A simple and reliable assay for detecting specific nucleotide sequences in plants using optical thin-film biosensor chips


  • Su-Lan Bai,

    1. Peking–Yale Joint Center for Plant Molecular Genetics and Agro-Biotechnology, and The National Laboratory of Protein Engineering and Plant Genetic Engineering, College of Life Sciences, Peking University, Beijing 100871, China,
    Search for more papers by this author
  • Xiaobo Zhong,

    1. Peking–Yale Joint Center for Plant Molecular Genetics and Agro-Biotechnology, and The National Laboratory of Protein Engineering and Plant Genetic Engineering, College of Life Sciences, Peking University, Beijing 100871, China,
    2. Department of Pharmacology, Toxicology and Therapeutics, University of Kansas Medical Center, Kansas City, KS 66160, USA,
    Search for more papers by this author
  • Ligeng Ma,

    1. Peking–Yale Joint Center for Plant Molecular Genetics and Agro-Biotechnology, and The National Laboratory of Protein Engineering and Plant Genetic Engineering, College of Life Sciences, Peking University, Beijing 100871, China,
    Search for more papers by this author
  • Wenjie Zheng,

    1. Tianjin Custom Inspection and Quarantine Bureau of the People's Republic of China, Tianjin 300201, China, and
    Search for more papers by this author
  • Liu-Min Fan,

    1. Peking–Yale Joint Center for Plant Molecular Genetics and Agro-Biotechnology, and The National Laboratory of Protein Engineering and Plant Genetic Engineering, College of Life Sciences, Peking University, Beijing 100871, China,
    Search for more papers by this author
  • Ning Wei,

    1. Department of Molecular, Cellular and Developmental Biology, Yale University, New Haven, CT 06520-8104, USA
    Search for more papers by this author
  • Xing Wang Deng

    Corresponding author
    1. Peking–Yale Joint Center for Plant Molecular Genetics and Agro-Biotechnology, and The National Laboratory of Protein Engineering and Plant Genetic Engineering, College of Life Sciences, Peking University, Beijing 100871, China,
    2. Department of Molecular, Cellular and Developmental Biology, Yale University, New Haven, CT 06520-8104, USA
    Search for more papers by this author

(fax +1 203 432 5726; e-mail xingwang.deng@yale.edu).


Here we report the adaptation and optimization of an efficient, accurate and inexpensive assay that employs custom-designed silicon-based optical thin-film biosensor chips to detect unique transgenes in genetically modified (GM) crops and SNP markers in model plant genomes. Briefly, aldehyde-attached sequence-specific single-stranded oligonucleotide probes are arrayed and covalently attached to a hydrazine-derivatized biosensor chip surface. Unique DNA sequences (or genes) are detected by hybridizing biotinylated PCR amplicons of the DNA sequences to probes on the chip surface. In the SNP assay, target sequences (PCR amplicons) are hybridized in the presence of a mixture of biotinylated detector probes and a thermostable DNA ligase. Only perfect matches between the probe and target sequences, but not those with even a single nucleotide mismatch, can be covalently fixed on the chip surface. In both cases, the presence of specific target sequences is signified by a color change on the chip surface (gold to blue/purple) after brief incubation with an anti-biotin IgG horseradish peroxidase (HRP) to generate a precipitable product from an HRP substrate. Highly sensitive and accurate identification of PCR targets can be completed within 30 min. This assay is extremely robust, exhibits high sensitivity and specificity, and is flexible from low to high throughput and very economical. This technology can be customized for any nucleotide sequence-based identification assay and widely applied in crop breeding, trait mapping, and other work requiring positive detection of specific nucleotide sequences.


Rapid and reliable identification of a specific nucleotide sequence in plants is desirable in essentially all aspects of plant science research and associated applications. Distinguishing between species or ecotypes at the genome level or tracing patterns of inheritance are routine laboratory procedures. In addition, custom and government inspectors, crop breeders, and commercial and food-processing sectors often need to detect genetically modified (GM) crops or microbial pathogens in plant product shipments to comply with regulatory requirements. The large and growing demand for a means to detect foreign sequences in plant products has led us to develop an inexpensive, highly specific and easy-to-use assay that is suited to a broad range of settings and applications.

There are already many detection techniques available for positive identification of a specific nucleotide sequence or polymorphism. These detection assays may be classified into three broad types. The first type is based on PCR. PCR is the most commonly used method for amplification of a specific DNA sequence. The basic technique for demonstrating the presence of amplified DNA sequences is gel electrophoresis, a technique that allows the quantity, size and even the sequence of the DNA to be determined (Chiueh et al., 2002; German et al., 2003; Gilliland et al., 1990; Holst-Jensen et al., 2003; Miraglia et al., 2004; Rogers and Parkes, 1999; Su et al., 2003). Real-time PCR is a commonly used technology for quantification of specific DNA fragments. The amount of product synthesized during the PCR reaction is measured in real time by detection of a fluorescence signal produced as a result of amplification. Real-time PCR requires special thermal cycling machines and specific fluorescent probes. It is rapid and sensitive, but expensive and prone to generating false-positive signals, and such mis-identifications can be very costly (Baric and Dalla-Via, 2004; Hernandez et al., 2004; Shibata et al., 1998; Stubner, 2002).

The second type of assay is based on molecular markers. DNA markers have now become a popular means for identification or authentification of plant species (Andersen and Lübberstedt, 2003). The commonly used molecular markers include restriction fragment length polymorphisms (RFLPs), random amplified polymorphic DNA (RAPD), simple sequence repeats (SSRs) and single nucleotide polymorphisms (SNPs). Identification of RFLPs is based on hybridization of cloned DNA to enzymatically digested DNA fragments from a test sample (Bai et al., 2000; Cavallotti et al., 2003; Yu et al., 1995; Zhang et al., 1995). It involves restriction enzyme digestion, gel electrophoresis, transfer to the membrane and hybridization to a labeled DNA probe. This method is accurate but lengthy and laborious. Use of RAPD is based on PCR amplification of random sequences in the plant genome. Extensive PCR amplification is required and the result is often contingent upon the experimental conditions for each PCR reaction (Bai et al., 2000; Ilbi, 2003; Luo et al., 2002). SSR markers or microsatellites are unique tandem repeats interspersed throughout the genome and can be PCR-amplified using primers that flank these regions (Cordeiro et al., 2000; Edwards et al., 1998). In a mixture of denatured DNA, SSR tends to re-associate quickly owing to the low complexity of the nucleotide composition. Primers for SSR analysis can be designed by consulting GenBank for SSR loci of related species or by screening genomic libraries.

