Tracing genetic diversity captures the molecular basis of misfolding disease

Zhao, Pei; Wang, Chao; Sun, Shuhong; Wang, Xi; Balch, William E.

doi:10.1038/s41467-024-47520-0

Download PDF

Article
Open access
Published: 18 April 2024

Tracing genetic diversity captures the molecular basis of misfolding disease

Nature Communications volume 15, Article number: 3333 (2024) Cite this article

1196 Accesses
1 Altmetric
Metrics details

Subjects

Abstract

Genetic variation in human populations can result in the misfolding and aggregation of proteins, giving rise to systemic and neurodegenerative diseases that require management by proteostasis. Here, we define the role of GRP94, the endoplasmic reticulum Hsp90 chaperone paralog, in managing alpha-1-antitrypsin deficiency on a residue-by-residue basis using Gaussian process regression-based machine learning to profile the spatial covariance relationships that dictate protein folding arising from sequence variants in the population. Covariance analysis suggests a role for the ATPase activity of GRP94 in controlling the N- to C-terminal cooperative folding of alpha-1-antitrypsin responsible for the correction of liver aggregation and lung-disease phenotypes of alpha-1-antitrypsin deficiency. Gaussian process-based spatial covariance profiling provides a standard model built on covariant principles to evaluate the role of proteostasis components in guiding information flow from genome to proteome in response to genetic variation, potentially allowing us to intervene in the onset and progression of complex multi-system human diseases.

Genome-wide association studies

Article 26 August 2021

Discovery of potent inhibitors of α-synuclein aggregation using structure-based iterative learning

Article Open access 17 April 2024

Proteome-scale discovery of protein degradation and stabilization effectors

Article 20 March 2024

Introduction

Genome variation in the human population provides a vast reservoir for the protein fold to explore and evolve conformational plasticity required for form and function in changing complex biological environments- but also poses the risk for protein misfolding and/or aggregation that could lead to human genetic disease^1,2. We now appreciate that many aspects of the evolving protein sequence-to-function-to-structure relationships are dynamically managed by a ubiquitous protein folding system referred to as proteostasis^3,4,5,6. Proteostasis is composed of a collection of chaperone/co-chaperone components, trafficking components, degradation machineries, and signaling pathways such as the heat shock response (HSR) and unfolded protein response (UPR) that manage the fold in response to sequence variation and environmental stress across a lifespan^7,8,9. The potential role of proteostasis in disease management is just beginning to be unraveled^{10,11,12,13,14,15}.

To address the central problem of information flow from the genome to the proteome in human biology in the context of the complex folding dynamics impacted by heat shock protein 90 (Hsp90) family members¹⁶ that constitute a central ATP-dependent chaperone system in the cell, we developed a Gaussian process (GP) regression-based machine learning^{1,17,18,19,20,21} approach to learn the basal state of the protein fold and its response to proteostasis management. GP can use a sparse collection of variants across the worldwide population and their associated phenotypes as a collective input to generate spatial covariance (SCV) maps of the response of every residue in the protein sequence and its residue–residue responses to environmental changes, including pharmacological intervention^{1,17,18,19,20,21}. GP-SCV principled modeling captures as a collective the variant changes that report on the evolutionary trajectory of the entire wild-type (WT) protein fold at atomic resolution that impacts health and disease of the host in response to the environment and natural selection^{1,17,18,19,20,21}.

