High levels of αB-crystallin are present in the cardiomyocyte, yet little is understood about the function and importance of this protein. Like many other small heat shock proteins, αB-crystallin forms large oligomeric complexes whose size can be regulated by posttranslational modifications. The size of these complexes can modify the function of the protein. A naturally occurring COOH-terminal mutant has many detrimental effects in the lens of the eye and altered oligomerization. Therefore, we mutated the two COOH-terminal lysines of αB-crystallin to glycines (K174/175G) and adenovirally mounted them to transduce cardiomyocytes. We analyzed the effect of this mutation on oligomerization, microtubular stabilization, and ischemic outcome. A nearly 45% downward shift in complex size was observed with the mutant by native PAGE followed by immunoblotting. The overexpressed protein no longer protected the tubulin cytoskeleton against ischemic stress by confocal analysis. The mutant caused a 30% increase in cytosolic enzyme release with ischemia compared with control, whereas a 33% decrease was associated with wild-type αB-crystallin overexpression. We conclude that the COOH terminus of αB-crystallin is crucial to its proper function.
- heat shock protein
- recombinant adenovirus
- confocal microscopy
mammalianαB-crystallin is a member of the family of small heat shock proteins (smHSPs), which includes the lens α-crystallins and HSP25. αB-crystallin is an HSP by homology criteria and the classic induction pattern with thermal stress (20). An important protective function of the smHSPs including αB-crystallin has been well documented. Aoyama and coworkers (1) showed that αB-crystallin overexpression increased thermoprotection in fibroblasts. αB-crystallin has long been recognized as a major component in the lens of the eye contributing to the ability of this protective shield to remain transparent despite extensive oxidative insults. More recently, it was discovered that αB-crystallin is present at significant levels in nonlenticular tissues, especially the heart and skeletal muscles (8, 19). We previously demonstrated (25) that the expression of exogenous αB-crystallin correlated with protection against ischemic damage in cardiac myocytes. Additionally, it was demonstrated that αB-crystallin associates with actin and desmin, with the desmin interaction increasing in affinity with a decrease in pH, as during ischemia (3, 5). Also, we showed (6) that ectopic αB-crystallin preserves the tubulin cytoskeletal structure against ischemia-induced disruption in cardiomyocytes. Further evidence of the importance of αB-crystallin for proper cytoskeletal arrangements was shown in the cosegregation of an αB-crystallin missense mutation (R120G) with a desmin-related myopathy in a French family (35).
The smHSPs contain an evolutionarily conserved α-crystallin domain, which is preceded by an NH2-terminal domain. The NH2-terminal domain is highly variable in size and is generally buried in the aggregate formed by multiple individual HSPs constituting a large oligomeric structure. The COOH-terminal region is located toward the outside of the structure containing oligomerized smHSPs (22). The COOH-terminal region of smHSP appears relatively unstructured and is known to undergo numerous modifications. These include the relatively unexplored addition of O-linkedN-acetylglucosamine to the COOH-terminal threonine (T170), preliminarily thought to regulate the protective function and/or subcellular localization of αB-crystallin (4, 30). Also, a naturally occurring truncation of the four COOH-terminal amino acids of αB-crystallin in the lens of the eye (which includes lysines 174/175), the degree of which increases with age, correlates with a decay in the protective shield and an increase in cataract formation (18). Furthermore, Takemoto and coworkers (33) demonstrated by immobilizing the normally flexible COOH-terminal region that this flexibility is necessary for chaperone function. Using yeast two-hybrid screening with different sections of the crystallins, Fu and Liang (11) found that the COOH-terminal portion of αB-crystallin interacted and oligomerized far better than the NH2-terminal portion. Other posttranslational modifications, in particular phosphorylation, do play an important role in modifying smHSP and αB-crystallin function. αB-crystallin is phosphorylated via the mitogen-activated protein (MAP) kinase pathway at serines 19, 45, and 59, which has been demonstrated to decrease the oligomer size (10,15).
