Five genes with homology to the floral homeotic genes deficiens of Antirrhinum and agamous of Arabidopsis were isolated from tomato. Each of the five genes is unique in the genome and could be localized to a different chromosome by RFLP mapping. Four of the tomato genes (hereafter TM) are flower-specific with distinguishable temporal expression. TM4 and TM8 are 'early', while TM5 and TM6 are 'late' genes. TM4 is homologous to squamous and TM6 is similar to deficiens, which are, respectively, 'early' and 'late' bona fide homeotic genes in Antirrhinum. The proteins encoded by the five tomato genes, like several known homeotic genes from other plants, contain within their N-terminus a highly conserved DNA-binding domain, the MADS box. All known plant MADS box genes also share, however, other properties. They all contain a central, moderately conserved, and rather basic domain, and a highly divergent or even missing C-terminal domain. Furthermore, molecular modelling predicts the presence of a conserved amphipatic alpha helix, at a constant distance from the MADS box in each of these proteins. The common properties of eight MADS box proteins from three plant families indicate that all their domains were coded for by the same ancestor gene. The sequence homology between pairs of MADS genes from different species indicates that the MADS ancestor gene multiplied and diverged in an ancestor plant common to several dicotyledon families.

Phylogenetic analyses of angiosperm MADS-box genes suggest that this gene family has undergone multiple duplication events followed by sequence divergence. To determine when such events have taken place and to understand the relationships of particular MADS-box gene lineages, we have identified APETALA1/FRUITFULL-like MADS-box genes from a variety of angiosperm species. Our phylogenetic analyses show two gene clades within the core eudicots, euAP1 (including Arabidopsis APETALA1 and Antirrhinum SQUAMOSA) and euFUL (including Arabidopsis FRUITFULL). Non-core eudicot species have only sequences similar to euFUL genes (FUL-like). The predicted protein products of euFUL and FUL-like genes share a conserved C-terminal motif. In contrast, predicted products of members of the euAP1 gene clade contain a different C terminus that includes an acidic transcription activation domain and a farnesylation signal. Sequence analyses indicate that the euAP1 amino acid motifs may have arisen via a translational frameshift from the euFUL/FUL-like motif. The euAP1 gene clade includes key regulators of floral development that have been implicated in the specification of perianth identity. However, the presence of euAP1 genes only in core eudicots suggests that there may have been changes in mechanisms of floral development that are correlated with the fixation of floral structure seen in this clade.

MADS-box transcription factors play crucial roles in organ and cell differentiation in organisms ranging from yeast to humans. Most of the work on plant MADS-box proteins focused on their roles in floral development whereas less information is available on their function in fruit maturation. We cloned three distinct tomato cDNAs using a RT-PCR approach, encoding LeMADS1, LeMADS5 and LeMADS6 factors and whose mRNAs mostly accumulate in tomato flowers and fruits. Phylogeny analysis indicates that LeMADS1, 5 and 6 belong to the MEF2A-like family. When transiently expressed in tobacco leaves or in human cells, LeMADS1, 5 and 6 are targeted to the cell nucleus. As the endogenous target genes of these putative transcription factors are unknown, the transcriptional activity of these proteins was characterized in a heterologous system and we showed that, when fused to a Gal4-DNA-binding domain, they repress the transcription of heterologous reporter genes. Since histone deacetylases control MEF2 transcriptional activity and since a putative histone deacetylase binding site was present in LeMADS1, 5 and 6, we tested the potential interaction between these factors and HDAC5 deacetylase. Surprisingly, in this heterologous system, LeMADS1, 5 and 6 interacted with HDAC5 N-terminal region. Our data suggest that, like mammalian MEF2A, plant MADS-box transcriptional activity might be regulated by enzymes controlling chromatin acetylation.