Interactome Changes Quantified to Identify the ER Proteostasis Network to Fight Amyloid Diseases

Abstract

In this issue of Cell Chemical Biology, Plate et al. (2019) used quantitative interactome proteomics to define the molecular mechanism by which ATF6 activation reduces amyloidogenic protein secretion. These results shed light on preventing the amyloid formation at the very early step to treat devastating amyloid diseases.

Proteostasis plays an essential role in ensuring physiological function (Balch et al., 2008; Jayaraj et al., 2019). Proteostasis imbalance underlies numerous disorders related to protein misfolding, such as amyloid diseases (Chiti and Dobson, 2017). In such diseases, a critical pathogenic step is the secretion of destabilized amyloidogenic proteins, which leads to their extracellular aggregation into toxic species. The endoplasmic reticulum (ER) is the location where the folding, trafficking, and degradation of proteins in the secretory pathway are regulated. Consequently, manipulating the ER proteostasis network holds the promise to prevent the secretion of amyloidogenic proteins to treat devastating amyloid diseases. In this issue Plate et al. (2019) used quantitative interactome proteomics to determine how operating the ER proteostasis network reduces the secretion of the destabilized amyloidogenic immunoglobulin light chain (ALLC).

The unfolded protein response (UPR) is the major cellular pathway that remodels the ER proteostasis network to enhance the ER folding capacity by activating transcription factors such as XBP1s (spliced XBP1) and ATF6 (Hetz and Papa, 2018; Walter and Ron, 2011). The UPR senses the ER proteostasis environment and adapts the ER proteostasis network. Since many of the physiologically important proteins fold in the ER, modulating the UPR offers great promise to change the fate of pathogenic proteins that are associated with various protein conformational diseases, such as lysosomal storage diseases, transthyretin amyloidosis and light chain amyloidosis, retinitis pigmentosa, α1 antitrypsin deficiency, and familial forms of amyotrophic lateral sclerosis. Previous efforts from the authors demonstrated that ATF6 activation substantially reduces the secretion and aggregation of ALLC without affecting the secretion of an energetically normal light chains (Cooley et al., 2014). In addition, ATF6 activation does not influence the ER degradation pathway of ALLC. In contrast, XBP1s activation only modestly influences the secretion of ALLC while enhancing its ER degradation pathway. Therefore, distinct mechanisms exist for ATF6- and XBP1s-dependent reduction of ALLC secretion, possibly through interactions between ALLC and ER proteostasis network differentially induced by ATF6 and XBP1s.

To elucidate such distinct mechanisms between ATF6 and XBP1s, Plate et al. (2019) employed elegant proteomics platforms to quantify the changes of the ALLC interactomes after the genetic activation of ATF6 and XBP1s. The authors applied both tandem mass tag (TMT)-based and stable isotope labeling with amino acids in cell culture (SILAC)-based affinity purification mass spectrometry proteomics approaches to improve the throughput and reliability without compromising sensitivity to study the interactomes under various stimulus conditions (Pankow et al., 2015). This enables a deep proteome coverage and a global view of the altered ER proteostasis network components that interact with ALLC. As such, the authors identified 72 high-confidence interactors of the light chain proteins, serving as the basis for their interactome analysis under various conditions.

Impressively, the authors demonstrated distinct interactome changes after ATF6 and XBP1s activation, accounting for their different effects on the reduction of ALLC secretion. ATF6 activation enhances the interactions between ALLC and a subset of ER proteostasis network components, including ER lumen Hsp70 chaperone BiP and its co-chaperones ERdj3 and HYOU1, ER lumen Hsp90 chaperone GRP94, and a protein disulfide isomerase PDIA4. Importantly, they act as pro-folding factors and enhance the ER folding capacity to retain ALLC in the ER, reducing its secretion to form toxic extracellular aggregation (Figure 1). In contrast, XBP1s activation globally attenuates the interactions between ALLC and ER proteostasis network components, suggesting a weakened folding capacity for ALLC and its enhanced ER degradation. Interestingly, ATF6 and XBP1s activation does not produce synergy to reduce the secretion of destabilized ALLC.

