A sensitive array for microRNA expression profiling (miChip) based on locked nucleic acids (LNA)

Abstract

MicroRNAs represent a class of short (approximately 22 nt), noncoding regulatory RNAs involved in development, differentiation, and metabolism. We describe a novel microarray platform for genome-wide profiling of mature miRNAs (miChip) using locked nucleic acid (LNA)-modified capture probes. The biophysical properties of LNA were exploited to design probe sets for uniform, high-affinity hybridizations yielding highly accurate signals able to discriminate between single nucleotide differences and, hence, between closely related miRNA family members. The superior detection sensitivity eliminates the need for RNA size selection and/or amplification. MiChip will greatly simplify miRNA expression profiling of biological and clinical samples.

FIGURE 1.

Mixed DNA/LNA capture probes display increased sensitivity for miRNA detection. miRNA expression was assessed in murine liver using a test set of LNA-modified (left) or unmodified DNA oligonucleotide capture probes (right). Decreasing amounts of total RNA were used as input material for miRNA analysis. Data are presented as median intensity (four replicas per miRNA capture probe; a representative experiment is shown).

FIGURE 2.

Accurate miRNA detection in murine heart and liver samples using T_m normalized LNA-modified capture probes. Detection of miRNA expression by PM (perfect match; 100% complementarity to the miRNA sequence), 1MM (single nucleotide mismatch at the central position of the nucleotide sequence of the mature miRNA), and 2MM (two nucleotide mismatches at the central position) LNA-modified captures probes. Data are presented as median intensity ± SD (four replicas per miRNA capture probe; a representative experiment is shown).

FIGURE 3.

Confirmation of miChip data by Northern blotting. Eight miRNAs were selected for comparative analysis as described in the text. The Y-axis of the miChip graph refers to the background-corrected averaged mean fluorescence intensities of four replicas normalized for U6 RNA expression, while the Y-axis in the Northern blot graph refers to background-corrected intensities (as calculated by ImageJ). For both the miChip and the Northern blot data, the signal corresponding to the miRNA with the highest signal intensity (miR-1) was set to 100%, and signals for the additional miRNAs analyzed were calculated as a percentage thereof. Ethidium bromide stained 5S RNA (*) is shown as a loading control. Arrows indicate the position of the pre-miRNAs (top arrow, ∼65 nt) and of the mature miRNAs (bottom arrow, ∼22 nt).

FIGURE 4.

Genome-wide miRNA profiling using miChip. miRNA expression profiles were monitored across four mouse tissues (duodenum [“Duod.”], spleen, heart, and liver). Data were organized according to the expression levels of individual miRNAs (condition tree). The key color bar indicates miRNA expression levels (dark red indicates high expression, while dark blue indicates no detectable expression; the Genespring data file is presented in Supplemental Table 5).