Microarray-based analysis of gene expression in very large gene families: the cytochrome P450 gene superfamily of Arabidopsis thaliana.

Cytochrome P450 (P450s) are heme-thiolate protein products of a very large gene superfamily, present in all kingdoms and involved in a variety of metabolic reactions. P450s are classified according to the degree of amino acid sequence identity, with P450s of the same family defined as having >40% identity, and P450s of the same subfamily having >55% identity. Currently, 273 P450 genes distributed over 45 families have been identified in Arabidopsis, and its genome is estimated to contain as many as 286. Genome-wide DNA microarrays make it possible to broadly correlate P450 gene activity with alterations in physiological or developmental states. A potential problem with microarray research is that sequence similarity between and within these families of closely related genes may lead to cross-hybridization. We designed experiments to systematically evaluate the specificity of P450 microarrays, and showed that conditions could be optimized to provide a very high degree of hybridization specificity. Under these conditions, and employing a 20% intensity value of maximum hybridization intensity as a cut-off, labeled P450 genes exhibited essentially no cross-hybridization between families and within subfamilies. We also compared the gene transcription levels of microarray probes derived from EST clones and from genomic DNA sequences for which ESTs were not available, using cDNA produced from RNA from various Arabidopsis tissue as the target. Many of the P450 genes displayed tissue-specific expression, leading to hypotheses as to the function of individual genes and their regulation. We also observed that several of the genomic sequences reported high levels of expression, highlighting the limitations of expression analysis based on ESTs alone.

Intron-exon organization and phylogeny in a large superfamily, the paralogous cytochrome P450 genes of Arabidopsis thaliana.

The cytochrome P450 gene superfamily is represented by 80 genes in animal genomes and perhaps more than 300 genes in plant genomes. We analyzed about half of all Arabidopsis P450 genes, a very large dataset of truly paralogous genes. Sequence alignments were used to draw phylogenetic trees, and this information was compared with the intron-exon organization of each P450 gene. We found 60 unique intron positions, of which 37 were phase 0 introns. Our results confirm the polyphyletic origin of plant P450 genes. One group of these genes, the A-type P450s, are plant specific and characterized by a simple organization, with one highly conserved intron. Closely related A-type P450 genes are often clustered in the genome with as many as a dozen genes (e.g., of the CYP71 subfamily) on a short stretch of chromosome. The other P450 genes (non-A-type) form several distinct clades and are characterized by numerous introns. One such clade contains the two CYP51 genes, which are thought to encode obtusifoliol 14a demethylase. The two CYP51 genes have a single intron that is not shared with CYP51 genes from vertebrates or fungi, or with any other Arabidopsis P450 gene. Only a few of the Arabidopsis P450 genes are intronless (e.g., the CYP710A and CYP96A subfamilies). There was a relatively good correlation between intron conservation and phylogenetic relationships between members of the P450 subfamilies. Gene organization appears to be a useful tool in establishing the evolutionary relatedness of P450 genes, which may help in predictions of P450 function.