Both of the smHSPs αB-crystallin and HSP25 form large oligomeric complexes that are thought to be intrinsic to protection against noxious stresses, possibly providing a “safe haven” for cellular components. A number of studies have examined the link between oligomeric size and chaperone function for the small HSPs and the role that different domains of smHSPs play in oligomerization and chaperone function (for reviews, see Refs. 24 and 28).
Several studies have explored the relative importance of various regions of αB-crystallin by in vitro or prokaryotic assays. In one such study, the mutagenesis of the COOH-terminal lysines was associated with a decrease in thermoprotection in a prokaryotic thermokill assay but no change in oligomerization (29). The authors suggest that this electropositive region (K174/175) may be involved in the interaction with unfolded proteins. Thus altering this domain would decrease the chaperone activity of αB-crystallin but not necessarily the oligomerization. However, the authors tested only bacterially expressed recombinant proteins for oligomeric changes (and found no difference). The significance of this basically charged region in an in vivo mammalian cell model is unknown, especially with regard to a physiological stress like ischemia. Therefore, we sought to examine the importance of the COOH-terminal region of αB-crystallin by mutating the two terminal COOH-terminal lysines (174/175) to neutral glycines. The COOH-terminal mutant was adenovirally expressed in cardiomyocytes, and subsequently we compared it with the similarly expressed native protein on ischemic outcome, oligomeric formation, and cytoskeletal stabilization.
MATERIALS AND METHODS
Isolation of neonatal ventricular cardiomyocytes.
Neonatal cardiomyocytes were isolated via collagenase/pancreatin digestion and Percoll gradient separation from 1- to 2-day old rats as previously described (23). They were plated at a density of ∼500 cells/mm2 in 6-cm dishes precoated with gelatin in plating media (DMEM-M199, 4:1 vol/vol, with 10% horse serum and 5% fetal bovine serum supplemented with penicillin and streptomycin) for 18 h before adenovirus infection. For confocal studies, cells were plated at a density of 200 cells/mm2 on fibronectin-coated four-well chamber slides (Labtek; Fisher Scientific, Pittsburgh, PA).
Construction of adenoviral vectors.
The rat αB-crystallin adenovirus construction was previously described (23). We used the shuttle vector pACCMVpLpASR-(provided by Dr. Robert D. Gerard, University of Texas Southwestern, Dallas, TX), which contains the 5′ end of the adenovirus serotype 5 genome (map units 0–17) in which the E1 region has been replaced with the human cytomegalovirus (CMV) enhancer-promoter followed by the pUC19 multiple cloning site and the SV40 polyadenylation region. For a control virus we used a previously constructed non-transgene-containing adenovirus named SR (23). For the COOH-terminal mutant, the sense strand primer ACT GCA GCC CCT GGG GGG TAG ATT CCC TTT and the complementary antisense strand were synthesized as well as primers to the flanking regions of the pACCMVpLpASR− plasmid for PCR-based mutagenesis. The resultant mutated cDNAs were sequenced to confirm the mutation before construction of the recombinant adenovirus. Adenoviruses were generated via standard cotransfection protocols with the mutant αB-crystallin-SR plasmid and pJM17, a plasmid containing the entire adenovirus genome, into 293 cells (26). Virus was plaque purified, amplified, CsCl gradient purified, and dialyzed before spectrophotometric titering (OD260 × dilution × 1010 = particles/ml). Cells were infected with 100 particles/cell for 2 h in maintenance medium (DMEM-M199, 4:1 vol/vol) with 2% FCS. The medium was then changed to no-serum maintenance medium. Experiments began ∼40 h later.
Simulated ischemia and enzyme quantitation.