SNPs are single nucleotide differences in alleles that contribute to overall genomic diversity and natural variation in plants (Maloof et al., 2001). SNP-based assays can distinguish sequences that have single nucleotide differences (Lovmar et al., 2003; Rafalski, 2002; Zhong et al., 2003). It is the most common type of molecular marker used in identification of distinct strains or cultivars of the same plant species. For example, on average, there is one SNP every 3.3 kb in the Arabidopsis genome (a total of >37 000 SNP in 125 Mb) when comparing the sequences of two major Arabidopsis ecotypes, Col and Ler (http://www.arabidopsis.org/cereon/). Rice has possibly ten times as many SNP markers per unit DNA sequence length as Arabidopsis, between the sequenced genomes of the two major sub-species, indica 9311 (Yu et al., 2005) and japonica Nipponbare (International Rice Genome Sequencing Project, 2005). A survey of the existing SNP assays and platforms reveals that high-throughput technologies such as the InvaderTM and SNiPerTM assays require expensive high-tech instrumentation to detect polymorphic differences (Mashima et al., 2004; Pati et al., 2004). The low-throughput technologies, although they do not require expensive equipment, are nonetheless time- and labor-consuming.

The third and more recently developed assay type is based on microarray techniques. This type of assay relies on the capacity of a microarray to simultaneously identify a large number of specific DNA or molecular markers (Li et al., 2005, 2006; Ma et al., 2003, 2005). Selected probes are attached to a solid surface with each spot containing numerous copies of a single probe. The array is subsequently hybridized with PCR-amplified DNA labeled with a fluorescent marker that has been isolated from the sample of interest. During the hybridization phase, the labeled fragments associate with probes that have complementary DNA sequences. This method has been widely applied in many research fields, but it requires expensive equipment and highly trained researchers to perform, thus limiting its general usage for breeding or routine sample identification.

We have adapted and optimized a method for the detection of specific nucleic acid sequences and molecular markers in plants on the surface of optical thin-film biosensor chips (Zhong et al., 2003). One of the advantages of this technology is that, due to the optical characteristics of the thin-film biosensor chip surface, experimental results can be visualized by the unaided human eye. This technology eliminates a large initial investment in expensive instrumentation and may be widely distributed to any individual laboratory or research station that has a basic molecular biology facility with PCR capability. Here we report the detection of specific foreign sequences in GM crops and plant SNPs using this method. Our results demonstrate that this method is rapid, robust, highly sensitive and specific, can be extensively multiplexed, and is considerably less expensive compared with the existing technologies.


Strategy for detection of specific nucleic acids on a thin-film biosensor chip

The detection of genetically modified organisms/food (GMO) described here involves hybridizing biotinylated PCR fragments with capture probes covalently attached to thin-film biosensor silicon chips (see Experimental procedures and Figure 1a,b) (Zhong et al., 2003). For detection of each gene or a known DNA fragment, we synthesized a forward and reverse primer for each PCR reaction, and a capture probe with forward strand sequence for spotting on the chip surface. The reverse primer for PCR carries a covalently attached biotin at the 5′ end for future detection usage (Table 1 and Figure 1b). The capture probe has an aldehyde group at the 5′ end of ten deoxyadenosine (dA) residues (as a spacer) followed by a 40 nucleotide sequence that perfectly matches to the PCR amplicon sequence. This capture probe is covalently attached through its 5′ terminus to the chip surface (Table 1 and Figure 1a) (Zhong et al., 2003). Target DNA hybridization reactions begin with a 10 min incubation, during which the biotinylated reverse strand of the PCR amplicon binds to the capture probe through standard DNA base pairing. After simple washes with 1 × SSC to remove all mismatched molecules, attachment of the biotinylated PCR fragment is detected by incubation with an anti-biotin IgG horseradish peroxidase (HRP) conjugate and tetramethylbenzidine (TMB), a substrate that can be converted into a precipitable product by HRP to increase the thin-film thickness. This change in film thickness due to mass deposition from the substrate results in a distinguishable color change from gold to blue/purple on the chip surface, which can be visualized by the unaided eye or recorded by a simple digital-imaging documentation system depending on the density and size of probe spots.

Figure 1.

 Principles underlying thin-film biosensor chip assays for detecting specific nucleotide sequences.
(a) Immobilization of capture probes on thin-film biosensor chip surface by a chemical reaction between an aldehyde group at the 5′ terminus of capture probes and a hydrazine group at the chip surface.
(b) Strategy for GMO detection on thin-film biosensor chips. A 5′ biotinylated reverse primer was used to generate PCR amplicon targets. A biotinylated DNA strand complementary to the probe sequence on biosensor chips can then specifically hybridize, and the subsequent identification reaction requires an anti-biotin antibody as described in Experimental procedures.
(c) Strategy for SNP assay on thin-film biosensor chips as described in Experimental procedures. TMB, tetramethylbenzidine; ppt, precipitates.

Table 1.   Oligonucleotide sequences of PCR primers and capture probes for GMO detection
 GenesPrimers and probesFragment length (bp)Accession number
  1. F, forward PCR primer; R, reverse PCR primer; P, probe; ALD, aldehyde modification.

