Structural basis for the oligomerization of the MADS domain transcription factor SEPALLATA3 in Arabidopsis

Abstract

In plants, MADS domain transcription factors act as central regulators of diverse developmental pathways. In Arabidopsis thaliana, one of the most central members of this family is SEPALLATA3 (SEP3), which is involved in many aspects of plant reproduction, including floral meristem and floral organ development. SEP3 has been shown to form homo and heterooligomeric complexes with other MADS domain transcription factors through its intervening (I) and keratin-like (K) domains. SEP3 function depends on its ability to form specific protein-protein complexes; however, the atomic level determinants of oligomerization are poorly understood. Here, we report the 2.5-Å crystal structure of a small portion of the intervening and the complete keratin-like domain of SEP3. The domains form two amphipathic alpha helices separated by a rigid kink, which prevents intramolecular association and presents separate dimerization and tetramerization interfaces comprising predominantly hydrophobic patches. Mutations to the tetramerization interface demonstrate the importance of highly conserved hydrophobic residues for tetramer stability. Atomic force microscopy was used to show SEP3-DNA interactions and the role of oligomerization in DNA binding and conformation. Based on these data, the oligomerization patterns of the larger family of MADS domain transcription factors can be predicted and manipulated based on the primary sequence.

Figure 1.

Amino Acid Sequence of SEP3 and Truncation Constructs. (A) SEP3 sequence colored by domain, with the M domain in green, the I domain in yellow, the K domain in blue, and the C domain in pink. The domain structure is depicted schematically below the amino acid sequence. (B) Sequence of the SEP3^75-178 construct used for all crystallization studies spanning a portion of the I domain, the complete K domain, and a portion of the C domain. (C) Sequence of the SEP3^1-110 construct used in the AFM studies.

Figure 2.

Overview of Structural Quality. (A) At left, SEP3 tetramer depicted as a cartoon and colored by temperature factor (B-factor) with dark blue (lowest) and red (highest). The average B-factor for the structure was 69 Ų. At right, view as per left with each monomer colored uniquely and one monomer displayed with 2Fo-Fc electron density contoured at 1.5 sigma. The loop region is circled in red and the dimerization region in yellow. (B) At left, close-up of the loop region corresponding to the red circled region in (A), right. At right, electron density for the dimerization region corresponding to the yellow circled region in (A), right. Based on the quality of the electron density map, the protein backbone and side chains could be positioned unambiguously.

Figure 3.

Structure of SEP3 Oligomerization Domains. (A) SEP3 tetramer depicted as a cartoon, with each monomer A to D colored uniquely in light green, dark green, light purple, and dark purple, respectively, with the N and C termini labeled. Helix 1 and helix 2 are labeled and indicated by arrows. (B) The hydrophobic kink region is shown for one monomer with the view as per (A). Residues are labeled and drawn as sticks colored by atom. Hydrogen bonds are shown as dashed red lines. (C) Dimerization of SEP3 is mediated by leucines and tyrosines in helix 1 and intermolecular salt bridges via the N-terminal portion of helix 2. Residues are labeled and depicted as sticks colored by atom. Hydrogen bonds are drawn as dashed red lines. For clarity, residues are labeled for one monomer. (D) View down the 2-fold crystallographic axis that forms the tetramerization interface. The intermolecular water-mediated hydrogen bonding network is shown. Residues are depicted as sticks and colored by atom, water molecules are in dark blue, and residues labeled for a single monomer for clarity.

Figure 4.

Tetramerization Interface of SEP3. (A) Sequence alignment of representative MADS TFs spanning the sequence of the crystallized SEP3^75-178 construct described here with the I domain in yellow, the K domain in blue, and the C-terminal domain in pink. Residues involved in dimerization and tetramerization are highlighted in light blue and light green, respectively. Mutants are marked with a star, and residues corresponding to deletion mutant SEP3^Δ161-174 are boxed in black. All proteins are from Arabidopsis with SEP3, SEP2, SEP1, AP1, AG, SOC1, SVP (SHORT VEGETATIVE PHASE), FLC, PI, and AP3. (B) Close-up of the tetramerization interface of SEP3. Interacting residues are depicted as sticks and colored by atom with monomers colored uniquely. Hydrogen bonds are show as dashed red lines. Residues from the green monomer are labeled. (C) Residues that are deleted in SEP3^Δ161-174 are shown in cartoon colored gray. Point mutants that affect tetramerization are depicted as sticks and colored gray. Labels are as per (B).

