Human Molecular Genetics, 2003, Vol. 12, No. 13 1609-1620
DOI: 10.1093/hmg/ddg173
© 2003 Oxford University Press
Distinct chaperone mechanisms can delay the formation of aggresomes by the myopathy-causing R120G 
B-crystallin mutant
Centre de recherche en cancérologie de l'Université Laval, L'Hôtel-Dieu de Québec, 9 rue McMahon, Québec, Canada G1R 2J6
Received March 10, 2003; Revised April 30, 2003; Accepted May 7, 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
A familial form of desmin-related myopathy (DRM) is associated with a missense mutation (R120G) in
B-crystallin (B) and is characterized by intracellular desmin aggregation. Because B is a molecular chaperone that participates in the assembly of desmin filaments, it has been suggested that the desmin aggregation might be due to the loss of B function. We report here that BR120G has indeed impaired in vivo function and structure as reflected by a highly reduced capacity to protect cells against heat shock and by an abnormal supramolecular organization even in cells not expressing desmin. In many cells, BR120G accumulated in inclusion bodies that had characteristics of aggresomes concentrating around the centrosome following a microtubule-facilitated process. Three distinct chaperone mechanisms could reduce or even prevent the formation of the BR120G aggresomes. Wild-type B and Hsp27 prevented aggresome formation by co-oligomerizing with BR120G. Hsp70 with its co-chaperone Hdj-1 or Chip-1 but not a mutant of Chip-1 lacking ubiquitin ligase activity, reduced the frequency of aggresome formation likely by targeting BR120G for degradation. Finally, HspB8 interacted only transiently with B but nonetheless rescued the BR120G oligomeric organization, suggesting that it acted as a true chaperone assisting in the folding of the mutant protein. Hence, the formation of inclusion bodies in BR120G-mediated DRM is probably due to the misfolding of BR120G per se and can be delayed or prevented by expression of the wild type B allele or other molecular chaperones, thereby explaining the adult onset of the disease. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
A missense mutation in
B-crystallin (B) changing arginine 120 to glycine (BR120G) has been identified in a French family as the cause of an autosomal-dominant desmin-related myopathy (DRM) also called B-crystallinopathy. This neuromuscular disease is characterized by the formation of large aggregates containing desmin and B in skeletal and cardiac muscles (1,2). The reason for the development of the disease is not clearly understood. B is a heat shock protein (Hsp) member of a group of nine related proteins collectively known as small Hsp (sHsp) of which, A-crystallin (A), B and Hsp27/HspB1 are the best-studied members (3–5). B is expressed constitutively at high level in lens and muscle and, like many other Hsp, its expression increases several-fold after heat shock in most cell lines and tissues (6–8). Also like many Hsp, B has chaperone properties in vitro, protecting proteins against heat-induced denaturation and aggregation (9), and in vivo confers cellular protection against heat shock and other stress (10,11). It was suggested that, in -crystallinopathy, desmin collapses with BR120G due to a reduced chaperone activity of BR120G, which would be necessary for the proper organization of the desmin filaments (1,12–16). In support of the idea that the disease was related to a loss of function of B and to the disorganization of desmin filaments, a similar form of DRM has also been associated with missense mutations in desmin (17,18). In desmin mutant- as well as BR120G-induced DRM, the disorder has an adult onset and is characterized by the accumulation of protein aggregates of desmin and B. Furthermore, forced expression of BR120G in desmin-expressing cells as well as targeted expression of BR120G in the heart of mice, in both cases, led to formation of inclusion bodies containing both desmin and B and, in the animal, to the development of myopathy (1,19). Finally, B-null mice develop a form of myopathy. The last result was, however, difficult to interpret because the mice also suffered from a disruption of the gene coding from another small heat shock protein highly expressed in muscle, MKBP/HspB2 (20).
Intriguingly, desmin- and B-null mice and as well as mice expressing desmin mutants all have a much less severe myopathy than those expressing BR120G, suggesting that BR120G-induced myopathy might result from more than a loss function of desmin or even B (19). B is known to be quite unstable. In the lens, B precipitates into inclusion bodies in the absence of its oligomerization partner, A-crystallin (21). A possibility is that the B-desmin inclusion bodies that form in B-crystallinopathy result from the misfolding and progressive aggregation of B to which subsequently associate desmin filaments and that the toxicity results directly from the accumulation of aggregates in the cells. Misfolding as a result of mutation followed by the aggregation of the misfolded proteins into inclusion bodies to which associate intermediate filaments is not unique to DRM and accompanies many protein conformational diseases (22,23). These aggregates can by themselves be toxic, inhibiting the ubiquitin–proteasomal system of protein degradation (24).
Hsp chaperones, including wild type B, are up-regulated and accumulate into inclusion bodies in many protein conformation diseases. For example, B is also found in Rosenthal fibers of Alexander disease, in Mallory bodies of alcoholic liver disease, and in cortical Lewy bodies and Alzheimer disease plaques and neurofibrillary tangles (5,8,25–27). The exact reason for the frequent association of B with these structures is not totally clear but probably results from the chaperone activity of B. Hsp chaperones are part of the quality control mechanism of the cells that aims to prevent the aggregation of proteins, which may either be induced as a result of thermal or oxidative stress, occur spontaneously in the process of translation or supramolecular organization, or result from misfolding caused by mutation. For example, Hsp70 can prevent heat-shock-induced aggregation of protein, thereby protecting against heat-shock-induced cell death (28,29). Furthermore, in experimental models of degenerative diseases, Hsp70 with the co-chaperones Hdj1 or Hdj2 can suppress the aggregation of the misfolded proteins and protect against toxicity (30–33). The mechanism of Hsp70-mediated protection appears to involve either the assisted refolding or the ubiquitin-dependent degradation of proteins (34–36). In addition to Hdj1/2, which associates to Hsp70 to regulate its ATPase activity (37), a key co-chaperone of Hsp70 in this process is Chip-1, an ubiquitin E3 ligase that uses the Hsp70–Hdj1 chaperone complex as an adaptor to detect misfolded proteins and tag them for proteasomal degradation by ubiquitylation (38–40). Although not yet clearly demonstrated, it is possible that sHSP like B accomplish in vivo a similar role to limit the formation or the growth of the aggregates (41).