Endogenous gene groupLectin (soybean)F: 5′-gccctctactccacccccatcc
R: 5′-biotin-gcccatctgcaagcctttttgtg
P: 5′-ALD-aaaaaaaaaacatttgggacaaagaaaccggtagcgttgccagcttcgcc
Ivr1 (maize)F: 5′-ccgctgtatcacaagggctggtacc
R: 5′-biotin –tgtagagcatgacgatcc
P: 5′-ALD-aaaaaaaaaaacactggctgcacctaccgctggccatggtgcccgatcacc
Accg8 (canola)F: 5′- gagaatgaggaggaccaagctc
R: 5′-biotin-ggcgcagcatcggctctt
P: 5′-ALD-aaaaaaaaaagacgaacacctattagacattcgttccattggtcgatgga
Sad1 (cotton)F: 5′-ccaaaggaggtgcctgttca
R: 5′-biotin-ttgctcatgaaatccatca
P: 5′-ALD-aaaaaaaaaagattgagatctttaaatctttggagggctgggctgagaac
Screening gene groupCaMV 35S promoterF: 5′-gctcctacaaatgccatcat
R: 5′-biotin-gatagtgggattgtgcgtca
P: 5′-ALD-aaaaaaaaaacccacccacgaggagcatcgtggaaaaagaagacgttcca
nos terminatorF: 5′-gaatcctgttgccggtcttg
R: 5′-biotin-ttatcctagtttgcgcgcta
P: 5′-ALD-aaaaaaaaaaatgacgttatttatgagatgggtttttatgattagagtcc
nptIIF: 5′-ctcaccttgctcctgccgaga
R: 5′-biotin-agtcgatgaatccagaaa
P: 5′-ALD-aaaaaaaaaatcgcatcgagcgagcacgtactcggatggaagccggtctt
GUSF: 5′-tcagcgcgaagtctttatac
R: 5′-biotin-ttcagttcgttgttcacacaaacggtga
P: 5′-ALD-aaaaaaaaaacggcaaagtgtgggtcaaataatcaggaagtgatggagca
Identifying gene groupCryIAbF: 5′-atggacaacaaccacaac
R: 5′-biotin-gatgtcgatgggggtgtaa
P: 5′-ALD-aaaaaaaaaagagtgcatcccgtacaactgcctcagcaaccctgaggtcg
CryIAcF: 5′-gttcgttctcggactagttgaca
R: 5′-biotin-tcggcttcccactctctgaagctat
P: 5′-ALD-aaaaaaaaaagagcagttgatcaaccagaggatcgaagagttcgccagga
cp4-epSPSF: 5′-gcgaagatcgaactctccg
R: 5′-biotin-tcaatcttaagaaactttattgcc
P: 5′-ALD-aaaaaaaaaaatgcctgatgagctcgaattcgagctcggtaccggatccaa
BAR (pat)F: 5′-actcggccgtccagtcgta
R: 5′-biotin-atcgtcaaccactacatc
P: 5′-ALD-aaaaaaaaaataggcgatgccggcgacctcgccgtccacctcggcgacga

This biosensor chip can be further adapted for an SNP assay (see Experimental procedures and Figure 1c). To determine the presence or absence of an SNP, one pair of oligomers (P1 capture probes) containing a 10-base space linker followed by 40 nucleotides complementary to the DNA sequence on one side of the corresponding SNP in the target gene, is covalently attached to the chip's surface through their 5′ termini (Figure 1a). The two sequences differ only in their 3′ terminal nucleotides which correspond to different forms of the SNP alleles of interest. A second oligonucleotide probe (P2) contains a sequence complementary to the region adjacent to the SNP on the opposite side, as well as a biotin group at the 3′ end for detection and a phosphate group at the 5′ end for ligation. Target DNA hybridization and P1–P2 ligation reactions are allowed to proceed simultaneously during a 20 min incubation in the presence of a thermostable DNA ligase, as hybridization with the target facilitates ligation of the P1 and P2 probes. After a stringent wash with NaOH to remove all unligated molecules, immobilized biotinylated P2 is retained only when the P1 3′ SNP nucleotide perfectly complements its counterpart in the DNA target sequence. Ligated biotinylated P2 probe is detected by incubation with an anti-biotin IgG HRP conjugate and a precipitable HRP substrate as described above.

Defining sensitivity and specificity parameters of biosensor chips

In order to establish the suitable probe density for spotting on a biosensor chip, we chose probes for the well-described lectin gene and CaMV 35S promoter, and diluted them to various concentrations. As shown in Figure 2(a, left), 200 nl of solution containing each of the two probes at concentrations of 0.001, 0.01, 0.1 and 1 μm each were manually spotted onto the chip surface. Identical chips were hybridized for 10 min at 42°C with CaMV 35S promoter PCR products at concentrations of 0, 0.1, 1, 10 and 100 fmol in a 100 μl total reaction volume. The chips were washed once with 0.1 × SSC at room temperature, followed by 5 min incubation with an anti-biotin IgG HRP conjugate and 5 min incubation with tetramethylbenzidine (TMB) substrate. The chips were then washed, dried and visualized (Figure 2a, right panels). As expected, only the spots containing CaMV 35S promoter probe, but not the lectin gene, were detected, indicating the specificity of this assay. Although the signal intensity decreases as the concentration of target sequence is lowered, as little as 0.1 fmol of target can be detected for the probe spotted at high concentration (1 μm). The signals can be detected at the probe concentration of 0.01 μm if 100 fmol of the target PCR product is used in the assay. Spotting of probes from stock solutions higher than 1 μm did not increase detection sensitivity in the range of target concentrations tested, leading us to adopt the 1 μm probe concentration for subsequent spotting.

Figure 2.

