Supplementary Data Figure 1 - PDF File
Strategy to Identify Novel Genes Related to CIITA.
Portions of the CIITA primary amino acid sequence corresponding to the N-terminus (residues 1-336), the NBD (residues 336-886), or the C-terminus (LRR, 886-1130) were used as a query sequence for BLAST searches of genome sequence databases. As "hits" were identified, these sequences (and novel domains, i.e. CARD and Pyrin) were used as queries to identify additional related sequences.
Supplementary Data Figure 2 - PDF File
Clustal Alignment of CATERPILLER NBDs.
The NBDs of the nucleotide binding domains of CATERPILLER family members were aligned using CLUSTAL with minor manual adjustments and shaded using the BOXSHADE algorithm. Positions with greater than fifty percent identical residues are shaded in black, similar residues are shaded in grey. Note that in some instances residues that may actually be isofuctional (e.g. F, L, V, I, M; all hydrophobic) are not shaded as similar. The motifs identified (I-XII) correspond to those summarized in Fig. 2 of the main text. 1, 2, 3 Correspond to the NACHT motif numbers VI-VII. Asterisks indicate NACHT motif defining residues.
Supplementary Data Figure 3 - PDF File
Alignment of Selected LRRs
To explore the nature of the LRRs in this family The LRRs of CIITA, MATER, Nalp1, Nalp2, Nod1, and Nod2 were aligned using CLUSTAL with manual adjustments1. MATER (Maternal Antigen That Embryos Require) is a mouse NBD/LRR protein required for early cell divisions in embryogenesis2 and is a target of autoantibodies during premature autoimmune ovarian failure in mice3 (to the best of our knowledge 19.8 is the closest human relative to MATER). It should be noted that LRR sequences start in the NBD containing exon for the examined sequences (see Suppl. Data Fig. 3A). Residues residing within LRRs that fit the motif are shaded black (LXXLXLXX(N/C)XL(X)7LXXXL; where L = any hydrophobic residue, C = cysteine, N = asparagines, and X = any residue). Each LRR-containing region is marked with a solid bar. All of these sequences contain LRRs with varying degree of identity to one another within a sequence and between sequences. Likewise, some LRR completely match the consensus motif and others match less well, therefore we scored the LRR based on identity to the motif. Scoring results are diagrammed in Suppl. Data Fig. 3A.
Supplementary Data Figure 4 - PDF File
Schematic of LRR and LRR-like Regions and Structural Predictions
Scoring the LRRs for degree of motif conservation yielded a picture where given motifs are highly conserved (intact), poorly conserved but identifiable (pseudo), or unidentifiable. For unidentifiable sequences flanked by known intact or pseudo-LRR it appears that the spacing (20-30 AA) has been preserved, suggesting that while the LRR motif is not observed, the space that it occupied has been conserved. The preservation of the motif with observed divergence in non-motif residues, the presence of pseudo-LRR, and the conserved LRR-sized "space" support the hypothesis that LRRs provide some functional specificity (see Discussion).
