Arrays

Arrays have become a ubiquitous technology platform that offers the ability to conduct many parallel experiments. For example, DNA arrays enable the interrogation of many target sequences simultaneously. There are a multitude of array formats ranging in size from macroarrays (features >100 µm) to microarrays (features 0.1–100 µm) down to nanoarrays (features <0.1 µm), with a variety of readout mechanisms including optical-, electrochemical-, thermal-, and mass-based. The number of features and the size of these features determine the array size. The nature of the readout mechanism dictates the substrate comprising the array.

Optical Fibers

One very useful array substrate is based on optical fibers (1,2,3,4,5,6). Optical fiber arrays are comprised of thousands of individual glass or plastic fibers that are manufactured by an iterative process in which individual fibers are bundled, melted, and pulled through a fiber drawing tower to create arrays of fibers in which all the individual fibers have fused into a unitary substrate. By this fabrication process, the individual fibers in the array maintain their relative position throughout the length of the bundle, such that their relative locations remain fixed wherever the fiber is cut. The individual fibers in the array are comprised of two types of glass—a central core glass with a higher refractive index than the surrounding clad glass. The clad prevents light from leaking out of the core, so that each fiber acts as an independent channel of light. Typical fiber arrays have an overall diameter of <1 mm and contain between a few thousand and 100,000 individual fibers with individual fibers ranging in size from 2–10 µm in diameter (Figure 1A).

Figure 1. Typical fiber arrays. (Click to enlarge)

It was observed nearly 10 years ago that the fiber cores could be selectively etched when a polished array was placed into an acid solution (7). The etching process creates an array of microwells at one end of the fiber (Figure 1B). A typical microwell has a volume of a few tens of femtoliters (10⁻¹⁵ L), with the bottom of the well defined by the polished optical fiber core. This structure enables each well to be addressed by the optical fiber defining its base, providing a high-density array of microwells that can be simultaneously and individually interrogated by light. In one manifestation of the arrays, beads matched in size to the wells are loaded into the microwells by either wet or dry loading (Figure 1C) (8). The beads are first modified with chemistry that enables them to respond to various analytes (e.g., DNA probes, antibodies, protein receptors). A number of different biosensor arrays have been created for different applications.

In this article, fluorescence is used to detect the signals on the array. A conventional fluorescence microscope is readily coupled to the fiber arrays. Excitation light is introduced into the proximal (unfunctionalized) end of the fiber. Fluorescence is excited in the sensor elements located on the distal end of the fiber. The fluorescence signal from each sensor returns through the fiber and is magnified and projected onto a charge-coupled device (CCD) camera, enabling all the sensors to be observed simultaneously.

DNA Analysis

The most developed use of the optical fiber arrays is for DNA analysis (9,10,11,12,13,14,15,16). In this application, single-stranded DNA is first attached to microspheres (beads). Different DNA sequences are attached to different batches of beads. The beads are combined and loaded into the arrays. By using dyes to encode the different bead types, the positions of each different bead can be identified from its color. Alternatively, the intrinsic sequence information can be used to decode the arrays using a series of hybridization reactions (5,15). This latter approach is used commercially by Illumina, Inc. (San Diego, CA, USA) to create arrays containing 1536 probe sequences per fiber that can be assembled into an Array of Arrays^™ capable of performing hundreds of thousands to millions of hybridizations simultaneously (www.illumina.com). This platform can be used for high-density genotyping and gene expression and was used for a majority of the HapMap Project.

Single Cell Arrays

Cell biosensors can be created by using the wells to contain individual cells (17,18,19,20,21,22,23). Wells can be prepared from optical fiber arrays comprised of fibers with diameters matched to the sizes of cells. In this manner, individual cells can be loaded into the etched fiber by simply allowing them to settle in the wells or by gentle centrifugation. Cells remain viable for many hours to several days. For cell-based biosensing, cells are typically engineered to express a fluorescent reporter protein or an enzyme that catalyzes the formation of a fluorescent product. In cell-based arrays, the fluorescent signals coming from individual cells are carried through the fibers—it is not possible to collect information about subcellular structures such as organelles. In contrast to most biosensors that only report on binding, cells provide the ability to perform functional sensing—that is, cells respond upon ligand binding but they also provide information about bioavailability and efficacy in activating cellular pathways. This type of information is much more valuable than knowing simply whether or not a compound binds to its receptor. The cells in the array are in intimate contact with medium, so toxic metabolites (e.g., H⁺) do not build up, and there is no depletion of essential nutrients (glucose, O₂) because they are replenished rapidly by fresh medium. When conducting cell-based biosensing, it is desirable to use different cell types with different sensitivities to assay multiple parameters simultaneously. To identify the different cells, they can be labeled with different dyes or they can be engineered to express different fluorescent proteins. In this way, multiple cell types can be included in a single array, such that control cell lines or cells with different sensitivities can be simultaneously interrogated. For example, when using the array for a yeast two-hybrid assay (19), both a negative control strain and the wild-type strain can be included in the array by including the test cells and the controls in the same array with the different cell types labeled with different dyes. This ability to screen multiple cell types simultaneously avoids problems with trying to compare different experiments in which contamination, sampling and dispensing errors, or environmental variations may perturb one experiment and lead to erroneous results. If such problems occur, they affect all the cell types identically.