 Specificity and sensitivity of GMO DNA target detection on thin-film biosensor chips.
(a) Specificity and sensitivity of GMO detection on a chip with capture probes spotted by hand at various concentrations. Capture probes for the Lectin gene and CaMV 35S promoter were spotted by hand at a volume of 200 nl and concentrations of 0.001, 0.01, 0.1 and 1 μm, respectively (left panel). M, Biotin-dA20 at 1 μm, positive control. PCR amplicons of CaMV 35S promoter with concentrations of 0, 0.1, 1, 10 and 100 fmol, respectively, in 100 μl reaction solution were hybridized to five identical chips (right panel).
(b) Specificity and sensitivity of GMO detection on a chip with capture probes spotted by a computer-controlled dispenser. Each spot comprised 40 nl of 1 μm probe solution. Capture probes were printed in the order shown in the left panel: M, biotin-dA20 (positive control marker); spots 1-4, endogenous genes: 1, lectin (soybean); 2, Ivr1 (maize); 3, Accg8 (canola); 4, Sad1 (cotton); spots 5–8, screening markers: 5, CaMV 35S promoter; 6, nopaline synthase (nos) terminator; 7, nptII (neomycin phosphotransferase II); 8, GUS (uidA) gene; spots 9–12, identifying markers: 9, CryIAb (resistant to European corn borer); 10, CryIAc (resistant to lepidopteran insects); 11, BAR (pat, phosphinothricin acetyltransferase); 12, cp4-epSPS (a microbe-derived herbicide resistance gene). PCR amplicons of Accg8 at concentrations of 0, 0.1, 1, 10 and 100 fmol, respectively, in 100 μl reaction solution were applied to five identical chips.

We then examined the effects of spot size, which in part determines the number and density of spots per chip, as described previously (Zhong et al., 2003), and of target DNA concentration on the sensitivity of probe detection (Figure 2b). We used chips spotted by a computer-controlled nanoliter dispenser (BioDot dispense arrayer AD3200) with 40 nl per spot at a concentration of 1.0 μm (Figure 2b, left panel). All probes were spotted three times. The probes are divided into three groups according to the characteristics of the target genes. The first group includes endogenous genes for identifying plant species, including the Lectin gene from soybean, the invertase gene (Ivr1) from maize, an ACC synthase gene (Accg8) from canola, and the stearoyl-ACP (fiber-specific acyl carrier protein) desaturase gene (Sad1) from cotton. The second group includes probes for detecting the promoter, selectable marker gene and terminator of commonly used transgenes, such as the CaMV 35S promoter, nptII, GUS (uidA) and nos terminator. The third group includes probes for identifying trait genes (such as herbicide tolerance or insect resistance), including the Round-up herbicide-resistant gene from Agrobacterium strain CP4 encoding 5-enol-pyruvylshikimate-3-phosphate synthase (cp4-epSPS), the BAR gene encoding phosphinothricin acetyl transferase (pat) and Bt toxin genes (CryIAb, CryIAc) (see also Table 1). A representative target gene (Accg8) at concentrations from 0.1–100 fmol in a 100 μl reaction mixture was used for hybridization. Clear signals were detected when using 1–100 fmol of target DNA, reaching saturation at 100 fmol (Figure 2b, right panels). This experiment indicated that 10–100 fmol target DNA in a 100 μl reaction mixture would constitute a suitable concentration range for this assay. Again, no cross-reaction of target DNA at any concentration to other probes was observed, demonstrating the extraordinary specificity of the assay.

Detection of transgenes among GM crops using optical thin-film biosensor chips

The same chip shown in Figure 2(b, left) was used to detect transgenes from samples of genetically modified soybean, maize, canola and cotton tissues. The expected DNA targets were successfully amplified by PCR (Figure 3a) and hybridized to the chip. For example, Roundup Ready soybean samples can be used to amplify five DNA targets of expected sizes corresponding to Lectin, CaMV 35S promoter, nos terminator, cp4-epSPS and nptII (Figure 3a). A mixture of these five PCR DNA fragment targets (by equal volume of PCR product) was used for hybridization in a 100 μl reaction, and the assay resulted in five sets of colored dots by chip detection (Figure 3b, bottom left). Five expected DNA targets were amplified from a maize Bt11 sample corresponding to Ivr1, CaMV 35S promoter, nos terminator, BAR (pat) and CryIAb (Figure 3a) and were mixed for an assay on the chip (Figure 3b). Very similar results were obtained for Canola Roundup Ready samples and Bt cotton SGK9708 samples (Figure 3a,b). In each case, all five probes hybridized to expected DNA targets and gave rise to positive identification. No false positives were observed among these tests. In all four cases, DNA targets from non-transgenic control samples only gave products corresponding to the endogenous gene in the chip assay results (data not shown). Our data indicate that one custom-designed chip can be readily used to detect the presence of these four GMOs. This chip can be further modified to accommodate all commercial GMO detection needs.

Figure 3.

 GMO detection on biosensor chips.
(a) Amplification of target DNA fragments for transgenic and endogenous genes by PCR from the seeds of each individual GMO crop. PCR products specific to the genes labeled were checked by agarose gel electrophoresis.
(b) Detection of the transgenic and endogenous genes in GMO products on thin-film biosensor chips. A mixture of the PCR targets was applied on a GMO detection chip.

The thin-film biosensor chip assay described here is advantageous over gel electrophoresis of gene-specific PCR fragments in several ways. The biosensor chip assay is far more sensitive (at least a thousand-fold) than gel electrophoresis, and therefore is able to detect the transgenes with small amount of material. It also filters out non-specific PCR amplifications, which would otherwise give false-positive identification in gel electrophoresis and real-time PCR methods.

Detection of indica- and japonica-specific SNPs using a biosensor chip

To demonstrate the ability of the thin-film biosensor chips technique to discriminate between indica and japonica rice by genome-specific SNPs, four SNP markers were selected from the indica/japonica SNPs database at http://plantgenome.agtec.uga.edu/snp. These four SNPs have SNP identification numbers 33400 G/A (indica versus japonica), 33402 C/T, 481007 G/C and 479709 A/T. Genomic DNA was isolated from two rice sub-species (9311, a cultivar of indica, and Nipponbare, a cultivar of japonica) whose complete genome sequences are available. SNP-specific target sequences were amplified by PCR with specific primer pairs in a multiplex PCR reaction with four sets of primers together (see Experimental procedures). The four selected SNP P1 capture probe pairs (Table 2) were manually arrayed on the thin-film biosensor chips with indica probes on the left and japonica probes on the right (shown in Figure 4a, left panel). When PCR amplicons from indica or japonica genomic DNA were applied to the chip, signals were generated only on the four spots corresponding to correct genome-specific SNP P1 capture probes (see representative images in Figure 4a, right). Very weak cross-reaction of the two indica-specific SNP probes (spots 1 and 5) with the japonica DNA targets was observed. This cross-reaction can be easily eliminated by increasing the stringency of hybridization and washing (data not shown). Once a PCR-amplified DNA target has been obtained, the entire procedure takes about 30 min, and can be performed using the simplest laboratory equipment (60°C water bath for incubation and a heating block at 95°C for denaturing DNA). The results are visible by the unaided human eye and do not require any image-reading instrument.

