Human αA-crystallin cDNA cloned into pET-23d (+) vector (Novagen, Madison, WI, United States) was used as a template to generate G98R mutation (Murugesan et al., 2007). The mutant and wild-type proteins were expressed in E. coli BL21(DE3)pLysS cells (Invitrogen, Carlsbad, CA, United States) and purified as described earlier with minor changes (Santhoshkumar and Sharma, 2006). Briefly, bacterial cell pellet obtained from one-liter culture was suspended in 5 ml lysis buffer containing 50 mM Tris-HCl (pH 8.0), 0.1 M NaCl, and 2 mM EDTA and treated with 50 μl of protease inhibitor cocktail III (Calbiochem, San Diego, CA, United States), 1 mg lysozyme (Worthington Biochemical Corp. Lakewood, NJ, United States) and 10 mM DTT. The cell suspension was treated with 1 μl of benzonase (Sigma, St. Louis, MO, United States) and incubated at 37°C on a shaking platform for 30 min. The extract was centrifuged at 17,000 g for 1 h. The αAG98R protein partitioned into insoluble fractions was washed and dissolved in Tris-HCl (pH 7.2) buffer containing 1 mM EDTA and 6M urea. The urea dissolved supernatant was filtered and loaded into a Q-Sepharose Fast Flow ion-exchange column equilibrated with Tris-EDTA buffer. The protein was eluted using a stepwise gradient of 1 M NaCl in 50 mM Tris-HCl (pH 7.2) containing 1 mM EDTA at 1 ml/min flow rate. The αA-WT protein partitioned into the soluble fraction was initially purified on a Hiload 16/60 superdex 200 gel filtration column equilibrated with 0.05 M PO₄ buffer containing 0.15 M NaCl (pH 7.4). The αA-WT peak was pooled, concentrated, treated with solid urea (6M), and purified using an anion exchange column as described for the mutant protein. The purity of the proteins was checked by SDS-PAGE, and the molecular mass was determined by mass spectrometry. The concentration of the mutant and wild-type proteins was estimated using Bio-Rad protein assay reagent.

Frozen calf lenses were thawed, decapsulated, and added to 5.0 ml of 20 mM tris buffer, pH 7.9. The mixture was gently stirred for 30 min at 4°C and was then centrifuged at 12,000 rpm for 30 min. The supernatant was loaded onto a Sephadex G200 column to separate the β_L-crystallin. The βB₂-crystallin was isolated from the β_L fraction as described previously (Slingsby and Bateman, 1990). Briefly, the β_L fraction was concentrated to 5 mg/ml and equilibrated in tris buffer (pH 7.4) containing 6M urea and 2 mM tris(2-carboxyethyl) phosphine hydrochloride (Buffer A). The protein was then applied to Q Sepharose FF anion exchange column (GE Biosciences) equilibrated with buffer A, and the proteins were eluted using a linear salt gradient (0%–15%) of buffer A containing 1 M NaCl. The tubes containing the βB₂—crystallin were pooled and concentrated. A small fraction of the protein was digested with trypsin, and the peptides were subjected to qTOF MS/MS analysis to confirm the protein.

Protein samples from various incubations (see fig legend) were injected (maximum volume 200 μl) onto a TSKgel G5000PW_XL (Tosoh Bioscience) size-exclusion column equilibrated with 0.05 M sodium phosphate buffer containing 0.15 M NaCl (pH 7.4). The flow rate was set to 0.75 ml. The column was attached to an HPLC system connected with UV and refractive index (RI) detectors and coupled to a static multi-angle laser light scattering (DAWN-EOS) and dynamic quasi-elastic light scattering (QELS) detectors (Wyatt Technology, Santa Barbara, CA, United States). The molar mass (M_w), and hydrodynamic radius (R_h), were determined using ASTRA (6.1) software developed by Wyatt Technology (Santa Barbara, CA, United States).

The chaperone activity of αA-WT and αAG98R was compared using insulin, alcohol dehydrogenase (ADH), citrate synthase (CS) and βΒ₂-crystallin (βΒ₂) as substrates. The extent of aggregation was measured by monitoring the light scattering at 360 nm on a Shimadzu UV-VIS spectrophotometer equipped with a thermal controller. All assays were done in the absence or presence of different amounts of wild-type and mutant proteins in a 1 ml buffer.

The ability of αA-WT or αAG98R to stabilize SmaI restriction enzyme activity (New England Biolabs, USA) was tested as described by us earlier (Santhoshkumar and Sharma, 2001). SmaI (1 unit) was incubated in the presence or absence of 0.2 μg of αA-WT or αAG98R proteins at 37°C for 90 min. The incubation volume was adjusted to 10 μl using the buffer supplied by the enzyme manufacturer. After thermal inactivation, 200 ng of a plasmid containing three SmaI cleavable sites was added to the tube and incubated for 90 min at 25°C. The digested mixture was analyzed on 1% agarose gel.

