PMC

Site-specific K_A/K_S analysis for V2Rs. (A) Ribbon diagram showing ω⁺ sites. (B) Representation of the extracellular domain of metabotropic glutamate receptors with mapped ω⁺ sites. Predicted ω⁺ sites are mapped as described in and colored blue. Residues are mapped to the structure of rat metabotropic glutamate receptor subtype 1 complexed with glutamate here colored purple (PDB 1EWK, ). ω⁺ sites in the carboxy-terminal domain were mapped to the secondary structures of bovine rhodopsin 1EWK (for full details see Suppl. Table 2).

Site-specific K_A/K_S analysis of MHC-M10 genes. Codons that were predicted to be under positive selection with a posterior probability >0.90 by one codeml model, and >0.5 by at least one other model, are mapped to the crystal structure of mouse MHC class I H-2DD, chain: A in complex with the HIV-1 derived peptide p18–110, shown in green (PDB identifier 1BII (). ω⁺ sites, shown in blue, are predominantly located along the α-helices of the peptide recognition region. No ω⁺ sites were mapped to the immunoglobulin domain of the protein, shown here in orange.

(A) Site-specific K_A/K_S analysis of OBPs. ω⁺ codons with a posterior probability >0.90 by one codeml model, and >0.5 by at least one other model, were mapped to the crystal structure of cow OBP (PDB: 1OBP; ) and colored blue. Three ω⁺ site side-chains, shown in ball-and-stick format and colored red, are predicted to project into the interior of the OBP β-barrel. Two of these side-chains (1OBP:T38 and V69) are within 5 Å of the ligand that was cocrystalized in the 1OBP X-ray structure, whereas the third side-chain (1OBP:E84) is more distant from the ligand (∼8 Å), and closer to the cavity edge. (B) Positively selected sites shared by both MUP/α_2u-globulins and OBPs. A multiple alignment of MUP/α_2u-globulins and OBPs was used to identify nine ω⁺ codons shared by both families. These sites are spatially clustered and therefore represent a hypervariable region likely to be important for functional diversity. The nine ω⁺ codons were mapped to the crystal structure of mouse MUPs (PDB 1MUP) and are colored red. 2-(Sec-butyl) thiazoline is shown in purple.

Site-specific K_A/K_S analysis of olfactory receptors, candidate testis odorant receptors, and V1Rs. ω⁺ codons that were predicted to be under positive selection are mapped to a ribbon representation of the structure of bovine rhodopsin (PDB 1L9H; ; for full details see Suppl. Table 2). (A) Ribbon diagram of bovine rhodopsin chain A colored by secondary structure succession. (B) Schematic representation of secondary structures. Ribbon diagram (C) and secondary structure schematic (D) of ω⁺ sites for olfactory receptors. ω⁺ sites predicted for OR family A are highlighted in green, B in black, C in purple, D in blue, and E in red. The analysis of these ORs indicates that positive selection appears to be confined largely to the lumenal half of the molecule. Seventy-five percent of the ω⁺ sites are located in the amino terminal region, extracellular loops, and the extracellular half of the transmembrane helices. Ribbon diagram (E) and secondary structure schematic (F) of ω⁺ sites for the candidate testis-specific odorant receptors (posterior probability > 0.99, in a single model). Three ω⁺ sites identified when analyzing the rat genes alone are highlighted in green. A single ω⁺ site identified in analysis of both the mouse and rat lineages is highlighted in blue. (G) Schematic of ω⁺ sites for the V1R family (posterior probability > 0.90 in one model and > 0.5 in at least one other model) highlighted in blue. The V1R proteins could not be reliably aligned to the rhodopsin sequence, and so positions of ω⁺ sites are relative to published predictions of secondary structures (; ). Swiss-PDBviewer (; ) was used for all structural manipulations, and POVRAY () was used to generate images.