Table 2.   Oligonucleotide sequences of P1 and P2 probes and PCR primers for SNP detection
  1. The genomic locations of four rice SNP markers are mapped to the current version (4.0) of japonica assembly at http://www.TIGR.org.

 SNP-479409 at chromosome 10 position 16 406 758P1-A indica: ALD-aaaaaaaaaactacaaggagacaaataaacaaggtgttgt atataacaaa
P1-T japonica: ALD-aaaaaaaaaactacaaggagacaaataaacaaggtgt tgtatataacaat
P2: 5′-phosphate-aataaacacatacctggata-biotin-3′
Primer forward: tgccgagcctgtgtattatg
Primer reverse: gctaaactattttgtgtccccc
PCR product: 150 bp
 SNP-33400 at chromosome 4 position 27 582 119P1-G indica: ALD-aaaaaaaaaaaaagaaatttgacctatctaccataaacaa ttccaaagcg
P1-A japonica: ALD-aaaaaaaaaaaaagaaatttgacctatctaccataaac aattccaaagca
P2: 5′-phosphate-cttcgccgcaacaaatatct-biotin-3′
Primer forward: ggttgcatcatctgtatctcag
Primer reverse: cagcatcaatgacattaaccac
PCR product: 199 bp
 SNP-33402 at chromosome 4 position 27 582 637P1-C indica: ALD-aaaaaaaaaacaactgtgtagctttatatagcagaaaaata tagtaaagc
P1-T japonica: ALD-aaaaaaaaaacaactgtgtagctttatatagcagaaaaa tatagtaaagt
P2: 5′-phosphate-gcatatagaattgaaataat-biotin-3′
Primer forward: gcagaataggatcatgagtagc
Primer reverse: agtcctggatctgttcaaaatc
PCR product: 126 bp
 SNP-481007 at chromosome 4 position 27 385 971P1-G indica: ALD-aaaaaaaaaaaccaggaagctccatttcttgcagaaatcag catggaatg
P1-C japonica: ALD- aaaaaaaaaaaccaggaagctccatttcttgcagaaat cagcatggaatc
P2: 5′-phosphate-tgagcctttagcatcaagtt-biotin-3′
Primer forward: cagccaaggttggtagttc
Primer reverse: agtaggccatacaccatgatac
PCR product: 161 bp
 SNP COP1 position 690P1-G Columbia: ALD-aaaaaaaaaaaatagatttataccgagctagggacag atattctgtatag
P1-A Landsberg: ALD-aaaaaaaaaaaatagatttataccgagctagggaca gatattctgtataa
P2: 5′-phosphate-ttgcggatgctcggagatga-biotin-3′
 SNP COP1 position 1722P1-C Columbia: ALD-aaaaaaaaaaacacaagaaagcagtttcctatgttaa atttttgtccaac
P1-T Landsberg: ALD- aaaaaaaaaaacacaagaaagcagtttcctatgttaaatttttg tccaat
P2: 5′-phosphate-aacgagctcgcttctgcgtc-biotin-3′
 cop1-4 mutationP1-C wild type: ALD-aaaaaaaaaagggctaccaaagaaggatgcgctgagtgggtca gattcgc
P1-T mutant in cop1-4: ALD-aaaaaaaaaagggctaccaaagaaggatgcgctgag tgggtcagattcgt
P2 5′-phosphate-aaagtttgaatcagtcaact-biotin-3′
 cop1-6 mutationP1-G wild type: ALD- aaaaaaaaaagcacagattgcctaattctgttaaagtgtcttgtc ttgtg
P1-A mutant in cop1-6 ALD- aaaaaaaaaagcacagattgcctaattctgttaaagt gtcttgtcttgta
P2: 5′-phosphate-gttcaatgatttacaagaat-biotin-3′
 PCR primersForward: 5′-tgccgttgagagacatagaatag-3′
Reverse: 5′-gtgtgctatctgtggacgcag-3′
PCR product: 1123 bp
 SNP TG576P1-G L. esculentum: ALD- aaaaaaaaaatgaccaggttctatctctctcattctctttct ttgatgtg
P1-T L. pimpinellifolium: ALD- aaaaaaaaaatgaccaggttctatctctctcattct ctttctttgatgtt
P2: 5′-phosphate-ctggttattgtttctgaaac-biotin-3′
Primer forward: tcatcacttggatggtaatgc
Primer reverse: tgaaactaggcagaaaagcag
PCR product: 150 bp
 SNP BAC33RP1-A L. esculentum: ALD- aaaaaaaaaaaaattttaaattttgaatccgcgagcataa ataatgtcga
P1-G L. pimpinellifolium: ALD- aaaaaaaaaaaaattttaaattttgaatccgcgag cataaataatgtcgg
P2: 5′-phosphate-agagtgatatgtgttacaac-biotin-3′
Primer forward: aaaacattaactacttcatccg
Primer reverse: ttttccccagaggagagtac
PCR product: 133 bp
Figure 4.

 Detection of SNP markers in rice, Arabidopsis and tomato on the thin-film biosensor chips.
(a) Spotting positions of capture probes from four SNPs (left panel) and representative images of SNP detection from indica and japonica rice varieties (right panels).
(b) Spotting positions of capture probes from two ecotype-specific SNPs (1–4) and two point mutations in the COP1 gene (5–8) are shown on the left panel. Right panels show representative images of detection results using PCR targets from Col, Ler, cop1-4 and cop1-6 as indicated.
(c) SNP genotype images of the two parents and six F2 progenies in a tomato breeding program.