Figure 1. An Overview of the Molecular Mechanism by which ATF6 Activation Reduces the Secretion of Destabilized ALLC.

Under basal condition, ALLC, a destabilized amyloidogenic protein, is secreted into extracellular space and aggregates into toxic species. ATF6 activation retains ALLC in the ER by increasing its interactions with ATF6-dependent ER proteostasis factors, such as pro-folding chaperones, without apparent influence on its ER degradation. This substantially reduces the secretion of destabilized ALLC. XPB1s activation globally attenuates the interactions between ALLC and ER proteostasis network components and enhances the ER degradation of ALLC. This only modestly influences the secretion of ALLC.

Furthermore, the authors evaluated how ATF6 activation differentiates the destabilized ALLC from energetically normal light chain JTO. Analysis of the interactomes of ALLC and JTO identified a subset of ER proteostasis factors that display increased interactions with ALLC compared to JTO. Although AFT6 activation also increases the interactions between JTO and ER proteostasis factors, such interactions are weaker than those between ALLC and ER proteostasis factors. Taken together, the results indicate that enhanced interactions between ALLC and a subset of ER proteostasis factors after ATF6 activation selectively target destabilized ALLC to prolonged ER folding cycles and thus ER retention, leading to reduced extracellular secretion.

Because ATF6 and XBP1s upregulate the mRNA/protein expression levels of numerous ER proteostasis factors, the authors determined whether there is a correlation between their increased expression levels and their increased interactions with ALLC. Interestingly, there is a general correlation for ATF6-dependent ALLC interactors, whereas in sharp contrast, no correlation is observed for XBP1s-dependent ALLC interactors. This also suggests that compared to XBP1s activation, ATF6 activation would likely be more effective in controlling the secretion and aggregation of ALLC. Furthermore, the authors hypothesized that the correlation between protein expressions and interactions for ATF6-dependent ALLC interactors is functionally important for many pro-folding factors, including BiP, GRP94, and ERdj3. To test that, the authors demonstrated that overexpression of these pro-folding factors selectively reduces the secretion of destabilized ALLC without influencing its ER degradation relative to the energetically normal JTO. This result indicates that the core ATF6-dependent ALLC interactors can recapitulate the function of ATF6 activation in reducing ALLC secretion. However, such a recapitulation by these pro-folding factors is only partial, because even the most efficient ALLC secretion reduction afforded by BiP overexpression is still substantially less than that afforded by ATF6 activation. This result underlies the importance of the global remodeling of the ER proteostasis network induced by ATF6.

In summary, the results in this study (Plate et al., 2019) present two important findings. First, remodeling the ER proteostasis network by stress-independent activation of ATF6 represents a promising therapeutic strategy to ameliorate light chain amyloidosis, a currently uncurable medical condition. Such a strategy works by retaining the mutant proteins in the ER and preventing them from reaching the harmful environment. This operation interferes with the very early step of the disease progression and thus would likely prevent the pathogenic effects associated with the early aggregation of toxic species. Furthermore, this strategy could be extrapolated to many protein conformational diseases resulting from the aggregation of mutant proteins, such as retinitis pigmentosa caused from mutant rhodopsin aggregation and genetic epilepsy caused by the aggregation of mutant neurotransmitter receptors (Kang et al., 2015). Second, from the perspective of mechanism of action, ATF6 activation selectively reduces the secretion of destabilized ALLC by increasing its interaction with ATF6-dependent ER proteostasis network components, including pro-folding factors BiP, GRP94, and ERdj3 (Figure 1). This further holds the destabilized ALLC in the ER by prolonged ALLC chaperoning. This binding-and-holding mechanism is different from another unconventional secretion mechanism for misfolded cytosolic proteins, where the ER-associated deubiquitylase UPS19 initiates the export of misfolded proteins for extracellular degradation (Ye, 2018). It appears that cells develop delicate machinery to sense the misfolded substrates and determine how to handle them for optimal cell survival.