In silico structure-function analysis of pathological variation in the HSD11B2 gene sequence

Abstract

11beta-Hydroxysteroid dehydrogenase type 2 (11betaHSD2) is a short-chain dehydrogenase/reductase (SDR) responsible for inactivating cortisol and preventing its binding to the mineralocorticoid receptor (MR). Nonfunctional mutations in HSD11B2, the gene encoding 11betaHSD2, cause the hypertensive syndrome of apparent mineralocorticoid excess (AME). Like other such Mendelian disorders, AME is rare but has nevertheless helped to illuminate principles fundamental to the regulation of blood pressure. Furthermore, polymorphisms in HSD11B2 have been associated with salt sensitivity, a major risk factor for cardiovascular mortality. It is therefore highly likely that sequence variation in HSD11B2, having subtle functional ramifications, will affect blood pressure in the wider population. In this study, a three-dimensional homology model of 11betaHSD2 was created and used to hypothesize the functional consequences in terms of protein structure of published mutations in HSD11B2. This approach underscored the strong genotype-phenotype correlation of AME: severe forms of the disease, associated with little in vivo enzyme activity, arise from mutations occurring in invariant alignment positions. These were predicted to exert gross structural changes in the protein. In contrast, those mutations causing a mild clinical phenotype were in less conserved regions of the protein that were predicted to be relatively more tolerant to substitution. Finally, a number of pathogenic mutations are shown to be associated with regions predicted to participate in dimer formation, and in protein stabilization, which may therefore suggest molecular mechanisms of disease.

Fig. 1.

Outline of 11β-hydroxysteroid dehydrogenase type 2 (11βHSD2) mechanism. 11βHSD2 prevents activation of the mineralocorticoid receptor (MR) by cortisol, allowing aldosterone to bind with consequent downstream effects on transcription. ENaC, epithelial sodium channel. Figure credit: Louise Evans.

Fig. 2.

Multiple protein sequence alignment of 11βHSD2 enzymes and their immediate neighbors, the 17βHSD2s, adapted from the alignment for the Ensembl protein family ENSFM00500000270244. Figure prepared with Jalview (64) with ClustalX (60) coloring option.

Fig. 3.

Pairwise sequence alignment of human 11βHSD2 with the template used for modeling, 17βHSD1 [PDB ID: 1JTV (25), chain A], showing only those template residues solved in the structure. Secondary structure is shown below the alignment, derived from the sequence via the secondary structure prediction server JPred (15) and from the template structure via Stride (23). JPred confidence scores for assignment of residues to secondary structure element are shown for the target sequence. Red, α-helices; green, β-strands. Figure created with Jalview (64), with ClustalX coloring. Numbers above sequences refer to alignment position, not sequence position.

Fig. 4.

Dimeric 3-dimensional model of human 11βHSD2 based on human 17βHSD1. One monomer is shown in dark gray and the other in white, and both are depicted in cartoon representation. NAD is shown in stick form with carbons in white, while cortisol is shown in gray stick form. Other atoms are in standard color scheme (red, oxygen; blue, nitrogen; orange, phosphorus). Red cartoon regions highlight portions of the model with 1 or more identified problems in modeling.

Fig. 5.

The model 11βHSD2 (right) compared with the template 17βHSD1 (left) with location of mutations mapped. One chain is split by color to separate major groups of secondary structure elements. The main cofactor-binding domain is shown in red at the NH₂ terminus, 2 long helices that support much of the active site are shown in green, and the COOH-terminal region is shown in blue. Note: a short segment (amino acids 191–197) was not solved in the template [PDB ID: 1JTV (25)], seen at top left. Spheres indicate the α-carbon positions of pathogenic mutations or their equivalents in the template—only mapped on 1 chain for clarity. White spheres indicate potential problems at these positions during modeling (see text).

Fig. 6.

Residues important for function or associated with apparent mineralocorticoid excess (AME) on mutation around the substrate and cofactor binding regions. The side chains of residues involved in pathogenic mutation are shown in yellow stick form. Dashed yellow lines indicate predicted hydrogen bonds. The location of this region with respect to the overall model dimer is shown in the inset. Protein chains are colored as in Fig. 5 and ligands as in Fig. 4.

Fig. 7.

Residues involved in AME-associated mutations that may affect protein stability are shown labeled. The dimer is rotated 180° relative to Fig. 5. Color scheme and formatting as in previous figures.

Fig. 8.

Residues involved in AME-associated mutations that occur at the dimer interface are indicated. The dimer is rotated 180° relative to Fig. 5. Color scheme as in previous figures.

Fig. 9.

Residues involved in AME-associated mutations that were not classified are shown. As indicated in the inset, the dimer is shown in the same orientation as Fig. 5.