Detection of SNPs and point mutations in Arabidopsis

We also demonstrated that the SNP assay platform on the thin-film biosensor chips can be used to identify Arabidopsis ecotypes of Columbia (Col) and Landsberg erecta (Ler), and detect specific single nucleotide point mutations in a gene of interest. As an example, detection of SNPs and point mutations in Arabidopsis COP1 genes was carried out (Figure 4b). The COP1 gene (constitutive photomorphogenic locus 1) encodes an essential regulatory protein that plays a role in photomorphogenic development in Arabidopsis. The COP1 gene is located in the middle of the lower arm of chromosome 2, and several ecotype-specific SNP and point mutations have been reported (McNellis et al., 1994). Here we optimized a procedure to detect the two ecotype SNPs at 690 G/A and 1722 C/T (Col versus Ler, counted from the A base of the COP1 gene start codon), and two cop1 mutations, cop1-4 at 889 C/T (Col wild-type versus mutant) and cop1-6 at 945 G/A. Four pairs of SNP P1 capture probes were manually spotted on the thin-film biosensor chip surface in the pattern shown in Figure 4 (b, left panel). Target DNA samples from Col, Ler, cop1-4 and cop1-6 mutants were prepared by PCR to produce a single 1.1 kb fragment covering all four COP1 SNPs. In all cases, the DNA target always reacted with the matching P1 probes and enabled definitive identification of all four SNP markers (Figure 4b, right panel).

SNP assay as a tool for marker-assisted breeding in tomato

Using the thin-film biosensor chip SNP detection platform, we demonstrated a simple, fast and inexpensive procedure for monitoring hereditary transmission of molecular markers in a tomato breeding scheme, as an example. Two tightly linked tomato SNP markers TG576 and BAC33R were selected for a sample genotyping analysis in the F2 progenies (Figure 4c). The P1 capture probes were manually spotted in duplicate spots on the top row (Figure 4c, left) with TG576 P1-G (left) for the genotype of L. esculentum and TG576 P1-T (right) for genotype of L. pimpinellifolium. On the bottom row, BAC33R P1 probes were spotted in duplicate on the left for L. esculentum (P1-A) and on the right for L. pimpinellifolium (P1-G). Multiplex PCR products for both SNP target DNA regions from each of the two parents of Sun1642 (L. esculentum) and LA1589 (L. pimpinellifolium) and representative F2 progenies were applied to the silicon chips for genotyping. Representative images of the genotypes of the two parents and six F2 progenies for both SNP markers are shown in Figure 4(c). This simple SNP assay can definitively determine the genotype of all F2 progeny individuals at these two SNP positions.

Quantification of the relative abundance of SNP alleles in a mixture of DNA targets

In the optical biosensor chip SNP assay, we noted that there is a wide range of target concentrations for which the signal intensity correlates with the concentration of the DNA targets. Therefore, we explored the feasibility of using this biosensor chip SNP assay to quantify target molecule abundance by measuring the signal intensity. To test this, we first PCR-amplified DNA fragment targets from DNA samples taken from two rice sub-species that cover SNP-470409, and artificially mixed them at different ratios (from 0.1:100 to 100:0.1). Rice SNP-479409 capture probes P1-A (top, japonica) and P1-T (bottom, indica) were spotted by hand on chips in four replicates (200 nl per spot) using 1 μm stock solution. A mixture of the two PCR products for SNP-479409 from indica and japonica rice within the above ratio range was applied to the chips. As shown in Figure 5, direct inspection of the signal intensities (Figure 5, top panels) clearly indicates a strict correlation between signal intensity and target concentration, and complete absence of cross-reaction. We quantified the signal intensity by scanning the images and plotting the relative intensity of different samples (bottom panel). It is evident that, within a 100-fold ratio range (1:100 or 100:1) for the two allele-specific DNA fragment targets, our biosensor chip-based SNP assay was able to quantify the respective abundance of the two SNP populations. Therefore, our assay will be able to detect the presence one specific sequence among 100 in the population in real sample analysis.

Figure 5.

 Correlation of the signal intensity of the biosensor chip SNP assay with the abundance of the specific SNP target molecules in the sample.
Chip assay images are shown at the top and the quantification data are shown at the bottom. The signal intensity of the chip assay is proportional to the relative amount of target PCR products over a 10 000-fold range of ratio between the two SNP forms.


We report here a rapid, accurate and inexpensive method for the detection of specific nucleic acid or SNP markers using optical thin-film biosensor chips. The highly sensitive and accurate identification of target PCR fragments can be completed within 30 min, making this the most rapid assay available. The assay is inexpensive due to the low cost of the chemical and biological reagents and the small size of the biosensor chips used. Furthermore, this assay is extremely robust, exhibits high sensitivity and exquisite specificity, can be adapted for low, medium or high throughput, and, most importantly, is significantly less expensive compared to other high-throughput alternatives. This technology can also be customized for any specialized assay based on unique nucleotide sequences. Thus it can be adapted for a variety of purposes in laboratory research (such as mutation mapping), plant product inspection, marker-assisted breeding, SNP marker-based plant strain identification, and pathogen detection.

