Sequence alignment of BTB domains of ZBTB transcription factors identifies a unique central region in PATZ1 that is conserved in mammals. (a) Sequence alignment of PATZ1 BTB domains from selected vertebrate species. The unique central sequence of the A2/B3 loop that is conserved in mammals and is missing in fish and amphibians is indicated in bold. (b) Sequence alignment of selected human ZBTB proteins and their predicted secondary structure. The sequences of the human PATZ1 and cKrox BTB domains with their unique extra region between the A2 helix and B3 strand are shown above. The PATZ1 amino acids that correspond to this region without electron-density assignments from the crystal structure are shown in bold. Arrows and rods identify predicted conserved β-strand and α-helical regions. The eight BTB domains with solved structures are annotated on the left with their common names in addition to the ZBTB nomenclature. Shading, asterisks, colons and periods identify conserved residues according to the Clustal format. A consensus sequence is shown at the bottom, with the three predicted degron residues involved in BTB domain stability indicated by arrows.

Structure of the PATZ1 BTB domain. (a) Crystal structure of the BTB domain of the mouse PATZ1 protein (PDB entry 6guv) in cartoon representation (front view). The crystallographic asymmetric unit contains a PATZ1 BTB domain monomer (blue) with a second monomer (green) created by crystallographic symmetry. Secondary structures are indicated in capital and Greek letters. (b) Top view of the mouse PATZ1 BTB domain structure. N- and C-­termini are indicated on one monomer. The coordinates of 31 residues in a central region, unique to mammalian PATZ1 BTB domains, could not be assigned (indicated by a dotted loop). (c) Crystal structure of the BTB domain of the zebrafish PATZ1 protein (PDB entry 6guw) with individual monomers coloured yellow and red (front view). (d) Top view of the zebrafish PATZ1 BTB structure. The coordinates of seven residues in the zebrafish PATZ1 BTB domain could not be assigned (indicated by a dotted loop). (e) Superimposition of the mouse (blue and green) and zebrafish (yellow and red) PATZ1 BTB domains (r.m.s.d. of 0.62 Å). (f) A space-filling representation of the mouse PATZ1 BTB domain structure. The predicted structure (in purple) of the central region was generated by homology modelling followed by conformation equilibration using MD simulations. Note that the modelled structure contains a predicted short β-strand. The three conserved degron residues, annotated at the bottom of Fig. 1 ▸(b) and predicted to play a role in BTB dimer degradation, are highlighted in red. Numbering refers to the residues in the crystal structure.

Identification of residues involved in homodimer interaction interfaces. Contact maps of inter-chain residue interactions of the mouse (a) and zebrafish (b) PATZ1 BTB domain crystal structures. Hydrophobic interactions (green), hydrogen bonds (blue) and ionic bonds (red) are indicated. The relevant elements of the secondary structures are shown for orientation. A highly charged dimerization interface mediates PATZ1 BTB homodimerization. (c) A split homodimer view in surface representation and completed with the modelled loop highlights the residues involved in the interaction interface; positively (blue) and negatively (red) charged residues are annotated and neutral residues are shown in purple. Inter-chain salt bridges that persist above the threshold are indicated by straight lines and those that do not persist by dotted lines. (d) The zebrafish PATZ1 dimerization interface is also shown as a split homodimer view for comparison. Numbering refers to the crystal structure files.

Comparison of the lateral grooves of PATZ1 and BCL6 BTB domains. The structures of the BCL6 (PDB entry 1r2b) (a) and the energy-minimized modelled mouse (b) and zebrafish (c) PATZ1 BTB domains are shown in surface representation viewed from the front and bottom. The surface area of the residues in the lateral groove of the BCL6 BTB domain is buried upon formation of the BCL6–SMRT complex (Ahmad et al., 2003 ▸). The SMRT peptide binding to the BCL6 lateral groove is shown in cartoon representation in orange. All residues in the lateral groove are labelled for one monomer. Colours indicate residue type: positively charged, blue; negatively charged, red. All other residues in this region are coloured purple. The positions of the mutations that affect the binding to the co-repressor peptides in BCL6 and in PATZ1 are indicated in black (references are given in the text). A sequence alignment of BCL6 and PATZ1 BTB residues located in the lateral groove region is shown in (d). Residues are numbered according to the BCL6 and mouse PATZ1 structure files, with the charged residues coloured as in (a)–(c). Apart from positions 29 and 40, where the two alternatives are indicated, mouse and zebrafish PATZ1 contain the same residues in these structurally corresponding positions. Although the residue conservation for BCL6 and PATZ1 in this region is low, SMRT/NCOR peptides are predicted to bind the BTB domain of PATZ1 in the same region.