Nucleotide sequences recognized by the AGAMOUS MADS domain of Arabidopsis thaliana in vitro

The AGAMOUS gene of Arabidopsis thaliana is a homeotic gene involved in the development of stamens and carpels. This gene encodes a putative DNA-binding protein sharing a homologous region with the DNA-binding domains, MADS boxes, of yeast MCM1 and mammalian SRF. To examine the DNA-binding activity of the AGAMOUS protein, double-stranded oligonucleotides with random sequences of 40 bp in the central region were synthesized and mixed with the AGAMOUS MADS domain overproduced in Escherichia coli . Oligonucleotides which bound to the MADS domain were recovered by repeated immunoprecipitation with an antibody which recognizes the overproduced protein. From a comparison of the recovered DNA sequences, the consensus sequence of the high-affinity binding-sites for the AGAMOUS MADS domain was determined to be 5'-TT(A/T/G) CC(A/T)₆GG(A/T/C)AA-3'. DNase I footprinting and methylation interference experiments showed that the MADS domain binds to this motif. Comparisons with the binding-site sequences of other MADS-box proteins revealed that the MCM1 binding-sites in a-mating type-specific promoters of Saccharomyces cerevisiae show similarities with the binding-site sequence of the AGAMOUS MADS domain. A synthetic MCM1 binding-site in the upstream region of the STE2 gene is recognized by the AGAMOUS MADS domain.     

Specific interactions between the K domains of AG and AGLs, members of the MADS domain family of DNA binding proteins

MADS domain (for M CM1, A G, D EFA and S RF) proteins are regulatory proteins found in all major eukaryotic kingdoms. Plant MADS domain regulatory proteins have a region of moderate sequence similarity that has been designated as the K domain, and its predicted coiled-coil structure suggests a role in establishing a protein—protein interaction. In vivo studies with the Arabidopsis AGAMOUS (AG) protein have indicated that the K domain is important for AG function. Using a bait fusion protein containing the K domain and the C-terminal region of AG in a yeast two-hybrid selection, 156 clones that encode potential AG-interacting proteins were identified. These clones each encode one of four highly related MADS domain proteins: AGL2, AGL4, AGL6 and AGL9. Additional analysis showed that the K domain of AG alone was able to bind the K domains of these AGLs. This binding was further confirmed by immunoprecipitation experiments using in vitro synthesized AG and AGL K domains. These results strongly suggest that AG interacts with AGL2, AGL4, AGL6 and AGL9 in vivo. Based on these results and previous observations, it is proposed that the AG function requires interaction with at least one of these AGL proteins, and such interactions contribute to the functional specificity of the AG protein.     

Genetic and molecular interactions between BELL1 and MADS box factors support ovule development in Arabidopsis

In Arabidopsis thaliana and many other plant species, ovules arise from carpel tissue as new meristematic formations. Cell fate in proliferating ovule primordia is specified by particular ovule identity factors, such as the homeodomain factor BELL1 (BEL1) and MADS box family members SEEDSTICK (STK), SHATTERPROOF1 (SHP1), SHP2, and AGAMOUS. Both in the bel1 mutant and the stk shp1 shp2 triple mutant, integuments are transformed into carpelloid structures. Combining these mutants in a bel1 stk shp1 shp2 quadruple mutant, we showed that the bel1 phenotype is significantly enhanced. We also demonstrate that ovule differentiation requires the regulation of the stem cell maintenance gene WUSCHEL, repression of which is predominantly maintained by BEL1 during ovule development. Based on yeast three-hybrid assays and genetic data, we show that BEL1 interacts with the ovule identity MADS box factors when they dimerize with SEPALLATA proteins. We propose a model for ovule development that explains how the balance between carpel identity activity and ovule identity activity is established by a MADS box homeodomain protein complex.     

Identification of Fructose-1,6-bisphosphate aldolase cytosolic class I as an NMH7 MADS domain associated protein

We are interested in identifying proteins that interact with the MADS domain protein NMH7 of Medicago sativa. We use an affinity column with a synthetic peptide derived from the MADS domain of NMH7 which has been reported to mediate protein-protein interaction with non-MADS domain interacting proteins. We identified ∼40 and ∼80 kDa specifically bound proteins as the monomeric and dimeric forms of Fructose-1,6-bisphosphate aldolase cytosolic class I. NiNTA pull down assays revealed that K- and C-terminus regions of NMH7 are not required for the interaction with aldolase. Aldolase enzymatic activity is not required for the interaction with NMH7. NMH7 and aldolase were coimmunoprecipitated from non-inoculated seed and seedlings extracts. Colocalization studies using confocal microscopy showed that aldolase and NMH7 are localized in the cytoplasm and the nucleus of the cortical cells. These data together show that M. sativa aldolase is a novel MADS domain binding protein, and suggest a broader functional repertory for this enzyme, as has been proposed for other glycolytic enzymes.     

Defining subdomains of the K domain important for protein–protein interactions of plant MADS proteins

The MADS proteins APETALA3 (AP3), PISTILLATA (PI), SEPALLATAI (SEPI), SEP2, SEP3, AGAMOUS, and APETALA are required for proper floral organ identity in Arabidopsis flowers. All of these floral MADS proteins conserve two domains: the MADS domain that mediates DNA binding and dimerization, and the K domain that mediates protein protein interaction. The K domain is postulated to form a several amphipathic c-helices referred to as K1, K2, and K3. The K1 and K2 helicies are located entirely within the K domain while the K3 helix spans the K domain-C domain boundary. Here we report on our studies on the interactions of the B class MADS proteins AP3 and PI with the E class MADS proteins SEP1, SEP2, and SEP3. A comparative analysis of mutants in the K domain reveals that the subdomains mediating the PI/AP3 interaction are different from the subdomains mediating the PI/SEP3 (or PI/SEP1) interaction. The strong PI/SEP3 (or PI/SEP1) interaction requires K2, part of K3, and the interhelical region between K1 and K2. By contrast, K1, K2 and the region between K1 and K2 are important for strong AP3/PI interaction. Most of the K3 helix does not appear to be important for either the PI/AP3 or the PI/SEP3 (or PI/SEP1) interaction. Conserved hydrophobic positions are most important for the strength of both PI/AP3 and PI/SEP3 dimerization, though ionic and/or polar interactions appear to play a secondary role.