Use of the biosensor chip for assaying the presence of transgenes in plants

A number of GM crops have entered the global market since the first GM crops were commercialized in 1996 (http://www.agbios.com, http://www.isaaa.org, Burke, 2005; Conner et al., 2003; Nap et al., 2003; Pray et al., 2002). Although the introduction of GM crops has resulted in many documented benefits, there is an urgent need for a reliable, fast and inexpensive method for detecting the presence of transgenes associated with all known commercial GM crops in order to comply with regulations. Now that many countries require labeling and separation of GM and non-GM foods, an identification assay is needed to confirm the identity of shipments as GM or non-GM, as well as to detect possible cross-pollination between GM and wild crop species in the environment (Cockburn, 2002; Engel et al., 2002; Halford and Shewry, 2000; Konig et al., 2004). At present, there are three basic methods for detecting the presence or absence of a particular transgene. Many current DNA-based methods that involve amplification of a specific DNA sequence by PCR followed by verification of the PCR product are either time-consuming, expensive or prone to false positives (Chiueh et al., 2002; Hernandez et al., 2004; Su et al., 2003). Other methods are based on the detection of proteins specifically expressed in GM crops. Direct double-antibody (preferred) and indirect triple-antibody sandwich ELISA formats are most frequently employed to detect and measure the levels of novel proteins produced in GM plant varieties (Allen, 1990; Lipp et al., 2000). The main advantages of these methods are their simplicity, high specificity and ease of quantification, although the sensitivity is rather low and they may not be suitable for a processed food product. For example, some genes expressed in green tissues (e.g. CryIAb in maize) could not be detected in kernels using the protein-based method. Microarray-based methods allow the simultaneous identification of a large number of transgenes associated with known GM crops in a single sample (Birch et al., 2001; Miraglia et al., 2004; Taton et al., 2000). However, the expensive equipment required for this assay may be beyond the budget of many laboratories. The thin-film biosensor chip assay described in this paper is superior to the other available assays in terms of cost, sensitivity, ease of use, and time required. The same assay has many other possible applications as well. For example, it may be easily adapted to screen for the presence of harmful pathogens in food products, as long as some unique genomic sequence information for the pathogens in question is available.

Biosensor chip applications for genotyping SNP markers in plants

The vast number of genome-wide SNP markers present in all plant species examined so far make them ideal markers for molecular breeding, trait mapping and identity determination. For example, the presence of a specific set of SNP markers can be used for positive identification of any isogenic strain or the parental strain of hybrid progeny. It is also the most popular marker for analyzing genetic diversity, conducting association studies of genetic traits, or mapping mutations or traits to chromosomal loci. Therefore a simple, reliable, fast and inexpensive assay for SNPs will significantly facilitate many aspects of plant biology research, crop breeding, and strain or line identification.

Given the abundance of SNP markers all over the genome, the described thin-film biosensor chip SNP assay platform can be easily adapted as a standard assay for a genome-wide SNP scoring system. This is because this SNP assay can rapidly identify target molecules with high specificity and sensitivity. The custom-designed assay chips possess excellent chemical stability, and can generate a quantitative signal that can be detected without expensive instrumentation. Using a custom-designed chip containing SNP markers that strategically span the whole genome, one can rapidly genotype the progeny during crop breeding, as the process of progeny genotype and selection is the limiting step. As SNPs are co-dominant markers and our assay can positively identify all SNP forms at any given position in genomic DNA, adaptation of this new assay could significantly speed up the crop breeding process. Moreover, due to its ease of use, low cost, and no need for expensive equipment, this assay can be carried out by breeding research stations, particularly in developing countries, that cannot afford the expensive instrumentation associated with other available SNP assays.

A similar customized chip covering a large number of well-spaced SNP markers on the entire genome of model research plants should facilitate rapid chromosomal mapping of mutations for map-based gene cloning. Individual mutants in an F2 progeny population resulting from a mapping cross will retain the genotype of the original mutation-generating strain at the locus of the mutation. Thus, if a large number (in the hundreds) of mutants from the F2 progeny are pooled for DNA isolation and SNP marker identification, the SNP markers around the mutation locus will be biased toward the strain containing the original mutation (see Figure 6). The other regions of the genome will have a relatively equal chance of having both parental types of SNP markers. As demonstrated in Figure 5, our SNP assay can quantitatively detect the abundance of both parental SNP markers from a single locus in the DNA target. It can further define one location (a small genomic region) where SNP markers are biased towards the original mutant strain of the two parental strains, thus locating the mutation locus in that region. Obviously, further experimentation will be needed to optimize this assay for this mapping purpose and define the resolution limit that this mapping strategy can achieve.

Figure 6.

 Mapping of a new mutation using the described SNP detection assay.
The COP1 locus at the bottom of Arabidopsis chromosome 2 was used as example.
(a) Illustration of the mapping process and collection of segregating mutant individuals in the F2 population. The mutant collection from the F2 population is pooled for DNA isolation and genome-wide SNP markers segregation analysis.
(b) The frequency of a SNP marker being Col-specific or Ler-specific alone chromosome 2. Note that, when approaching the COP1 locus, the SNP markers will be biased towards those of the mutation-originating parent due to the pooled F2 mutant DNA used.

Experimental procedures

Custom design of optical thin-film biosensor chips

Thin-film biosensor chips are capable of transducing specific molecular interactions into signals that can be visualized by the naked eye. This occurs when mass deposited on the thin-film surface undergoes enzymatic catalysis, altering the wavelength of light reflected by the optical layer and resulting in a perceived color change on the surface (Zhong et al., 2003). The biosensor chips used here were purchased from Inverness Medical-Biostar (Louisville, CO, USA), and prepared according to the procedure described by Zhong et al. (2003), except that the oligo probes were specifically designed based on the requirements of each specific application outlined in this study.

For GMO detection, oligonucleotides were synthesized by Invitrogen (Carlsbad, CA, USA). PCR primers and capture probes were designed based on published sequence entries available in the GenBank database. The PCR primers were chosen from four transgenic crops – soybean, maize, canola and cotton – and the sequences are listed in Table 1. The reverse primers for PCR were synthesized with a biotin group at the 5′ end for detection. The 5′ termini of capture probes were modified with an aldehyde group that can react with a hydrazine group on the thin-film biosensor chip surface to covalently immobilize the P1 capture probe to the chip, and followed by ten deoxyadenosine (dA) residues as a spacer. The spacer is followed by 40 nucleotides complementary to the corresponding target sequence.

For SNP detection, the oligonucleotide probes were designed based on the standard approach described previously (Zhong et al., 2003), with detailed sequence information provided in Table 2. For each SNP, a pair of 50-nucleotide P1 capture probes was synthesized with only one base difference at the 3′ terminal nucleotide sequence corresponding to different SNP forms. The 5′ terminal nucleotide was modified for reactivity with the chip surface, and followed by an additional ten deoxyadenosine (dA) residue spacer and then a 40-nucleotide sequence complementary to the corresponding target sequence. A second oligonucleotide probe (biotin-P2) contains a 20-nucleotide sequence immediately adjacent to the SNP nucleotide (Figure 1c). Its 3′ terminus was modified with biotin for detection, while its 5′ terminus was modified with a phosphate for allele-specific ligation. The P1 and P2 probes were synthesized by Invitrogen at a 50 nmol scale without post-synthesis purification.

