Interaction of C-Terminal Truncated Human αA-Crystallins with Target Proteins

Background Significant portion of αA-crystallin in human lenses exists as C-terminal residues cleaved at residues 172, 168, and 162. Chaperone activity, determined with alcohol dehydrogenase (ADH) and βL-crystallin as target proteins, was increased in αA1–172 and decreased in αA1–168 and αA1–162. The purpose of this study was to show whether the absence of the C-terminal residues influences protein-protein interactions with target proteins. Methodology/Principal Findings Our hypothesis is that the chaperone-target protein binding kinetics, otherwise termed subunit exchange rates, are expected to reflect the changes in chaperone activity. To study this, we have relied on fluorescence resonance energy transfer (FRET) utilizing amine specific and cysteine specific fluorescent probes. The subunit exchange rate (k) for ADH and αA1–172 was nearly the same as that of ADH and αA-wt, αA1–168 had lower and αA1–162 had the lowest k values. When βL-crystallin was used as the target protein, αA1–172 had slightly higher k value than αA-wt and αA1–168 and αA1–162 had lower k values. As expected from earlier studies, the chaperone activity of αA1–172 was slightly better than that of αA-wt, the chaperone activity of αA1–168 was similar to that of αA-wt and αA1–162 had substantially decreased chaperone activity. Conclusions/Significance Cleavage of eleven C-terminal residues including Arg-163 and the C-terminal flexible arm significantly affects the interaction with target proteins. The predominantly hydrophilic flexible arm appears to be needed to keep the chaperone-target protein complex soluble.


Introduction
The major proteins in the vertebrate eye lens are a-, b-, and ccrystallins the predominant one being the a-crystallin. a-Crystallin consists of two nearly homologous subunits, namely, aAand aBcrystallins and both having a molecular mass of 20 kDa in the monomer form and contain 173 and 175 amino acid residues respectively [1][2][3]. Both aAand aB-crystallins belong to the class of small heat shock proteins [4] and function as molecular chaperones having the ability to prevent aggregation of partially unfolded proteins [5][6][7]. The model structure of a-crystallin consists of a globular N-terminal domain and a C-terminal domain containing an exposed C-terminal arm rich in hydrophilic amino acids, whereas the C-terminal stretch of 80-100 residues known as the 'a-crystallin/sHsp domain' are highly conserved [8].
C-terminal cleavage of aA-crystallin at residues 162, 168, and 172 has been reported earlier [9][10][11][12][13][14][15]. The major post-translational modification which occurs in human aA-crystallin is the loss of the C-terminal serine residue [9,14,15]. Enhanced cleavage of the Cterminal residue of aA-crystallin in diabetic human lenses has been reported in our earlier study, the average level of the truncated aA-crystallin increased from 30% to 50% [15]. Aziz et al [16] have recently reported modification in the oligomeric structure and chaperone function of the various truncated human aA-crystallins. Interestingly, the truncated aA 1-172 exhibited significant increase in its oligomeric size as well as chaperone activity. The oligomeric size of aA  was similar to that of aA-wild type (aA-wt) whereas the chaperone activity was moderately decreased. aA 1-162 , on the other hand, showed substantial decrease in the oligomeric size as well as the chaperone activity. If indeed chaperone to target protein binding is an essential step for aA-crystallin to operate as a molecular chaperone, characterization of the interaction of the truncated aA-crystallins with target proteins should show why their chaperone function is altered as a result of the truncation. In this study, we have used fluorescence resonance energy transfer (FRET) to study chaperone -target protein interaction using ADH and bL-crystallin, two widely used target proteins [16,17], and recombinant aA-wt and C-terminal truncated aA-crystallins.

Results
Levels of fluorescence labeling of human aA-wt, Cterminal truncated aA-crystallins and the target proteins The level of subunit exchange between two different proteins was determined by FRET. In order to determine the in vitro FRET level two fluorescent dyes with overlapping fluorescence spectra are necessary. SITS and LYI are widely used fluorescent dyes with spectral overlaps and safe from structural alterations due to fluorescent tags [18][19][20]. So, in the present study the amine specific fluorescent probe SITS was attached to human aA-wt and the Cterminal truncated aA-crystallins and the cysteine specific LYI fluorescent probe was attached to the target proteins ADH and bL-crystallin. The level of labeling was determined spectrophotometrically using molar extinction coefficients of 47,000 mol 21 cm 21 at 336 nm for SITS and 11,000 mol 21 cm 21 at 426 nm for LYI. The level of labeling was about 0.8, 1.12, 1.17, 1.04 and 7.1 and 1.56 moles for aA-wt, aA 1-172 , aA 1-168 , aA 1-162 , ADH and bL-crystallin respectively. Differences in the level of labeling of the two different fluorophores were taken into account while computing the data.