B-crystallinopathy is thus a special case of protein conformation diseases in which the destabilizing mutation at the origin of the disorder occurs in a molecular chaperone, itself potentially involved in the protein quality control of the cells. Here we studied the effect of the R120G mutation using as experimental systems non-muscular cell lines expressing little endogenous B and no desmin. We showed that BR120G does indeed suffer severe supramolecular disorganization that affects its protective role in the cells, however, BR120G forms aggresome-type of inclusion bodies independently of the presence of desmin. The extent and rate of BR120G aggresome formation is modulated by the presence of the normal B allele and also by other molecular chaperones, which either assist in the correct folding and supramolecular organization of BR120G or target it for degradation. These results provide a molecular basis explaining the adult onset of the B-crystallinopathy.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Altered structure and function of
BR120GRecombinant
BR120G purified from bacterial expression systems forms in vitro polydispersed structures of molecular weight higher than 1000 kDa, which is in contrast with the uniform ca 600 kDa multimer formed by the wild-type protein (14–16). To determine whether such an anomaly also occurred in vivo, the wild-type and mutant proteins encoded by the plasmid pCRYSBWT and pCRYSBR120G, respectively, were expressed in NIH3T3 (a cell line that expresses no detectable level of small heat shock proteins) (Fig. 1A) or PTK2 cells (Fig. 6). In both cell lines, B was found entirely in the soluble fraction, whereas 50% of BR120G was in the insoluble fraction (data not shown and Fig. 5). The soluble extracts were ultracentrifuged on a glycerol gradient and the distribution of the proteins along the gradient was determined by western blot using the anti-B antibody. B sedimented mainly with an apparent molecular mass of some 540 kDa, corresponding approximately to the size predicted for a 24-mer of a 22 kDa molecule. BR120G showed no distinctive peak in that region. Instead, the protein sedimented as highly heterogeneous species of molecular weight of 2000 kDa and larger (Fig. 1A). These major structural alterations observed in BR120G were reflected at the functional level. In contrast to expression of B, which provided a strong resistance to heat shock, expression of BR120G produced very little protection (about 70% reduction) (Fig. 1B). The lack of protective activity of BR120G is consistent with its reduced chaperone activity measured in vitro (14–16).
|
|
|
Aggregate formation by the
BR120GThe R120G missense mutation is responsible for a form of DRM characterized by the accumulation of aggregates containing desmin. Transfection of
BR120G in desmin expressing cells was previously found to cause the formation of aggregates of desmin and B (1). This could be due either to the failure of desmin to correctly organize in the absence of a functional B or to the aggregation of the misfolded BR120G to which desmin subsequently associates. To test this idea, the plasmids pCRYSBWT and pCRYSBR120G were transfected in NIH3T3, CCL39 and PTK2 cells to express wild-type B and BR120G, respectively. None of these cells express desmin. The cells were fixed at various times after transfection and then processed for immunofluorescence as described in the Materials and Methods. In all cell lines investigated wild-type B distributed uniformly within the cytoplasm whereas BR120G mutant formed aggregates in some 30–40% of the cells (Fig. 2). At early times, the aggregates tended to be small and dispersed all over the cytoplasm. With times, larger and larger aggregates were observed and at 72 h large amorphous masses were found in the perinuclear area of most cells often distorting the nuclear envelope (Fig. 2A and also Fig. 4). To support the conclusion that the formation of the aggregates was a direct consequence of the misfolding of BR120G, we examined the consequence of a similar mutation in another small heat shock protein, Hsp27. Ectopic expression of the Chinese hamster Hsp27 mutant R148G in NIH3T3 or CCL39 cells similarly resulted in the formation of perinuclear aggregates in some 30% of the transfected cells (data not shown).
|
|
BR120G forms aggresomesThe
BR120G aggregates resembled aggresomes, amorphous structures that form around the centrosome by the accumulation misfolded proteins when the proteasomal degradation system is saturated. Aggresomes often contain ubiquitin, heat shock proteins and proteasomal proteins and are wrapped by intermediate filaments (42–44). The position of the BR120G perinuclear aggregates relative to the centrosome was determined using 
-tubulin as a localization marker. Double-immunofluorescence for B and -tubulin indicated that the large perinuclear aggregates of BR120G were at or around the centrosome in more than 90% of the cases (Fig. 3A). We also determined that the BR120G aggresomes contained ubiquitin by co-transfecting Myc-tagged ubiquitin and BR120G. Immunofluorescence microscopy revealed a co-localization of myc-ubiquitin with BR120G (Fig. 3B). This is consistent with the fact that ubiquitin is present in the protein aggregates observed in almost all degenerative diseases (32,42,45–49). The effect of BR120G aggregation on intermediate filaments was investigated in PTK2 cells, a cell line of epithelial origin that has two independent intermediate filament networks, one formed by vimentin and the other by keratin 8 and 18. Analyses by optical sectioning on a confocal microscope indicated that both keratin (Fig. 3C and D) and vimentin (Fig. 3E and F) formed a cage around the aggregates of BR120G. However, filament proteins were not present within the aggregates, and neither filament network was disturbed outside of the aggregate. Actin and -tubulin filaments were also displaced by the aggregates but neither filament was closely associated with the surface of the aggregates as seen for intermediate filaments (data not shown). The results demonstrated that the aggregates formed by BR120G can be surrounded not only by intermediate type III filaments like vimentin but also by type I and II filaments like keratins, suggesting a general response of intermediate filaments networks to the aggregated proteins.