Initial simulated ischemia experiments were performed by placing the cells in a hypotonic balanced salt solution (in mM: 1.3 CaCl2, 5 KCl, 0.3 KH2PO4, 0.5 MgCl2, 0.4 MgSO2, 69 NaCl, 4 NaHCO2, and 0.3 Na2HPO4) without glucose or serum. The plates were then placed in an airtight jar containing the oxygen-consuming GasPak system (BBL Microbiology Systems, Franklin Lakes, NJ) and flushed for 5 min with argon to rapidly achieve <0.2% O2. After 14 h, the dishes were removed from the chamber with the medium, and the cells were each assayed separately for enzyme content. Subsequent experiments, using normotonic ischemic buffer, produced identical ischemic protection results. After the medium was removed, the cells were scraped into 1 ml of cold PBS and sonicated (ultrasonic homogenizer 4710, Cole-Parmer, Vernon Hills, IL) for 15 s followed by 10 min of centrifugation at 150 g. The creatine phosphokinase (CK) in the medium and from the sonicated cells was quantified with a CK determination kit (Sigma, St. Louis, MO). From these values the percentage of CK released was calculated, and then these values were normalized to the enzyme released by control cells that were also made ischemic. This allowed us to pool data from several experiments that might have variable total CK release, because of cell plating variability and differences in the severity of the ischemic stress.
A whole cell toxicology assay kit (Sigma) was used to permit the measurement of cell viability as a function of mitochondria activity. This method is based on 2,3-bis [2-methoxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-carboxanilide inner salt (XTT) cleavage. Functional mitochondrial dehydrogenase in the viable cells cleaves the tetrazolium ring of XTT, which produces orange formazan crystals that are measurable spectrophotometrically. Fifty thousand cells were plated in each well of a 96-well plate, infected, and stressed in the same manner as described above except that after ischemia the buffer was changed to 80 μl of isotonic balanced salt buffer containing 10 mM glucose and 20 μl of the XTT solution was added. The plates were placed back in the incubator and read at 450 nm 4 h later with a water-containing well for a blank. These data were analyzed in the same manner as the CK values.
For denaturing PAGE, cells were harvested in solution B (in mM: 20 NaCl, 20 Tris pH 7.5, and 0.1 EDTA) containing 1% Triton X-100, 0.5% deoxycholate, and 5 μM 2-mercaptoethanol. The samples were vortexed vigorously and placed on ice for 15 min before 15-min centrifugation at 12,000 g. The concentration of the supernatant was determined by Bradford assay, and equal amounts were loaded on either Bio-Rad minigels for Western analysis or larger (16 cm) Hoefer setups for labeled extracts and native PAGE. For the [35S]methionine-labeled extracts equal trichloroacetic acid-precipitable counts were fractionated through a 12% SDS-PAGE, which was fixed, enhanced, dried, and exposed to film for the appropriate length of time at −70°C. The phosphospecific rabbit polyclonal anti-αB-crystallin (serine 59) was generously provided by K. Kato (Institute for Developmental Research, Aichi Human Service Center, Aichi, Japan) (15) via P. Eaton (Rayne Institute, St. Thomas' Hospital, London, UK) (10).
For the native PAGE, cells were washed with cold PBS and then harvested in 10 mM HEPES pH 7.5, 130 mM NaCl, 0.5 mM phenylmethylsulfonyl fluoride, 20 μg/ml aprotinin, and 20 μg/ml leupeptin. These extracts were then sheared repeatedly through a 27-gauge needle, and their concentrations were determined by Bradford assay. Twenty micrograms of protein were loaded on a 5% PAGE for 1,000 V-h and then electrotransferred to nitrocellulose for immunoblotting. The primary antibody used was the rabbit polyclonal anti-αB-crystallin (SPA 223; Stressgen, Victoria, BC, Canada) with a horseradish peroxidase-conjugated secondary anti-rabbit IgG and an ECL detection kit (Amersham, Piscataway, NJ). The calculation of the size of the αB-crystallin complexes was based on the pooling of four independent native gels. For the linear extrapolation, Coomassie blue-stained apoferritin (450-kDa monomer, 900-kDa dimer, 1,350-kDa trimer) and urease (272-kDa trimer, 545-kDa hexamer; Sigma) were used as standards.
Microtubular integrity assay.
