02 Apr 2002

Though the successful development of DNA chip technology has facilitated the simultaneous measurement of thousands of mRNA transcripts in a single sample, the technology is presently incapable of distinguishing between alternatively spliced transcripts. As much of the complexity of the mammalian proteome is due to the occurrence of alternatively spliced mRNAs, an automated method to detect these transcripts would be immensely valuable. This is especially true for neurodegenerative diseases, where some of the major suspect genes occur in different splice forms; tau, for example, has six.

In this month's Nature Biotechnology, researchers in Xiang-Dong Fu's lab at the University of California, San Diego, report the development of just such a tool. Their method uses a combination of polymerase chain reaction (PCR) followed by a solid-state detection assay. At the heart of the technique are multifunctional "smart" oligonucleotides. These are designed in pairs to match sequences on the donor and acceptor side of known splice sites. Prior to amplification, mRNA samples are incubated with a mixture of these oligos, which also serve as a template for PCR primers, and those that fall on either side of a junction are ligated together. This facilitates amplification and ensures that only splice sites that occur within the RNA sample are amplified, as in the absence of a given junction its two oligos will not ligate.

Following amplification, splice sites are detected by a fiber-optic microarray that contains thousands of DNA "addresses"-microbeads that have unique nucleotide sequences attached. The "smart" oligos also have a complementary "address" that allows the DNA amplified from a splice junction to be captured on the microarray. The presence of the captured DNA can then be detected by fluorescence.

Yeakley et al. used the technique to measure alternative splicing of six different genes in five different cell types, and the data compared well with the profile obtained using conventional reverse transcription-PCR. The technique promises to be useful for detecting changes in splicing patterns that may arise when, for example, a normal cell turns cancerous, as Paula Grabowski, University of Pittsburgh, points out in an accompanying News and Views.

The authors also point out the method's limitations. It fails to detect distinct isoforms of the receptor tyrosine phosphatase, PTPRC, known to be expressed in U-937 and Jurkat cells (see also comment below by George Church). However, it may be possible to overcome this limitation by the use of "smarter" oligonucleotides.—Tom Fagan