The chaperone proteins were incubated with substrate proteins in aggregation assay buffers (250 μl) as defined earlier. A separate tube was set for each time point of analysis. For ADH (100 µg) and CS (20 µg) incubations, equal amounts of the chaperone proteins were used [1:1 ratio (w/w)]. For βB₂-crystallin (150 µg) incubations, 25 µg of the mutant protein or 50 µg of αA-WT was used. The concentrations of the proteins selected were empirically determined. The incubation mixtures were analyzed at different time points as specified under the results section.

Chaperone proteins (1 mg/ml) in PBS were incubated with or without purified MGO (50 µM) (In-kind gift from Dr. Nagaraj) at 37°C in the dark under sterile conditions. The chaperone activity of the MGO modified αAG98R (25 µg) was analyzed using ADH (described in detail under results). Small aliquots (5 µg) were drawn at regular intervals and analyzed by SDS-PAGE analysis to check for crosslinking or autolysis. Another aliquot (25 µg) of the sample was analyzed by multi-angle light scattering, as described previously, to look for changes in the oligomeric mass distribution.

Previously, studies have shown that αAG98R is prone to aggregate rapidly at near-physiological temperatures. Comparing the chaperone activity of αA-WT and the mutant protein we reported earlier that the mutant protein shows a substrate-dependent chaperone activity (Murugesan et al., 2007; Raju et al., 2011). This study involved multiple concentrations of the αA-WT and mutant protein and different client proteins and assay conditions to estimate relative chaperone efficiency (EC₅₀) of both chaperone proteins. In thermal aggregation assays, where the aggregation of substrate proteins occurred slowly, the mutant was more efficient than the αA-WT in preventing the aggregation of substrate proteins (Figure 1). Based on the EC₅₀ values determined by comparing the scattering at 60 min assay time point, αAG98R (EC₅₀—16 µg) (Figure 1A) was nearly three times more efficient than the WT protein (EC₅₀—45 µg) in preventing ADH aggregation. For CS, fifty percent protection from aggregation was obtained with 5 µg of the mutant protein (Figure 1B). The mutant was over four times better in protecting CS aggregation when compared to the αA-WT which, exhibits similar protection of CS at 22 µg. Similarly, 30 µg of αA-WT was required to suppress the aggregation of βB₂ by 50%, whereas 4 µg of the mutant protein was able to provide the same level of protection (Figure 1C). However, in assays where the substrate protein aggregates rapidly, like the DTT-induced insulin aggregation, αA-WT showed better chaperone activity than the mutant protein (Figure 1D). While the DTT-induced insulin aggregation was entirely prevented by αA-WT when used at a 1:1 (w:w) ratio (Figure 1D, Curve 2), the G98R mutant protein at the same amount could not provide any protection against insulin. Instead, it increased the light scattering by insulin (Figure 1D, Curve 3), and when doubling the amount of mutant protein, a 25% suppression of insulin aggregation was seen (Figure 1D, Curve 4). Although αAG98R was better than the WT protein in suppressing ADH aggregation during the assay period (Figure 1A), the substrate protein complexes formed with the mutant (in tubes showing 100% protection during chaperone assay) precipitated when the chaperone assay samples were left overnight (16 h) at room temperature (Figure 1E, Lane 2). The tubes with αA-WT showing a similar level of protection did not show any visible deposits when left overnight (16 h) at room temperature (Figure 1E, Lane 4). Next, we tested if αAG98R can protect from the thermal inactivation of restriction endonuclease, SmaI, at 37°C. The mutant protein completely protected the SmaI from heat inactivation (Figure 1F, Lane 2), just like the αA-WT (Figure 1F, Lane 1). In the absence of chaperone protein, SmaI incubated at 37°C failed to cleave the DNA (Lanes 3 & 4). Our data confirm the substrate-dependent chaperone activity of αAG98R mutant. The results also suggest that αAG98R protein, which is unstable at near-physiological temperatures, gets transiently stabilized upon interaction with specific substrate proteins.