In brief, after 14 h of normotonic simulated ischemia, cells were methanol fixed and then washed with PBS before 20-min blocking with 10% goat serum-3% BSA in PBS. After being washed with PBS the cells were incubated for 1 h in 1:1,000 diluted monoclonal anti-tubulin (T5168, Sigma) in PBS containing 3% BSA. After additional PBS washing the cells were incubated with FITC-conjugated goat anti-mouse antibody at 1:100 for 1 h. After final PBS washes the slides were aspirated dry and sealed with coverslips and Vectashield mounting medium (Vector Labs, Burlingame, CA). Confocal microscopy was performed with a Bio-Rad MRC 1024 confocal microscope. The 488-nm-wavelength line of the krypton-argon laser was used for excitation, and the emitted light was collected at 522 nm. Image acquisition software settings (laser intensity, iris size, gain) were strictly maintained for all images taken on the same day to allow quantitative, paired comparisons. The Laser Sharp software of the confocal microscope was used for image analysis. The width of the pixel intensity histogram was found to be a reliable and sensitive measure of microtubule integrity. Further details were published previously (6).
One-way ANOVA with post hoc Bonferroni t-test was used to determine statistic significance of CK release (see Fig. 2) and XTT bioreduction (see Fig. 3). Microtubular stability (see Fig. 4) comparisons utilized repeated-measures ANOVA, the equivalent of paired comparisons for multiple groups. Paired statistics were used because crucial parameters such as myocyte plating or the severity of ischemia were matched for all cells analyzed on any one day but were subject to day-to-day variation.
Because of the increased interest in the role of αB-crystallin in the cardiovascular setting and particularly the importance of its COOH terminus in protective function, we mutated the two terminal lysines to glycines and then mounted the mutant and wild type in adenovirus for use in cardiomyocytes. Figure1 A compares the expression and size of denatured mutated and wild-type αB-crystallin by Western blot analysis of a minigel-PAGE. In Fig. 1 B, we used [35S]methionine-labeled cell extracts to show the slight decrease in the size of the denatured αB-crystallin associated with the mutation of these two lysines to glycines on a larger-resolving PAGE. Figure 1 C depicts the size of the native αB-crystallin complexes as examined by 5% nondenaturing PAGE and Western immunoblot. There is a dramatic downward shift with the COOH-terminal mutant from 1,100 ± 138 to 610 ± 125 kDa (n = 5). The broad nature of the bands may be due to the heterologous nature of the complexes.
After having shown equivalent overexpression and a dramatic decrease in the size of the native complex with the lysine mutant, we next examined it in our simulated ischemia model. Figure2 confirms the 33% decrease in CK release we saw previously (25) with overexpression of αB-crystallin. However, the expression of the COOH-terminal mutant produced a 30% increase in ischemic damage compared with control and a 63% increase compared with wild-type αB-crystallin-overexpressing cells. To further confirm these results, we looked at another marker of cellular damage, the reduction of XTT, an indicator of functional mitochondrial dehydrogenase. These results are displayed in Fig. 3 and again show a complete loss of any ischemic protection with the lysine to glycine mutant.
We next immunoprecipitated extracts from control, αB-crystallin-, and mutated αB-crystallin-expressing cardiomyocytes with antibody to αB-crystallin to examine any differences in association with the intermediate filament proteins desmin, tubulin, and vimentin. There is evidence that the interaction of αB-crystallin with these proteins may be altered under ischemic conditions in perfused hearts (3). We could not detect any difference with the mutant via immunoprecipitation and subsequent Western blot analysis even with extracts from ischemic cells (data not shown). This may be due to the particular models, isolated cells vs. whole heart, or a limitation of the immunoprecipitation methods used, or perhaps this mutation does not affect this interaction.
Therefore, we next used a highly sensitive assay we recently developed (6), a quantitative method for analyzing cytoskeletal integrity based on immunostaining and confocal microscopy. This method examines the integrity of the entire cytoskeletal array. In our previous study (6) we found the ischemia-induced microtubular disruption to be similar to that caused by colchicine treatment as assessed by this method of quantifying immunostaining by measuring the pixel intensity scatter. A series of experiments was performed comparing the microtubular integrity of control, αB-crystallin-, and COOH-terminal mutant-overexpressing cardiomyocytes after ischemic stress. The results are summarized in Fig. 4. This demonstrates that the additional stabilization of the microtubular array against ischemic stress with wild-type αB-crystallin overexpression (vs. control) is lost with the lysine to glycine mutations.
