Interaction of βA3-Crystallin with Deamidated Mutants of αA- and αB-Crystallins

Interaction among crystallins is required for the maintenance of lens transparency. Deamidation is one of the most common post-translational modifications in crystallins, which results in incorrect interaction and leads to aggregate formation. Various studies have established interaction among the α- and β-crystallins. Here, we investigated the effects of the deamidation of αA- and αB-crystallins on their interaction with βA3-crystallin using surface plasmon resonance (SPR) and fluorescence lifetime imaging microscopy-fluorescence resonance energy transfer (FLIM-FRET) methods. SPR analysis confirmed adherence of WT αA- and WT αB-crystallins and their deamidated mutants with βA3-crystallin. The deamidated mutants of αA–crystallin (αA N101D and αA N123D) displayed lower adherence propensity for βA3-crystallin relative to the binding affinity shown by WT αA-crystallin. Among αB-crystallin mutants, αB N78D displayed higher adherence propensity whereas αB N146D mutant showed slightly lower binding affinity for βA3-crystallin relative to that shown by WT αB-crystallin. Under the in vivo condition (FLIM-FRET), both αA-deamidated mutants (αA N101D and αA N123D) exhibited strong interaction with βA3-crystallin (32±4% and 36±4% FRET efficiencies, respectively) compared to WT αA-crystallin (18±4%). Similarly, the αB N78D and αB N146D mutants showed strong interaction (36±4% and 22±4% FRET efficiencies, respectively) with βA3-crystallin compared to 18±4% FRET efficiency of WT αB-crystallin. Further, FLIM-FRET analysis of the C-terminal domain (CTE), N-terminal domain (NTD), and core domain (CD) of αA- and αB-crystallins with βA3-crystallin suggested that interaction sites most likely reside in the αA CTE and αB NTD regions, respectively, as these domains showed the highest FRET efficiencies. Overall, results suggest that similar to WT αA- and WTαB-crystallins, the deamidated mutants showed strong interactionfor βA3-crystallin. Variable in vitro and in vivo interactions are most likely due to the mutant’s large size oligomers, reduced hydrophobicity, and altered structures. Together, the results suggest that deamidation of α-crystallin may facilitate greater interaction and the formation of large oligomers with other crystallins, and this may contribute to the cataractogenic mechanism.

During aging and cataract development, various mutations and age-related post-translational modifications (PTMs) occur in the crystallins. Examples of such PTMs include photooxidation, deamidation, disulfide bond formation, and cleavage [6,7]. The PTMs result in incorrect interactions, oligomerization, aggregation, cross-linking, and insolubilization of crystallins, which may lead to the development of lens opacity [6][7][8][9][10][11]. Misfolding, deletion, and premature termination of crystallins have been demonstrated to be associated with the human inherited autosomal, dominant, congenital zonular, or nuclear sutural cataracts [12][13][14]. Some mutations such as splice site-, point-, or nonsense mutations have also been reported in various autosomal dominant-, congenital zonular-, and nuclear sutural cataracts in human and mouse models [15,16]. PTMs such as truncations of the crystallins can lead to altered solubility, oligomerization, and supra-molecular assembly, which are believed to be causative factors for cataract development. For example, truncation of 51 residues from the C-terminal region of the CRYBB2 gene mutant (Q155) have been shown to cause cerulean cataract [17]. Studies have shown that altered crystallin structures could lead to abnormal interactions with other crystallins and to cataract development.
α-crystallin has 3 distinct domains, i.e. N-terminal domain (NTD), core domain (CD), and C-terminal extension (CTE). The N-terminal domain (NTD) and core domain (CD) of α-crystallin have been reported as substrate binding sites for chaperone activity [33][34][35][36][37][38]. The CTE of α-crystallins has been reported to be involved in the recognition and selection of unfolded protein substrates; the CTE of αB-crystallin has also been identified as a substrate binding site [39]. In addition to chaperone activity, α-crystallin has also been reported to inhibit trypsin, elastase [40], caspase-3 [41,42], and an endogenous lens proteinase [43]. The C-terminal extension has been assumed to be an inhibitor of trypsin [44].