We followed the interactions of αA-WT and the mutant with the substrate proteins (ADH, CS, and βB₂) on a MALS detector at regular intervals to analyze the complex formation during chaperone action at 37°C. Excess of chaperone proteins was used to ensure that the substrates did not scatter light (A360 nm) during the duration of the experiment. The incubations were carried out in appropriate aggregation buffers (0.25 ml) and in separate tubes for each time point of analysis. All samples were filtered through 0.45 µm Whatman® Mini-UniPrep® syringeless filters (Sigma, St Louis, MO, United States) before fractionating on a TSKgel G5000PW_XL column connected to an HPLC coupled with MALS-DLS detectors. Under the defined incubation conditions, CS aggregated slowly, so we decided to follow the complex formation for up to 20 h. ADH and βB₂ samples were followed up to 90 min. ADH and βB₂ showed a scatter of (Absorbance 360 nm) about 0.35 and 0.06, respectively, in the absence of chaperones at 90 min.

Figure 2 shows the elution profiles of the samples from the UV detector. The average molar mass (M_w) and the hydrodynamic radii (R_h) at the complex peak determined from the light scattering detectors using the ASTRA software are given in Table 1. A slight shift in the αA-WT peak with ADH was seen in samples analyzed at 90 min (Figure 2, Peak 8) when compared to the peak at 0 min (Figure 2, Peak 5), which correlated with a corresponding increase in M_w from 824 kDa (0 min) to 1,192 kDa (90 min) and a rise in R_h from 7.8 to 12.8 nm (Table 1). We also found some high molecular weight complexes eluting at 7–9 min. However, the protein concentration was too low to analyze the M_w. In contrast to the WT, the mutant protein formed large oligomers with ADH, and the complex size (R_h) increased with time from 12.5 nm at 0 min to 38 nm at 90 min. We could only determine the Mw at 0 min, after which the light scattering signal maxed out, hampering M_w analysis.

With CS, αAG98R showed a similar effect, and the R_h increased from 13 to 35 nm, and M_w increased from 2,452 kDa to 161,900 kDa in 0–7 h (Table 1). The sample at 20 h could not be analyzed due to high scattering and low protein concentrations. Most of the aggregates were likely removed during the initial filtering of the sample. In contrast to αAG98R, the interaction with CS did not increase the oligomeric mass of αA-WT. Instead, the M_w at the peak apex reduced from 822 to 589 kDa in 3.5 h, and it decreased further to 378 kDa in 7 h (Table 1). However, further incubation of the sample up to 20 h showed the complex peak mass increasing to 851 kDa. The reduction in the CS + αAWT complex peak mass is supported by delay in the elution of complex peak (Peak 8, Figure 2 lower center) with time. Since the data points from the DLS detector did not fit well, the R_h values reported by the ASTRA software for these samples were not considered.

The UV profile of αAG98R and βB₂ at 0 min showed the mutant peak (Figure 2, Peak 1, the upper panel right) appearing early at 9.72 min, unlike in ADH and CS incubations, the mutant was seen eluting at 10.2 min (Table 1). The early appearance of the mutant protein in the βB₂ sample is also supported by an increase in M_w (9,520 kDa) and R_h (17.83 nm) (Table 1) at the peak apex, suggesting that αAG98R and βB2-crystallin interact during the duration (<5 min) of sample preparation. Surprisingly, continuing the incubations further did not result in high molecular weight species as seen with ADH and CS. Instead, there was a significant decrease in M_w and R_h of the complex peak (Table 1). The αAG98R-βB₂ peak intensity also progressively increased with time, unlike the peak intensities of complexes with the other two substrates (Figure 2, upper panel). The interaction of αA-WT with βB₂ did not significantly alter the M_w and R_h of the complex peak up to 60 min. However, at 90 min, the M_w slightly increased from 688 to 822 kDa (Table 1).

Using multi-angle light scattering, we have investigated whether incubating αAG98R with the water-soluble fraction of young human lenses (donor lenses obtained from the Lions Eye Tissue Bank, Columbia, MO) can alter the mass distribution during sample aging. The water-soluble fraction from a 16-year-old human lens (250 µg protein) (Santhoshkumar et al., 2008) was incubated under aseptic conditions with 50 µg of αA-WT or mutant protein in 0.25 ml PBS containing 0.03% sodium azide and protease inhibitors (Sigma, St. Louis, MO, United States) at 37°C for 7 days. At the end of the incubation, the sample (150 µl) was processed for MALS analysis as described previously. Another set of samples was prepared and analyzed after 30 min incubation at 37°C. The mass distribution across the refractive index (RI) profiles of the incubation mixtures at 30 min and 7 days was compared (Figure 3). As expected, the lens proteins treated with αA-WT or mutant protein for 7 days (red curve) eluted early from the column compared to the freshly analyzed (30 min) sample (blue curve), suggesting the formation of high molecular weight aggregates during sample aging in vitro. In sample(s) with αA-WT, the mass distribution across the RI profile of both time points was more or less overlapping, indicating no significant change in the mass of the complexes during the incubation period. The elution profile and mass distribution in incubations of lens extract alone (data now shown) were like in lens extract incubated with αA-WT. In contrast, samples incubated with αAG98R showed a differential mass distribution with the 7-day incubation mixture showing higher molar mass across the RI profile (Figure 3 bottom). The aged sample with αAG98R also had large oligomers eluting in 7–9 ml. This caused the light scattering detector to max out so we could not show the mass distribution in that region. Overall, the magnitude of aggregates was increased when αAG98R was added to the lens extract. The aggregates’ size was comparable to those found in water-soluble aged lens extracts (Santhoshkumar et al., 2008), implying that G98R mutation in αA could accelerate lens aging.

