the transcription level of these was shown in Figure 5, which is consistent with previous reports 36 . The results indicated that the AUDP PCR can be successfully applied to investigate different gene expression patterns using only one AUDP probe.

In the ms1 mutant of Arabidopsis, one "G" base was replaced by "A" in the genome, resulting in a loss of splicing of the first intron 34 . One AUDP PCR assay was employed to discriminate between the ms1 mutant and wild type. The AUDP PCR detection results showed that fluorescent signal could be obtained in the reaction of the homozygous and heterozygous ms1 plants. A mean Ct value of 25.29 was obtained in 100 ng homozygous ms1 genomic DNA reactions with 9 repeats, the mean Ct value of 26.41 in 100 ng heterozygous ms1 genomic DNA reactions (Figure 6), and the estimated copy number of homozygous ms1 mutated genomic DNA were about twice that of heterozygous ms1 mutated genomic DNA. From these results we believed that the AUDP PCR method can be further used for genotype discrimination.

From the unique principle of fluorescent signal generation in AUDP PCR, we hypothesize that this PCR assay has the ability to quantify target DNA with a range of template sizes. To test this, four AUDP PCR assays with different amplicon lengths (195, 546, 1032, and 1630 bp) were designed based on the pBI121 DNA. Likewise, four TaqMan and UT-PCR assays with similar amplicon sizes were designed, respectively, based on pBI121 sequence (Table 3). For each PCR assay chemistries, the PCR conditions were optimized with high efficiency (above 0.90), and the limits of detection (LOD) were tested. Fluorescent signals were obtained in all the four AUDP real-time PCR assays, regardless of amplified DNA length, and the LOD was 10 copies for the 195-bp and 546-bp amplicons, and 50 copies for the 1032-bp and 1630-bp amplicons (Table 4). However, fluorescent signals were detected for only 210-bp and 510-bp amplicons in TaqMan PCR assays, and no fluorescent signal was observed in the 992-bp and 1596-bp PCR assays. The LODs of the TaqMan PCR assays were 10 and 20 copies for the 210-bp and 510-bp amplicons, respectively (Table 4). Also we found that the UT-PCR assay could not be used for quantifying longer DNA fragments, as no fluorescent signal was observed in 1032-bp and 1630-bp amplicon assays (Table 4). These results demonstrate that the AUDP PCR assay described in this study can be successfully used for quantifying longer target DNAs with higher efficiency and sensitivity.

Transgenic Bt176 and Bt11 maize developed and supplied by Syngenta Seeds, Inc. (Greensboro, NC, USA), GM maize lines (MON863, MON810, and GA21) and Roundup Ready soybean (RRS) developed and supplied by Monsanto Company (St Louis, MO), and non-transgenic maize and soybean (obtained from Shanghai local market) were used in this study. Wild type Arabidopsis Landsberg erecta and the ms1 mutant were from University of Nottingham. The pBI121 plasmid was purchased from BD Biosciences Clontech (Palo Alto, CA, USA).

Plant genomic DNA was extracted using Plant DNA Mini-Prep kit (Ruifeng Agro-tech Co., Ltd, Shanghai, China). Plasmid DNAs were isolated using Plasmid Mini Kit (Watson Biotechnologies, Inc, Shanghai, China). DNA quantity and quality was evaluated using absorbance measurements at 260 nm and 280 nm, and copy numbers were estimated by comparison to the size of plant and plasmid genomic DNA 27 28 .

Arabidopsis total RNA was isolated from flower buds using Trizol reagent (Invitrogen). After treatment with DNase (Promega), 0.3 μg RNA was used to synthesize the oligo(dT) primed first-strand cDNA using the ReverTra Ace-a-First Strand cDNA synthesis kit (TOYOBO). Three microliters of the reverse transcription products were subsequently used as template for PCR.

In this study, some candidate UT sequences were created to avoid significant sequence similarity between the UT sequence and any known genomic sequences especially those of pBI121, maize, canola and soybean, as well as the sequence of relevant transgenes and universal regulatory elements 18 . One UT sequence is selected and attached to the 5' end of the target specific primers. The universal duplex probes are composed of two complementary single-labeled oligonucleotides of differing length. The fluorescent probe (FP) labeled with a fluorophore at its 5' end has the UT nucleotide sequence with about 30 nucleotides (nt) in size, while the quenching probe (QP) is labeled with a non-fluorescence quencher (dabcy1) at its 3' end with about 20 nt that are complementary to the 5' end of FP.

