Mice were sacrificed by a rising concentration of CO2 and livers were immediately removed for preparation of RNA. Total RNA was isolated from mouse liver with a phenol-guanidine isothiocyanate reagent, TRIzol (Invitrogen) and further purified with RNeasy Mini Kit (Qiagen Ltd) in accordance with the manufacturer’s instructions. RNA was pooled from three animals of each genotype for each experiment, and the experiments were carried out twice. Results from the two experiments were exactly the same. The A260/280 ratio of total RNA used was typically ³ 1.9. The quality of RNA was assessed by using the Agilent 2100 Bioanalyzer. The probe labelling and hybridisation procedures were conducted by following the Affymetrix Technical Manual (Affymetrix, Santa Clara, CA). cDNA was synthesized from the total RNA by using Superscript ds-cDNA Synthesis Kit (Invitrogen Ltd, Paisley, U.K.) with a T7-(dT)24 primer incorporating a T7 RNA polymerase promoter. The cRNA was prepared and biotin-labelled by in vitro transcription by using BioArray High Yield RNA Transcript Labelling Kit (ENZO Biochemical). Labelled cRNA was fragmented by incubation at 940C for 35 min in the presence of 40 mM Tris acetate, pH 8.1, 100mM potassium acetate, and 30 mM magnesium acetate. The samples were tested with hybridization to GeneChip Test3 arrays and analysed by Agilent Bioanalyzer. Fragmented cRNA (15 mg) was hybridized 16h at 450C to an MG_U74Av2 array (Affymetrix). After hybridization, the gene chips were washed and stained with streptavidin-phycoerythrin by using a fluidics station (Affymetrix). The chips were scanned in an Agilent G2500A scanner. Affymetrix oligonucleotide microarrays utilize multiple perfect-match and mismatch oligonucleotides to determine expression levels, so Affymetrix GCOS software was used to scan, determine the presence and the average difference value, and assess signal intensity of each probe set. Chip fluorescence was normalized by scaling total chip fluorescence intensity to a common value of 100 prior to comparison, and a normalisation value was set at 1. Data and statistical analyses were performed with Genespring Ver 6.1 (Silicon Genetics, Redwood City, CA) bioinformatics algorithms.