It is now well established that PTMs of crystallins including deamidation affect their interaction and result in the loss of lens transparency [5]. Therefore, it is important to characterize the effect of deamidation on crystallin-crystallin interactions. We previously demonstrated that βA3-crystallin isolated from α-crystallins fraction exhibited protease activity after it's dissociate from α-crystallin [45]. This result implied that the α-crystallin is most likely acting as an inhibitor of the βA3-crystallin's protease activity. Furthermore, we also demonstrated the interaction of βA3-crystallin with αAand αB-crystallins and identified the interaction sites of βA3-crystallin [46]. To further extend our previous studies, here we have analyzed in vitro interaction of WT and deamidated αAand αB-crystallins with βA3-crystallin by SPR method, and also in vivo interaction by FLIM-FRET method in HeLa cells. SPR analysis is selected because it identifies even the weak interactions and also quantifies molar association and dissociation rates. Similarly, the FLIM-FRET method could identify interactions among crystallins in vivo under the physiological condition in cells. As mentioned above, NTD, CD and CTE of α-crystallins are known to be involved in substrate binding, recognition and selection of unfolded substrates therefore, it is valuable to identify the interacting regions of α-crystallins with βA3-crystallin. Using the FLIM-FRET method, we identified αAand αB-crystallins domains (NTD, CD and CTE regions) that interact with the βA3-crystallin.

Materials and Methods Materials
The mammalian expression vectors pAm Cyan1-N1 and pZS Yellow1-N1, the Hi-Fi PCR mix, and the infusion enzyme were obtained from Clontech Laboratories (Mountain View, CA) to generate fluorescently-tagged fusion proteins. Cell culture reagents were obtained from Invitrogen (Carlsbad, CA), and a CM-5 chip was purchased from GE-Biosciences (Arlington Heights, IL). HeLa cells were a kind gift from Dr. Vincenzo Guarcello (Department of Vision Sciences, University of Alabama at Birmingham, Birmingham, AL, USA), and Dr. Scott W. Blume (Department of Medical Hematology and Oncology, University of Alabama at Birmingham, Birmingham, Al, USA). Primers used in the study were synthesized by Sigma-Aldrich (St. Louis, MO). Profinity IMAC Ni-charged resins, DNA, and protein markers were from Bio-Rad (Hercules, CA), and the site-directed mutagenesis kit was from Stratagene (Agilent Technologies Inc, CA). Restriction enzymes and T4 DNA ligase were obtained from New England Biolabs Inc (Ipswich, MA) and Thermo Fisher Scientific (Atlanta, GA). Unless indicated otherwise, all other chemicals used were purchased from Thermo Fisher Scientific (Atlanta, GA).

Construction of Recombinant Proteins
Ligation method. Genes were cloned in pET 28b and pZS Yellow1-N1 vectors using a ligation method as described below. Desired genes were amplified by polymerase chain reaction (PCR) with appropriate primers in a 25 μl reaction mixture containing 2.5 μl of Taq buffer (10X), 20 pmoles of forward and reverse primers, 25 ng of template DNA, 0.2 μl of DNTPs (10 mM), and 0.5 μl of Taq DNA polymerase. The following PCR conditions were used: predenaturation at 95°C for 5 min, 30 cycles of denaturation at 95°C for 30 sec, annealing at 55-60°C for 30 sec (depending upon the Tm of the primers), and extension at 72°C for 1 min with a final extension at 72°C for 5 min. PCR products were purified by a gel-elution method using a gel extraction kit (Qiagen, Hilden, Germany) and treated with Nhe I and Xho I for 1 h at 37°C. Simultaneously, the vector was also linearized with Nhe I and Xho I, and all the reactions were inactivated. The restriction enzyme-treated DNA was gel-purified. Further, a 3:1 insert vector ratio was used for the ligation with T4 DNA ligase, and the ligation mixture was incubated at room temperature for 1 h. Next, the ligation mixture (2-5 μl) was transformed to 20 μl E. coli XL-10. All constructs were confirmed by DNA sequencing at the Genomics Core Laboratory of the University of Alabama at Birmingham.
