Evolution and divergence of the MADS-box gene family based on genome-wide expression analyses

MADS-box genes encode transcription factors involved in various important aspects of development and differentiation in land plants, metazoans, and other organisms. Three types of land plant MADS-box genes have been reported. MIKCC- and MIKC*-type genes both contain conserved MADS and K domains but have different exon/intron structures. M-type genes lack a K domain. Most MADS-box genes previously analyzed in land plants are expressed in the sporophyte (diploid plant body); few are expressed in the gametophyte (haploid plant body). Land plants are believed to have evolved from a gametophyte (haploid)-dominant ancestor without a multicellular sporophyte (diploid plant body); most genes expressed in the sporophyte probably originated from those used in the gametophyte during the evolution of land plants. To analyze the evolution and diversification of MADS-box genes in land plants, gametophytic MADS-box genes were screened using macroarray analyses for 105 MADS-box genes found in the Arabidopsis genome. Eight MADS-box genes were predominantly expressed in pollen, the male gametophyte; all but one of their expression patterns was confirmed by Northern analyses. Analyses of the exon/intron structure of these seven genes revealed that they included two MIKCC-type, one M-type, and four MIKC*-type MADS-box genes. Previously, MIKC*-type genes have been reported only from a moss and a club moss, and this is the first record in seed plants. These genes can be used to investigate the unknown ancestral functions of MADS-box genes in land plants. The macroarray analyses did not detect expression of 56 of 61 M-type MADS-box genes in any tissues examined. A phylogenetic tree including all three types of Arabidopsis MADS-box genes with representative genes from other organisms showed that M-type genes were polyphyletic and that their branch lengths were much longer than for the other genes. This finding suggests that most M-type genes are pseudogenes, although further experiments are necessary to confirm this possibility. Our global phylogenetic analyses of MADS-box genes did not support the previous classification of MADS-box genes into type I and H groups, based on smaller scale analyses. An evolutionary scenario for the evolution of MADS-box genes in land plants is discussed.