Originally Published IVD Technology May 2005
Detection Technologies
Electrochemical detection for microarray IVD applications
Alternative methods offer advantages over traditional fluorescence.
Amit Kumar and Kilian Dill
 |
| Figure 1. Detection of DNA via capacitance measurements (a). Hybridization allows the accumulation of charge near the electrode due to the phosphate groups on each strand. Detection of DNA via ferrocyanide/ferricyanide couple (b). Differences can be observed between single- and double-stranded DNA in the presence of ferrocyanide/ferricyanide (click to enlarge). |
Microarrays can analyze large numbers of biological molecules using small amounts of material that are spatially segregated on a substrate (e.g., plastic, glass, semiconductor). The biological materials deposited or synthesized on the chip surface include nucleic acids, proteins, peptides, and carbohydrates. The methods used for placing materials onto the chip surface rely on physical spotting, piezoelectric deposition, and in situ synthesis (primarily for DNA). DNA synthesis methods include photolithography and electrochemical synthesis.
For more than a decade, microarrays, especially those chips containing DNA, have been used in pharmaceutical research and development. Such devices have leveraged the efforts of the human genome sequencing programs and have shown tremendous utility in understanding gene expression, single nucleotide polymorphism (SNP), and functional genomics. The ultimate goal of most microarray studies is to develop better drugs through a detailed understanding of genetics. In addition, microarrays are used for IVD applications to diagnose diseases, identify individuals who are susceptible to particular diseases, and identify drug response profiles. This latter issue is timely considering the recent incidents involving Vioxx and Celebrex.1
Most discussions of microarrays examine applications and fabrication methods. Although this article will touch upon these topics, it will primarily focus on detection methods, specifically emerging electrochemical techniques. This article will also discuss why certain electrochemical detection (ECD) methods should enable the development of smaller, less-expensive detection systems and provide better performance. IVD manufacturers must keep such factors in mind as microarrays continue to have an increasingly important role in molecular diagnostics.
Detection Methods
The most common method for analyzing hybridization events on DNA microarrays is fluorescence. However, many different methods have been demonstrated including other optical techniques. Some techniques use an optical label that is attached to probes in the sample during sample preparation. Additional esoteric methods have also been developed, such as surface plasmon resonance, atomic force microscopy, and techniques that measure mass increases upon hybridization (e.g., quartz crystal microbalance, cantilevers).2-8 While each methodology has advantages and disadvantages, the only technique that has been commercially successful is fluorescence.
 |
| Figure 2. Detection hybridized DNA via oxidation of guanine. A signal is observed only when a guanine residue is oxidized. This electrochemical reaction can be undertaken without an intercalator (click to enlarge). |
ECD is one such alternative method that has been demonstrated in laboratories but has not been successfully implemented in commercial settings. This is due to complex ECD schemes and correspondingly complex instrumentation requirements. However, there are ECD methods that are simple to implement, while maintaining its advantages over fluorescence.
Fluorescence output measurements in traditional optical systems are monitored 90 degrees from the excitation beam using expensive systems that reduce or eliminate the excitation frequency from the detector. For microarrays, similar types of optical systems are incorporated to obtain enhanced signal while reducing the background light from the excitation beam. Because of this, costly optical systems are needed with better raster/scanning systems using laser beams. The smaller the width of the laser beam, the higher the resolution scans. This issue becomes crucial when the microarray spot sizes become smaller with higher densities. For microarrays that contain more than one fluorophore label, multiple laser frequencies are required with additional optical components. All of these requirements mandate large and expensive scanners for microarray analysis.
Microarray Application:
Gene Expression
In a typical gene expression experiment, RNA is isolated from tissue or cells and converted to cDNA using reverse transcriptase–polymerase chain reaction (RT-PCR). RNA polymerase amplifies the sample in the presence of nucleotides, including several that have been chemically linked to fluorescent molecules. A number of the fluorescent-labeled nucleotides are incorporated into the RNA that is produced by the polymerase during amplification. After the amplified and labeled RNA is fragmented by chemical means and exposed to the array, it is hybridized to complementary sequences on the array. Fluorescence at specific spots, as measured by a fluorescence scanner, indicates a labeled RNA has bound to a specific complementary DNA present at that site.