Infusion method. All the constructs in mammalian expression vectors, i.e. pAm Cyan1-N1 and pZS Yellow1-N1, were generated by an infusion method. Vectors were linearized with Nhe I and Xho I and purified by the gel-elution method. The desired genes were PCR amplified with the appropriate primers in a 25 μl reaction mixture containing 12.5 μl of Hi Fi PCR mix, 10 pmoles of forward and reverse primers, and 5-20 ng of template DNA. Amplifications were performed under the following PCR conditions: denaturation at 98°C for 10 sec, annealing at 55°C for 15 sec, and extension at 72°C for 5 sec for 30 cycles. Amplicons were purified by the gel-elution method. Further, 10 μl reaction mixtures containing 100 ng of linearized vector, 50 ng of insert, and 2 μl of 5X Infusion HD enzyme premix were incubated at 50°C for 15 min, and placed on ice. Transformations were performed in 20 μl stellar competent cells (Clontech, Mountain View, CA) with a 2 μl infusion mixture using a standard transformation method. Recombinant bacteria were selected on a kanamycin agar plates, and plasmids were isolated from 4 random colonies. Constructs were verified by DNA sequencing at the Genomics Core Laboratory of the University of Alabama at Birmingham.
Site-directed mutagenesis. Deamidation of Asn to Asp residue in αAand αB-crystallins was performed using a site-directed mutagenesis kit (Quickchange; Stratagene, La Jolla, CA) by following the manufacturer's instructions. The pET 28b constructs and pZS Yellow1-N1 constructs of αAand αB-crystallins were used as templates with the appropriate primers. The desired mutations were confirmed by DNA sequencing at the DNA core facility as described above.
In Vitro Interaction of βA3-Crystallin with αAand αB-Crystallins and Their Deamidated Mutants Cloning, expression and purification of proteins. Recombinant His-tagged WT βA3-, WTαA-, and WTαB-crystallins were generated in a pET 28b vector using the ligation method as described above. The deamidated mutants of αAand αB-crystallins (αA N101D, αA N123D, αB N78D and αB N146D) were generated by site-directed mutagenesis using appropriate primers (Table 1). Proteins were expressed in E. coli pLys S BL-21 cells at 37°C using the IPTG method, and the expressed proteins were released by cell lysis. All the proteins were purified using the Ni 2+ -affinity column chromatographic method as described earlier [47]. Purified proteins were dialyzed against the phosphate buffer (50 mM sodium phosphates, pH 7.8, 150 mM NaCl), and concentrated using 10 kDa cut off centricon tubes (EMD-Millipore, Billerica, MA). Freshly purified proteins with >80-90% purity were confirmed by SDS-PAGE. The minor bands were identified as oligomers by western blot analysis, and were used for the SPR study.
Individual analytes αA-(WT, αA N101D, αA N123D) as well as αB-(WT, αB N78D, αB N146D) were flowed over the immobilized βA3-crystallin surface at varying concentrations (5 μM-25 μM) at a flow rate of 20 μL/min. The interactions between the various components were recorded and after each binding cycle, the chip surface was regenerated with 1 M NaCl. All experiments were carried out in duplicates at room temperature with 50 mM sodium phosphate buffer, pH 7.8 containing 150 mM NaCl as the running buffer. The concentration of the proteins was a limiting factor in these experiments, as they tended to aggregate readily even at very low concentrations. Given the limitations, these experiments were carried out and the curve fitting were analysed with BIA evaluation software 8.1 [48]. The observed variations in RU for any given concentration indicated that there could be additional non-specific adhesion events that occur during and after the initial interaction. While many different binding models were utilized to fit the curves, we have chosen to report here the simplistic 1:1 Langmuir kinetics, with the cautionary note that one cannot overly emphasize the fitted kinetic parameters due to the complex nature of the interactions that exist between βA3and WTαA-/WTαBcrystallins and their deamidated mutants. Currently there exists no curve fitting protocol to delineate and determine kinetics of specific and non-specific interactions. Therefore instead of reporting fitted affinities, we have analysed the adherence propensities. In this analysis, the average RU values after injection of the analyte between 260-270 seconds (where the peak plateaus towards saturation for each concentration), were analyzed and plotted for each concentration. The utilization of the same immobilized chip surface provided the basis for this type of analysis.

Primers
Forward Reverse were used as negative control. The CFP-tagged protein (CFP βA3-crystallin) was used as a donor, and the YFP-tagged proteins were used as acceptors in the FLIM-FRET analysis.
