The recent discovery of microRNAs (miRNAs) has redefined our understanding of the flow of genetic information in eukaryotes. MicroRNAs are 22-nucleotide non-coding RNAs that impact cell function and development by targeting the mRNA sequences of protein-coding transcripts, resulting in either mRNA cleavage or repression of productive translation [1-3]. MicroRNAs were first discovered in mutants of the nematode Caenorhabditis elegans that lacked the ability to control the timing of specific cell fate switches during development [4, 5]. Since then, several hundred miRNAs from C. elegans, plants, Drosophila melanogaster, and mammals have been identified through computational and cloning approaches [6, 7].

Figure 1

Full Figure and legendFigures & Tables index

Figure 2

Full Figure and legendFigures & Tables index

Figure 3

Full Figure and legendFigures & Tables index

There are currently estimated to be ~500-1000 miRNAs in the human genome. In C. elegans, each cell contains >1000 copies of any individual miRNA, with some cells exceeding 50,000 copies [7]. In humans, each cell is predicted to have between 100 and 240,000 copies of any specific miRNA. Scientists estimate that miRNAs account for 1%-5% of expressed genes in animal genomes. Therefore, miRNAs are one of the most highly abundant gene regulators in higher eukaryotes [8, 9].

Though hundreds of miRNAs have been discovered in a variety of organisms, little is known about their cellular function. The recent discovery of miRNAs has led to the development of several species-specific, high-throughput detection methods. In several reports, spotted oligonucleotide microarray technology has proven to be effective.  Most approaches to labeling miRNAs for microarray analysis have made use of traditional reverse transcription reactions. However, reverse transcription is not optimal for efficient labeling of such short templates. In addition, strong conservation between miRNA family members makes it difficult to design probes that are specific at the level of a single nucleotide out of a 20-nucleotide sequence. 

The NCode^™ Multi-Species miRNA Microarray v2 consists of ~1100 unique probes printed in triplicate for interrogating validated human, mouse, rat, C. elegans, D. melanogaster, and Zebrafish miRNAs, plus predicted human miRNAs. The array has all the recently released content from Sanger 9.0.  Additionally, the array contains internal controls to assess dye performance, labeling efficiency and sample normalization.   Also included on the array are probes for pre-miRNAs.  Array probes were designed using a proprietary algorithm to optimize hybridization at a uniform melting temperature, often allowing for single-mismatch discrimination.

Total RNA samples are enriched for small RNA molecules (<200 nucleotides) prior to microarray analysis using the PureLink^™ miRNA Isolation Kit or Centricon size exclusion. The NCode^™ direct labeling method polyadenylates the enriched small RNA molecules and ligates a capture sequence to the tailed RNA. The tagged and tailed miRNAs are subsequently hybridized to the multi-species array.  Bound miRNAs are detected by the hybridization of branched DNA structures containing approximately 900 Alexa Fluor^® 3 or Alexa Fluor^® 5 dye molecules. The 1:1 ratio of branched DNA structure to tagged and tailed hybridized miRNAs on the array, along with the signal amplification from 900 dye molecules per branched structure, provide a large linear range of signal intensities with minimal background.

The NCode^™ microarray platform is sensitive to sub-femtamolar levels of microRNAs using conventional static hybridization methods, but sensitivity can be significantly increased to attamolar levels using dynamic hybridization methods like the MAUI^® Hybridization System (BioMicro Systems). Figure 1A demonstrates the overall performance of the platform using a universal RNA reference standard. A universal reference RNA was diluted from 20 fmol to 18 amol and labeled with the NCode™ miRNA Labeling System. Samples were hybridized to an NCode™ Multispecies miRNA Microarray. Data were background corrected and mean intensities were averaged for all mammalian features. The number of positive features was defined as the number of features with intensities greater than 5X median background intensity in both channels.

The histogram  compares the mean signal in both channels (A3 and A5) from duplicate samples spiked with decreasing quantities of NCode™ universal reference RNA. Error bars represent 1 standard deviation.  Additionally, the figure compares the number of features positive in both A3 and A5 channels on both arrays versus the spike quantity.  More than 50% of the universal reference miRNA molecules were detectable (>5X median background) at concentrations as low as 0.6 fmol per reaction, which corresponds to 100 copies/cell for a given microRNA. Using the MAUI^® Hybridization System, >90% of the universal reference miRNAs were detected from the same spike quantity, while >50% were detectable when spiked at concentrations as low as 0.018 fmol/reaction.

The heat map and table in Figure 1 demonstrates the reproducibility of the NCode^™ platform using 17 replicate homotypic arrays with 10 g of mouse heart RNA in each channel.  The data were background corrected and normalized. The %cv was calculated from mean log₂ expression values from array to array.  The log₂%cv for both channels was approximately 7.7% using static hybridization. The array-to-array reproducibility improved to 3.5% using active hybridization.  The % cv log₂ RFU did not exceed 17% for any human feature significantly above background.

The biggest limitation of most microarray platforms is the quantity of starting material, typically 1–10 g of total RNA. In many instances, it is difficult or impossible to obtain this amount of RNA, especially from laser capture microdissections, fine-needle aspirates, and cell-sorted samples. The NCode^™ miRNA Amplification System is a robust linear miRNA amplification system that employs modifications to previously documented mRNA amplification methods, facilitating microarray profiling from as little as 50 ng of total RNA.

