A multiplexed analysis of single-nucleotide variations without target amplification.
The Verigene system by Nanosphere Inc. (Northbrook, IL) is a nanotechnology-based molecular diagnostic system for nucleic acids and proteins.
In recent years, the number of clinically relevant singlenucleotide polymorphisms (SNPs), in combination with linkage disequilibrium information, has increased exponentially.1 Consequently, molecular analysis of allelic variations has been gaining momentum in diagnostics. Among the genes that contain clinically relevant variations are apoE4 (Alzheimer's disease), Nod2 (inflammatory bowel disease), PTPN22 (rheumatoid arthritis and type 1 diabetes), Factor V Leiden, and Factor II Prothrombin (deep vein thrombosis), to name a few.2–5
Systems that test for singlenucleotide variations and mutations differ substantially in the way they work. These systems include mass-spectrometry-based procedures such as matrix-assisted laser desorption/ ionization, denaturing highpressure liquid chromatography, oligonucleotide ligation assays, and solid-phase-array-type systems.6–11 Most of these approaches include some form of enzymatic DNA amplification such as polymerase chain reaction (PCR), ligase chain reaction, or rolling-circle amplification.12–14 Other methods, such as the Invader assay by Third Wave Technologies Inc. (Madison, WI), amplify the generated signal probe rather than the target DNA.15
Although they differ in the way they analyze DNA, all of these approaches, with the exception of the Invader assay, share one goal—reducing the complexity of nucleic acids either during or upstream of the actual genotyping reaction. Simplifying the human genomic DNA becomes inevitable as the huge number of human genome base pairs—approximately 3 billion—exacerbates the direct analysis of a singlenucleotide variation.16 A reduction of complexity is commonly achieved by enzymatic amplification and enrichment of short stretches of DNA harboring the variable sites. The amplification step also overcomes the relatively low sensitivity associated with classical detection methods (e.g., fluorescent excitation, which is used in most microarray-based genotyping reactions).
Figure 1. (click to enlarge) The genomic SNP detection assay. Human genomic DNA is hybridized to microarray-immobilized capture oligonucleotides harboring a single-nucleotide variation. After removing unbound gDNA, oligonucleotide-modified gold nanoparticles hybridize to the captured target nucleic acid. Following a wash step, signal amplification is performed. During this step, elementary silver is deposited on the gold nanoparticles, greatly enhancing their light-scattering ability.
Recently, Nanosphere Inc. (Northbrook, IL) published a novel, enzyme-free approach capable of analyzing single-nucleotide variations directly from human genomic DNA.17 The gold-nanoparticle-based assay relies on two consecutive hybridization steps. First, genomic DNA is hybridized to allele-specific microarray-bound oligonucleotides. Next, DNA-modified gold nanoparticles hybridize to a sequence in close vicinity to the SNP. Finally, a signal-amplification step is performed during which elementary silver is deposited on the gold nanoparticles and the light scattering induced by an evanescent wave in the glass substrate is measured and quantified (see Figure 1). It is possible to unambiguously determine the allelic occurrence of three genetic variants or mutations (Factor V Leiden, Factor II Prothrom bin, and MTHFR 677C→T) in multiple genomic DNA samples.17
In this article, the automation of the genomic detection assay is described. Ultrasonically fragmented DNA in hybridization buffer is injected into a hybridization cartridge that is mounted on top of a microarray. The hybridization cartridge is then inserted into the prototype automated processing system (APS). The APS delivers the assay reagents to the hybridization cartridge and provides the environmental conditions (i.e., heating and cooling) necessary for the genotyping reaction. After the assay is finished, the hybridization cartridge is removed and the glass microarray is inserted into the scanning unit (in this case, a Verigene ID by Nano sphere Inc.). The Verigene ID measures the light scattering of the silver-amplified gold nanoparticles and performs the data analysis, which includes multiple scans at various exposure times, grid alignment, and the output of a genotype call. Together, the APS and the Verigene ID make up the Verigene system.
Verigene System Description. The Verigene system is capable of genotyping multiple genes in a DNA sample in approximately 1 hour. The amount of DNA used with the Verigene system can range from 1 to 50 μg. This quantity of DNA can be readily obtained from a few hundred microliters of whole blood using commercially available DNA extraction kits as well as automated extraction systems.