Tissue culture and transfection. HeLa cells were grown in a modified Eagle's medium with high glutamate (MEM Glutamax, Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS) at 37°C with 5% CO 2 . For transfection, 1X10 5 cells were seeded in 35 mm μ-dishes (35μ-dish, Ibidi, Germany) in 800 μl medium and were grown for 18 h or until they reached up to 80% confluency. For transfection, 3 μl lipofectamine 2000 (Invitrogen, Carlsbad, CA), 2 μg of donor DNA (CFP βA3-crystallin), and 2 μg of acceptor (YFP constructs) were mixed with 200 μl of Opti-MEM medium (Phenol red free DMEM) and incubated at room temperature for 25 min. After 25 min, the volume of the lipofectamine-DNA mixture was maintained up to 800 μl. Cells were washed with phosphate buffer saline (PBS), and 800 μl transfection mixtures were added to the cells and incubated at 37°C with 5% CO 2 for 5-6 h. Next, the transfection mixture was supplemented with 10% FBS and incubated for 18-20 h at 37°C with 5% CO 2 . Expressions of the CFP-tagged and YFP-tagged proteins were examined using a Zeiss Axioplan2 fluorescence microscope at the Vision Sciences Research Centre Ocular Phenotyping and Molecular Analysis core facility at the University of Alabama at Birmingham.
Western blot analysis. After 24 h of transfection, cells were lysed with 100 μl radio immunoprecipitation assay (RIPA) buffer (Fisher-Thermo Scientific) containing a cocktail of protease inhibitors (Roche). Twenty μl of lysates were then loaded onto 15% SDS-PAGE gels, and after electrophoresis, electro-blotted to a PVDF membrane by the Trans-Blot Turbo transfer system (Bio-Rad). Next, the blots were blocked with 3% bovine serum albumin prepared in PBST (Phosphate buffer saline supplemented with 1% v/v Tween 20) for 1 h, and subsequently incubated with a primary polyclonal anti-βA3-crystallin antibody (1:500 dilution, Santa Cruz Biotechnology, Dallas, TX), or monoclonal anti-αA-or anti αB-antibodies (1:1000 dilution, Abcam, Cambridge, MA) for 1 h at room temperature. Blots were washed three times with PBST and incubated in the dark with an appropriate secondary antibody (IR-dye conjugated, Li-Cor, Lincoln, NE) for 1 h at room temperature. Signals were detected by exposing the blots to 700 and 800 channels using a Li-Cor Oddessy instrument (Lincoln, NE). β-actin was used as a loading control, and the membrane was probed with a rabbit polyclonal antibody against βactin (1:1000 dilution, Cell Signaling Technology, Danvers, MA) using the protocol described above.
FLIM-FRET imaging. Live HeLa cells plated onto μ-dishes were subjected to confocal FLIM imaging using a Becker and Hick GmbH pulsed diode 405 nm with a simple Tau Time correlated single photon counting module attached to a Zeiss LSM 710 confocal microscope at the High Resolution Imaging core facility of the University of Alabama at Birmingham. Confocal imaging was performed with the Zeiss microscope to detect the localization of CFP-tagged Table 2. List of primers used for generation of recombinant crystallins in mammalian expression vectors.

Primers
Forward Reverse

Results
Purification of βA3-, WT αAand WT αB-Crystallins and Their Mutants Recombinant proteins were purified using the Ni-affinity chromatographic method, and the purity of each protein was monitored by SDS-PAGE analysis (Fig 1). Each of the protein preparation with >80-90% purity was recovered and concentrated. α-crystallins formed oligomers which appeared as minor bands and was confirmed by western blot (data not shown). These >80-90% pure proteins were used for the SPR analysis.  βA3-crystallin was studied using 5 μM to 25 μM of analytes. S1 Fig display sensograms for the observed interactions of WT αA-, αA N101D and αA N123D with βA3-crystallin, respectively. The fitted curves for WT αA-, αA N101D and αA N123D indicated very high affinity interactions with βA3-crystallin (K D values of 1.63X10 -9 M, 5.20X10 -9 M and 5.09X10 -9 M, respectively). As explained earlier in the Methods section, utilization of various binding models did not yield better fitting to the observed results. The currently existing fitting protocols were ineffective in delineating the complex interactions (specific/non-specific) between the molecules, and we therefore report binding propensities based on average RU's observed of the analyte between 260-270 sec where the RU's plateau (or saturate) for any given concentration (Fig 2). As shown in Fig 2, WT αA-adherence was relatively higher at each concentration compared to αA N101D and αA-N123D mutants. The adherence propensities of WT αAand it's deamidated mutants are ranked as WT αA> αA N101D> αAN123D. This suggested that deamidation of Asn101 and Asn123 to Asp decreased the binding affinity of WT αA-crystallin for βA3-crystallin.
