Nuclear receptors, nuclear-receptor factors, and nuclear-receptor-like orphans form a large paralog cluster in Homo sapiens.

We studied a human protein paralog cluster formed by 38 nonredundant sequences taken from the Swiss-Prot database and its supplement, TrEMBL. These sequences include nuclear receptors, nuclear-receptor factors and nuclear-receptor-like orphans. Working separately with both the central cysteine-rich DNA-binding domain and the carboxy-terminal ligand-binding domain, we performed multialignment analyses that included drawings of paralog trees. Our results show that the cluster is highly multibranched, with considerable differences in the amino acid sequence in the ligand-binding domain (LBD), and 17 proximal subbranches which are identifiable and fully coincident when independent trees from both domains are compared. We identified the six recently proposed subfamilies as groups of neighboring clusters in the LBD paralog tree. We found similarities of 80%-100% for the N-terminal transactivation domain among mammalian ortholog receptors, as well as some paralog resemblances within diverse subbranches. Our studies suggest that during the evolutionary process, the three domains were assembled in a modular fashion with a nonshuffled modular fusion of the LBD. We used the EMBL server PredictProtein to make secondary-structure predictions for all 38 LBD subsequences. Amino acid residues in the multialigned homologous domains--taking the beginning of helix H3 of the human retinoic acid receptor-gamma as the initial point of reference--were substituted with H or E, which identify residues predicted to be helical or extended, respectively. The result was a secondary structure multialignment with the surprising feature that the prediction follows a canonical pattern of alignable alpha-helices with some short extended elements in between, despite the fact that a number of subsequences resemble each other by less than 25% in terms of the similarity index. We also identified the presence of a binary patterning in all of the predicted helices that were conserved throughout the 38-sequence sample. Our results fit well with a recently proposed evolutionary model that combines protein secondary structure and amino acid replacement. We propose a new hypothesis for molecular evolution, in which chaperones--acting as an endogenous cellular device for selection--play a crucial role in preserving protein secondary structure.