|
Aggresomes are formed from the precipitation of denatured proteins at multiple foci in the cell. These small nuclei of denatured proteins are then retro-transported along the microtubules to end up into large peri-centrosomal masses (42–44). We looked in more details at the kinetics of formation of the
BR120G aggregates in the presence or absence of colchicine (Fig. 4). PTK2 cells were transfected with pCRYSBR120G and treated for 24 or 48 h with 10 µg/ml colchicine. At this concentration, colchicine caused a severe disruption of the microtubule. With times after transfection, presumably as the concentration of the transfected proteins increased, there was an increasing number of cells showing precipitates of BR120G, going approximately from 20% at 24 h to 50% at 48 h (top panel). These numbers were not affected in a major way by colchicine, or perhaps only slightly increased. There was, however, a major effect of colchicine on the dynamics of formation of the aggresomes. As illustrated in Figure 2 and quantified in Figure 4 (lower panel), in control situations at 24 h after transfection, precipitates were present as small spots scattered all over the cytoplasm in more than 70% of the cells in which precipitates of BR120G were detected. At that time, large perinuclear aggregates were seen in less than 30% of the cells. The situation was at the opposite at 48 h. Then, 70% of the cells had one or a few large aggregates located on one side of the nucleus, and small spots were seen in less than 10% of the cells. Colchicine produced a major effect on the growth and relocalization of the small nuclei into the large perinuclear aggregates. In the presence of colchicine, large aggregates were seen in only 30% of the cells at 48 h and most of the cells had small foci or a combination of small and large spots still scattered in the cytoplasm. The result thus strongly suggested that nuclei of denatured BR120G first appear as small spots that coalesce into one or two large amorphous perinucleosomal aggregates with the help of the microtubule. In comparison, cytochalasin D at a concentration that results into the de-polymerization of actin filaments had little effects on the growth or localization of the aggregates (data not shown).
Modulation of aggresome formation
In the pathogenic situation in vivo, a single allele of B is mutated in such a way that the wild-type protein is co-expressed with BR120G. In the transfection experiments, the level of expression of the mutant largely surpassed that of the wild-type endogenous protein. We therefore examined the effect of expressing wild-type B on aggresome formation by BR120G. In order to determine the relative level of expression of the two proteins, we used a myc-tagged version of wild type B, which was distinguishable from BR120G due to its reduced electrophoretic mobility. In the absence of B, 50% of BR120G was recovered in the insoluble fraction. At a level of expression of B equivalent or even lower than that of BR120G, BR120G was expressed almost exclusively in the soluble fraction (Fig. 5A) and the formation of large aggregates, which was the most frequent type of precipitates seen at 48 h in the cells transfected with BR120G alone, was totally inhibited (Fig. 5B). Some small precipitates were, however, still present in a reduced proportion of the cells (Fig. 5C). The results suggested that, in the presence of the wild-type proteins, the misfolded mutant could start to precipitate but that the process of coalescence and growth of the foci into large aggregates was considerably slowed down. This was likely due to the formation of mixed oligomers between the two proteins since the sedimentation profile of BR120G was completely rescued in the presence of B (Fig. 6A) and the two proteins co-immunoprecipitated and co-sedimented (Figs 5D and 6A). The reason for the higher apparent molecular weight of the B oligomers in PTK2 as compared with 3T3 cells (Fig. 1) was not investigated.
Another protein of the sHsp family, Hsp27, can form hetero-oligomers with native B (50–52) and, as expected, was found here to be immunoprecipitated with wild-type B (Fig. 5D). We therefore examined the effect of expressing Hsp27 on the formation of BR120G aggresomes. Hsp27 produced an effect identical to that of wild-type B. Hsp27 caused a solubilization of BR120G (Fig. 5A) and a drastic reduction in the frequency of formation of BR120G aggresomes (Fig. 5B). In fact, the formation of large perinuclear aggregates was totally inhibited, and in the few cells where precipitates of BR120G were still seen, the aggregates were small and scattered throughout the cytoplasm (Fig. 5C). Furthermore, BR120G was also rescued to form oligomers of some 600 kDa in the presence of Hsp27 (Fig. 6B). The two proteins co-sedimented and co-immunoprecipitated suggesting that they formed hetero-oligomers (Figs 5D and 6). Note that in the presence of Hsp27 a small proportion of BR120G sedimented at sizes corresponding to dimers or tetramers suggesting an association with the small oligomers of Hsp27 (53).
Although Hsp27 and B have in vitro chaperone activities, the fact that they interact with wild-type B as well as with BR120G and remain associated with the final oligomeric structures of BR120G makes it impossible to determine whether this activity is involved in the refolding of BR120G. By definition a chaperone should not remain associated with the final structure (54). We therefore looked at the effect of another member of the small heat shock protein family, HspB8 (4,55). In contrast to Hsp27 and B, HspB8 did not form mixed oligomers and co-immunoprecipitated only very weakly with wild type B or Hsp27 (Fig. 5D and data not shown). However, HspB8 co-immunoprecipitated quite strongly with the mutant BR120G (Fig. 5D), suggesting a capacity for HspB8 to recognize badly folded proteins. Interestingly, HspB8 also solubilized BR120G (Fig. 5A), was as efficient as Hsp27 to prevent the formation of large aggregates (Fig. 5B and C), could totally escue BR120G in its capacity to form normal 600 kDa oligomers and was not found in the final BR120G oligomers (Fig. 6C). Transfected alone as well as with BR120G, HspB8 sedimented with the apparent size distribution consistent with the formation of dimer or tetramers.
