The cataract-associated V41M mutant of human γS-crystallin shows specific structural changes that directly enhance local surface hydrophobicity

Abstract

The major crystallins expressed in the human lens are γS-, γC- and γD-crystallins. Several mutations in γS-crystallin are associated with hereditary cataracts, one of which involves the substitution of a highly conserved Valine at position 41 to Methionine. According to a recent report, the mutant protein, V41M, shows lower stability and increased surface hydrophobicity compared to the wild-type, and a propensity for self-aggregation. Here we address the structural differences between the two proteins, with residue-level specificity using NMR spectroscopy. Based on the structural model of the mutant protein, our results clearly show that the mutation creates a major local perturbation almost at the junction of the first and second "Greek-key" motifs in the N-terminal domain. A larger section of the second motif (residues 44-86) appears to be mainly affected. Based on the sizeable chemical shift of the imino proton of the indole side-chain of Trp46 in V41M, we suggest that the sulphur atom of Met41 is involved in an S-π interaction with Trp46. This interaction would bring the last β-strand of the first "Greek-key" motif closer to the first β-strand of the second motif. This appears to lead to a domino effect, towards both the N- and C-terminal ends, even as it decays off substantially beyond the domain interface. During this process discreet hydrophobic surface patches are created, as revealed by ANS-binding. Such changes would not affect the secondary structure or cause a major change in the tertiary structure, but can lead to self-aggregation or aberrant binding interactions of the mutant protein in vivo, and lead to lens opacity or cataract.

Keywords: 4,4′-Dianilino-1,1′-binaphthyl-5,5′disulfonic acid; 8-Anilinonaphthalene-1-sulfonic acid; ANS; CSP; Cataract; Chemical shift perturbation; Crystallin; HGS; HSQC; Human γS crystallin; Hydrophobicity; NMR spectroscopy; S–pi interactions; V41M; bis-ANS; heteronuclear single quantum coherence; the Val41Met mutant of human γS crystallin.

Figure 1

2D ¹⁵N-¹H HSQC spectrum of [U-,¹⁵N] HGS (black contours), overlaid with [U-,¹⁵N] V41M mutant (red contours), in 100 mM PBS buffer, pH 6.0. Residues with large average chemical shift perturbation (< 0.1ppm) between the two proteins are marked by a circle and their shifts indicated by arrows. Large shifts in residues close to the mutation site, namely E42, G43, G44, T45 and W46 are clearly depicted – see also Fig. 2. The shift in the side chain of W46 (W46e in the lower-left corner) is also observed, while the remaining three tryptophan residues do not show a significant corresponding shift.

Figure 2

Averaged amide chemical shift perturbation (CSP) between the mutant V41M and the wild-type HGS. Inset shows a model of the mutant on which the CSPs are mapped as a “putty” using Pymol (Delano Scientific). The highest value of CSP is shown in red and lowest in blue, with radius of the ribbon also increasing from low CSP to high CSP. The secondary structural elements (strands and helices) are superimposed on the ribbon-putty, and are coloured according to the “Greek-key” motifs they belong to: i.e. green, orange, cyan and magenta for motifs 1 to 4 respectively. The maximum perturbation is observed in residues near the mutation site (see E42 and G44) in the first “Greek key” motif (residues 5 to 43) of the N-terminal domain. Residues 45 to 49, forming the first strand of the second “Greek-key” motif are also significantly perturbed. Residues 60 (see Inset) to 66 (green thick spaghetti) also show high perturbation, as they are adjacent to residues 42 and 44. Clearly the structure of the second motif (residues 44 to 86) of the N-terminal domain is substantially perturbed. On the other hand, motif 3 (cyan) containing residues 93–133 is totally unaffected as it is far removed from the mutation site. Residues on either end of the mutation site are affected, but the effect appears to be transmitted in a discreet pattern, probably propagated by strand-strand interaction. Residues 148 to 150 and 174–175 are part of two adjacent strands, which show minimal perturbation in motif 4. Overall, it appears that the perturbation is largely carried over from the mutation site by strand-strand interactions.

Figure 3

Amide chemical shift perturbations (CSP) in the ¹⁵N-¹H HSQC spectrum of [U-, ^15N] V41M (black contours), due to ANS binding (A); CSP in the presence of 5X molar excess of ANS (data from panel A) in [U-,^15N] V41M mutant (B). ); CSP in the presence of 5X molar excess of ANS (data not shown) in [U-,^15N] HGS (C). A horizontal line corresponding to the CSP of 0.005 has been drawn in panels B and C as an arbitrary cut-off for a significant CSP. Since the data in panels B and C are plotted using identical scales, it is clear that the CSPs, particularly in the N-terminal domain, are significantly higher for the mutant (panel C); specifically, residues clustered around residue 20 and residue 60, and individual residues E42, W72, G91 and C129. In panels D, E, and F CSPs of L61, E42 and I175 respectively, are shown in an expanded scale. Only one contour per residue is shown for clarity.