Subunit exchange between labeled aA-crystallins and the target protein ADH The interaction between the SITS -labeled human aA-wt and the C-terminal truncated aA-crystallins and the LYI-labeled target protein ADH was initiated by mixing equimolar concentration of the proteins in 20 mM MOPS buffer with 100 mM NaCl and 10 mM EDTA at 37uC. EDTA was used for unfolding ADH so that aAcrystallin will bind to ADH under the same condition as used for the chaperone assay. The rate of subunit exchange with ADH was determined by FRET analysis. Figure 1 shows the fluorescence spectra showing the time dependent decrease in SITS emission intensity at 426 nm and a concomitant increase in LYI fluorescence at 515 nm. After 30 min at 37uC, there was no remarkable change in the emission intensity at 515 nm due to the achievement of stable equilibrium (data not shown). The fluorescence emission spectra illustrate that the highest increase in the acceptor spectrum at 515 nm was observed in aA-wt when compared to the C-terminal truncated aA-crystallins (Fig. 1). We have calculated the rate of subunit exchange from the increase in acceptor fluorescence intensity after taking into account differences in the levels of tagging of the various proteins by the probes. Figure 2A shows the plot of Ft/ F0 of LYI at 515 nm as a function of time, where Ft and F0 are the emission intensities at time t and zero respectively. The rate constant was obtained by fitting the data to the exponential function Ft/ F0 = A1+A2 e-kt, where A1 and A2 are constants and k is the rate constant for subunit exchange. The increase in the relative fluorescence intensity at 515 nm is due to fluorescence resonance energy transfer from donor SITS-labeled human aA-wt and Cterminal truncated aA-crystallins' fluorophore to the acceptor LYIlabeled ADH during the interaction. The maximum relative fluorescence intensity was seen with aA-wt whereas aA 1-172 had lower fluorescence intensity, aA 1-168 showed further decrease in the fluorescence intensity, and aA 1-162 showed the lowest. The subunit exchange rates or the k values are summarized in Table 1 which is in agreement with the above observation. For instance, aA-wt had the highest k value (2.073) whereas the k values of aA 1-172 , aA 1-168 , and aA 1-162 , respectively, were 6, 24, and 43% lower. Figure 2B illustrates the decrease in the relative fluorescence at 426 nm when the fluorescence resonance energy is transferred to the acceptor target protein. The rate constant (k) values (Table 1) were higher when collected and calculated at donor energy (at 426 nm) than when calculated at acceptor energy (at 515 nm). This is believed to be was due to fluorescence quenching because no direct transfer of fluorescence energy occurs where there is no sufficient contact between the residues carrying the two probes. The amount of loss  Subunit exchange between aA-crystallins and the target protein bL-crystallin The interaction between the SITS-labeled human aA-wt and the C-terminal truncated aA-crystallins and the LYI-labeled target protein bL-crystallin was initiated by mixing equimolar concentration of the proteins in 20 mM MOPS buffer with 100 mM NaCl at 62uC. Higher temperature was necessary to unfold bL-crystallin. Figure 3 shows the fluorescence spectra and Figure 4A and 4B show the increase and decrease in relative fluorescence intensity or Ft/F0 at 515 nm and 426 nm, respectively as a function of time. The k values which reflect the data in Figure 4A and 4B were summarized in Table 2. Interestingly, the acceptor gain rate constant (k) value for aA  was maximal at 2.422 and it was about 8% higher than the value for aA-wt and about 41% higher than the values for aA  and aA 1-162 . The same trend was noticed in donor k values determined at 426 nm also. The C-terminal truncated aA 1-172 was maximal at 6.391 which was about 24% higher than that of aA-wt and 34% and 54% higher than