Plant materials and PCR conditions

GMO crop seeds and negative samples of soybean, maize, canola and cotton were provided by the Tianjin Customs Inspection and Quarantine Bureau of the People's Republic of China. GMO crops used here are soybean (Glycine max) of the brand Roundup Ready, Maize (Zea mays) BT11, canola (Brassica napus) Roundup Ready, and cotton (Gossypium hirsutum) SGK9708. The genomic DNA samples were isolated from seeds by the Wizard® kit (Promega, Madison, WI, USA) according to the protocol outlined in the technical manual.

The genomic DNA samples from indica and japonica rice were provided by Professor Lihuang Zhu at the Institute of Genetics and Developmental Biology, Chinese Academy of Sciences. The genomic DNAs from Arabidopsis ecotypes Col and Ler and two cop1 mutants were isolated from seedlings using a DNA isolation kit from Qiagen (Valencia, CA, USA). The genomic DNA samples from the tomato sub-species L. esculentum and L. pimpinellifolium were provided by Dr Esther van der Knaap of Ohio State University, USA.

Target DNA sequences were prepared by PCR amplification from the corresponding genomic DNA. PCR primer sets were designed using DS Gene software (Accerlrys, San Diego, CA, USA) with product sizes ranging from 110 to 350 bp. The PCR reactions were carried out in a 20 μl solution containing PCR reaction buffer (AmpliTaq Gold, Applied Biosystems, Foster City, CA, USA), 2.5 mm MgCl2, 0.2 mm dNTP, 200 nm each of forward and reverse primers, 100 ng of genomic DNA, and 1 unit of AmpliTaq Gold DNA polymerase. The PCR program used was 95°C for 10 min, followed by 40 cycles of 94°C for 15 sec, 56°C for 30 sec, 72°C for 30 sec, and a final extension at 72°C for 5 min. The PCR products were then quantified by Quant-iT Picogreen dsDNA Kits (Molecular Probe/Invitrogen, Carlsbad, CA, USA) and quality checked by electrophoresis on 2% agarose gels.

Detection of transgenes from GMO seeds on thin-film biosensor chips

Aldehyde-labeled oligonucleotides (probes) were spotted by manual pipette (200 nl aliquot per spot) or a computer-controlled nanoliter dispenser (BioDot, Inc., Irvine, CA, USA dispense arrayer AD3200) (40 nl aliquot per spot) onto biosensor chips from the 1.0 μm oligonucleotide stocks in 0.1 m sodium phosphate buffer, pH 7.8. PCR amplicon targets, from 100–350 bp in length, at a concentration of 100 fmol each per 100 μl reaction, were denatured and hybridized on the chip for 10 min at 45°C in hybridization buffer (5 × SSC and 5 mg ml−1 acid-treated casein, ATC). After washing three times in 0.1 × SSC, the chips were incubated with an anti-biotin IgG HRP conjugate (Jackson ImmunoResearch, West Grove, PA, USA; 1:1000 dilution from a 1 mg ml−1 stock in a buffer containing 5 × SSC/5 mg ml−1 ATC/10% glycerol) for 5 min in hybridization buffer. The chips were rinsed with 0.1 × SSC three times, and then 100 μl of tetramethylbenzidine (TMB) formulation from BioFx Laboratories (Owings Mills, MD, USA) was added to the chips and incubated for 5 min at room temperature. The chips were rinsed in double-distilled water (ddH2O), air-dried and scored readily by eye or photographed under a dissection microscope (Olympus, Center Valley, PA, USA SZX12).

SNP detection using thin-film biosensor chips

SNP detection was optimized from a reported standard assay for SNP genotyping (Zhong et al., 2003). P1 capture probes were manually spotted in the format shown in Figure 4 on hydrazine-activated thin-film biosensor chips (ThermoBioStar) from a mixture consisting of 0.2 μl of 1 μm P1 in 0.1 m phosphate buffer, pH 7.8, with 10% glycerol. After incubation for 2 h at room temperature in a humid chamber, the chips were washed with 0.1% SDS and ddH2O, and then dried. A ligation mixture, containing 20 mm Tris-HCl, pH 8.3, 25 mm KCl, 10 mm MgCl2, 0.5 mm nicotinamide adenine dinucleotide, 0.01% Triton X-100, 5 mg ml−1 acid-treated casein, 10 nm P2-biotin probes and 0.04 unit μl−1 mutant Ampligase (Lys294Arg of T. thermophilus ligase), was applied to each chip and pre-warmed to 60°C. The PCR amplicons from plant samples at a concentration of about 100 fmol in 10 μl of ddH2O were denatured at 95°C for 3 min. After denaturation, 10 μl of this solution was immediately added to the pre-incubated ligation mixture and incubated at 60°C for 10 min. A stringent wash with 0.01 m NaOH at 60°C was applied three times, followed by three brief rinses with 0.1 × SSC at room temperature. The chips were then incubated at room temperature for 5 min with 100 μl of an anti-biotin IgG HRP conjugate (Jackson ImmunoResearch Lab, 1:1000 dilution from a 1 mg ml−1 stock in a buffer containing 5 × SSC and 5 mg ml−1 ATC). After three brief washes with 0.1 × SSC, 100 μl TMB was added to each chip and incubated for 5 min at room temperature, then rinsed in ddH2O and air-dried. SNP genotypes were determined visually by eye, and the results were recorded using a dissection microscope fitted with a digital camera.


We thank Dr Lihuang Zhu at the Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, for the genomic DNA samples from indica and japonica rice, and Dr Esther van der Knaap at Ohio State University for the genomic DNA samples from two tomato sub-species (L. esculentum and L. pimpinellifolium) and their F2 progeny. We also thank Valerie J. Karplus for critical reading and comments on this manuscript.