Labelling of targets for cDNA microarrays; comparison of commercial and non-commercial methods

Anne-Marie Cleton-Jansen, Tom van Wezel, Feija van Bokhoven, Thirza Broekhuizen-van de Berg, Hans Baelde

Dept. Pathology, Leiden University Medical Center, Leiden  

Labelling of cDNA for hybridisation is a critical step in the cDNA microarray technique. The fluorescent labels are expensive and the target cDNA is often unique and limited. The original method that is applied in Stanford uses micrograms of mRNA for labelling with Cy5 and Cy3 in a direct reverse transcription, which is an unfeasible approach for most primary patient samples, simply because it is not possible to obtain that amount of RNA from such material. In our hands this method was extremely inefficient because much RNA is lost when purifying mRNA.

We have tested several protocols for direct incorporation of Cy5 and Cy3 into total RNA. Using Real Time PCR we assessed which amount of RNA and which reverse transcription conditions results in the best yield of cDNA.

For our specific applications direct incorporation was not satisfactory and therefore we tested several amplification methods. Two types of amplification can be applied: signal amplification or cDNA amplification. For signal amplification we have tested two commercially available methods: the Tyramid Signal Amplification of Perkin Elmer/NEN and the 3DNA Submicro detection kit from Genisphere. For cDNA amplification we applied in vitro transcription on double strand cDNA and subsequent direct incorporation of Cy-dyes in a second RT on the resulting cRNA. Furthermore we have tested the protocol by TIGR for aminoallyl directed labelling and the ARES Alexa Fluor labelling kit from Molecular Probes.

In conclusion, few labelling methods give satisfactory results and although companies are supplying kits, these do not always guarantee high signals and low background. Much depends on the quality of the input RNA but also on the coating of the glass slides.