Over a decade of work has led to the identification of domains that are responsible for modulating levels of expression and partial responsible for conferring cell-specific expression. As yet no explicit description exists that offers an explanation of the exquisite distribution of any single GPCR. This is due in part to the inherent multifactorial complexity of the problem but is probably due in equal measure to an over-reliance upon use of transient reporter gene analysis which takes little heed of the organized chromatin environment of the endogenous gene. It can be no coincidence that the GPCR gene about whose transcriptional control we know most, the rhodopsin gene, has been subject to relatively intense transgenic analysis. Wholesale application of transgenic mouse strategies and translation of piecemeal biochemical assays into array based whole-genome analyses will go some way to addressing these deficiencies. In this latter respect, following the well-worn historical precedent of borrowing technology from the yeast community will continue to provide the technical impetus. Witness, for example, the application of array-based technology to the genome-wide identification of transcription factor targets ofyeast cell cycle genes, SBF and MBF (Iyer etal. 2001), Gal4 (Ren etal. 2000) and the more recent identification of mammalian targets of p53 (Wang et al. 2001). Obvious developmental biological constraints place a natural limitation on the pace with which transgenic mouse strains can be generated and analysed. It is likely that the rapidly advancing improvements in bioinformatic analyses of gene prediction, gene structure and diagnosis of genomic regulatory elements will provide the necessary focus to expedite in vivo analysis. As more genome sequences become available in the public domain, such in silico analysis of gene structure, promoter localization and enhancer/repressor identification will largely replace the tedium of RNAse protection analysis, primer extension and RACE analysis that have so often been the rate limiting step in defining potential genomic regulatory elements. Bioinformatic 'context-dependent' comparative studies (Kel et al. 2001) across different phyla and genes will allow explicit identification of real transcription factor binding sites, something that is not possible at present using only short motifs that occur stochastically several thousands of times per genome. Maybe then, will we be able to answer the question 'Why does a hippocampal pyramidal cell express M1 but not M5 muscarinic cholinergic receptors?'.
Was this article helpful?