Figure 4

Modelling of homeodomain-DNA interactions in human LEUTX and transactivation potential of point mutations at positions 47 and 54. (A) Structure of the Q50K Engrailed homeodomain-DNA complex (PDB:2HDD, chain A), which was used as template for modelling the human LEUTX 3D structure. Alpha helices 1, 2 and 3 are coloured red, green, and blue (=recognition helix), respectively. (B) Scheme of protein-DNA interactions and relative positions in the sequences (see alignment). The left panel shows the homeodomain residues of Engrailed (red; 2HDD) and DLX5 (purple; 4RDU), while the right panel shows those of human LEUTX (yellow). Interactions with the DNA are marked with arrows - hydrogen bonds, solid; hydrophobic interactions, dashed - in the corresponding colours. Green dots represent water molecules. The nucleotides are numbered from 1 to 6 for the motif 5′-TAATCC-3′, and asterisks denote the antisense strand. (C) Amino acid sequences of human LEUTX and the templates used for the modelling. Due to missing structural data in the PDB entry 2HDD (Engrailed) at the N-terminus, the structure of human DLX5 (4RDU, chain D) was used for modelling residues 1–4 of the LEUTX homeodomain. (D–I) Structural models of wild-type and mutated human LEUTX homeodomains. Water molecules are shown as red spheres; hydrogen bonds as dotted grey lines; hydrophobic interactions as dotted grey lines highlighted in cyan; and non-favourable interactions as dashed red lines. (D) Close-up view around residue K50; lysine can form many possible interactions and assumes two conformations in the 2HDD X-ray structure. (E) Structural model with the K50A mutation; alanine is unable to interact with DNA like K50. (F) Wild type LEUTX homeodomain showing interactions of the specificity-determining residues I47, N51, A54, and R58. (G), Model of the I47T mutation (H), Model of the loss of function mutation A54V (I) Double mutation of I47T and A54V, which restores function. (J,K) Residue and codon usage comparison between primates and Laurasiatheria at position 47 (J) and 54 (K) of the LEUTX homeodomain. Dotted lines represent possible paths for transitions, which are more frequent than transversions, between two amino acids. (L-N) RNL for different promoter reporter constructs and different LEUTX protein expression vectors. “-”, “4 × 11 bp”, “4 × 36 bp”at the bottom “Promoter” line denote co-transfection of plain firefly reporter (pGL4.25), “4 × 11 bp” reporter, or “4 × 36 bp” reporter, respectively. X-axis labels (marked LEUTX) indicate the different LEUTX TFs used: “-”: no co-transfection of pFastBac-based TF over-expression vector; “Wild type”: co-transfection of wild-type human LEUTX; I47T, A54V, and I47T + A54V: cotransfection of mutant LEUTX proteins. Asterisks indicate statistically significant (p < 0.05, t-test) differences of RNL compared to the control without TF (“-”, labelled also by “ref.”) for each promoter construct. Batch 1 to 3 represent the three experimental replicates. (M) RNL fold change derived from “with” and “without” LEUTX protein expression constructs. Fold change values were calculated from (L), where either 4 × 11 bp or 4 × 36 bp was co-transfected. (N) RNL fold change derived from “with” and “without” promoter reporter constructs. Fold change values were calculated from (L), where TF over-expression vector was co-transfected.