In summary, WTαA-, WTαB-crystallin and their deamidated mutants diplayed higher affinity interactions with βA3-crystallin. However, deamidated mutants of αA-crystallin exhibited comparatively lower binding propensities for βA3-crystallin than WT αA-crystallin, whereas the αB N78D mutant had a higher binding affinity relative to the WT αB and the αB N146D mutant. In Vivo Interaction of βA3-Crystallin with αAand αB-Crystallins and Its

Mutants by FLIM-FRET Method
The in-vivo interaction of WTβA3-crystallin with WTαA-and WTαB-crystallins was determined by measuring the life-time of a donor by exciting the fluorescent dye of the donor (CFP βA3-crystallin) in the presence and absence of an acceptor (YFP WT αAor WT αB-crystallins or their mutants). When the acceptor and donor are at <10 nm distance and they interact, then the energy of the donor is transferred to the acceptor, and a decrease in the life-time of the donor could be observed. An YFP-CFP fusion protein separated by a 12-amino acid linker was used as a positive control. CFP co expressed with YFP, and CFP βA3-crystallin co-expressed with YFP were used as negative controls. Donor and acceptor proteins co-expressed in cytoplasm were selected as the region of interest for FRET analysis.
Expression of proteins in HeLa cells. CFP βA3-crystallin was expressed in both cytoplasm and in the nucleus (Fig 4A, top panel in a), and no bleed-through was observed in the YFP channel ( Fig 4A, panel b). αAand αB-crystallins were expressed in cytoplasm (Fig 4A,  panel b). The CFP βA3and YFP αA-/YFP αB-crystallins were expressed together (Fig 4A,  panel c) and were selected as a region of interest for FRET analysis. Western blot analysis showed that βA3-crystallin was expressed as~51 kDa protein (Fig 4B lane 5 in a and b). αAand αB-crystallins were expressed as~46 kDa protein (Fig 4B lane 3 and 4 in a and b). As shown in Fig 4B, degraded βA3-(lane 5), αA-, and αB-crystallins (lane 3 and 4 in a and b) were also observed during western blot analysis. A significant decrease in degraded product of βA3-crystallin was observed in the presence of αAand αBcrystallins (Fig 4B lanes 1 and 2). The deamidated mutants of αAand αB-crystallins, YFP αA N101D, YFP αA N123D, YFP αB N78D, and YFP αB N146D were also expressed in the cytoplasm (Fig 5A and 5B, panel b).
Among αA-domain mutants, YFP αA NTD was expressed as a cytoplasmic protein, whereas YFP αA CD and YFP αA CTE were expressed in both cytoplasm and the nucleus ( Fig  6A, panel b). Among αB domain mutants, YFP αB NTD was expressed mostly around the nucleus and in the cytoplasm, whereas YFP αB CD and YFP αB CTE were expressed in both cytoplasm and the nucleus (Fig 6B, panel b).  The positive control showed a decrease in the mean life-time of the CFP to 2.2 ± 0.1 ns from 2.8±0.2 ns, which indicated that the energy was transferred from the CFP to the YFP. The life-time of the CFP βA3-crystallin was 2.8±0.2 ns, which did not change in the presence of YFP (Fig 7A). The mean life-time of the CFP was also 2.8±0.2 ns when co-expressed with the YFP (Fig 7A). These results showed that the life-time of the donor was 2.8±0.2 ns, which was used in the FRET efficiency calculations.
When the CFP βA3-crystallin was co-expressed with the WT YFP αA-crystallin, the mean life-time of the CFP βA3-crystallin was reduced to 2.4±0.1 ns from 2.8±0.2 ns, which suggested that the energy was transferred to the YFP WT αA-crystallin (Fig 7B). Similarly, in the presence of the WT YFP αB-crystallin, the CFP βA3-crystallin transferred energy, and the life-time was decreased to 2.4±0.2 ns from 2.8±0.2 ns (Fig 7C). The positive control showed 22±4% FRET efficiency. The CFP WT βA3-crystallin transferred an almost equal level of energy to the YFP WT αA-crystallin and the YFP WT αB-crystallin with a FRET efficiency of 18±4% for both (Fig 8A and 8B). The results suggested that the YFP WT αAand YFP WT αB-crystallins were almost at an equal distance from the CFP WT βA3-crystallin, and, therefore, similar level of energy was transferred.
