Genes carry information that controls the physical development and behavior of every biological organism. Gene expression changes dynamically in living cells and manifests as patterns with numerous combinations, reflecting the physiological states. Therefore, understanding gene expression and its variations would provide critical insights into the mystery of life. Gene expression analysis allows functional characterization of genes, molecular diagnostics, pharmacogenomic and toxicogenomic screening, delineation of molecular pathways, genotyping and mutation detection. These findings would pave the way for biomarker profiling, drug discovery, disease control and the development of individualized medicine. Traditional methods of assaying genes are usually based on the concept of “one gene, one assay,” which is inherently low throughput and can’t measure the whole scheme of gene functions comprehensively. Real-time quantitative polymerase chain reaction (PCR) is one of the most well-known and widely utilized examples of this type of assay. However, the genetic information network is complex and consists of thousands of genes that are interconnected in a highly regulated fashion, even for lower organisms such as bacteria. Despite efforts to broaden the multiplexing capability of quantitative PCR, it can screen only as many as four to six genes per assay — far from enough to cope with the challenge of modern quantitative biology. For this reason, many techniques for global gene expression profiling (or so-called high-throughput methods) have emerged on stage in the past decade. Among them, the DNA microarray, which has become the most popularly used approach. Similar to the initial step of quantitative PCR, in which messenger RNA is first reverse-transcribed into cDNA and subsequently amplified, a DNA microarray involves labeling amplified DNA molecules with a tag molecule and hybridizing them to nucleic acid probes. The tags can be quantitated in a secondary detection system, usually by reading their intensity of fluorescence. DNA microarrays allow massively parallel gene expression analysis (up to a million probes featuring tens of thousands of genes in a given organism) on a single microchip. Each probe representing a gene is assigned a location on the array by either physical delivery — mechanical spotting, ink-jet printing — or in situ synthesis — photolithography, electrochemical synthesis. The technique allows molecular biologists to analyze every gene present in a genome and, in recent years, has expanded into areas such as proteins, small molecules, polysaccharides, lipids, metabolites and even tissue samples. Despite their wide applications, many drawbacks have been described, including variable reliability of differential expression data, expensive equipment for manufacturing and reading the chips, and discrepancies in calculation for changes in a given gene. Photobleaching also is a major problem for fluorescence-based microarrays, severely limiting sensitivity and quantitation. Lastly, the microarray platform uses a fixed assay and is limited to the assays on the chip, and it is inconvenient to change the gene panels. Therefore, it is not economical to perform assays on a small number of genes. These shortcomings prevent the DNA microarray from reaching its potential. Many detection assays are performed with reaction volumes in the microliter range and RNA samples of micrograms. Detecting biological samples at such extremely low amounts is a major challenge for clinical diagnosis and drug tests. Thus, the next-generation tools for gene analysis must be more sensitive, more flexible and less costly, while maintaining data accuracy and reliability. Label detection is a key determinant of sensitivity. The most commonly used labels in biological diagnostics are organic fluorescent dyes. Although new types of fluorescent microscopes and techniques have pushed the resolution well below the diffraction limit of light, fluorescent probes have not followed the same impressive evolution trend. Organic dyes still suffer from notorious limitations such as photobleaching and discrete excitation bands that preclude their use in many applications. Quantum dots for biolabeling Nanotechnology based on inorganic semiconductor quantum dots might overcome some of these limitations and provide a new scheme of biolabeling. Quantum dots can be formed by a core of cadmium-selenide nanocrystals (diameters of 2 to 10 nm) encapsulated with a zinc-sulfide shell so that they are quantumly confined, leading to a fluorescence quantum yield of more than 50 percent (for review, see ref. 1). The extinction coefficient of the nanocrystals is many times higher than that of an organic dye molecule, making quantum dots incredibly bright. The size of the core determines the absorption and emission wavelengths by means of the quantum-mechanical confinement of the optical excitation energy; hence, the name “quantum dot.” The smaller the nanocrystal, the greater the confinement energy, and the higher the energy, the shorter the wavelength of the light it emits. Quantum dots possess many unique optical and electronic properties that make them attractive for biolabeling: a broad excitation spectrum, narrow and precisely tunable emission, and improved signal brightness. In addition, their colors can be readily changed by simply