Role of molecular chaperones in protein folding diseases

National Heart, Lung, And Blood Institute

Investigators

Linked publications, trials & patents

Abstract

Proteins that have prion-like domains, which are regions of low sequence complexity residues enriched in polar amino acid residues, tend to aggregate. These domains are responsible for the folding of prion proteins into amyloid fibers, rich in beta sheet, and contribute to the phase separation of proteins. We have been studying the properties of the yeast prion protein Sup35, an essential protein that functions in translation termination. In addition to the properly folded conformation, this protein can form into condensates when yeast are depleted of energy or into an amyloid prion called PSI+ when infected with PSI+ seeds. Condensates occur either by acidification of the cytosol or energy depletion, conditions that lower intracellular pH and ATP. However, Sup35 does not form condensates when yeast are depleted of energy at neutral pH, which shows that in addition to low intracellular ATP, acidification of the cytosol is necessary for condensate formation. This suggests that raising the intracellular pH would solubilize the condensates, but this is not the case. Condensates are only solubilized upon addition of glucose, which increases both pH and ATP levels. The lack of reversibility indicates that the two phases, the condensed-phase and the dilute-phase, are not in thermodynamic equilibrium. Instead, dissolution of the Sup35 condensates is an energy dependent process. Since Sup35 binds to several proteins, we examined whether these binding partners of Sup35 also form condensates when yeast are depleted of energy. Sup35 binds to Sup45 to form the translation termination complex. Like Sup35, Sup45 form condensates at low pH, but not at neutral pH in energy depleted yeast. Glucose addition, which increases the internal pH an ATP, causes dissolution of the Sup45 condensates, whereas the condensates persist in yeast when only the internal pH is increased. Furthermore, the Sup45 colocalizes with the Sup35 in the condensates. In contrast, Pab1, another protein that binds Sup35, forms stress granules in energy depleted cells with different properties from the Sup35 condensates. The Sup35 condensates show only partial colocalization with Pab1 stress granules and different conditions cause the phase separation of Pab1 and Sup35. The Sup35 condensates also appear to contain RNA since the addition of cycloheximide during energy depletion inhibits the formation of Sup35 condensates. This suggests that the Sup35 condensates contain Sup45 and RNA, but not Pab1. In another project related to the Sup35 prion, we have examined the effect of different Hsp70 isozymes on the curing of PSI+ by Hsp104 overexpression. Based on our previous data, we proposed a dissolution mechanism for curing by Hsp104 overexpression in which Hsp104 dissociates monomers from the ends of the amyloid fibers in a process we called trimming. To further the curing by Hsp104 overexpression, we have been examining the curing of PSI+ in yeast expressing different Hsp70 isozymes. Even though the number of prion seeds is the same in PSI+ yeast expressing either Ssa1 or Ssa2, the different Hsp70 isozymes affect the rate of curing of PSI+ by Hsp104 overexpression. This difference in curing rate, but not in seed number, suggests Hsp104 has two binding sites for Hsp70. In addition to the well-characterized binding site of Hsp70 on the M-domain of Hsp104, which is necessary for activation of Hsp104, another binding site for Hsp70 has been identified located in the N-terminal domain of Hsp104. When this N-terminal putative Hsp70 binding site was mutated, PSI+ is not cured by overexpression of the mutated Hsp104, even though the mutant propagates PSI+. Consistent with the lack of curing, there is no trimming of the prion seeds by the overexpressed mutated Hsp104. These results further support a mechanism in which overexpression of Hsp104 cures by dissolution of the prion seeds with a novel role for Hsp70 in curing PSI+. Our research on the curing of yeast prions has been extended to examining the curing of URE3 by overexpression of Btn2, Cur1, Hsp42, or Ydj1. Our previous research showed these proteins cause aggregation of the prion seeds, indicating that asymmetric segregation of the seeds is contributing to the curing of URE3. With overexpression of each of these proteins, the Ure2 seeds colocalize with the Hsp42 aggregates, which suggests that by binding to the Ure2 seeds, Hsp42 causes asymmetric segregation of the seeds, leading to URE3 curing. This, in turn, predicts that if Hsp42 is deleted, the Ure2 seeds should not aggregate and any curing that does occur should be due to inhibition of severing of Hsp104. Surprisingly however, in the hsp42 deletion strain, there is still aggregation of the Ure2 seeds in yeast overexpressing Btn2, Cur1 or Ydj1 although both the rate of aggregation of the seeds and the fraction of cells with aggregates are reduced when Hsp42 is deleted. Therefore, in addition to Hsp42, other factors also cause aggregation of the Ure2 seeds. As expected, the rate of URE3 curing was reduced in the hsp42 deletion strain. The most marked effect was that there was no significant curing of URE3 by overexpression of Btn2, indicating that Btn2 cures primarily by asymmetric segregation. On the other hand, overexpression of Cur1 and Yd1 still cures URE3 in the hsp42 deletion strain, but at a slower rate. Imaging shows that this curing is due to a combination of seed aggregation along with inhibition of severing of the non-aggregated seeds, which causes a progressive loss of the non-aggregated seeds over time. Like overexpression of Btn2, overexpression of Hsp42 cures URE3 by asymmetric segregation due to the binding of all the Ure2 seeds to large Hsp42 aggregates, but, this binding is dependent on endogenous expression of Btn2 and Cur1; in the absence of the latter proteins, overexpression of Hsp42 by itself does not cure URE3. Therefore, Cur1, Btn2, and Hsp42, all proteins that function in protein quality control by sorting and sequestering misfolded non-amyloid proteins, also function in URE3 curing by aggregating the Ure2 seeds thus causing their asymmetric segregation. In addition to prion proteins, we have been examining the aggregation of huntingtin fragments with expanded polyglutamine repeat regions (HttpolyQ) in yeast. Like prion proteins, these Htt fragments assemble into amyloid fibrils. The aggregation of these fragments in yeast have been reported to depend on the presence of a prion with an amyloid conformation. As a result of this relationship, HttpolyQ aggregation indirectly depends on Hsp104 for prion propagation. We find that HttQ103 aggregation is directly affected by Hsp104 with and without the presence of RNQ+ and PSI+ prions. When we inactivate Hsp104 in the presence of prion, yeast have only one or a few large HttQ103 aggregates rather than numerous smaller aggregates. When we inactivate Hsp104 in the absence of prion, there is no significant aggregation of HttQ103, whereas with active Hsp104, HttQ103 aggregates slowly accumulate due to the severing of spontaneously nucleated aggregates by Hsp104. We do not observe either effect with HttQ103P, which has a polyproline-rich region downstream of the polyglutamine region, because HttQ103P does not spontaneously nucleate and Hsp104 does not efficiently sever the prion-nucleated HttQ103P aggregates. Therefore, in contrast to HttQ103P aggregation in which the only function of Hsp104 is to propagate yeast prion, Hsp104 severs the HttQ103 aggregates. This leads to the proliferation and transmission of the HttQ103 aggregates to the daughter cells.