Previous studies have shown that MGO modification of arginine residues neutralizes the positive charge (Derham and Harding, 2002; Biswas et al., 2008). Since G98R mutation results in the gain of positive charge, we investigated whether incubation of the mutant protein with low MGO concentrations would improve the mutant protein’s stability. The αAG98R protein (1 mg/ml) was incubated at 37°C under aseptic condition (filtering through a 2 µm filter) without or with 50 µM MGO in PBS containing 0.03% sodium azide for 30 days. The samples were visually observed for sediments and were analyzed by SDS-PAGE every week to check for breakdown/crosslinking. In the tube without MGO, a large deposit was seen at the bottom at 30 days, while the sample with MGO remained transparent (Figure 4A). SDS-PAGE analysis showed that the mutant protein shows autolytic activity without MGO (Figure 4B). More than 50% of the mutant protein band was reduced in 15 days without MGO, and at 30 days, the band for full-length αAG98R protein was completely absent and a prominent band for a cleavage product was seen. In contrast, the mutant protein incubated with MGO did not show any breakdown at 15 days, but a small amount (<5%) of crosslinked proteins were seen. In the 30-day sample, the crosslinks increased slightly, and some faint bands below the prominent intact protein band were seen. In general, the data suggest that low concentrations of MGO can stabilize αAG98R mutant protein.

We next investigated the effect of MGO modifications on the molar mass distribution of αAG98R by MALS analysis. Figure 5 shows the mass distribution across the UV profiles of αA-WT and αAG98R proteins incubated with or without 50 µM MGO at 37°C for 7 days. The M_w and R_h calculated from the MALS-DLS signals using the ASTRA software are given in Table 2. In the absence of MGO, αAG98R mainly precipitated and was removed during sample filtering. Only 18% of the protein remained soluble and eluted as a broad peak (Figure 5 upper panel–red) consisting of G98R oligomers with mass ranging from 41,302–5,730 kDa (Table 2). In contrast, the mutant protein incubated with MGO showed a prominent peak (Figure 5 upper panel—blue), with 80% of the protein recovered in the soluble form (compared to the sample analyzed at 0 min). The average molar mass of the soluble αAG98R oligomers decreased from 9,386 kDa to 3,350 kDa in the presence of MGO at 7 days (Table 2). The hydrodynamic radii also showed a similar decrease in size (Table 2). The elution profile of αAG98R without MGO at 0 min (shown in black dash line for comparison) had an average oligomer mass of 2057 kDa (data not shown). Unlike as seen with αAG98R, MGO modifications of αA-WT did not show a significant reduction in M_w and R_h. The average M_w of αA-WT decreased by about 100 kDa, and the size was reduced by 0.1 nm in samples with MGO. The data suggest that MGO modifications reduce the aggregation propensity and improve the solubility of αAG98R mutant protein. The residues that MGO modifies in αAG98R are yet to be investigated.

Previous studies have shown that incubations with low concentrations of MGO will improve the chaperone function of α-crystallin (Nagaraj et al., 2003; Nagaraj et al., 2008). We have tested if αAG98R modification with MGO will alter the anti-aggregation property of the mutant protein. The mutant protein (1 mg/ml) was incubated with or without 50 µM MGO as described previously. The chaperone activity of the samples was measured immediately (0-time point) and at one and 2 days after the start of the incubation using ADH (250 µg) as the aggregating client protein. Before the chaperone assay, the free MGO in the sample was removed by quick buffer exchange (3x) using a 10 kDa Vivaspin 2 ultra-concentrator (Sartorius, Bohemia, NY, Unites States). Samples without MGO were also processed similarly. The protein concentrations were estimated using Bio-Rad protein assay. The anti-aggregation activity was tested using 25 µg of the mutant protein. The chaperone activity of unmodified αAG98R reduced with time, while that of the MGO-modified mutant protein remained the same at all time points (Figure 6). αAG98R treatment with MGO also slightly increased the anti-aggregation activity compared to the untreated protein. Our data suggest that the chaperone activity of αAG98R is improved by modification with MGO.