those of aA 1-168 and aA 1-162 respectively. As mentioned above, here also at the time of energy transfer from donor fluorophore to the acceptor fluorophore some energy loss was noticed. The amount of energy lost was about 54-66%. The subunit exchange rate constant (k) values clearly showed that in aA-crystallin and the counterparts with ADH at both emission intensities (at 426 & 515 nm) the aA-wt showed the highest interaction compared to the C-terminally truncated aA-crystallins and among the C-terminally truncated aA-crystallins, aA 1-172 showed higher k value followed by aA 1-168 and aA 1-162 in that order. However, the subunit exchange rate constant (k) values in aAcrystallin and counterparts with bL crystallin were higher in aA 1-172 followed by aA-wt, aA 1-168 and aA 1-162 . Nevertheless, the loss of energy (due to quenching) was least in aA-wt with both target   Chaperone activity of aA-wt and the C-terminal truncated aA-crystallins Although the chaperone activity of the truncated human aAcrystallins were reported earlier [16] the chaperone activity assay was repeated in the present study using the ratio of 1:1 which is the same ratio as used in the FRET analysis. With ADH as the target protein, aA 1-172 showed about 18% better chaperone activity than aA-wt, aA 1-168 had nearly the same chaperone activity as aA-wt and aA 1-162 showed nearly 80% loss (Fig. 5). With bL-crystallin as the protein, both aA 1-172 and aA 1-168 showed normal chaperone activity whereas aA 1-162 showed nearly 60% loss in chaperone activity (Fig. 6).
Labeling of human aA-wt, C-terminal truncated aAcrystallins and target proteins ADH and bL-crystallin with fluorescence probes Labeling of aA-crystallin and C-terminal truncated aAcrystallins with SITS and the target proteins ADH and bLcrystallin with LYI was done as described previously [20,21]. Briefly, 10 fold excess of solid SITS was added to 1 ml of a protein solution (3 mg/ml) in 20 mM MOPS buffer containing 100 mM NaCl (pH 7.9) and the reaction was allowed to proceed for about 16 h at room temperature (25uC) in the dark. The unlabeled fluorescent dye was removed from the fluorescently labeled protein Table 2. Subunit exchange rate constant (k) of aA-wt and its C-terminal truncated forms interacting with bL-crystallin when increase and decrease respectively (Mean6SE).  on a Sephadex G-25 column equilibrated with 20 mM MOPS buffer containing 100 mM NaCl (pH 7.9). Elution was performed with the same buffer, pooled and concentrated. LYI-labeled protein was prepared as described above with the same buffer in 20 fold excess of the reagent. The extent of labeling was determined spectrophotometrically using molar extinction coefficients of 47,000 mol 21 cm 21 at 336 nm for SITS and 11,000 mol 21 cm 21 at 426 nm for LYI and corrected for the contribution of the dye at 280 nm.
Measurement of the rate of subunit exchange among aA-crystallins and the target proteins The rate of subunit exchange was measured by the method of fluorescence resonance energy transfer (FRET). The subunit exchange reaction was initiated by mixing equal amounts (0.5 mg/ml) of SITS -labeled aA-wt and C-terminal truncated aA-crystallins (aA 1-172 , aA 1-168 & aA 1-162 ) with the same amount of LYI-labeled target proteins: 1)ADH in 20 mM MOPS buffer containing 100 mM NaCl and 10 mM EDTA (pH 7.9) at 37uC and 2) bL-crystallin in 20 mM MOPS buffer containing 100 mM NaCl (pH 7.9) at 62uC. The fluorescence emission spectra were obtained with Shimadzu RF-5301PC spectrofluorophotometer (Columbia, MD) at an excitation wavelength of 336 nm. Decrease in SITS emission intensity at 426 nm and increase in LYI emission intensity at 515 nm were recorded at every 2 min intervals and the subunit exchange rate constant (k) was calculated by curve fitting an exponential function F t / F 0 = A 1 +A 2 e 2kt where A 1 and A 2 are constants and k is the rate constant for subunit exchange.