We next examined whether overexpression of HSP70, a protein with demonstrated in vivo chaperone activity, could also block aggresome formation. In contrast to Hsp27, which was co-immunoprecipitated with B as well with BR120G, but similarly to HspB8, Hsp70 interacted only with denatured BR120G, not with native B (Fig. 5D). Transfection of HSP70 also reduced the amount of insoluble BR120G (Fig. 5A) and caused a major reduction in number of cells with BR120G aggresomes, an inhibition that amounted in several experiments to at least 50% (Fig. 5B). There were major differences, however, in the effect of Hsp70 compared with those of B, Hsp27 or HspB8. In contrast to the sHsp, Hsp70 produced an all-or-none inhibition. In cells that still had precipitates of BR120G, there were large perinuclear aggregates, which in many cases had trapped some HSP70 (Fig. 5C). Moreover, HSP70 did not facilitate the formation of a normal 600 kDa structure by the mutant protein. The fraction of BR120G that could be analyzed by sedimentation showed very little difference in sedimentation profile in the presence or absence of Hsp70 (Fig. 6D). This implied that the mechanism by which hsp70 interferes with aggregates formation was distinct from that involved in inhibition by B, Hsp27 or HspB8.
Hsp70 requires co-chaperone proteins to operate. Since it is expressed at a high constitutive level in the PTK2 cells (data not shown), it is possible that the failure of Hsp70 transfection to totally prevent aggresome formation was due to the low abundance of co-chaperones. We therefore tested whether the co-chaperones Hdj-1 and Chip-1 could enhance the protective effect of HSP70. Hdj-1 modulates the ATPase activity of Hsp70 whereas Chip-1 is an ubiquitin E3-ligase that ubiquitylates and targets for degradation badly folded proteins after they are recognized by Hsp70 (37,38). Transfection of Hdj-1 or Chip-1 alone, both reduced by some 80% the number of cells with BR120G aggresomes (Fig. 7A) and, as expected from their known role as co-chaperone of Hsp70, they produced the same all-or-none effect as Hsp70, i.e. in the few cells where aggresomes formed, their size were as large as in cells transfected with BR120G only (Fig. 7B). Furthermore, transfection of Hdj-1 or Chip-1 alone or together with Hsp70, like Hsp70 alone, did not normalize the sedimentation profile of BR120G (data not shown). A mutant of Chip-1 lacking the U-box, a domain required for ubiquitylation, was ineffective in preventing aggresome formation (Fig. 7A), suggesting that prevention of aggresome formation was a consequence of the targeted degradation of misfolded BR120G, rather the facilitated folding of the mutant proteins. However, we were unable to detect a reduction in the total expression of BR120G (insert in Fig. 7), or an increase in the degradation rate of the mutant proteins in the presence of Chip-1. Metabolic pulse labeling of total proteins followed by analyses of immunoprecipitated B and BR120G proteins during a chase period revealed that BR120G was degraded at a rate not significantly different from B (t1/2 of about 25 h) and that this rate was not significantly affected by the presence or absence of transfected Chip-1 (data not shown). The failure to detect changes in the global degradation rate of BR120G in the presence of Chip-1 might be explained if only a small proportion of BR120G was badly folded and therefore the target of Chip-1. This is a likely possibility considering that only 35–50% of the cells that expressed BR120G showed evidence of denatured proteins; hence the bulk of BR120G present in the other cells may be folded correctly enough so that it is not recognized by the chaperones.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The accumulation and aggregation of misfolded proteins in the cells is at the origin of several human degenerative diseases. In many examples, this occurs as a consequence of mutation, deletion or other alterations in the protein primary structure, which prevent correct folding or assembly into multimeric structures (22,23). It can also result from a defect in the cellular ability to eliminate misfolded proteins as they inadvertently accumulate in the process of protein translation or supramolecular organization (56). These processes are initiated by a group of molecular chaperones with the capacity to recognize and bind hydrophobic regions accidentally exposed in proteins. This binding of chaperones to denatured protein can prevent their aggregation, assist in their correct folding or else, it can mediate the targeting of the proteins for degradation before they aggregate (22,35,57).
B-crystallinophathy, the DRM caused by BR120G (1,2), is a very special case of degenerative disease in which the mutated protein causing the disease is itself a molecular chaperone whose wild-type version is already associated with aggregated proteins in several other degenerative diseases (5,8,25–27). For this reason, it was not clear what was responsible for the disease: the misfolding and aggregation of BR120G or the loss of the chaperone function of B that indirectly results from this misfolding and which may cause other proteins to misfold and aggregate. The last possibility was previously suggested because desmin filaments, which seem to require B to organize properly in the cells, accumulated with BR120G in the DRM inclusion bodies (1,14–16,19). Here, we confirmed the defective supramolecular organization and function of BR120G. BR120G failed to provide thermoprotection when overexpressed in vivo and furthermore formed huge masses larger than 2000 kDa instead of the 24-mer of some 600 kDa normally formed by wild-type B. Nevertheless, our results strongly suggested that the formation of the inclusion bodies is not due to a loss-of-function of B. None of the cell lines investigated here expressed desmin, indicating that formation of inclusion bodies in cell expressing the BR120G was not due to the failure of the B-crystallin mutant to assist in the organization of the desmin filaments. B may be involved in the organization of other intermediate filaments (12,13). However, neither the vimentin nor the keratin networks were disrupted in cells transfected with BR120G, excluding the possibility of a major defect in intermediate filament organization in the presence of the mutant B. Hence, the formation of the inclusion bodies appears to be a direct consequence of BR120G misfolding. Arginine 120 is a highly conserved residue in all small heat shock proteins and was suggested to be important for the structural integrity of these proteins. In vitro studies performed with purified BR120G or equivalent mutant (R116C) of A-crystallin revealed defects in the secondary, tertiary and quaternary structure of the proteins (14–16). The A-crystallin mutant caused congenital cataract in human (58). Moreover, equivalent conversion in Chinese hamster Hsp27 (arginine 148 to glycine) also destabilized the proteins that ended up in inclusion bodies when expressed in cell lines (data not shown). The fact that BR120G aggregates in the absence of desmin does not exclude the possibility that desmin may also be unstable in these cells and incapable of organizing into a normal filament network. The presence of desmin in the aggregates may occur if BR120G still associates with unpolymerized desmin and brings it down into the inclusion bodies.