Figure 2. (click to enlarge) (a) The Verigene system, which comprises the automated processing system (APS) and the Verigene ID. The APS controls the fluidic and temperature systems. The Verigene ID controls the APS and performs the microarray scanning and data analysis. (b) The disposable reagent cartridge, which contains the reagent and waste syringes. Reagents are transported to and from the hybridization cartridge via a fluid manifold. (c) An exploded view of the hybridization cartridge. The target gDNA is injected into the hybridization cartridge, which is then placed into the APS.
The Verigene system was designed and assembled at Nanosphere Inc. The APS prototype instrument contains fluidic and temperature sub systems (see Figure 2). The fluidic subsystem uses an array of stepper motors that interact with a disposable cartridge housing all the necessary assay reagents. The user inserts the reagent cartridge, and the instrument automatically engages and primes the reagents for use. Sensors detect the presence of the cartridge and alert the user when the reagents need replacement.
The hybridization cartridge comprises a glass CodeLink microarray slide made by GE Healthcare (Piscataway, NJ); a slide holder; a silicone hybridization gasket, which forms four individual 8-μl reaction chambers; and a polycarbonate housing. The housing holds four sample wells that are covered by a snap-on nozzle plate. The nozzle plate mates directly with the reagent cartridge manifold dispenser. During an assay, reagents are pumped from the storage syringes through the manifold and into the hybridization cartridge. The temperature subsystem is composed of resistive heating and thermoelectric cooling elements that have the ability to control fluid temperature in the hybridization cartridge between 15° and 60°C, within 1°C of accuracy.
Once a hybridization cartridge is inserted into the APS, the instrument automatically reads the bar code and alerts the Verigene ID of the requested hybridization. This action triggers the start of the automated assay. Up to four samples can be analyzed simultaneously per hybridization cartridge. Upon completion of the assay, the hybridization cartridge is removed from the APS and the top portion comprising the microfluidic channels and the hybridization chambers is removed from the substrate holder.
The substrate holder with the microarray slide is inserted into the Verigene ID, which automatically begins imaging and data analysis. Removing the hybridization cartridge automatically empties waste into a sealed container that is disposed of with the reagent cartridge. The reagent cartridge holds enough reagents to perform 32 genotyping reactions.
Automated Hybridization Assays. For the data presented in this article, DNA from 400 μl of peripheral blood was extracted using the Flexigene kit, a commercial DNA extraction kit by Qiagen NV (Venlo, The Netherlands). A number of DNA extraction kits—the Versagene DNA blood kit by Gentra Systems Inc. (Minneapolis), the QIAamp DNA mini kit by Qiagen, and the BioRobot EZ1 automated DNA extraction system by Qiagen—were evaluated with the Verigene system. Extracted DNA from these systems resulted in similar performance, suggesting that gDNA from various sources can be used with the Verigene system.
Human placenta DNA from Sigma-Aldrich Corp. (St. Louis) or patient genomic DNA samples from the Coriell Institute for Medical Research (Camden, NJ) as well as from 40 blood samples were fragmented by ultrasonication performed with a Microson XL processor by Misonix Inc. (Farmingdale, NY). Conditions were adjusted to yield a median DNA length of ~0.5 kb (i.e., 60 seconds at 20 kHz frequency, with an output of 3 W).
The target hybridization mixture contained 450 mmol NaNO3, 60 mmol NaCitrate (pH 7.4), Tween 20 (0.04% w/v), SDS (0.02% v/v), formamide (35% v/v), and 0.125–1 μg/μl human genomic DNA. The total target hybridization volume was 20 μl for all experiments (i.e., the total amount of gDNA ranged from 2.5 to 20 μg per assay). The mixture was injected into the hybridization cartridge after a 5-minute, 98°C heat denaturation step. The hybridization cartridge was inserted into the APS, which automatically started the assay according to a specific set of parameters.
Target hybridization was performed for 30 minutes at 38°C. Following a target wash with 0.5 mol NaNO3, the gold nanoparticle probe mix in hybridization buffer was pumped into the reaction chamber and incubated at 38°C for 10 minutes. Unbound gold nanoparticles were washed out using 0.5 mol NaNO3 followed by a 5-mmol Mg(NO3)2 rinse. Finally, the bound gold nanoparticles were reacted with a mix of silver enhancer A and B solutions from Nanosphere Inc. for 5.5 minutes at 19°C, then rinsed with ddH2O. The hybridization cartridge was removed from the APS, the top fluid-handling portion was detached, and the substrate holder was inserted into the Verigene ID, which automatically started the imaging and data analysis process.
