Mark McCormick and Emile F Nuwaysir Introduction

Toxicogenomics is based on the principle that changes in gene expression can be used as predictors of toxicity and perhaps to discern the mode of action of a particular compound. Under this paradigm, transcription profiles are generated from as many genes as possible, preferably the entire genome of the organism, when exposed to a given model toxicant. This transcriptome data can then serve as a training set to develop a more predictive and robust assay based on the expression profiles of a smaller number of genes, perhaps hundreds.

In the last several years, pilot studies have validated this approach in simple model systems using classical toxicants as test compounds. Some of these studies have suggested that differential expression of a very small number of genes, from 10 to 300, may be sufficient to serve as markers for toxicity [1-6].

Thus, the ideal tool to carry out toxicogenomics testing would be a microarray platform that can perform whole-genome analyses to identify the signature gene list, as well as run smaller focused arrays in extremely high throughput. The ability to run both genome-scale and focused arrays on the identical technology platform is critical to avoiding errors associated with inter-platform data comparison.

The standard method for high-density oligonucleotide microarray manufacture [7-9] takes advantage of three simple tools: (1) DNA phosphoramidite synthesis chemistry, (2) photolithography, and (3) photochemistry. This method has been used to synthesize hundreds of thousands of 25-mer oligonucleotides in parallel on solid supports.

Although the method is extremely powerful, the synthesis chemistry employed to date is relatively inefficient, and the required photolithographic masks are relatively expensive. These two characteristics make long-oligo microarrays (e. g. 60-mers) difficult and expensive to produce. Also, the photolithographic masks are inherently inflexible (changing a single probe requires the manufacture of an entire new set of masks), thus making the tool impractical for the design and redesign of smaller focused arrays for toxicogenomics.

A more recent technology developed by NimbleGen Systems employs photolithography to manufacture microarrays but obviates the need for photolithographic masks to pattern light [10-13]. The method instead utilizes the Digital Micromirror Device (DMD) developed by Texas Instruments [14, 15] to pattern light using only digital input.

This DMD-based approach to microarray manufacture has all the fundamental advantages of conventional photolithographic manufacturing techniques, including high density, high information content, parallel chemical synthesis, and manufacturing scalability. However, using the DMD, photolithographic masks are not required. This eliminates the fundamental cost and time barriers associated with the manufacture of custom high-density microarrays. Also, since the synthesis cycle for a given array is relatively short (approximately three hours) and the design process is digital (and therefore very low cost), an iterative approach to array design can be employed, in which experimental results can be used to improve the design of subsequent experiments.

The DMD is an array of 786 432 aluminium mirrors contained within a 17.4 x 13.1 mm area; each mirror is 16 ^m square and arranged on a 17 ^m centre-to-centre spacing [14, 15]. Figure 4.1 shows an external view of a DMD device, as well as micrographs of a section of the mirror array and the individual mirror substructure. Higher-density micromirror arrays are available from Texas Instruments, which offer 1.3 x 106 individual mirrors or more. Each mirror is mounted on a torsion hinge and can be deflected 10° in the positive or negative direction from the neutral state in a voltage-dependent manner. Mirrors tilted in the +10° orientation deflect light into the light path and onto the flow cell, while mirrors deflected in the -10° orientation deflect light out of the light path and onto an absorber. Using this method, sharply focused spatially resolved, digitally generated light patterns can be created.

Using these patterns of light in combination with photochemistry, DNA arrays can be manufactured. Figure 4.2 shows an overview of this synthesis process using

Fig.4.1 DMD structure. (A) Texas Instruments Digital Micromirror Device. (B) Photomicrograph of a section of the micromirror array, with a grain of table salt shown for comparison. Each mirror is 16 ^m square, on a 17 ^m pitch. (C) Diagram of two micromirrors, with one mir ror in the 'off position and one mirror in the 'on' position. (D) high-resolution photomicrograph of four micromirror structures, with one micromirror removed to show the underlying microelectromechanical architecture. (Images courtesy ofTexas Instruments Inc.).

Fig. 4.2 Method for manufacture of microarrays using the NimbleGen Maskless Array Synthesizer. The UV light source is projected by the mirror array onto locations where photodeprotection is required. To the right is a graphical depiction of the sequential spatially addressed addition of nucleotides to the growing array.

Fig. 4.2 Method for manufacture of microarrays using the NimbleGen Maskless Array Synthesizer. The UV light source is projected by the mirror array onto locations where photodeprotection is required. To the right is a graphical depiction of the sequential spatially addressed addition of nucleotides to the growing array.

the DMD. In the first step, two out of four of the possible synthesis sites are illuminated, resulting in deprotection of the surface-bound nucleotide and allowing for the subsequent coupling of the G nucleotide. In the next step, two different positions on the array are illuminated, resulting in coupling of the C nucleotide at those specific locations. By repeating this process, the desired oligonucleotide polymers can be synthesized in parallel.

Figure 4.3 shows the interior architecture of the NimbleGen Maskless Array Synthesizer (MAS). As shown in the figure, the fundamental components of the NimbleGen MAS instrument are the Texas Instruments DMD, an optical subassembly for illumination of the DMD, an imaging subassembly for refocusing the light projected from the DMD onto the activated substrate, a flow cell, and DNA synthesis chemistry including monomers with photolabile protecting groups. Not shown in this figure is the MAS fluidics delivery system, which is an Expedite DNA synthesizer from Applied Biosystems. With the exception of the fluidics station, a small optical shutter, and the micromirrors, there are no moving parts within the MAS unit. This results in a stable and robust instrument design that is amenable to desktop mi-croarray manufacture in the average laboratory.

The digital nature of the light patterns created by the DMD, combined with the fact that each mirror is independently addressable, allows for spots of varying sizes or centre-to-centre distances (pitch) to be synthesized. Virtually any pattern based on the 16 ^m mirror size can be created. Figure 4.4 is a hybridization image in which the same oligonucleotide probes have been synthesized in four different formats on the same array. In Figure 4.4 A, each probe was synthesized using four mirrors in a 2 x 2 block that were functionally grouped during synthesis. This results in a spot that is 33 ^m square (16 ^m mirror + 1 ^m space + 16 ^m mirror) on a 51 ^m pitch. As shown in the figure, a border of inactive mirrors surrounds each block of four

Fig. 4.3 NimbleGen maskless array synthesizer. reflects the required light patterns through the

Depicted is a cutaway view of the MAS in which the thick white line illustrates the light path from its origin at the lamp, through the optics, to the flow cell. The digital micromirror device (DM D)

optical pathway to project a 1 : 1 image on the flow cell. The system can produce very high-contrast images, allowing the synthesis of arrays of oligonucleotides with lengths of up to 80 nucleotides.

' ..- £ . /« JBäp .» i^i »| V ' ' Jt _ : . , ,.;.,.■;., i? V ¿i' I«! !' ' ~S

'¡I, :■■■.!■.. ■.■■:■;■..■.: ■: ¡-.; ..-v'.-'i.:.;.-r;: i.^-:^ ■ r ■: , ■ - ¡\.: : ■■■:■■ • V

0 0

Post a comment