(A) Gene order of Ly-6 homologous urinary protein (LUP) genes. Syntenic regions of mouse chromosome 9 (February 2003 assembly), human chromosome 11 (November 2002 assembly), and rat chromosome 8 (June 2003 assembly) are shown, approximately to scale. Genes are indicated in blue letters, and pseudogenes in red. Strand orientation is indicated by arrowheads. Homologous rat–mouse–human relationships are represented by pink lines. A single mouse–rat homologous relationship that does not have a detectable human counterpart is represented by a green line. Numbers represent genes not homologous to LUPs: 1, BOC an immunoglobulin superfamily member; 2, DEAD box polypeptide 25; 3, FLJ32915 hypothetical protein; 4, pseudouridine synthase 3; 5, checkpoint kinase 1; 6, oligosaccharyl transferase STT3 subunit (B5); 7, P53-induced protein 8; and 8, zygin I. LUP genes and pseudogenes scored greater than 0 bits using GeneWise, default parameters [excepting a substitution error rate (subs) of 0.1], and a hidden Markov model of LUP homologs' amino acid sequences. Pseudogenes, as opposed to genes, were assigned on the basis of incomplete sequences, frameshifts and in-frame stop codons. (B) Sitespecific K_A/K_S analysis of LUP homologs. ω⁺ codons with a posterior probability >0.80 by at least two codeml models are mapped to the NMR structure of human CD59 (PDB 1CDQ; ). Residues corresponding to ω⁺ codons are highlighted in blue. Several residues that have been experimentally shown to be important for species selectivity of human CD59 (i.e., H44, N48, D49, T51, T52, R55, and E58; ) are shown in green.

(A) Sequence divergence of rat α_2u-globulin genes and pseudogenes. Genomic DNA from 22 rat α_2u-globulin genes, including 12 suspected of being pseudogenes, identified as described in Methods. Nucleotide sequences were highly conserved (mean 93%±2% identity calculated from ungapped columns), and were aligned using HMMer () with manual adjustments. The proportion difference was calculated as the fraction of nucleotides in each column that differ from the consensus nucleotide. The average proportion difference from a sliding window of 10 nt is plotted against initial nucleotide position. Alignment columns with >50% gaps are shown as diamonds and were ignored for these calculations. Exons are shown as solid horizontal lines, with intron-exon boundaries defined by dashed vertical lines. (B). Site-specific K_A/K_S analysis of the MUP/α_2u-globulin family, lipocalin homologs. ω⁺ codons with a posterior probability >0.90 by one codeml model, and >0.5 by at least one other model, are shown mapped to the crystal structure of murine MUP complexed with 2-(Sec-butyl) thiazoline (PDB 1MUP; ). Residues corresponding to ω⁺ codons are highlighted in blue; 2-(Sec-butyl) thiazoline is shown in orange. A single interior ω⁺ site corresponds to 1MUP:Y124 (shown in green). Although there are no data available to support a direct role in binding, this tyrosine residue neighbors, and has a similar orientation to, 1MUP:R126, which is required for thiazoline-binding (). Y124 is as close to 2-(Sec-butyl) thiazoline as 1MUP:F60, a proposed ligand-binding residue (). All mouse, and most rat, sequences contain a tyrosine in the position equivalent to 1MUP:Y124. However, two rat genes (mup_rn11 and mup_rn20) possess substitutions (to leucine and tryptophan, respectively) that might modify the geometry of the internal ligand-binding cavity. Consequently, these two genes might have evolved either to bind a ligand different from those of their paralogs, or else the same ligand with a different affinity. We also identified three codons (1MUP:S69, E79, and F91) that have been subject to rat lineage-specific positive selection. These are shown here in ball-and-stick format and colored red. These sites are the only alignment positions where amino acid substitutions are not conserved in rat α_2u-globulins but are conserved in mouse MUPs. One rat α_2u-globulin (mup_rn5) possesses unusual substitutions at all three positions (for S69, two substitutions at second and third codon positions [R→T]; for E79, a first position transversion [P→A]; and for F91, a second position transition G→E). mup_rn18 shares the R→T substitution at S69 with mup_rn18. This suggests that at least three sites in α_2u-globulins may be under positive selection in the rat lineage alone, possibly reflecting the functional divergence of two genes (mup_rn5 and mup_rn18) from the other rat paralogs.