Data Analysis. The Verigene ID generates a 16-bit grayscale image and aligns a grid in which the imaging control (IC) spots are used as initiator points and the remaining spots of the array are aligned relative to the IC spots. Median signal intensities (S) of the triplicate spots are averaged and subtracted from the local background. They are then used to calculate a genotype score, or discrimination factor (DF), according to the following algorithm: DF = (Swt – Smut)/(Swt + Smut).
Figure 3. (click to enlarge) Screen shot of the Verigene ID data output. After scanning and analyzing, the Verigene ID calculates the DF values for each locus. The genotype information, together with the microarray bar code and the patient identification, are stored in a database. The Verigene ID software is laboratory information management system (LIMS) compatible.
The Verigene ID software uses the same DF threshold values to make genotype calls as reported previously: DF values greater than 0.4 were assigned major allele; DF values below –0.4 were called minor allele.17 A sample with a DF value between 0.2 and –0.2 was called heterozygous. In order to increase the confidence level of the genotype calls, a no-call zone between ±0.4 and ±0.2 was introduced. Although never observed in this study, a DF value in the no-call zone would require a repeat of the assay. A screen shot of the Verigene ID data output is shown in Figure 3.
Figure 4. (click to enlarge) Results of a genotyping assay with 500 ng/μl weight placenta gDNA. (a) The image generated by the Verigene ID shows strong signals at the capture oligonucleotides complementary to the major allele of the Factor V (F5) and Factor II (F2) allele, respectively. (b) A schematic drawing outlining the array layout. (c) A graph displaying the discrimination factors (DF) for both alleles in the four reaction chambers.
In a first attempt to transfer the genomic SNP detection assay into the Verigene system, 500 ng/μl of fragmented human genomic DNA in hybridization buffer was loaded into the hybridization cartridge. Clearly visible in the results are the major allele signals in triplicate, as indicated by the schematic outline of the array pattern (see Figure 4). The average DF values for Factor V and Factor II were 0.95 and 0.89, respectively. The threshold values for the genotype numbers are indicated by solid lines for the major and minor alleles (±0.4) and by dotted lines for the heterozygous alleles (between 0.2 and –0.2).
Figure 5. (click to enlarge) (a) A 16-bit grayscale image created by the Verigene ID after various gDNA samples (500 ng/μl) with known Factor II and Factor V genotypes were subjected to a standard 1-hour assay in the Verigene system. The expected allelic status is depicted to the right of the image. (b) A graphical representation of the discrimination factor values shows correct distinction between the different genotypes.
Next, the system was tested for allelic discrimination capability by introducing three different DNA samples with known genotype from the Coriell Institute—GM16000 (Factor II minor allele, Factor V major allele), GM14899 (Factor II major allele, Factor V minor allele), and GM16028 (Factor II and Factor V het)—along with wt placenta gDNA (Factor II major allele and Factor V major allele). The gDNA concentration of the samples used in this experiment was 500 ng/μl. All genotypes were assigned correctly by the Verigene ID due to clearly distinct DF values (see Figure 5).
Figure 6. (click to enlarge) (a) An image created by the Verigene ID after human placenta gDNA was diluted and assayed. (b) Mean signal intensities of the weight capture oligonucleotides for both genes (subtracted by the local background) show a linear dose response. (c) The average discrimination factor (DF) values of 10 target titration experiments. The error bars depict 1 standard deviation of the signal intensity over the 10 repeats of the target titration experiment. The threshold values for the three genotypes are shown as solid lines for the homozygous alleles (≥ 0.4 = wt; ≤ –0.4 = mut) and as dotted lines for heterozygous genotypes (between ± 0.2).
In order to determine the limit of detection, a target titration experiment was performed (see Figure 6). Sigma wt placenta gDNA was diluted from 1000 to 125 ng/μl in hybridization buffer, and the genotyping assay was performed under standard conditions (i.e., 1 hour total assay time). The Verigene ID was able to make 100% of the calls, even at the lowest DNA concentration tested. To assess the stability of the system, the target titration experiment was repeated 10 times. The call rate remained 100% with the average DF values at the lowest gDNA concentration still significantly exceeding the threshold for a major allele call of 0.4.
