100-250 ng mRNA was isolated from total RNA for each patient using the Oligotex kit (Qiagen) according to the manufacturers instructions. Prior to hybridization, 20ng of Arabidopsis thaliana DNA template was labeled with a-33P dCTP (10 microCi/microL; Amersham) during a random priming labeling reaction (pd(N)6, Amersham) (Feinberg and Vogelstein, 1983). Reversed transcribed cDNA from patient targets were prepared in a final volume of 30microL as described in detail previously, with the exception that the labeling reaction was carried out at 42°C for 50 min (Eickhoff et al. 2000). RNA was hydrolyzed under alkaline conditions by the addition of 3.5 microL of 3M NaOH and incubation for 20 min at 68°C. The samples were neutralized with the addition of 10 microL of 1M Tris-HCl (pH 7.4) and 10.5 microL of 1M HCl. Unincorporated radioactive nucleotides were removed from both control and patient targets by filtration through Microspin G-50 columns according to the product manual (Amersham).
Prior to each hybridization reaction, unused filters were initially washed in 1 X SSC; 0.1% SDS at room temperature for 10 min (X 2) followed by three washes in 0.1 X SSC; 0.1% SDS for 20 min at 80°C. These washes eliminated the potential problem for unbound or loosely bound probe cDNA from interfering with the hybridization reaction. Each cDNA filter was pre-wetted with deionized water and placed in a hybridization bottle (3.5 X 25 cm). The water was completely poured off and the filter was pre-hybridized for 2 hrs at 50°C in a 10 mL of hybridization solution containing 7% SDS, 50% formamide, 5 X SCC, 2% blocking reagent (Roche Diagnostics), 50mM sodium phosphate (pH 7.0) and denatured herring sperm DNA (50microg/microL). The labeled target was added to the filters in 5 mL of the same preheated solution and hybridization of the filters was performed overnight (24 hrs) at 42°C. Following overnight hybridization, membranes were removed from the bottles and washed in 1 X SSC/0.1% SDS buffer for 10 min at room temperature followed by two washes for 30 min in 0.2 X SSC/0.1% SDS at 65°C. Excess moisture was removed by briefly placing the membranes on filter paper (Whatman) and the damp filters were wrapped in commercial grade plastic wrap. Filters were exposed ~24 hrs to imaging plates (BAS-MS 2325, Fujifilm, Japan) and scanning was carried out at a resolution of 50microns on a FLA-3000G imaging system (Fujifilm, Japan). After hybridization, filters were individually stripped in a boiling solution of 0.5% SDS for 5 min, taken off the heat and agitated for an additional 3 hrs at room temperature and stored damp wrapped in plastic wrap at 20ºC for 7-8 weeks (half-life of 33P is 25.4 days). Prior to re-use, filters were re-scanned as above to ensure that residual radioactivity was negligible. For the purposes of the microarray comparison study, all filters were chosen from the same spotting production run and all filters were used for the first time.
Spot quantitation: An exposure time of ~24 h was sufficient to yield signals from cDNA clones, yet keeping the data overflow to a mimimum. Images were quantified using VisualGrid® software V3.4 (http://www.gpc-biotech.com). Each grid was manually adjusted to the image such that the linear correlation coefficient between duplicate spots was greater than 0.984. The following output values from VisualGrid were used for subsequent data analysis: SIGNAL_RAW1 (the weighted mean pixel intensity within the node mask for the first duplicate), SIGNAL_RAW2 (the weighted mean pixel intensity within the node mask for the second duplicate), BKD_MEAN (the mean of the weighted means of all background spots in a 6 X 6 block). For this dataset, the weight of each pixel was set to one.