Ton Raap1, Renate Hagedoorn1, Soemini Kasanmoentalib2, Paul Eilers2, Jeffrey Killian3 and Richard Joseph3
Departments of 1Molecular Cell Biology and 2Medical Statistics, Leiden University Medical Center, 2333AL Leiden, the Nederland; 3Perkin-Elmer Life Sciences, Boston, MA, USA.
The advent of microarray and chip hybridization technology has enabled large scale gene expression analysis at the RNA level by hybridizing fluorochromized complementary DNA (cDNA) of an mRNA sample to large numbers of gene probes and taking the resulting fluorescence hybridization signals as measures of mRNA concentrations.
In the great majority of such gene expression profiling studies, oligodeoxythymidilate (oligo-dT) primed reverse transcription (RT) of polyadenylated (poly-A) mRNA is applied to label the hybridization target. In a substantial fraction of these studies, the oligo-dT primer is extended at the 5' end with a sequence that promotes cRNA transcription by a bacterial phage RNA polymerase thus permitting, after 2nd strand cDNA synthesis, a linear target amplification that serves to compensate for limited sample size and sensitivity.
A well-established consequence of oligo-dT primed RT is that due to limited processivity of the reverse transcriptase and secondary RNA structure, many of the mRNA species are represented in the labeled target only with their 3' ends, a bias that will also be reflected in amplified
cRNA.
It is for this reason of 3'-end bias that in oligonucleotide probe based microarray - or chip hybridization, the probe sequences are often selected from the first couple of hundred nucleotides from the 3'-end of the mRNAs. As a consequence of this bias, detection of splice variants by oligonucleotide microarray hybridization is essentially limited to the 3'-end of the genes. Also the sensitivity potential of cDNA microarray hybridizations is not fully exploited as the labeled cDNA target will generally not cover the full mRNA complexity.
Here we present a chemical labeling procedure that overcomes the 5'-end representation problem. The method is based on fluorescent cis-platin derivatives that bind to DNA or RNA monofunctionally, with preference for the N7 of the guanosine residues. Such compounds are referred to here as MICROMAXÔ ASAP Labeling Reagents. They allow, in principle, labeling over the full length of mRNA, as evidenced here by the efficient detection of hybridization signals from labeled mRNA sequences hybridized to oligonucleotidearrays with the 3'-ends, middle parts and 5'-ends, represented as 70-mer oligonucleotides. Apart from its general utility in large scale expression profiling by cDNA microarray hybridization, this mRNA labeling method will be valuable for verifying expressio of exons by hybridizing mRNA labeled without bias to arrays of tiled genomic oligonucleotides, large scale detection of alternative splice variants by hybridisation to splice junction oligonucleotides and the study of long-range heterogeneity at 3'-ends. It may also find application in expression profiling non-polyadenylated RNAs.