Sequence alignment and mutation mapping of CIAS1 NBD. (A) Alignment of the CIAS1 sequence with 4 NBD sequences of known 3D structures corresponding to proteins of the AAA+ ATPases superfamily. p97: d1 AAA domain of membrane fusion ATPase p97 (PDB identifier, 1e32^[chain A])³⁶  nsf: D2 hexamerization domain of N-ethylmaleimide-sensitive factor (PDB 1d2n [chain A])^34,35 ; Cdc6p (PDB 1fnn [chain A])³³ ; and RuvB (PDB 1hqc [chain A]).40 Alignment was generated by threading and PSI-BLAST procedures and was carefully refined and extended using the sensitive hydrophobic cluster analysis (HCA) (“Patients and methods”). Vertical arrows indicate the positions of the CIAS1 mutations. Amino acids of nsf involved in intersubunit contacts³⁵  are boxed in red. Identical amino acids are shaded in black, whereas similar amino acids are shaded in gray with white and black letters for hydrophobic (or amino acids that can substitute for them) and nonhydrophobic amino acids, respectively. Green circles indicate the positions most frequently occupied by hydrophobic amino acids in the NACHT family of domains. These residues mainly correspond to amino acids that, in the aligned NBDs, are buried within the considered structures and that can serve as anchors for the alignment procedure. Moreover, positions of the observed regular secondary structures (underlined and labeled under the sequence (S indicates β-strand; H, helix) in most cases match those of the predicted secondary structures of CIAS1 with respect to its sequence (E indicates extended β-strand; H, helix; C, coil). No accurate structural alignment could be obtained for helix H5 (indicated in brackets). However, the N-termini of the corresponding sequences could be aligned, highlighting the conservation of 2 hydrophobic amino acids. The main original features of the CIAS1 NBD fold with respect to the NBD core structures shown in (A) are (1) the presence of another helix, H3C, after helix H3B (nsf labeling; in this respect, the predicted structure of CIAS1 is thought to be similar to that of Cdc6p,³³  in which a longer helix H3C is also present between helix H3B and strand S3 [yellow]) and (2) a large loop linking strand S3 to helix H4. This loop is 7 amino acids longer than the corresponding loop in nsf,^34,35  but it also contains an isoleucine-glycine-proline (IGP) sequence, which in the nsf structure forms a tight turn involved in hexamer interactions.^34,35  (B) Mapping of mutations on the 3D model of CIAS1 NBD. Two orthogonal views are shown in ribbon representation. The model was constructed based of the alignment shown in panel A. Strands and helices are labeled and colored as in panel A. Positions of mutations are shown and labeled, as are positions of the Walker A T231 (blue on helix H2) and Walker B D300 (orange in strand S3) motifs and as are positions of ATP and of the magnesium ion, as in the nsf structure.³⁵  Note that the conformation of the loop linking strand S3 to helix H4B is hypothetical. Cα-trace of the α-helical domain following NBD in AAA+ ATPases (nsf D2³⁵ ) is shown on right to illustrate its position with respect to the NDB. According to HCA (data not shown), T405 can be tentatively located in the C-terminal end of an extended structure, after 2 α-helices, lying near the ATP-binding site (gray circle).