Figure 5.

Size-Exclusion Chromatograms of Wild-Type and Mutant SEP3 Proteins. SEP3^75-178 (wt) is in black, SEP3^ΔV90 is in pink, SEP3^Δ161-174 is in dark blue, M150A is in green, L171A is in purple, and L154A is in yellow. The oligomerization state of the point mutants and SEP3^Δ161-174 was predominantly dimer as confirmed by multiangle laser light scattering. M150A gave a molar mass of 20,560 g mol⁻¹ (±7.3%), L154A 22,990 g mol⁻¹ (±9.5%), and SEP3^Δ161-174 18,790 g mol⁻¹ (±3.8%), all corresponding to predominantly a dimeric species in solution (calculated molecular mass of the dimer ∼24 kD). L171A was not measured with multiangle laser light scattering; however, its elution profile was the same as the other point mutants. The wild type and SEP3^ΔV90 eluted as a mixture of tetramer and dimer as shown in the chromatograms. All chromatograms were overlaid and the maximum absorbance at 280 nm normalized to 1.

Figure 6.

Comparison of MEF2A, SRF, and SEP3. (A) Partial sequence alignment of SEP3 from Arabidopsis, MEF2A from H. sapiens, and SRF from H. sapiens. The M domains span residues 1 to 58 of SEP3, residues 1 to 59 of MEF2A, and residues 141 to 197 of SRF. The SEP3 I domain (59 to 90), MEF domain (residues 60 to 89), and SRF SAM domain (198 to 227) were included in the structure-based sequence alignment. Helices are depicted as red cylinders, random coils as blue lines, and beta sheets as green arrows with the MEF2A secondary structure elements above and the SRF secondary structure elements below. (B) Alternate composite models of SEP3 MIK domains using the structure of MEF2A residues 1 to 89 (PDB 3KOV), left, and the structure of SRF structure residues 141 to 227 (PDB 1SRS), right. The DNA binding site is located at the distal extremes of the tetramer based on the SEP3^75-178 (I and K domains) structure determined here. The SEP3 structure is displayed as a surface colored by monomer and the MEF2A (left) or SRF (right) structure as a cartoon with protein in light and dark gray and DNA in orange and blue. The model with MEF2A requires an opening of helix 1 of the SEP3 K domain to accommodate the DNA.

Figure 7.

Atomic Force Micrographs of SEP3 in Complex with SOC1 Promoter DNA. Heights are color coded in nanometers at right of each image. (A) Full-length SEP3 in complex with a 1-kb DNA comprising two putative SEP3 binding sites. Arrows indicate bound SEP3 proteins. The bar at left is 200 nm and at right is 100 nm. DNA-protein complexes were formed at 2 to 5 nM protein and DNA and diluted to ∼1 nM for imaging. (B) Complex of SEP3 and DNA as per (A) with the complex formed at 10 to 15 nM protein and 5 nM DNA before dilution to 1 nM DNA concentration for imaging. Arrows indicate looping of DNA. Inset highlights the SEP3-DNA complex. The bar is 200 nm (left). Intermolecular interactions of SEP3 and DNA under the same conditions were observed (right). The bar is 100 nm. (C) SEP3^1-110 lacking the K domain in complex with DNA. The complex was formed at 5 nM protein and DNA and diluted to 1 nM for imaging. Arrows show protein bound to DNA. No inter- or intramolecular looping of DNA was observed. Image at right is a close-up view. Bars are 400 nm (left) and 200 nm (right). (D) SEP3^1-110 in complex with DNA as per (C) with a protein concentration of 25 to 5 nM DNA before dilution to 1 nM DNA concentration for imaging. Image at right is a close-up view with image masking to remove tailing. Bars are 200 nm (left) and 100 nm (right). Proteins bound to DNA are indicated by arrows.