RNA Amplification for Array Analysis

RNA amplification using the Van Gelder and Eberwine technique (Van Gelder 1990) utilizes an oligo(dT) primer containing the T7 RNA polymerase promoter for synthesis of first strand cDNA. The poly(A) tail at the end of mRNA sequences serves as the substrate for the binding of these primers. Since mRNA typically constitutes only 1-5% of the total RNA in the cell, only this fraction of the total RNA is amplified. The tissue type, its developmental state, and its health all influence the actual proportion of mRNA in a total RNA sample. Total RNA from brain, testes, and embryoic tissues may contain up to 4% mRNA, while RNA from many other tissues will have only 1% or less mRNA. The RNA isolation method can also influence mRNA content. The generally accepted average value for mRNA content is about 2% of a total RNA sample. When 1 µg of total RNA, 2% or 20 ng of which is mRNA, is amplified 1000-fold, yields of 20 µg aRNA (or cRNA) should be expected. You may observe higher fold amplification when starting with lower amounts of total RNA. This is because, in an in vitro transcription (IVT) reaction, a finite amount of RNA can be synthesized with the fixed amount of NTPs. When starting with less RNA, NTPs do not become limiting until the RNA is amplified beyond the typical 1000-2000-fold amplification levels seen with higher amounts of input RNA.

Glass microarray analysis experiments typically require 5-20 µg of total RNA per slide for sample labeling and hybridization. Thus, microarray-based gene expression analysis of very small samples [laser capture microdissection (LCM), tissue biopsies, or other clinical samples] is difficult due to the very low amounts of total RNA recovered from the samples. Linear amplification of RNA from small samples produces sufficient quantities of RNA for sample labeling and hybridization. Since the amplification technique is highly reproducible and maintains representation of the gene expression in the original sample, it is recommended for probe synthesis by most manufacturers of commercially available microarrays.

Direct labeling is incorporation of modified NTPs into amplification products during the IVT step of the amplification process. To make aRNA that is labeled with fluorescent dyes, a mixture of dye-modified and unmodified (or unlabeled) nucleotides are typically used in order to obtain an optimal ratio of dye-labeled to unlabeled nucleotide for maximal fluorescence. Usually ~200-400 µM of dye-labeled CTP is used with 1-3 mM unlabeled NTPs. Biotin-modified nucleotides are incorporated fairly well with T7 RNA polymerase. Ambion recommends using UTP:biotin-UTP ratios of 1:1 to 3:1. In general, labeled nucleotides are not incorporated as efficiently as unlabeled molecules during amplification, and therefore direct labeling does compromise sample yield. Furthermore, if both Cy5 and Cy3 are used in a direct labeling reaction, Cy5 is not incorporated as well as Cy3, and corrections during data analysis are necessary to adjust for this disparity.

Indirect labeling incorporates amino allyl UTP into amplification products during the IVT, and the amino allyl-modified aRNA produced is then chemically coupled to a detectable moiety such as a fluorescent dye or biotin. This method, though more time-consuming than direct labeling, can result in very highly labeled aRNA because amino allyl-modified UTP is incorporated very efficiently by T7 RNA polymerase.

A single round of amplification yields product sizes ranging from 200 bases to 6 kb. The majority of these products are approximately 1.5 kb in length. A second round of amplification will result in shorter products. We recommend using an Agilent 2100 bioanalyzer to visualize these products. Amplification products can be visualized by agarose gel electrophoresis; they will migrate as a smear. Although this data is still useful, it is less informative than bioanalyzer analysis.

Oligonucleotides are single-stranded and therefore only anneal to their complement. Probes for microarrays are either generated from an RT reaction (cDNA) or from linear amplification (aRNA). Both probe synthesis methods generate antisense sequences. Most oligonucleotide arrays, therefore, are designed with sequences in the sense orientation.