The process for gene expression studies in which hybridization is measured electrochemically is similar to the fluorescence case. However, with ECD, either no attached label or a label that generates a current is used. For example, nucleotides chemically linked to biotin may be used during amplification. Once they are incorporated into the amplified RNA and hybridized to the chip, an avidin-enzyme conjugate is exposed to the chip, followed by exposure to the enzyme substrate, resulting in a measurable current upon substrate turnover.
Electrochemical Detection
 |
| Figure 3. The redox enzyme amplification system by CombiMatrix (Multikeo, WA). A DNA capture probe is synthesized at the electrode. The complementary target is a PCR product containing a biotin molecule that may be attached at the end of the sequence or to bases within the sequence. Streptavidin-labeled horseradish peroxidase is then added to the sample, and HRP binds to biotin on the DNA strand. Addition of substrate allows HRP to produce a product and a current at the electrode (click to enlarge). |
ECD offers several advantages over conventional fluorescence detection in different areas such as the following:
• Power requirements. ECD measures small currents or voltages, so a large, expensive, and heavy power supply is not needed to power a light source.
• Components. ECD does not require optical components such as a light source, mirrors, filters, detectors, support mechanics, or movement mechanics for chip scanning. Consequently, ECD systems are simpler and less expensive, and the fewer component requirements result in smaller footprints and less weight.
• Portability. Since the instrument is small and light, and power can be supplied by batteries, ECD systems can potentially be portable. This characteristic will enable significant applications of ECD in the IVD market in which small, inexpensive, and portable systems are necessary.
• Performance. In optical detection schemes, the signal-to-noise ratio is limited by the amount of stray light from the incident beam that gets into the detector channel. ECD has no incident background. The only existing background comes from the inherent background currents in the measurement systems and the capacitive charging currents at the chip. Since these currents are small, a higher signal-to-noise ratio and greater sensitivity can be achieved with the ECD mode.
As ECD systems for microarray analysis are introduced, their portability, low cost, and improved performance could change the use of microarrays for not only IVD applications but also conventional research and development studies. A situation could emerge in which array readers can be taken anywhere and purchased for a few hundred dollars, as opposed to hundreds of thousands for top-of-the-line optical scanners.
The following is a discussion of four ECD methods for microarray applications that have been described in the literature.9-19 All of these methods have a spatial array of electrodes that are hardwired to be individually addressable. Scanning is performed by measuring an electrical property independently at each electrode in either a serial or parallel manner.
Capacitance Measurements
Capacitance at a surface is determined by the existence of charge and the ability of the region near the surface to charge or discharge. Capacitance at the surface with single-stranded DNA will be different from hybridized double-stranded DNA. By measuring capacity, the existence of a hybrid can be determined (see Figure 1a).
Although this ECD system is simple in concept and requires no label, it is rather insensitive and nonspecific due to the interference of salts and other materials in the solution. To observe differences between duplex and single-stranded DNA at the chip surface, salts and other solution components need to be removed or else they mask the signal.
Faradaic Current Measurements through DNA
 |
| Figure 4. A white light photograph of a portion of the 902 chip (a). The electrode density is greater than 1000 electrodes per square cm. The commercial 12K chip (b) (click to enlarge). |
DNA is comprised of organic and inorganic molecules that can inhibit current flow, and it is known that duplex DNA is a poorer insulator than single-stranded DNA. By introducing a redox couple (e.g., ferricyanide/ferrocyanide) and measuring the current at different electrodes with attached DNA, the existence of a duplex can be determined by noting an increase in current at a particular applied voltage (see Figure 1b). Similar to the method above, since this ECD system is very sensitive to interference by solution components, events occurring at the chip surface may be difficult to differentiate.