Recently Ito et al. (15) demonstrated that phosphorylation of αB-crystallin decreases its oligomerization and chaperone activity. Thus we examined our COOH-terminal mutant for alteration in basal phosphorylation by Western blot with a phosphospecific antibody to serine 59 of αB-crystallin. Figure5 demonstrates that the mutant appears phosphorylated, although in this particular experiment the ectopic expression of both the wild type and the mutant is lower than in Fig.1.
The major findings of this study are that 1) the COOH-terminal lysines of αB-crystallin are required for protection against ischemic damage in cardiomyocytes, 2) the mutation of these two lysines produced a significant decrease in the size of the native aggregate, and 3) this alteration also diminishes its stabilization of the microtubular cytoskeleton.
Previous work, primarily focused on the role of αB-crystallin in the eye, indicated that the integrity of the COOH terminus was important for its proper function. Bacterial thermokilling assays support the importance of the COOH-terminal lysines by mutagenesis for proper αB-crystallin function (29). Similarly, protease truncation experiments indicate this region is required to prevent heat-induced denaturation (33). More recently, a random mutagenesis approach followed by in vitro gal-lex binding experiments showed diminished αB-crystallin function with mutations in the COOH-terminal domain, suggesting that the charge of this region stabilizes the complex (7). These approaches all use nonmammalian systems to examine a mammalian protein, which may affect its native confirmation. The studies described in this present report examine the same region, by altering its basic nature to a more neutral one, in primary cardiac myocytes while determining its effect on ischemic outcome, oligomerization, and cytoskeletal stabilization. We determined that mutating the two COOH-terminal lysines to glycines was detrimental to cardiomyocyte survival of ischemic stress by cytosolic enzyme release (Fig. 2). Interestingly, the 30% increase in CK release with the mutant αB-crystallin is very similar to the increased release with the antisense to another smHSP (HSP27) intervention in this same model (25). Our initial experiments (examining CK release) even suggested a dominant-negative type function for the mutant (Fig. 2), but the results from the other ischemic end points (Figs. 3 and4) seem to indicate more a loss of the protective function associated with αB-crystallin overexpression than interference with endogenous αB-crystallin.
The oligomerization behavior of the smHSP also contributes in an important way to its function. In the current study we suggest that the charge of the COOH terminus may affect the size of the oligomer. Our COOH-terminal mutant had a significant downward shift in oligomeric size. It should also be noted that the native Western blots did not show a signal for the endogenous αB-crystallin (∼1,100 kDa) whenever the mutant αB-crystallin (∼610 kDa) was present. Whether this is due to endogenous αB-crystallin oligomerizing with the mutant αB-crystallin in complexes with fewer subunits or a limitation of the method used here we cannot say.
In contrast to our results, Plater et al. (29) found no difference in the oligomerization of their K174/175 αB-crystallin mutants. This may be due to the difference between utilizing bacterially expressed recombinant protein studies and our in situ cardiomyocyte approach. Also, we should point out that the murine αB-crystallin used by Plater and coworkers, somewhat uniquely, has an alanine rather than a threonine at 170, so it would not be glycosylated in the same manner. Thus, if our mutant affected the glycosylation state and this affected function and/or oligomerization, this would help explain the different outcome.
Phosphorylation and oligomerization.
Although our present study does not directly address the question of the phosphorylation of αB-crystallin and its oligomerization, it does deserve comment. Our COOH-terminal mutant, with decreased oligomerization, does not appear abnormally phosphorylated. Ito et al. (15) demonstrated that constitutively phosphorylated αB-crystallin retained a reduced chaperone function and decreased aggregate size. In contrast, Eaton and coworkers (10) found that increasing the phosphorylation of αB-crystallin with phenylephrine did not alter the aggregate size.