Accumulation of misfolded proteins as a result of the saturation of protein degradation system leads in several conformational diseases like Huntington and cystic fibrosis to the formation of inclusion bodies called aggresomes. The inclusion bodies formed by BR120G had the characteristics of aggresomes (42,59). By immunofluorescence, BR120G was first visible within a few hours of transfection as small spots scattered all over the cytoplasm, likely corresponding to nuclei of newly precipitated proteins. In the following 2 or 3 days, these small dots coalesced in a microtubule-assisted manner into one or a few large perinuclear aggregates located near the centrosome. In the presence of colchicine, the small foci of precipitated proteins were still seen, however the BR120G aggresomes grew more slowly. The results are thus consistent with the idea that the BR120G aggresomes develop as a result of the precipitation of misfolded BR120G all over the cytoplasm followed by the coalescence of these foci into large aggregates. The reason why the misfolded proteins become concentrated around the centrosome is not totally clear. The formation of aggresomes might constitute a cytoprotective mechanism (42). It has been suggested that after the misfolded proteins have developed into an insoluble form that cannot be degraded, their sequestration in one large mass may facilitate their elimination by autophagy (23). Alternatively, it may be simply a consequence of the concentration of the proteasomes, the structures responsible for degrading damaged proteins, in the pericentrosomal region. The failure of the proteasomal machinery to cope with the production of insoluble proteins would result into the accumulation of the doomed proteins at their site of degradation. Remarkably, the accumulation of misfolded proteins into aggresomes further inhibits the proteasome activity in an amplification loop that likely constitutes an important cause of lethality (24).
B-crystallinopathy has an adult onset, indicating that in the real physiological situation, the formation of the aggregates by BR120G is a slow and thus highly inhibited process. We documented three distinct mechanisms mediated by different Hsp that could modulate the formation of aggregates in BR120G-transfected cells and thus that might be involved in vivo in delaying the onset of the disease: co-polymerization with family members, assisted folding and targeted degradation.
First, misfolded BR120G could be rescued by wild-type B. This is interesting because in patients with DRM B is still expressed from the normal allele. Co-transfection of B with BR120G, which brought the concentration of B to a level of about 50% of that of BR120G, caused an almost 100% reduction in the percentage of the cells that show aggresomes. Some 50% of the transfected cells still showed nuclei of small precipitates. However, these precipitates very rarely grew into aggresomes over the period of times investigated. Two distinct mechanisms can in theory account for B-mediated protection. First, B has chaperone activity and therefore may assist in the correct folding BR120G. Second, B is an oligomeric protein and forms in cells structures of some 24 subunits that may tolerate the presence of some unstable mutant protein. The chaperone (first mechanism) and dilution (second) mechanisms are not exclusive and may occur concomitantly. However, the fact that B prevents the growth of the aggregates rather than the formation of the nuclei suggests that the dilution effect is more effective than the chaperone activity. At any rate, by definition, a chaperone should not remain associated with the final structure. The same discussion applies to Hsp27, a protein that is also highly expressed in muscular cells. Hsp27 can oligomerize with B and was shown here to very efficiently prevent the formation of aggresome. Although Hsp27 could in theory assist in the folding and thereby oppose to the formation of the BR120G precipitates, this activity is not clearly demonstrated here inasmuch as Hsp27 remained associated with the final BR120G complexes. For both B and Hsp27, a normalization of the structure by co-polymerization can probably fully explain the results.
The second mechanism regulating BR120G aggregate formation is mediated by Hsp70 and co-chaperones and likely involves true chaperoning activities. Hsp70 did not interact or form oligomers with native B, however, it was immunoprecipitated in complexes with BR120G and reduced the number of cell with BR120G aggresomes by more than 50%. Hsp70 has previously been shown to prevent the formation of aggresomes or inclusion bodies in other degenerative diseases (30–33). In cooperation with co-chaperones like Hdj-1, HSP70 recognizes opened hydrophobic structures in proteins and by successive binding and release, using energy generated from ATP, promote correct folding. HSP70 assisted by Hdj1 can in some circumstances attract the ubiquitin ligase Chip-1, thereby promoting the degradation rather than folding of the proteins (57). The mechanism regulating folding assistance versus targeting for degradation is not known. In PTK2 cells, Hsp70 is expressed at high basal level and accordingly transfection of Chip-1 or Hdj-1 alone was more efficient than transfection of Hsp70 to prevent aggresome formation. In neither cases using Hsp70, Chip-1 or Hdj-1, we were able to show a rescue of the organization of the mutant protein into normal oligomers as seen with B, Hsp27 or HspB8, and as it would have been expected if the proteins were assisting in folding. This seems to suggest that Hsp70 promoted protein degradation rather than re-folding in this system. This is further suggested by the fact that the Chip-1 mutant lacking ubiquitin ligase activity was inefficient in preventing aggresome formation. However, we were unable to show either an accumulation of ubiquitylated BR120G or a reproducible reduction in the concentration of BR120G in any of the conditions studied (data not shown). Direct measurements by pulse-chase metabolic labeling also failed to reveal any significant difference in the rates of synthesis or degradation of B as compared to BR120G in the presence or absence of Chip-1. This may be a detection sensitivity problem. In most experiments, only 30–50% of the cells expressing BR120G show visible foci of denatured proteins or aggresomes. In the other cells, a slightly lower level of expression of BR120G or a higher basal level of chaperone activity might contribute better conditions for the proper folding of the mutant protein, which would then not be recognized and degraded by Hsp70/Chip-1. Furthermore, a reduction in the concentration of the denatured protein may not need to be very high to result in a high reduction in the number of cell with aggresomes. In sum, a small increase in the degradation rate of BR120G only in the affected cells may not yield a detectable change in the degradation rate or the level of the protein for the whole cell population.