This ECD system gives the best results when coupled to an intercalator or groove binder. Intercalators bind to the DNA duplex but not single-stranded DNA. The intercalators or groove binders may or may not be sequence specific, and the intercalators may be an organic molecule, such as methylene blue, Hoechst 33258, or transition metal complexes composed of cobalt. The intercalator’s existence makes the duplex even more conductive, resulting in greater current at electrodes where hybridization has occurred. This method has had some commercial success.13,15 However, its main drawbacks are specificity and lack of signal amplification as the data are collected on a single acquisition cycle. In addition, while the intercalators bind to all DNA duplexes formed, some may prefer certain nucleotides in a specific order.
Direct Oxidation of DNA
Some ECD methods utilize metals such as osmium or ruthenium that oxidize a sample as a reading occurs (see Figure 2). Guanine is oxidized using a ruthenium complex, and the oxidized ruthenium is reduced at the electrode. The signal generated is amplified to a small degree based on the number of guanine residues present. By design, the capture probe contains limited guanine in the sequence, making single-stranded DNA and hybridized, double-stranded DNA distinguishable by this electrochemical technique. However, the sample DNA is destroyed, and data acquisition cannot be repeated.
Redox Enzyme Mediated Measurements
Another ECD method utilizes redox enzymes as the labels.9,19 The redox enzyme provides not only the signal indicating hybridization but also amplification comparable to optical methods (see Figure 3). Although this methodology has not been utilized for microarrays, it has been used for years in colorimetric enzyme-linked immunosorbent assays (ELISA). Redox enzymes have also been used in commercially successful glucose meters. This ECD approach is simple, commercially successful, and technically better than the methods discussed above. The signal is amplified by the enzyme’s action, and only minor modifications to conventional gene expression protocols are necessary.
 |
| Figure 5. ECD output of a lambda spike experiment (0.375pM, 0.75pM, 1.5pM, 3pM, 6pM, 12pM) into a complex sample of biotinylated cRNA from a leukemic cell line. The boxed areas in yellow are where the various lambda spike-in DNA should bind. An expanded version of one of the lambda binding-in areas is shown. Other probes on the chip are complementary to specific sets of genes, which are either expressed or not expressed in this particular cell line. Twenty-four repeats of each concentration range were measured (one range is shown in the expanded region). A direct correlation between fluorescence and e-chem data is currently under way (click to enlarge). |
CombiMatrix (Mukilteo, WA) is developing a commercial system that is based on this ECD approach and semiconductor-based microelectrode arrays. The CombiMatrix system can address each electrode individually and measure the signal present at that electrode site. One section of a chip has roughly 1000 electrodes per square cm (see Figure 4a). A newer 12K version of the chip has 13,000 electrodes per square cm (see Figure 4b). On both chips, each electrode is individually addressable and can have unique oligomers synthesized at each site. Hybridization analysis can be accomplished by either fluorescence or ECD methods.
 |
| Figure 6. An actual ECD prototype that has been developed (a). Computer generated image of a prototype of a handheld ECD that is under development (b). System uses microfluidics to aid in moving sample and solutions to the detector chip (click to enlarge). |
In the CombiMatrix signal amplification method, the attached reporter group is usually horseradish peroxidase. The enzyme oxidizes substrate to product in the presence of hydrogen peroxide. In return, the product is reduced at the electrode under the appropriate conditions. Because the enzyme continues to create product at an extremely fast pace, an amplified signal can be detected as current at the electrode. For example, the ECD output of a gene expression experiment for a leukemic cell line shows lambda spike-in controls (see Figure 5). The data were generated from a 13K chip in which given lambda capture probes were synthesized in specific regions of the chip. The detection limit was 0.375 pM of biotinylated lambda cRNA. Twenty-four replicates were taken for each lambda concentration with a given sequence.
CombiMatrix has produced a prototype and is developing a more compact, automated ECD system (see Figures 6a and 6b). Many of the reaction components, or precursors, can be stored in a dried format. Long-term stability studies will be investigated.