varying the size of the nanoparticles (Figure 1). Another unique advantage is that they are extremely resistant to photobleaching, allowing them to keep emission intensity for hours in contrast to minutes for regular organic dyes and making them suitable for quantitative analysis. Finally, by using various size CdSe/ZnS quantum dots, simultaneous excitation of multiple fluorescent colors can be achieved with a single light source in the blue to UV range (usually 405 nm), simplifying the requirement on light sources and instruments. Figure 1. Single nanocrystals of a CdSe-core, ZnS-shell semiconductor, or quantum dots (left), can be excited by a single light source (typically at 405 nm). Various size quantum dots can be excited simultaneously by the light source, and the emission wavelength can be tuned from 400 to 850 nm by increasing the size of the nanocrystal. Quantum dots can be attached to a wide variety of biomolecules, including DNA, proteins, antibodies and short peptides. For microarray applications, thousands of genes or more must be analyzed simultaneously, requiring multiplexing in the range of at least a few thousand. However, a drawback for quantum dot multiplex labeling is that up to only 12 colors can be used in the light-emitting range of 400 to 850 nm, with the narrower emission peaks of quantum dots. Even though this is an impressively improved number from that of the conventional organic dyes, it still is not enough to accommodate the multiplexity requirements at a global-analysis level. Barcoded beads Thus, Shuming Nie’s group at Georgia Institute of Technology and at Emory University School of Medicine, both in Atlanta, proposed using a combination of colors of quantum dots and various intensity levels as a solution to generate the necessary number of signals.2 They created barcodes using 1.2-mm polystyrene microspheres containing three colors of quantum dots in controlled ratios. Each type of quantum dot in the bead has 10 intensity levels, creating a quantum dot-based nanobarcode with ~1000 combinatorial possibilities. Because of the unique spectral properties of quantum dots, this technology has tremendous multiplexing capability to shed light on genomics and high-throughout screening. For example, a six-color/10-intensity-level combination can theoretically make up ~1 million codes. Figure 2. In the quantum-dot-based system, a nanobarcoded microbead is conjugated with a specific oligonucleotide probe. Each barcode has a magnetic microbead core to enable easy manipulation by robotics. Quantum dots of various colors are mixed at various ratios with a polymer and coated onto the magnetic bead. DNA oligonucleotide probes that can read genes are attached to the polymer surface. Each nanobarcode has a unique gene probe. Recently, we turned this concept into one that could be used in commercial instruments and developed a different nanobarcoded bead platform that can not only identify, but also accurately quantify the gene expression variations in a high-throughput and multiplexed format.3 We used 8-mm-diameter magnetic beads as the core to achieve convenient manipulation and automation during liquid handling and to lower the background signal because unbound samples can be washed off easily when subject to a magnetic field. Four fluorescent colors of quantum dots, with emissions at 525, 545, 565 and 585 nm, and 20-nm spacing between the peak wavelengths, were mixed with a polymer and coated onto the microbeads at controlled ratios (Figures 2 and 3). Figure 3. Four colors (525, 545, 565 and 585 nm) are used for barcoding the quantum dots, and a fifth quantum dot that emits at 655 nm is conjugated to streptavidin and used for quantification of the biotin-tagged cRNAs that are captured on individual beads (Figure 4) (A). The barcode can be magnetically sedimented on a microplate and imaged. The fluorescence intensity of individual wavelengths (525 to 585 nm) can be measured to allow the nanobarcodes to be identified (first four rectangular view fields in the front). In addition, the 655-nm-light intensity can be used to quantify the biotinylated-cRNA captured on the beads (B). A raw image shows one captured view field of barcodes (C). Optical coding is based on quantum dot color and on its intensity levels (525 to 585 nm) (D). We achieved 12 intensity levels for each color quantum dot, for a panel of hundreds of barcoded microbeads. With each bead conjugated to an oligonucleotide probe specific to a single gene, the panel allows hundreds of genes in a sample to be monitored simultaneously. Assaying genes is simple with this method. The messenger RNA of the analytic samples is reverse-transcribed, amplified and converted into the cRNA form, during which the cRNA is tagged by biotin. After the biotin-cRNA has been hybridized with gene probes on the microbeads, a fifth streptavidin 655-nm quantum dot binds to the biotin on the cRNA, acting as a quantification reporter. The biotin-cRNA is thus sandwiched between the barcoded microbead and the streptavidin quantum dot reporter (Figure 4), creating the “sandwich assay.” Figure 4. For gene analysis with a sandwich assay, the barcode-attached oligonucleotide probes capture biotin-tagged cRNA samples through sequence complementation (hybridization), and the 655-nm streptavidin quantum dots bind to the biotin on cRNA. Each gene’s cRNA can be