Determination of Chaperone activity
The ability of aA-wt and the C-terminal truncated aAcrystallins to prevent the extent of EDTA induced aggregation of ADH and thermal aggregation of bL-crystallin was determined as described before [16,17]. aA-wt and the C-terminal truncated aA-crystallins were mixed with equal amounts of target proteins (1:1 ratio) at a total concentration of 0.14 mg/ml. ADH was induced to unfold with 10 mM EDTA in 50 mM PBS buffer (pH 7.9) at 37uC and the bL-crystallin unfolding was thermally induced at 62uC. The extent of aggregation was measured by monitoring the light scattering at 360 nm using Shimadzu UV 160 spectrophotometer equipped with a temperature regulated cell holder.

Discussion
We have shown earlier that cleavage of 1, 5, and 11 C-terminal residues affects or improves chaperone activity depending on the number of residues cleaved and the chaperone assay system [16]. Chaperone activity of the truncated aA-crystallins can vary by two major factors: accessibility to the chaperone sites which are responsible for the chaperoning process or due to enhanced or decreased affinity to the substrate which is being unfolded. The present study focused on the latter aspect by studying the binding of the wild-type and the truncated aA-crystallins with the putative target proteins. The data show that the binding kinetics (k value) reflect the chaperone activity differences with the exception that aA 1-172 showed higher chaperone activity, but, lower k value. Thus, we can conclude that the binding affinity to target proteins does not always concur with the chaperone activity data. However, it is noteworthy that all the three truncated aAcrystallins have shown decreased target protein binding.
In an earlier study, we have determined the oligomeric size, secondary and tertiary structures and chaperone activity of recombinant human aA-wt and the various C-terminal truncated aA-crystallins [16]. aA 1-172 , which is the major form of the truncated human aA-crystallin, had a molecular mass of 866 kDa as compared to 702 kDa for aA-wt, when the molecular mass was determined by dynamic light scattering. The chaperone activity of aA 1-172 was higher than that of aA-wt, when ADH, insulin and bL-crystallin were used as target proteins and the aA-crystallin : target protein ratio varied between 1:1 and 1:20. In this study, we have used equal amounts of a-crystallin and target protein, ie, 1:1 ratio, in all the assays to avoid artificially exaggerated differences between aA-wt and the truncated aA-crystallins. As shown in  (2) bL-crystallin alone (¤) determined with ADH as the target protein at 1:1 ratio (0.14 mg/ml) at 62uC. doi:10.1371/journal.pone.0003175.g006 Figures 5 and 6, aA 1-172 had slightly better chaperone activity than aA-wt, which confirms our previous observation [16]. However, aA-crystallin-target protein binding studies gave conflicting results. For instance, the relative fluorescence intensity due to FRET and the k values were lower in aA   (Figs. 2 & 4;  Tables 1 & 2). Both the present study and the earlier study [16] have shown aA 1-168 having nearly the same chaperone activity as aA-wt. however, the FRET studies gave contradicting results as the relative fluorescence intensity as well as the k values were significantly lower (Figs. 2 & 4; Tables 1 & 2). With both the ADH and the bL-crystallin target protein methods, aA 1-162 showed the lowest chaperone activity (Figs. 5 & 6) as shown earlier.
Interestingly, aA 1-162 also showed the lowest level of relative fluorescence intensity and k value during FRET studies. Thus, the cleavage of 11 C-terminal residues of aA-crystallin, which is known to affect its secondary and tertiary structures [16], severely affects its binding to target proteins which in turn affect its ability to function as a molecular chaperone.
By using FRET analysis, we have recently investigated the effect of cleavage of the C-terminal residues of human aA-crystallin on subunit exchange with aB-crystallin forming heteroaggregates [21]. The subunit exchange rate or the k value for aA  interacting with aB-wt was decreased by 50% compared to aA-wt interacting with aB-wt. Likewise, the k value was decreased 40% when aA 1-168 interacted with aB-wt, whereas interaction of aA 1-162 with aB-wt showed 84% decrease in the k value. Thus, cleaving 11 residues including lysine-163 had shown the most effect. The importance of lysine-163 in maintaining the oligomeric structure of aA-crystallin has been demonstrated in a recent study [23]. In fact, in aA 1-162 the secondary and tertiary structures were significantly altered and the molecular mass was substantially decreased [16]. Such changes affect binding to aB-crystallin and heterooligomerization [21] and also seem to affect binding to target proteins decreasing the chaperone activity.