The CFP βA3-crystallin showed a life-time of 1.6±0.3 ns and 2.2±0.3 ns, respectively ( Fig  7C) in the presence of the YFP αB N78D and YFP αB N146D crystallins. YFP αB N78D showed a higher FRET efficiency of 36±5% relative to the 18±4% efficiency of WT YFP αBcrystallin (p<0.001). However, YFP αB N146D showed 22±4% efficiency, which was slightly higher than WT YFP αB crystallin (Fig 8B) but the difference was not statistically significant. This suggested that more energy was transferred from the CFP βA3-crystallin to the YFP αB N78D mutant compared to the level of energy transferred to the YFP αB N146D and YFP WT αB-crystallins.
In summary, the WT YFP αAand WT YFP αB-crystallins showed strong interaction to the CFP βA3-crystallin. Further, the deamidation of αAand αB-crystallins increased the interaction efficiency with the CFP βA3-crystallin.
Among the domain mutants of the αB-crystallin, the life-time of the CFP βA3-crystallin was reduced to 2.1±0.3 ns from 2.8±0.2 ns in the presence of the YFP αB NTD. However, in the presence of YFP αB CD and YFP αB CTE, the life-time was 2.5±0.2 ns (Fig 7C). The lifetime of the CFP βA3-crystallin was 2.4±0.2 ns in the presence of the WT YFP αB-crystallin, and the FRET efficiency of the YFP αB NTD was higher (32±6%) compared to the 14±4% of the YFP αB CD and YFP αB CTE and 18±4% of YFP WT αB-crystallins (Fig 8B) (p<0.05). The higher energy transfer to the αB NTD suggested that this region of αB might be involved in the interaction with βA3-crystallin.  Interaction of βA3-Crystallin with Deamidated αAand αB-Crystallins

Discussion
Lens transparency is maintained by the solubility and interactions among crystallins. The interaction among crystallins is delicate and can be perturbed by PTMs or stress [4,5]. These perturbed interactions result in non-uniform changes in lens protein density and refractive index and lead to a subsequent increase in light scattering. Specific genetic mutations in crystallins, reported during cataract development, also affect the protein-protein interactions [6][7][8][9][10]49].
The covalent changes in lens crystallins disturb the non-covalent and weak interactions among them and may alter association among crystallins, which eventually leads to the development of lens opacity. The increased interaction was observed between covalently modified γB-crystallin and α-crystallin, while interaction of α-γ-crystallins, and α-β-crystallins was decreased during aging [5,50,51]. In our present study, αAand αB-crystallins showed strong interaction with βA3-crystallin. WT αB-crystallin showed relatively higher binding with βA3-crystallin (Fig 3) than WT αA-crystallin (Fig 2). This is in confirmation with an earlier report when HMW β-crystallin, isolated from human lenses, was used in an interaction study [51]. However, when the interaction was studied under physiological condition using the FLIM-FRET method, both, αAand αB-crystallin showed almost equal interaction with βA3-crystallin. Further, FRET results were consistent with our earlier findings in which βA3-crystallin (βA3 fused to GFP) interaction with WT αAand WT αB-crystallins (fused to pDS red) was examined using a photobleaching FRET method [46]. In the previous and present studies, FRET efficiencies were almost equal even though FRET calculation methods by the two techniques were different. In the present study, FRET was calculated by measuring a decrease in donor life-time, whereas in the previous study [46], FRET was calculated by an increase in the intensity of the acceptor after photobleaching.
Deamidation in crystallins has been shown to cause structural changes, which have profound effects on protein-protein interactions [5,23,25,26]. Therefore, in this study, we examined the effects of deamidation of αAand αB-crystallins on their interaction with βA3-crystallin in vitro and in vivo. The binding propensities of deamidated αA-crystallins were lower than WT αA-crystallin for βA3-crystallin (Fig 2). This result was expected since our earlier study showed that deamidation in αA-crystallin decreased the tryptophan spectra intensity, surface hydrophobicity, and formed more compact, but higher molecular weight oligomers, suggesting their altered conformation and assembly [27]. Therefore, we suggest that steric hindrance caused by the increased oligomer sizes is likely to be responsible for the lower binding of deamidated αA-crystallin relative to WT αA-crystallin. Similarly, relative to the WT αA-crystallin, the increased oligomer size and altered structures of the deamidated αA-crystallin mutants could be responsible for their closer proximity to βA3-crystallin, and for the relatively higher energy transfer to them compared to smaller size oligomers of WT αA-crystallin.
