Fig 1.
Structural comparison of Ankyrin proteins.
Proteins from the non redundant dataset of Ankyrin proteins were structurally pairwise-aligned using TopMatch. The sequence identity (relI) and structural similarity (relS) parameters are compared. The distributions for each variable are shown in the upper and right side respectively.
Fig 2.
(A) Tileability values for the ANKs in the dataset. Higher values correspond to more regular structures. Lower values correspond to proteins containing structural perturbations that break the propagation of symmetry at higher length-scales by modifying the spatial arrangement of the repeating units. (B) Per residue tileability for Notch1 (PdbID: 2he0,A), the most symmetric protein in the dataset. (C) Per residue tileability for SWI6 (PdbID: 1sw6,A), one of the least symmetric proteins, due to a high occurrence of structural modifications in the repeats and insertions between them. The x-axis marks the residues in the structure, the y-axis represents the length of the fragments being considered in each case to tesselate the structures. The gray-scale represents the fraction of times in which a residue is part of a fragment that tessellates the protein.
Fig 3.
Defining the ANK repeat phase.
(A) Frequency of structural repetitions as characterized by the δ-functions for ANK proteins. Most of the ANKs have their highest peak at a tile length of 33 residues. (B) Distribution of typical phases for the highest scoring fragments according to their tiling scores. The phase is shown relative to the ANK HMM in Pfam. These phases are defined by applying the tiling algorithm over the set of ANK structures, thus it is related to the geometry of the molecules. (C) Distribution of phases for the highest scoring fragments according to their relative foldability values. The phase is shown relative to the ANK HMM in Pfam. These phases are defined by calculating the foldability of the different tiles of the structure, thus it is related to the energetics of the different fragments.
Fig 4.
Overview of ANK domains structure.
(A) Distribution of lengths for the repeats detected in structures. (B) Distribution of lengths for the N-terminal, internal and C-terminal repeats. (C) Distribution of deletions along the canonical positions in ANKs relative to the phase of the structural based HMM and their length. The grayscale for the dots is indicative of many points being superimposed. (D) Distribution of insertions along the canonical positions in the ANK repeat relative to the phase of the structural based HMM and their length.
Fig 5.
Conservation of local frustration in the Ankyrin Repeat Domains.
A) Sequence logo corresponding to the hidden Markov models derived from indirect structural alignments for N-terminal repeats, Internal repeats and C-terminal repeats. B) Information Content for the conservation of frustration states according to the single residue level frustration index for N-terminal repeats, Internal repeats and C-terminal repeats. The total height of the stack represents the overall information content value of it. Contributions to the overall value of each frustration state at each position is represented by the proportion of each color to the stack.
Fig 6.
Local frustration and sequence conservation.
A) Relation between information content per characteristic position in the ankyrin repeat derived from the conservation of the sequence and the conservation of the frustration state. There exists a positive correlation between the two variables with significant p-values (<0.05) for the Internal and C-terminal repeats. The correlation is not significant in the case of N-terminal repeats. B) Configurational frustration index plotted over representative structures of the ANKs family (the consensus protein 4ANK, PdbID: 1n0r,A; Ankyrin1, PdbID: 1n11,A; LegA1 protein from Legionella pneumophila, PdbID: 2aja,A and the P19INK4D CDK4/6 inhibitor, PdbID: 1bd8,A). The green lines correspond to interactions that are minimally frustrated, red lines correspond to highly frustrated interactions.
Fig 7.
Conservation of local frustration at the contact-level.
A) Contact map for a consecutive pair of ANK repeats. In the upper matrix the most informative state from the frustration IC calculations is shown. Red represents that the most informative state is the highly frustrated, gray is for the neutral state and green when the minimally frustrated state is the most informative one. Those values are calculated over the absolute counts that occur for each canonical interaction in the dataset. In the lower matrix the relative frequency of having a contact between two canonical positions in the repeats pair is shown in grey-scale where darker points mean that the interaction is present in a bigger fraction of the repeats being analyzed. B) Conservation of interactions in and between repeats: Contacts within the ankyrin repeat with a IC value higher than the highest interaction where the neutral state is the most informative. Residues involved in these conserved interactions are marked: Most conserved amino acid(s): Yellow: Position 8; G. Blue: Positions 11–14;TPLH. Orange: Positions 15–16; AA. Red: Positions 23–24; IV. C) Conservation of most informative contacts at the interface between repeats.
Fig 8.
Quaternary assemblies in ANKs.
Quaternary assemblies that involve ANKs are shown with the ANK molecule in orange and the rest of the complex in yellow. Residues from the ANK structure that are involved at the interface are coloured in magenta. A) Myosin phosphatase targeting subunit 1 (MYPT1) in complex with the ser/thr phosphatase-1 (delta) (PdbID: 1s70). B) IκBβ/NF-κB p65 homodimer complex (PdbID: 1k3z) C) YAR1 when binding the 40S ribosomal protein S3 (PdbID: 4bsz) D) Tankyrase-1, (PdbID: 3utm). E) Change in the configurational frustration index between the unbound and bound states for the ANK proteins. In the x-axis the configurational frustration for the monomeric unbound state is shown. In the y-axis the difference between the complexed state and the monomeric state is shown. A positive change means that the frustration is decreased upon interaction.