Conclusion
The purpose of this article was not to review the field comprehensively, but rather to provide insights into new detection methods that have the potential to change microarray applications and hasten their use in the IVD market. The advantages enabled by ECD have the potential to make microarrays ubiquitous tools in not only research and development but also clinical IVDs.
References
 |
 |
| Amit Kumar, PhD, is president and chief executive officer, and Kilian Dill, PhD, is director of intellectual property at CombiMatrix Corp. (Mukilteo, WA). |
1. A Tsao, “Pfizer and Merck: Different Strokes,” Business Week Online, December 21, 2004.
2. BJ Cheek et al., “Chemiluminescence Detection for Hybridization Assays on the Flow-Thru Chip, a Three-Dimensional Microchannel Biochip,” Analytical Chemistry 73 (2001): 5777–5783.
3. SPA Fodor, “DNA Sequencing—Massively Parallel Genomics,” Science 277 (1977): 393–403.
4. JR Epstein, I Biran, and DR Walt, “Fluorescence-Based Nucleic Acid Detection and Microarrays,” Analytical Chemistry 469 (2002): 3–36.
5. JM McDonnell, “Surface Plasmon Resonance: Towards an Understanding of Mechanisms of Biological Recognition,” Current Opinions in Chemical Biology 5 (2001): 572–577.
6. JJ Storhoff et al., “One-Pot Colorometric Differentiation of Polynucleotides with Single Base Imperfections Using Gold Nanoparticle Probes,” Journal of the American Chemistry Society 120 (1988): 1959–1964.
7. J Wang, M Jiang, and E Palecek, “Real Time Monitoring of Enzymatic Cleavage of Nucleic Acids Using a Quartz Crystal Microbalance,” Bioelectrochemical Bioenergetics 4 (1999): 477–480.
8. J Fritz et al., “Translating Biomolecular Recognition into Nanomechanics,” Science 288 (2000): 316–318.
9. K Dill et al., “Immunoassays and Sequence Specific DNA Detection on a Microchip Using Enzyme Amplified Electrochemical Detection,” Journal of Biochemical Biophysical Methods 59 (2004): 181–187.
10. TG Drummond, MG Hill, and JK Barton, “Electrochemical DNA Sensors,” Nature Biotechnology 21 (2003): 1192–1198.
11. J Wang, “Survey and Summary: From DNA Biosensors to Gene Chips,” Nucleic Acid Research 28 (2000): 3011–3016.
12. JJ Gau et al., “A MEMS Based Amperometric Detector for E. coli Bacteria Using Self Assembled Monolayers,” Biosensors and Bioelectronics 16 (2001): 745–755.
13. K Hashimoto and Y Ishimori, “Preliminary Evaluation of Electrochemical DNA Array for Detection of Single Base Mismatch Mutations,” Lab on a Chip 1 (2001): 61–63.
14. JS Rossier and HH Girault, “Enzyme Linked Immunosorbent Assay on a Microchip with Electrochemical Detection,” Lab on a Chip 1 (2001): 153–157.
15. SO Kelly et al., “Single-Base Mismatch Detection Based on Charge Transduction through DNA,” Nucleic Acid Research 27 (1999): 4830–4837.
16. CJ Yu et al., “ Electronic Detection of Single-Base Mismatches in DNA with Ferrocene-Modified Probes,” Journal of the American Chemistry Society 123 (2001): 11155–11161.
17. IV Yang and HH Thorp, “Kinetics of Metal-Mediated One-Electron Oxidation of Guanadine in Polymeric DNA and in Oligonucleotides Containing Trinucleotide Repeat Sequences,” Inorganic Chemistry 39 (2000): 4969–4976.
18. ZP Aguilar and I Fritsch, “Immobilized Enzyme-Linked DNA-Hybridization Assay with Electrochemical Detection for Cryptosporidium parvum hsp70 mRNA,” Analytical Chemistry 75 (2003): 3890–3897.
19. K Dill et al., “Immunoassays Based on Electrochemical Detection Using Microelectrode Arrays,” Biosensors and Bioelectronics 20 (2004): 736–742.
Copyright ©2005 IVD Technology