Deamidation of αB-crystallins has been shown to alter the secondary and tertiary structures, which results in the partitioning of deamidated αB-crystallin into water-insoluble fractions and formation of aggregates with other crystallins in aging and cataractous lenses [29,49]. In SPR analysis the αB N78D mutant showed a higher binding propensity to βA3-crystallin compared to WT αBand the αB N146D mutant (Fig 3). Similar results were also observed during FLIM-FRET analysis. This result was surprising since our past study showed that αB N78D did not have any significant changes in structural and functional properties compared to WT αBcrystallin [29]. Therefore, the reason for the increased association of αB N78D to βA3-crystallin needs to be further investigated. In contrast, deamidation of αB-crystallin at position 146 has been shown to increase the surface hydrophobicity and an increase the oligomeric size relative to WT αB-crystallin [29]. This could be one of the reasons for comparatively lower binding affinity (Fig 3) and increased FRET efficiency of the αB N146D crystallin. Reports have shown that deamidated mutants of αAand αB-crystallins (isolated from lenses) increased their surface hydrophobicity with higher molecular weight aggregates [20,52,53]. Therefore, we speculate that under physiological conditions, an increased surface hydrophobicity might have exposed the surface binding sites of the deamidated mutants of αAand αB-crystallins, and, therefore they showed a stronger association with the βA3-crystallin. However, the discrepancies in binding affinity in our in-vivo and in-vitro studies need further investigation.
Based on these data and previous studies, we hypothesized that increased oligomer size of deamidated mutants may play an important role in interaction with βA3-crystallin as shown schematically (Fig 9). During in vitro interaction studies of deamidated mutants of α-crystallin, steric hindrance caused by their larger oligomers size and reduced surface hydrophobicity might result in loose binding with βA3-crystallin. However, under in vivo conditions, these larger oligomers are in close proximity with the βA3-crystallin in cells, which may have resulted in increased interaction of deamidated mutants of α-crystallin with βA3-crystallin.
With respect to the interactions of the NTD, CD and CTE domains of αA and αB-crystallin with βA3-crystallin, the FLIM-FRET data showed that αA CTE had a higher binding efficiency to βA3-crystallin compared to the binding efficiency of WT αA-crystallin to βA3-crystallin, whereas αA NTD and αA CD had almost equal FRET efficiency (Fig 8A). A previous study has shown that αA NTD and αA CD did not show any significant structural changes and retained Interaction of βA3-Crystallin with Deamidated αAand αB-Crystallins their chaperone activity [47], therefore, the binding strength of both the domains to the βA3-crystallin did not vary. However, the CTE domain of αA-crystallin has been shown to be involved in oligomeric assembly [54,55], and it might have formed an oligomeric complex with βA3-crystallin that resulted in the higher FRET efficiency. Among the three domains of αB crystallin, αB NTD showed higher FRET efficiency compared to αB CD, αB CTE, and WT αBcrystallins. The NTD domain of αB crystallin was identified as an interacting region with the substrate, and it is more flexible and solvent accessible with higher chaperone activity [47,56]. Therefore, it may have shown higher affinity for βA3-crystallin in comparison to αB CD and αB CTE.
In summary, our study demonstrated that WT αAand WT αB-crystallins as well as their deamidated mutants had strong interaction with βA3-crystallin. However, under in vivo conditions on deamidation in αA-crystallin at positions Asn 101 and Asn 123 and in αB-crystallin at positions Asn 78 and Asn 146 relatively increased the interaction with βA3-crystallin. Further, the results showed that all the three domains (NTD, CD, and CTE-) of WT αAand αB-crystallins were involved in the interaction with βA3-crystallin, but αA CTE and αB NTD showed greater affinities for βA3-crystallin. The residues of α-crystallins (WT and deamidated mutants) involved in the interaction with βA3-crystallin will be investigated in the future studies.