First, small RNAs are enriched from 50–500 ng of total RNA using the PureLink^™ miRNA Isolation Kit or Centricon size exclusion columns. Then, using the amplification kit, the small RNA is polyadenylated and used in a first-strand cDNA synthesis reaction. The first-strand cDNA is subsequently modified at the 3' end to incorporate a T7 promoter and amplified via  in vitro transcription. After purification, the amplified RNA is ready for microarray labeling and detection or cloning applications.

The NCode^™ miRNA Amplification System routinely yields approximately 30 gs of amplified miRNA from as little as 15 ng of enriched small RNA with greater than 0.8 r² value in comparison to unamplified direct labeled material.

Figure 2 demonstrates the accuracy of the amplification system. Three-hundred nanograms of total RNA from large-cell carcinoma tissue and adjacent normal tissue from three separate patients was enriched and amplified. The amplified samples were compared to 10 g of direct-labeled total RNA from the same samples by microarray analysis. The array data were background corrected, normalized, and averaged across patients. The ratios of all known human miRNAs in both tumor and  healthy tissues were calculated, and statistically significant differentially expressed miRNAs were validated by qRT-PCR. The histogram shown in figure 2 shows the mean miRNA expression in tumor samples in comparison to healthy tissue.  ΔC_T values from the qRTPCR validation are shown next to the corresponding miRNA.  Several differentially expressed microRNAs associated with the disease state were identified in both amplified and direct-labeled samples.

Quantitative RT-PCR has become the standard for microarray data validation, as well as an invaluable tool for profiling subsets of miRNAs with the greatest sensitivity. Most commercially available miRNA qRT-PCR systems use proprietary, pre-designed miRNA-specific primer sets for reverse transcription. Unfortunately, these systems require known publicly available miRNA sequences and a commercial qRT-PCR assay, limiting their use for rare or recently discovered miRNAs. The NCode^™ miRNA SYBR^®  qRT-PCR Kit overcomes this limitation, combining a carefully optimized polyadenylation reaction with the market-leading reverse transcriptase, SuperScript^™ III , in a "universal" first-strand cDNA synthesis reaction that does not require a proprietary miRNA-specific primer set.

The miRNA-specific amplification occurs during the qPCR reaction, using a PCR primer that the user can order for any miRNA sequence of interest. The miRNA-specific forward primer is paired with a universal reverse primer in a qPCR reaction with Platinum^® SYBR^® Green qPCR SuperMix-UDG. The SuperMix combines Platinum^®Taq DNA polymerase with SYBR^® Green fluorescent dye, delivering excellent sensitivity in the quantification of target sequences.

The performance of the NCode^™ SYBR^® miRNA qRT-PCR is demonstrated in figure 3.  1 10⁸ to 1 10² copies of a synthetic miRNA were spiked into each reaction. Prior to cDNA synthesis samples are polyadenylated.  Universal cDNA synthesis generates template from all polyadenylated tails using a primer specific for the first 24 adenosine nucleotides from the PAP reaction.  The same primer contains a universal sequence for PCR amplification.  miRNA specificity is conferred during PCR amplification using a forward primer specific for the mature miRNA sequence. The assay demonstrated a wide dynamic range of 7 orders of magnitude and sensitivity down to 100 copies of template.  Additionally, the NCode^™ SYBR^® Green miRNA qRT-PCR Kit is able to provide single nucleotide discrimination between closely related family members with only small losses to overall assay sensitivity.

The NCode^™ miRNA profiling platform enables researchers to cost effectively profile miRNA expression accurately and reproducibly from very small quantities of total RNA.  In addition, the flexibility, sensitivity and large dynamic range of the NCode^™ SYBR^® miRNA qRT-PCR Kit allows researchers to validate low or high abundant miRNA targets identified by microarray analysis in small quantities of total RNA.  These tools can easily be used to quickly discover relevant miRNAs associated with a disease state or cellular differentiation state.  It is anticipated that future use of these technologies will lead to a greater understanding of gene-regulating RNA molecules.  

References

1. Lagos-Quintana, M. Identification of novel genes coding for small expressed RNAs. Science  294 853–858 (2001).
2. Lagos-Quintana, M. , Rauhut, R. , Meyer, J. , Borkhardt, A. & Tuschl, T. New microRNAs from mouse and human. RNA  9 175–179 (2003).
3. Ambrose, V. microRNAs: tiny regulators with great potential. Cell  107 823–826 (2001).
4. Ambrose, V. MicroRNA pathways in flies and worms: growth, death, fat, stress, and timing. Cell  113 673–676 (2003).
5. Bartel, D.P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell  116 281–297 (2004).
6. Ambrose, V. The functions of animal microRNAs. Nature  431 350–355 (2004).
7. Ambrose, V. , Lee, R.C. , Lavanway, A. , Williams, P.T. & Jewell, D. MicroRNAs and other tiny endogenous RNAs in C. elegans. Curr Biol  13 807–818 (2003).
8. Bartel, D.P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell  116 281–297 (2004).
9. Nakahara, K. & Carthew, R.W. Expanding roles for miRNAs and siRNAs in cell regulation. Curr Opin Cell Biol  16 127–133 (2004).