Average DF values did not change significantly between the different gDNA concentrations (average DFF2 = 0.84, DFF5 = 0.88) due to normalization of signal intensity ratios. By increasing the target hybridization time to 2 hours, it was possible to successfully genotype at a gDNA concentration of 50 ng/μl, corresponding to 15,000 genome copies per μl.
Remarkably, the sensitivity of the Verigene system exceeds a previously reported direct genomic SNP analysis from whole-genome amplified DNA 50–100-fold (5–6 μg/μl versus 50–100 ng/μl).11 This level of sensitivity was reached after 30 minutes of target hybridization for the gold nanoparticle assay, whereas the target hybridization time for the fluorescent-based bead assay ranged from 16 to 18 hours. The approximately 3-log greater detection sensitivity of nanoparticles over fluorophores is in concordance with previously reported data.18
Table I. (click to enlarge) Summary of the genotype calls made by the Verigene ID. Out of 80 possible calls, one Factor II call was rejected (no call) by the Verigene ID verification algorithm due to a high coefficient of variation value of the triplicate spots. After retesting, the sample was correctly assigned with a major allele genotype. Accuracy of the genotypes was verified by restriction-fragmentlength polymorphism analysis.
To define the performance of the Verigene system under clinical conditions, genomic DNA was prepared from 40 blood samples using a commercial DNA extraction kit. Genomic DNA from 400 μl of peripheral whole blood was prepared according to the manufacturer's recommendations, ultrasonically fragmented, and then subjected to an automated geno typing assay in the Verigene system (see Table I).
The overall call rate was 98.75%. Out of the 40 Factor V calls, five had a heterozygous genotype, whereas only one Factor II call was heterozygous. No minor allele for either Factor V or Factor II was detected. The one no-call result was associated with a high-signalintensity coefficient of variation of the triplicate spots per capture oligonucleotide and was appropriately rejected by the Verigene ID's filter and verification algorithms. This sample was reprocessed and called correctly, which allowed for an overall call rate of 100%. The accuracy of the genomic SNP detection assay was confirmed by restriction fragment-length polymorphism analysis, and the results showed 100% concordance between the two assay formats.19–21
The Verigene system presents an automated, enzyme-free genotyping setup that can use target nucleic acid of extreme complexity for analyzing single-nucleotide variations. Although shown for just two SNPs in this article, the gold-nanoparticlebased assay is capable of a significantly higher degree of multiplexity. The genomic DNA can be prepared using a variety of commercially available extraction kits, which suggests that the quality of the resulting gDNA is sufficient for genotyping using the Verigene system. The overall sample-to-result time is approximately 90 minutes, with 15–30 minutes for the gDNA extraction (depending on the kit used), 1 minute for sonication, and 1 hour for the actual genotyping, scanning, and data analysis.
In addition, the Verigene system overcomes the following key disadvantages of other genotyping approaches, especially PCR-based systems:
A possible limitation of the Verigene system, although not observed with the currently employed tests, stems from the fact that the genotype discrimination depends solely on allele-specific hybridization. This might be limiting in the case of special sequence conditions (e.g., extreme G-C–rich sequences). However, by adjusting the assay parameters, it is possible to reach sufficient discrimination power, even with G-C contents greater than 60%.
The versatility of the Verigene system can potentially be expanded beyond the analysis of nucleotide variations in genomic DNA to gene expression experiments. Differential gene expression analysis using gold nanoparticles has been demonstrated by using RNA directly as target nucleic acid.22 Total RNA is hybridized to microarray-bound oligonucleotides and detected by T20-modified gold nanoparticles that target the polyA-tail of bound mRNA molecules.
Although not described in this article, changing the gene-specific probes employed for the genomic SNP detection assay to a T20-modified gold nanoparticle probe solution, in combination with the appropriate microarrays, should enable the Verigene system to perform gene expression analysis directly from total RNA.
Planned design improvements of the Verigene system include additional sample-preparation features by including sonication/denaturation onboard. The purified gDNA could be directly inserted into the Nanosphere automated process without any further manual treatment.
Martin Huber, PhD, is a senior scientist at Nanosphere Inc. (Northbrook, IL). He can be reached at firstname.lastname@example.org.
William Cork is chief scientific officer at Nanosphere Inc. (Northbrook, IL).
Mark Weber is director of the software department at Nanosphere Inc. (Northbrook, IL).
Tim Patno is director of the systems engineering department at Nanosphere Inc. (Northbrook, IL). He can be reached at email@example.com.