decoded by the barcodes in the sandwiched complex, and the quantity of the cRNA can be determined by the emission intensity of the 655-nm quantum dot. There are many advantages of the nanobarcode microbead system over the DNA microarray and quantitative PCR platforms. Quality control is easier and cheaper. Each nanobarcode can be made in 1-g quantities, enough to perform at least 109 assays. Each panel can be assembled with precalibrated barcodes, meeting most stringent FDA requirements, while each microarray chip requires careful calibration before use. In the microbead system, the relative quantification level for gene expression is a single gene copy per cell, providing the sensitivity of 103 to 104 detectable target molecules. That level is higher than 105, achievable with a commercial high-density microarray system and equal to what is usually observed for quantitative PCR. High precision The technique has a dynamic range of 3.5 logs, which, although not as good as the 6 logs achievable by quantitative PCR, is better than the 2 to 3 logs observed on various microarray platforms. Because the hybridization reaction in the microbead system is performed in liquid phase, it can be completed in one hour. This is at least one order of magnitude faster than microarray-based hybridizations, which are slowed by the physical interaction between solid phase-immobilized probes and a liquid-phase gene target. Detectable fold change is lower than 1.4, according to spike-in experiments, showing high precision even at close to a single-copy-per-cell level. Reproducibility for this proof-of-concept study is close to that of an Affymetrix UK Ltd. GeneChip microarray, with an R2 value between two repeats at 0.984 and an interwell coefficient of variation of less than 5 percent. In a validation study with the GeneChip platform, we investigated gene-expression fold changes for a panel of ~100 genes and found the correlation coefficient >0.90, which is satisfactory and similar to what we observed when comparing quantitative PCR with the microarray. In addition, the nanobarcode system uses only 1/20 the amount of sample RNA of that used in the microarray system, suggesting better efficiency to detect traces of starting materials. Figure 5. A nanobarcode is competitive with quantitative PCR and a microarray in cost, speed and sensitivity. The nanobarcode system is flexible. A new barcoded genetic probe can be made within hours and “dialed in” to the existing panel of nanobarcodes. Consequently, users can quickly assemble their own panel of gene probes. Recently, flexible microarrays have been developed, but they require expensive instruments. As another popular gene expression analysis tool, quantitative PCR lacks the ability to normalize between genes and internal precision and resolution to allow for measurement of small changes of mRNA expression. Therefore, neither microarrays nor quantitative PCR technologies provide enough internal technical replicates, calibrators or controls to monitor the quality of experimental data. The flexibility could allow costs for the nanobarcode assay to be much lower; a set of 100 genes can have a price comparable to a multiplexed quantitative PCR assay. Some commercial bead assays use a randomly assembled array in wells that code the beads with oligonucleotide zip codes. Even though this approach has been shown to have whole-genome-encoding capacity, it requires a decoding hybridization for each chip. Another platform uses dual-color bead assays involving antibodies, enzymes, toxins and nucleic acids; however, the photoinstability of dye in the beads makes quantification less reproducible, and the multiplexity is difficult to expand because of the broad bandwidth of the conventional fluorophores used in the beads. Future applications of the quantum dot microbead system include genotyping, especially single nucleotide polymorphisms. When the oligonucleotide probes are replaced by peptides, antibodies, aptamers or other affinity capturing agents, we can perform multiplexed protein assays based on affinity capture. This newly developed technique may be routinely applied to clinical diagnostics, biomarker screening, toxicogenomics, gene expression screening or microbiology screening, even including biodefense, at much lower cost and greater accuracy than technologies currently in use. At the same time, more information from limited amounts of samples and compounds can be obtained because of its high sensitivity. The nanobarcode system provides an attractive alternative to conventional genetic analysis tools and promises to be a key biotechnology platform. Meet the authors Weiming Ruan is a research fellow at Children’s Hospital Oakland Research Institute in California.; e-mail: wruan@chori.org. P. Scott Eastman is senior scientist at Tethys Bioscience in Emeryville, Calif.; e-mail: eastmanps@comcast.net. Fanqing Frank Chen is a scientist at Lawrence Berkeley National Laboratory in California; e-mail: f_chen@lbl.gov. References 1. P. Alivisatos (January 2004). The use of nanocrystals in biological detection. Nature Biotechnol, Vol. 22, pp. 47-52. 2. M. Han et al (July 2001). Quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules. Nature Biotechnol, Vol. 19, pp. 631-635. 3. P.S. Eastman et al (May 2006). Qdot nanobarcodes for multiplexed gene expression analysis. NANO LETTERS, Vol.