Eye lens β-crystallins are predicted by native ion mobility-mass spectrometry and computations to form compact higher-ordered heterooligomers

Abstract

Eye lens α- and β-/γ-crystallin proteins are not replaced after fiber cell denucleation and maintain lens transparency and refractive properties. The exceptionally high (∼400-500 mg/mL) concentration of crystallins in mature lens tissue and multiple other factors impede precise characterization of β-crystallin interactions, oligomer composition, size, and topology. Native ion mobility-mass spectrometry is used here to probe β-crystallin association and provide insight into homo- and heterooligomerization kinetics for these proteins. These experiments include separation and characterization of higher-order β-crystallin oligomers and illustrate the unique advantages of native IM-MS. Recombinantly expressed βB1, βB2, and βA3 isoforms are found to have different homodimerization propensities, and only βA3 forms larger homooligomers. Heterodimerization of βB2 with βA3 occurs ∼3 times as fast as that of βB1 with βA3, and βB1 and βB2 heterodimerize less readily. Ion mobility experiments, molecular dynamics simulations, and PISA analysis together reveal that observed oligomers are consistent with predominantly compact, ring-like topologies.

Keywords: crystallin; dimerization kinetics; eye lens; native ion mobility-mass spectrometry.

Figure 1.

Native mass spectra acquired for a dilution series of individual β-crystallins and relative abundances of homooligomers detected (M = monomer, D = dimer, T = tetramer, H = hexamer). The native mass spectrum of each protein as labeled at the top of the figure is shown for the highest protein concentration used (11.5 μM, a, b, and c) and a lower protein concentration (1.15 μM or 144 nM as labeled in figure, d, e, and f) selected to illustrate disappearance of peaks corresponding to dimers. The detected abundance of each homooligomeric species is plotted in the bottom row (g, h, and i), with abundances for each protein at each concentration relative to that of the monomer.

Figure 2.

Native mass spectra of equimolar mixtures of β-crystallin βA3 with either βB1 (a, c, e) or βB2 (b, d, f) acquired to investigate the formation kinetics of heterodimers. The native mass spectra shown in the top row were acquired for a small aliquot removed immediately upon mixing the sample (t = 0 min), while the middle row shows the native mass spectrum for the same samples after 3 hours (t = 180 min; see also mass spectra ranging from m/z 3000-6500 illustrating the presence of higher-order heterooligomers in Figures S2 and S3). Peaks corresponding to homo- and heterodimers are labeled according to the legend. Plots e and f illustrate differences in the formation of βB1/βA3 and βB2/βA3 heterodimers. The proportion of the total detected dimer abundance corresponding to the heterodimer is plotted as a function of time. Approximation by pseudo-first-order kinetics produces exponential fits with inverse rate constants (tau) labeled on each plot.

Figure 3.

Example illustrations of the various topologies of model structures for comparison with experimental ion mobility data, divided into categories. X-ray crystal structure of β-crystallin βB2 monomer (PDB ID: 1YTQ) is shown in surface representation in orange in the top right. Examples of model structure categories with various subunit arrangements are shown for tetramers. More detailed mesh illustrations of example structures can be found in the Supplemental Information (Figure S5).

Figure 4.

Comparison between experimental collision cross section (CCS) for each detected oligomeric species of β-crystallin proteins and those computed for simulated model structures of various topologies (see Figure 4 for cartoon representations). Percent differences were calculated as (computed CCS – experimental CCS) / experimental CCS. The established range of accuracy and precision of the simulation protocol (0 ± 4%) is represented by a green bar. Colors and shapes of symbols correspond to model structure categories and protein species as shown in the legend. For each homo- or heterooligomer, the unfilled symbol corresponds to the model structure CCS closest to experiment.

Figure 5.

Example ribbon structures identified by PISA analysis for βB2 tetramer (a, linear; b, ring; c, crossed). See Supplemental Information (Figure S5) for example ribbon structures of other structure types and oligomer sizes.