Golenhofen et al. (13) observed a slight increase in αB-crystallin phosphorylation with ischemia in the whole heart and suggested that it may be protective of the contractile apparatus. However, in a follow-up study this group did not find dephosphorylation of αB-crystallin concomitant with its displacement from ischemia-induced myofibrillar binding (12). Also, in plants the chloroplast-localized smHSP HSP21, which oligomerizes in vivo, does not appear to be phosphorylated at all (29). Two additional studies specifically addressed whether it is possible that this smHSP is an effector for ischemic preconditioning in the heart. Armstrong et al. (2) found an increase in αB-crystallin phosphorylation and translocation with ischemia-reperfusion injury in isolated cardiomyocytes that was not augmented with ischemic preconditioning. However, Eaton and coworkers (9) did find a correlation between the degree of αB-crystallin phosphorylation and translocation with ischemic preconditioning in perfused rat hearts. These models are all different from each other and from our own model, but they suggest that although the phosphorylation of αB-crystallin is an important posttranslational modification, other factors may also determine aggregate size.
Truncation and oligomerization.
A recent study of αB-crystallin from human cataracts found a disproportionately high amount of a truncated form, specifically missing the COOH-terminal lysine, which also had an altered oligomerization state yet was still effective in preventing the aggregation of purified lactalbumin (17). This naturally occurring truncation has properties similar to our lysine-to-glycine mutant except for its retained protective effect in an in vitro assay. How can we reconcile increased cataract formation with lysine truncation that is still able to prevent the malfolding of lactalbumin? Koretz et al. (21) suggested that one can separate the chaperone-like activities of α-crystallin (i.e., aggregation studies) from its actual in vivo function. They showed that the chaperone function is much reduced under more physiological conditions (i.e., in the presence of unchelated divalent cations) so it may not be as relevant within the cell. This is incentive enough to examine results from reconstitution assays via more applicable models.
Stabilization of cytoskeleton.
It has become apparent there are several smHSPs present at high levels in the cardiovasculature, with αB-crystallin having specific cytoskeletal associations (2, 3, 5, 6, 12, 13, 15, 23). It has been proposed via immunolocalization experiments that αB-crystallin relocates with ischemia to the Z lines of the myocardium, where it could serve as a stabilizer (13, 23). In isolated cardiomyocytes subjected to heat stress, αB-crystallin relocates to sarcomeric structures (34). There are also data suggesting that αB-crystallin interacts in a specific manner with various intermediate filaments in the cardiovasculature (3,5) and in other cell types (37). The protection of the ischemia-induced disruption of the microtubular cytoskeleton associated with ectopic expression of αB-crystallin is abrogated when the expressed αB-crystallin has its two COOH-terminal lysines mutated (Fig. 4). This also supports a scheme with αB-crystallin, particularly in the cardiovascular setting, playing a significant role in cytoskeletal stability, especially in responses to stress.
There is also some evidence that the COOH-terminal portion of αB-crystallin is in the part of the complex that interacts with malfolded cellular proteins (32, 33). It is possible that the lysine to glycine mutation leads to the loss of protection against ischemic injury via decreased ability to interact with partially malfolded proteins and thus potentially attenuate the denaturing process.
In conclusion, our studies indicate that the protective effect of αB-crystallin expression against ischemic stress, as assessed by two separate end points, is abrogated by mutating the two COOH-terminal lysines of crystallin. This same mutation caused a dramatic decrease in the oligomerization status of αB-crystallin. A consequence of this mutation was the loss of the preservation of the microtubular cytoskeleton against ischemia-induced disruption associated with ectopically expressed αB-crystallin. In summary, we believe the COOH terminus region of αB-crystallin is important for its proper function.
These studies were supported by the Ralph and Marian Falk Trust for Medical Research (J. L. Martin), National Heart, Lung, and Blood Institute (NHLBI) Grant F32-HL-10042 (W. F. Bluhm), American Heart Association Grant GIA 0050276N (R. Mestril), and NHLBI Grant R37-HL-49434 (W. H. Dillman).
Address for reprint requests and other correspondence: J. L. Martin, Cardiovascular Institute, Dept. of Medicine, Loyola Univ. Medical Center, 2160 S. First Ave., Maywood, IL 60153 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published March 14, 2002;10.1152/ajpheart.00512.2001
- Copyright © 2002 the American Physiological Society