The third mechanism is illustrated by the action of HspB8. Like Hsp70, HspB8 interacted only with BR120G, not with wild-type B, but in contrast to Hsp70, it fully rescued the supramolecular organization (sedimentation profile) of BR120G, suggesting that it is involved in folding assistance rather than degradation. Interestingly, unlike Hsp27 or B, HspB8 did not take part of the final structure and thus could be considered as playing a real chaperoning function. Little is known concerning the mechanism of chaperone function of sHsp in general and in fact, this result might represent the first direct evidence of an in vivo chaperone activity for any mammalian sHSP. These proteins have no ATPase activity and from in vitro studies with recombinant proteins, it is believed that they recognize and bind denatured protein, thereby preventing them from aggregating and keeping them in a folding competent state until other chaperones take action (60). However, HspB8 did not seem to act as Hsp70 and its co-chaperones. In contrast to them, HspB8 not only reduced the number of cells with precipitates, but it also totally blocked the growth of aggresomes from the small precipitates. This suggests that the growth of the precipitates into inclusion bodies in addition to the nucleation of the denatured proteins was affected. On a theoretical basis, a reduction in the concentration of the denatured protein caused by either assisted degradation or re-folding is expected to reduce the nucleation frequency, which is the rate-limiting step in inclusion body formation, but not as much the rate of growth of aggregation of the seeds once formed (61). The inhibition of growth observed with HspB8 may be a consequence of the oligomeric nature of B. With a multimeric protein, a chaperone may help in the correct folding of the monomer but also in its polymerization, which may then stabilize the mutant protein through intermolecular interactions. Furthermore, in the case of BR120G, it is possible that, once in the oligomer, the protein regains its chaperone activity and participates in the chaperoning of newly synthesized BR120G. It is noteworthy that the contrasting effect of Hsp70 and co-chaperones versus that of HspB8 may be cell type-specific (due for example to varying concentration of co-factors in different cells) or even substrate-specific. Baily et al. (33) obtained results very similar to ours using motor neuron-neuroblastoma hybrid cell lines showing that Hsp70 and Hdj-1/2 targeted a truncated androgen receptor containing glutamine repeats for degradation and reduced the number of cells showing aggregates formation (not their size). However, Sittler and co-workers (62) showed in the case of the aggregates formed in COS cells by the huntingtin exon 1 protein with glutamine repeats, that Hsp70 and Hdj-1 completely suppressed the aggregation of the protein leaving only spot-like dots as observed here with HspB8.
Our study has identified molecular mechanisms that can delay or prevent the formation of aggresomes and therefore that can, in vivo, contribute to the adult onset of B-crystallinopathy. The situation is, however, highly complex, owing in part to the fact that in this disease the altered protein is itself an Hsp, which is likely upregulated together with its normal allele in response to its own accumulation. This would partially dilute the beneficial induction and accumulation of the normal B allele as well as of other Hsp. However, the adverse effect of the stress-inducibility of BR120G may be compensated if the renatured protein once polymerized in a normally organized structure can then recover its chaperone activity.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Plasmids
pCRYSBWT contains the human
B-crystallin sequence inserted at the EcoR1 site of pCI-NEO (Promega). This construct was used to generate the mutant pCRYSBR120G by polymerase chain reaction using appropriate oligonucleotide primers to replace the arginine codon 120 (AGG) by a glycine codon (GGG). Human HspB8 cDNA was amplified by polymerase chain reaction of the IMAGE consortium clone ID 46388 and inserted into pCI-Neo to yield pCINHspB8. ßAprHsp70, which codes for human Hsp70, was a gift of Dick Mosser (63). pCINMycHU27, pCINMycHSPB8, pCINMycB and pCINMycHSP70 were constructed from the sequences contained in pCINHsp27 (64), pCINHspB8, pCRYSBWT and ßAprHsp70, cloned in a pCI-NEO vector in which the myc epitope (MEQKLISEEDL) had been inserted. They express human Hsp27, HspB8, B and Hsp70, respectively, tagged at their N-terminus with the myc epitope. The plasmids pCW7, which codes for His6-c-myc-tagged ubiquitin, was a gift from Ron Kopito (65). pCMV40, which codes for HdJ1/Hsp40, was a gift from Dr Harm Kampinga (28). Myc-CHIPWT (human Chip-1) and Myc-CHIP
E4 (Chip-1 196-303) were gifts from Dr Cam Patterson (40).
Antibodies
Anti-B is a rabbit polyclonal antibody raised against the c-terminal sequence IPITREEKPA of B-crystallin. The rat monoclonal anti-keratin 8 (TROMA-1) was described before (66). All other primary antibodies were mouse monoclonals; anti-Hsp70 (SPA-810) and anti-Hsp40 (SPA-450) were from Stressgen Biotechnologies Corp. (USA); anti-vimentin (clone V9), anti--tubulin (clone GTU-88) and anti--tubulin (clone B-5-1-2) were from Sigma-Aldrich (Canada); anti-Myc was prepared from hybridoma 9E10 cells (American Type Culture Collection).
