Manipulating polydispersity of lens β-crystallins using divalent cations demonstrates evidence of calcium regulation

Abstract

Crystallins comprise the protein-rich tissue of the eye lens. Of the three most common vertebrate subtypes, β-crystallins exhibit the widest degree of polydispersity due to their complex multimerization properties in situ. While polydispersity enables precise packing densities across the concentration gradient of the lens for vision, it is unclear why there is such a high degree of structural complexity within the β-crystallin subtype and what the role of this feature is in the lens. To investigate this, we first characterized β-crystallin polydispersity and then established a method to dynamically disrupt it in a process that is dependent on isoform composition and the presence of divalent cationic salts (CaCl₂ or MgCl₂). We used size-exclusion chromatography together with dynamic light scattering and mass spectrometry to show how high concentrations of divalent cations dissociate β-crystallin oligomers, reduce polydispersity, and shift the overall protein surface charge-properties that can be reversed when salts are removed. While the direct, physiological relevance of these divalent cations in the lens is still under investigation, our results support that specific isoforms of β-crystallin modulate polydispersity through multiple chemical equilibria and that this native state is disrupted by cation binding. This dynamic process may be essential to facilitating the molecular packing and optical function of the lens.

Keywords: assembly; divalent cations; lens; polydispersity; β-crystallin.

Fig. 1.

Correlations between tissue location, isoform composition, and β-crystallin quaternary structure. (A) Diagram of how each lens segment was prepared. Lens segments were lysed separately and purified in PBS, pH 7.3. (B) Representative chromatograms obtained from the nuclear, inner cortical, and outer cortical segments of the lens. In each case, 3 mg of protein was loaded. Elution volumes corresponding to each crystallin population are denoted on Top. (C) Comparison of changes in crystallins based on mass (%) across the lens showed a gradient of β-crystallins that was most concentrated in the lens cortex. Data points were averaged from biological replicates (N = 3) with error bars representing one SD. (D) Peptides corresponding to bovine crystallins were positively detected (≥ 95% confidence) and confirmed the presence of 15 crystallins in both β_H and β_L2. Above is a scale of the ratios of normalized PSM matching β-crystallin isoforms detected in both β_H and β_L2. From this, βB1 was most prevalent in larger oligomers (β_H) while βB2 was most prevalent in smaller β-oligomers (β_L2).

Fig. 2.

MgCl₂ alters oligomer distribution of β-crystallins. (A) Purifications of whole lens lysates eluted with 10 mM Tris, pH 7.3, with either 0.01, 0.1, or 1 M MgCl₂. Three milligrams of protein were loaded for each condition; absorbance has been normalized for more clear comparison. Retention times are similar for α- and β-crystallins across the three conditions. (B) Significant changes in β_H and β_L2, as well as a change in α-crystallin, were observed as the larger oligomers dissociated in increasing levels of divalent cations. Data were averaged for biological replicates (N = 3) with error bars representing one SD.

Fig. 3.

Oligomer dissociation through adsorption of divalent cations. Isoelectric points were determined by measurement of the zeta-potential as the sample was titrated with 0.1 M HCl. Crystallins were either purified in PBS or 10 mM Tris, pH 7.3 with 0.01 or 0.1 M MgCl₂. PBS-purified samples were diluted with phosphate-citrate-carbonate saline buffer. Each data point was averaged from five measurements with error bars representing one SD (one biological replicate, N = 1). All lines were fitted using a sigmoidal function (adjusted R² values between 0.9052 and 0.9964).

Fig. 4.

Reassociation of β-oligomers shows assembly pathway from β_L1 or β_L2 to β_H. In A–D, gray dashed lines represent the initial purification to recover select fractions. Solid black lines represent the experimental condition during reelution of selected subtypes or mixtures. (A) Whole-lens lysate was purified in PBS (gray dashed line), and the β_H fractions (gray highlighted) were collected. Then, β_H was reeluted with buffer containing 1 M MgCl₂ to dissociate β_H (solid black line). (B) Whole-lens lysate was purified in buffer containing 1 M MgCl₂ (gray dashed line) and both the β_L1 and β_L2 fractions (gray highlighted) were collected, following reelution with PBS + 5 mM EDTA (solid black line). (C) Same as B, except β_L1 (gray highlighted) was collected in 1 M MgCl₂ and reeluted into PBS (solid black line). (D) α-crystallin, purified in PBS, and β_L2, purified in 1 M MgCl₂, were mixed in a 1:1 mass ratio (total of 0.25 mg), and reeluted in PBS, pH 7.3. (E) AUC values for each of the previous chromatograms in B–D. Values are averages of biological replicates (N = 3) with error bars representing one SD.

Fig. 5.

The proposed multistep chemical equilibrium for higher order assembly of β-crystallin based on experimental evidence from dissociation and selective reassociation. Individual isoforms associate as dimers (β_L2) or trimers (β_L1), with possible low levels of interconversion between the two (K₅). Equilibrium will preferentially drive assembly further into β_H oligomers, forming hexamers, octamers, and potentially nonamers. All reactions are expected to be product favored. This schematic does not highlight the possibility of a newly synthesized isoform and an existing dimer (β_L2) to combine and form a trimer (β_L1). This combination is not supported nor refuted by our experimental data. We would expect that path to be a competing equilibrium with K₂.

Fig. 6.

Examination of hydrophobic interfaces in lens β-crystallins through the use of SYPRO Orange fluorescence. Tracking changes in exposed hydrophobic surface area in β_H and β_L2 through (A) 10 mM Tris, (B) 10 mM Tris + 0.1 M MgCl₂, and (C) 10 mM Tris + 1 M MgCl₂. Graphs represent the average fluorescent profile of six individual measurements from two biological replicates. Since β_H is completely dissociated in 1 M MgCl₂, only fluorescence of β_L2 was measured at this condition. (D) Comparison of all conditions, including PBS, for both β_H and β_L2. Bar graphs represent two batches of three lenses (N = 2) and the data are averaged with error bars representing one SD.