The target-specific PCR primers were analyzed and designed by modifying the analysis results of Primer premier 5.0 (PREMIER Biosoft International, Inc.). The 5' end of one target-specific PCR primer was then attached to the UT sequence. In addition, the PCR primers and duplex probes were designed to minimize the formation of primer dimers, and other types of secondary structures between the primers and probes. Primer premier 5.0 (PREMIER Biosoft International, Inc) and Oligo 6 software (Molecular Biology Insights, Inc.) were used to analyze candidate primers to avoid these problems. The GC content for the primers, UT sequence, and duplex probes were 30–80%. The melting temperature (Tm) of FP/QP duplex probes was about 5~10°C higher than that of the target DNA specific primers. To optimize the GC content and Tm values of the target PCR primers, a linker region of about 3~6 nt was added between the UT sequence and the target-specific primer region. This linker also served as a flexible region for the primer during PCR.

Two sets of universal duplex probes were designed and used for more than 20 different target DNA/RNA sequences quantification in this study (Table 4). One was labeled with FAM fluorescent reporter and another with HEX fluorescent reporter.

To reveal the advantages of AUDP technique, a series of target-specific primers were designed based on the plasmid DNA sequence of pBI121 to amplify target DNA sequences ranging from ~200 bp to ~1600 bp in length to validate the applicability of this real-time method. Meanwhile, TaqMan and UT-PCR primers and probes designed to amplify the DNA fragments with similar sizes to those of the AUDP PCR assays were produced for comparison. In the design of the four TaqMan real-time PCR assays with different sizes, the same TaqMan probe was designed for all the assays, the same sense primer was used except for the assay with 510 bp amplicon in size. The distances between the sense primer and TaqMan probe were less than 6 bp in these four real-time PCR assays according to the design principle of TaqMan PCR, and which might generate the similar efficiency of fluorescent signal production. The designed AUDP, UT, and TaqMan PCR primers and probes were listed in Table 3.

For GMOs quantification, the FAM-labeled probes were used for these exogenous genes, i.e. CaMV35s promoter, NOS terminator, cp4EPSPS gene, Cry1A(b) gene, and event-specific DNA fragments for Bt11, MON863, MON810 and GA21 maize, and the HEX-labeled probes were used for plant endogenous reference genes (Lectin and Invertase I), respectively. The primer pairs and probes are listed in Table 4.

To test the applicability of the AUDP PCR assay for gene expression and genotype discrimination, the FAM-labeled probe set were used for the following genes: Actin, SPOROCYTELESS (SPL) 29 , TAPETUM DETERMINANT1 (TPD1) 30 , DYSFUNCTIONAL TAPETUM1 (DYT1) 31 , ABORTED MICROSPORES (AMS) 32 , EXCESS MICROSPOROCYTES1 (EMS1) 33 , and ms1 mutant 34 ; their corresponding primers are also listed in Table 4.

In this study, all the real-time PCR assays (AUDP, UT, and TaqMan) were performed with the similar program and PCR reagents except for the primers, probes, and Mg Cl₂ conditions. The final volume of each real-time PCR assay was 25 μL, and the generated fluorescence was monitored in every PCR cycle at extension stage. The PCR assays contained the following mixture: 1 × PCR buffer, 100–300 nM primers, 200–400 nM FP, and 400–800 nM QP for different target sequences, 400 nM each of dATP, dGTP, dCTP, 800 nM dUTP, 1.5 U Taq DNA polymerase, 0.2 units of Amperase uracil N-glycosylase (UNG), and 6 mM MgCl₂. In terms of UT-PCR and TaqMan PCR, the PCR conditions were similar to duplex probe PCR except for the concentrations of primers (100–300 nM), probes (600–900 nM), and MgCl₂.

All real-time PCR assays were run using the following program: 2 min at 50°C, 10 min at 95°C, and 50 cycles of 30 s at 94°C, 30 s at 58°C, and 30 s at 72°C, except for the amplification of pBI121 with amplicon lengths greater than 500 bp, where 60 s at 72°C was used instead. Primers were synthesized by Invitrogen Co., Ltd (Shanghai, China), the fluorescent probes were synthesized by TaKaRa biotechnology Co., Ltd (Dalian, China), and other PCR reagents were purchased from Biocolors Co., Ltd (Shanghai, China).

All real-time PCR reactions were run on a fluorometric thermal cycler, Rotor-Gene 3000A (Corbett Research, Australia), and data was analyzed with the Rotor-Gene 3000A software (Corbett Research, Australia).

To test the applicability of other DNA polymerase in AUDP real-time PCR assay, Pfu DNA polymerase lacking 5'→3' exonuclease activity was used in AUDP real-time PCR assay with the similar reaction condition to those mentioned above except that an equal amount of Pfu DNA polymerase and Pfu PCR buffer (BioColor Co., Ltd, Shanghai. China) were substituted for Taq DNA polymerase related reagents.