Cell cultures and transfection
NIH3T3 and CCL39 cells were grown in Dulbecco's modified Eagle's medium high glucose (Gibco BRL) supplemented with 10% bovine calf serum or 5% fetal bovine serum, respectively. PTK2 cells were cultured in minimum essential medium (Gibco BRL) supplemented with 10% fetal bovine serum. Cells were plated 24 h before transfection. Transfection by calcium phosphate precipitation was done using 3.5–20 µg of plasmid. NIH3T3 and CCL39 cells were maintained in the presence of the precipitate and 50 µM of chloroquine for 5 h. PTK2 cells were transfected without chloroquine overnight. The cells were used at various times after transfection as indicated.
Immunofluorescence microscopy
Unless otherwise indicated all steps were carried out at room temperature. Cells in Petri dishes were fixed with methanol : acetone (50 : 50, v/v) for 10 min at -20°C, washed in PBS buffer (10 mM NaH2PO4, 130 mM NaCl, 2.7 mM KCl, 1 mM MgCl2, pH 7.5) and incubated with 5% bovine serum albumin in PBS for 5 min to block non-specific sites and with the specific primary antibodies diluted in PBS for 1 h. The cells were then washed six times with a 0.05% Tween-20 PBS solution and incubated for 1 h with the appropriate secondary antibodies labeled with either FITC, Texas-Red, Bodipy FL, Alexa 488 or Alexa 594 (Molecular Probes). Confocal microscopy was performed with a BioRad MRC-600 imaging system mounted on a Nikon Diaphot-TDM, equipped with a 60x objective lens.
Cell extracts
Total cell extracts were prepared by scraping cells in SDS–PAGE loading buffer. Soluble and insoluble extracts were prepared by sonication of the cells in Hepes buffer (25 mM Hepes, 1 mM EDTA and 1 mM DTT, pH 7.4). The solution was centrifuged at 18 000g for 5 min at 4°C to yield the soluble (supernatant) and insoluble (pellet) fractions.
Glycerol gradient centrifugation
Soluble cells extracts prepared 48 hours after transfection were loaded on a 12.0 ml linear gradient of glycerol (10–40%) made in the Hepes buffer. After an 18 h centrifugation at 30 000 rpm in a SW-40 rotor (Beckman) at 4°C, the gradient was fractionated and aliquots were analyzed by western or dot blot to determine the distribution of the protein of interest (53). The range of linearity of the detection methods was determined using varying amounts of the unfractionated extracts as controls.
Western and dot blot analyses
Samples were directly deposited onto nitrocellulose membranes (dot blot) or separated on SDS–polyacrylamide gel electrophoresis and transferred onto nitrocellulose membranes (western blot). After reacting the membrane with specific antibodies, antigen–antibody complexes were revealed using the Super-signal chemiluminescence substrate kit (Pierce) detected on the BioMax light-1 chemiluminescence film (Eastman Kodak), or revealed with iodinated secondary antibodies and quantified by phosphoimager analyses.
Non-denaturating immunoprecipitation
Forty-eight hours after transfection, cells plated in 100 mm2 were scraped into 0.5 ml of TEPT buffer (10 mM Tris–HCl pH 7.4, 1 mM EDTA, 10 mM NaF, 0.5% Triton X-100, 1 mM dithiothreitol, 0.2 mM PMSF). The extracts were vortexed and centrifuged at 18 000g for 15 min at 4°C. Supernatants were mixed with 1.5 vols of 1% low-fat dry milk diluted in TEBS buffer (10 mM Tris–HCl pH 7.4, 1 mM EDTA, 10 mM NaF, 150 mM NaCl, 0.1 mM PMSF 0.1% Triton X-100) and anti-Myc or anti-Hsp70. Antigen–antibody complexes were absorbed with Protein A Sepharose (Amersham Pharmacia Biotech). The beads were washed three times with TEBS buffer. The complexes were analyzed by western blot after SDS–PAGE.
Survival assay
NIH 3T3 cells in 75-cm2 flasks were transfected with various amounts (0–5 µg) of plasmids. Twenty-four hours after transfection, the cells were plated in duplicate at a concentration of 10 000 and 50 000 cells per 25 cm2 culture flask. Twenty-four hours later, the cells were submitted to a heat shock in water bath thermoregulated at 44°C for 2.5 h and returned to 37°C. Colonies of >30 cells were counted after 13 days. Results from different experiments were normalized for the quantity of the transfected proteins expressed at the time of heat shock.
|
ACKNOWLEDGEMENTS |
|---|
We thank R. Kopito, D. Mosser, L. Kampinga and K. Patterson for providing plasmids. This work was supported by the Canadian Institutes of Health Research grant MT-7088, by the Cancer Research Society Inc., Montréal, and by the Canada Research Chair in Stress Signal Transduction. A.T.C.Z. received a studentship from the Fundación Gran Mariscal de Ayacucho and the Universidad Centroccidental Lisandro Alvarado, Venezuela. J.R.T. received a studentship from the Cancer Research Society Inc.
|
FOOTNOTES |
|---|
* To whom correspondence should be addressed. Tel: +1 4186915281; Fax: +1 4186915439; Email: jacques.landry{at}med.ulaval.ca
Present address: Universidad Centroccidental Lisandro Alvarado, Decanato de Medicina, Departamento de Morfología, Sección de Anatomía Microscópica, Av. Libertador con Av. Andrés Bello, Barquisimeto, Estado Lara, Venezuela.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
- Vicart, P., Caron, A., Guicheney, P., Li, Z., Prevost, M.C., Faure, A., Chateau, D., Chapon, F., Tome, F., Dupret, J.M. et al. (1998) A missense mutation in the alphaB-crystallin chaperone gene causes a desmin-related myopathy. Nat. Genet., 20, 92–95.[CrossRef][ISI][Medline]
- Goebel, H.H. and Warlo, I. (2000) Gene-related protein surplus myopathies. Mol. Genet. Metab., 71, 267–275.[CrossRef][ISI][Medline]
- Arrigo, A.P. and Landry, J. (1994) Expression and function of the low-molecular-weight heat shock proteins. In Morimoto, R.I., Tissières, A. and Georgopoulos, C. (eds), The Biology of Heat Shock Proteins and Molecular Chaperones. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 335–373.