1. The International HapMap Consortium, “A Haplotype Map of the Human Genome,” Nature 437, no. 7063 (2005): 1299–1320.
2. WJ Strittmatter and AD Roses, “Apolipoprotein E and Alzheimer's Disease,” Annual Review of Neuroscience 19 (1996): 53–77.
3. Y Ogura et al., “A Frameshift Mutation in NOD2 Associated with Susceptibility to Crohn's Disease,” Nature 411, no. 6837 (2001): 603–606.
4. AB Begovich et al., “A Missense Single-Nucleotide Polymorphism in a Gene Encoding a Protein Tyrosine Phosphatase (PTPN22) Is Associated with Rheumatoid Arthritis,” American Journal of Human Genetics 75, no. 2 (2004): 330–337.
5. B Dahlback, “Resistance to Activated Protein C Due to Factor V R506Q Mutation as a Cause of Venous Thrombosis,” Revista de Investigación Clinica 49, supp. no. 1 (1997): 3–5.
6. TJ Griffin et al., “Direct Genetic Analysis by Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry,” Proceedings of the National Academy of Sciences of the United States of America 96 no. 11 (1999): 6301–6306.
7. S Sauer et al., “A Novel Procedure for Efficient Genotyping of Single Nucleotide Polymorphisms,” Nucleic Acids Research 28, no. 5 (2000): E13.
8. M Giordano et al., “Identification by Denaturing High-Performance Liquid Chromatography of Numerous Polymorphisms in a Candidate Region for Multiple Sclerosis Susceptibility,” Genomics 56, no. 3 (1999): 247–253.
9. MA Iannone et al., “Multiplexed Single Nucleotide Polymorphism Genotyping by Oligonucleotide Ligation and Flow Cytometry,” Cytometry 39, no. 2 (2000): 131–140.
10. JG Hacia, “Resequencing and Mutational Analysis Using Oligonucleotide Microarrays,” Nature Genetics 21, supp. no. 1 (1999): 42–47.
11. KL Gunderson et al., “A Genome-Wide Scalable SNP Genotyping Assay Using Microarray Technology,” Nature Genetics 37, no. 5 (2005): 549–554.
12. PM Lizardi et al., “Mutation Detection and Single-Molecule Counting Using Isothermal Rolling-Circle Amplification,” Nature Genetics 19, no. 3 (1998): 225–232.
13. M Huber et al., “Detection of Single Base Alterations in Genomic DNA by Solid Phase Polymerase Chain Reaction on Oligonucleotide Microarrays,” Analytical Biochemistry 299, no. 1 (2001): 24–30.
14. U Landegren et al., “A Ligase-Mediated Gene Detection Technique,” Science 241, no. 4869 (1988): 1077–1080.
15. JG Hall et al., “Sensitive Detection of DNA Polymorphisms by the Serial Invasive Signal Amplification Reaction,” Proceedings of the National Academy of Sciences of the United States of America 97, no. 15 (2000): 8272–8277.
16. ES Lander, “Array of Hope,” Nature Genetics 21 (1999): 3–4.
17. YP Bao et al., “SNP Identification in Unamplified Human Genomic DNA with Gold Nanoparticle Probes,” Nucleic Acids Research 33, no. 2 (2005): e15.
18. JJ Storhoff et al., “Gold Nanoparticle-Based Detection of Genomic DNA Targets on Microarrays Using a Novel Optical Detection System,” Biosensors and Bioelectronics 19, no. 8 (2004): 875–883.
19. M Ledford et al., “A Multi-Site Study for Detection of the Factor V (Leiden) Mutation from Genomic DNA Using a Homogeneous Invader Microtiter Plate Fluorescence Resonance Energy Transfer (FRET) Assay,” Journal of Molecular Diagnostics 2, no. 2 (2000): 97–104.
20. SR Poort et al., “A Common Genetic Variation in the 3'-Untranslated Region of the Prothrombin Gene Is Associated with Elevated Plasma Prothrombin Levels and an Increase in Venous Thrombosis,” Blood 88, no. 10 (1996): 3698–3703.
21. JG Chang et al., “Rapid Diagnosis of Beta-Thalassemia Mutations in Chinese by Naturally and Amplified Created Restriction Sites,” Blood 80, no. 8 (1992): 2092–2096.
22. M Huber et al., “Gold Nanoparticle Probe-Based Gene Expression Analysis with Unamplified Total Human RNA,” Nucleic Acids Research 32, no. 18 (2004): e137.