- Kappé, G., Verschuure, P., Philipsen, R.L., Staalduinen, A.A., Van de Boogaart, P., Boelens, W.C. and De Jong, W.W. (2001) Characterization of two novel human small heat shock proteins: protein kinase-related HspB8 and testis-specific HspB9. Biochim. Biophys. Acta, 1520, 1–6.[Medline]
- Clark, J.I. and Muchowski, P.J. (2000) Small heat-shock proteins and their potential role in human disease. Curr. Opin. Struct. Biol., 10, 52–59.[CrossRef][ISI][Medline]
-
Srinivasan, A.N., Nagineni, C.N. and Bhat, S.P. (1992) alphaA-crystallin is expressed in non-ocular tissues. J. Biol. Chem., 267, 23337–23341.
[Abstract/Free Full Text] -
Klemenz, R., Frohli, E., Steiger, R.H., Schafer, R. and Aoyama, A. (1991) Alpha B-crystallin is a small heat shock protein. Proc. Natl Acad. Sci. USA, 88, 3652–3656.
[Abstract/Free Full Text] - Iwaki, T., Kume-Iwaki, A., Liem, R.K. and Goldman, J.E. (1989) Alpha B-crystallin is expressed in non-lenticular tissues and accumulates in Alexander's disease brain. Cell, 57, 71–78.[CrossRef][ISI][Medline]
-
Horwitz, J. (1992) Alpha-crystallin can function as a molecular chaperone. Proc. Natl Acad. Sci. USA, 89, 10449–10453.
[Abstract/Free Full Text] -
Aoyama, A., Frohli, E., Schafer, R. and Klemenz, R. (1993) Alpha B-crystallin expression in mouse NIH 3T3 fibroblasts: glucocorticoid responsiveness and involvement in thermal protection. Mol. Cell. Biol., 13, 1824–1835.
[Abstract/Free Full Text] - Mehlen, P., Preville, X., Chareyron, P., Briolay, J., Klemenz, R. and Arrigo, A.P. (1995) Constitutive expression of human hsp27, Drosophila hsp27, or human alpha B-crystallin confers resistance to TNF- and oxidative stress-induced cytotoxicity in stably transfected murine L929 fibroblasts. J. Immunol., 154, 363–374.[Abstract]
- Nicholl, I.D. and Quinlan, R.A. (1994) Chaperone activity of alpha-crystallins modulates intermediate filament assembly. EMBO J., 13, 945–953.[ISI][Medline]
- Perng, M.D., Cairns, L., van den IJssel, P., Prescott, A., Hutcheson, A.M. and Quinlan, R.A. (1999) Intermediate filament interactions can be altered by HSP27 and alpha B-crystallin. J. Cell Sci., 112, 2099–2112.[Abstract]
-
Perng, M.D., Muchowski, P.J., van Den, I.P., Wu, G.J., Hutcheson, A.M., Clark, J.I. and Quinlan, R.A. (1999) The cardiomyopathy and lens cataract mutation in alphaB-crystallin alters its protein structure, chaperone activity, and interaction with intermediate filaments in vitro. J. Biol. Chem., 274, 33235–33243.
[Abstract/Free Full Text] -
Bova, M.P., Yaron, O., Huang, Q., Ding, L., Haley, D.A., Stewart, P.L. and Horwitz, J. (1999) Mutation R120G in alphaB-crystallin, which is linked to a desmin-related myopathy, results in an irregular structure and defective chaperone-like function. Proc. Natl Acad. Sci. USA, 96, 6137–6142.
[Abstract/Free Full Text] -
Kumar, L.V., Ramakrishna, T. and Rao, C.M. (1999) Structural and functional consequences of the mutation of a conserved arginine residue in alphaA and alphaB crystallins. J. Biol. Chem., 274, 24137–24141.
[Abstract/Free Full Text] - Goldfarb, L.G., Park, K.Y., Cervenakova, L., Gorokhova, S., Lee, H.S., Vasconcelos, O., Nagle, J.W., Semino-Mora, C., Sivakumar, K. and Dalakas, M.C. (1998) Missense mutations in desmin associated with familial cardiac and skeletal myopathy. Nat. Genet., 19, 402–403.[CrossRef][ISI][Medline]
-
Munoz-Marmol, A.M., Strasser, G., Isamat, M., Coulombe, P.A., Yang, Y., Roca, X., Vela, E., Mate, J.L., Coll, J., Fernandez-Figueras, M.T. et al. (1998) A dysfunctional desmin mutation in a patient with severe generalized myopathy. Proc. Natl Acad. Sci. USA, 95, 11312–11317.
[Abstract/Free Full Text] -
Wang, X., Osinska, H., Klevitsky, R., Gerdes, A.M., Nieman, M., Lorenz, J., Hewett, T. and Robbins, J. (2001) Expression of R120G-alphaB-crystallin causes aberrant desmin and alphaB-crystallin aggregation and cardiomyopathy in mice. Circul. Res., 89, 84–91.
[Abstract/Free&n