To mechanistically dissect Hsp90 function on a single residue basis, we used a selective Hsp90 inhibitor PU-WS13 that specifically reduces the ATPase activity of the endoplasmic reticulum (ER) Hsp90 paralog GRP94, serving as a chemical mimetic of the role of co-chaperones that normally manage its ATPase cycle^{22,23,24,25,26,27,28,29}. To trace the SCV principled dynamic profile of GRP94 ATPase activity on human variation, we examined its impact on the management of alpha-1-antitrypsin deficiency (AATD). AATD is caused by genetic variation in SERPINA1 gene that encodes AAT, a prominent member of the serine protease inhibitor family³⁰. AATD is an autosomal recessive inherited disease and the most prominent risk factor for the aging-related onset of chronic obstructive pulmonary disease (COPD) in response to smoking and airborne pollutants³⁰. AAT is synthesized and secreted from hepatocytes for delivery to the lung at grams per day where it binds human neutrophil elastase (NE) and prevents NE-induced degradation of the extracellular matrix (ECM) in the lung^30,31. Numerous variants in the AAT protein can cause protein misfolding during nascent synthesis in the ER, the first step of the secretory pathway^32,33,34, leading to aggregation with extended polymers that trigger liver disease phenotypes, such as fibrosis, cirrhosis and hepatocellular carcinoma^35,36,37. AAT aggregation in the liver results in reduced secretion of functional AAT to the plasma, leading to a loss-of-function in the lung manifested as the inflammatory diseases emphysema, bronchitis and COPD^30,38. AAT augmentation therapy using human plasma-derived purified AAT to re-supply the reduced level of AAT found in circulation has no impact on progression of liver aggregation disease and only modest impact on progression of lung disease³⁹. The distinct liver and lung pathologies pose unprecedented challenges to develop therapeutics that preserve lung function while preventing polymer formation^32,40,41.

Herein, we utilize GP-SCV principled relationships to understand the impact of GRP94 ATPase activity on human variation affecting the level of intracellular or secreted AAT monomer and polymer pools, as well as the secreted NE inhibitory activity, based on 76 AAT variants found in the population that differentially contribute to disease onset and progression³². Strikingly, we find that the inhibition of GRP94 ATPase activity improves AAT monomer secretion and its NE inhibitory activity, while reducing both intracellular polymer load and polymer secretion for most pathogenic AAT variants, including the prominent Z-variant affecting 95% of the AATD population. GP-SCV principled relationships reveal the impact of GPR94 ATPase activity to (re)direct the N- to C-terminal cooperative folding of AAT to correct AATD clinical phenotypes impacting both liver and lung disease. We suggest that GP-SCV principled relationships provide a standard computational framework to address information flow from the genome to proteome affecting protein fold conversion by a broad range of proteostasis components affecting health and disease on a residue-by-residue basis^1,20,21,34, providing a precision approach to mitigate disease.

Results

Assaying the diverse folding and functional features of AAT variants

We recently showed that pharmacological activation of ATF6, a specific branch of the unfolded protein response (UPR) pathway that regulates the proteostasis environment in the ER, can improve the function of AAT variants while simultaneously reducing aggregation in the ER³⁴. Preliminary results suggested that ER chaperones GRP78 and GRP94 are required for the restoration of function impacted by ATF6 activators³⁴. As a paralog of the cytosolic Hsp90, GRP94 is a highly abundant ATP-dependent chaperone found in the ER^{14,16,42,43,44}. It has previously been proposed to direct the misfolded AAT protein to ER associated degradation (ERAD) pathway^45,46,47. Like cytosolic Hsp90, how GRP94 recognizes and manages folding in response to genetic variation in the population, and how it could be modulated to manage ER misfolding diseases in response to variation impacted by ~33% of the protein load encoded by the human genome, remains unknown.

To probe the mechanism of how GRP94 manages the folding and function of AAT, we used AAT variants in the human population that harbor diverse folding conformations in the cell^32,33 including the most common Z-variant allele (E366K), leading to severe disease in 95% of the AATD population^41,48. The collection of AAT variants contains 44 pathologic variants associated with AATD liver and lung phenotypes reported in the literature^32,33,49 and those currently annotated in the ClinVar database⁵⁰. We include in addition 31 variants that are annotated as benign or of “uncertain significance” in the ClinVar database as controls⁵⁰. A total of 75 AAT variants are missense variants except for three variants that generate a truncated protein (Y62*, E281* and Null Hong Kong (NHK)) that severely disrupt function (Fig. 1a). This collection of variants includes the most common missense variants in AAT in terms of allele frequency (AF) that have been reported in the general population (gnomAD database⁵¹) including M1 (V237A; 22% AF), M2 (R125H; 15.6% AF), M3 (E400D; 27% AF), S (E288V; 2.3% AF), Z (E366K; 1.1% AF) and others that have >0.1% AF. The residues impacted by the variant collection are spread across the entire AAT sequence (Fig. 1a, b, lower panel, brown balls), thus providing molecular fiduciary markers¹ that enables us to probe the sequence-based folding and function space defining the ensemble of AAT structures found in the extant population in response to genetic variation.

**Fig. 1: Quantification of secreted monomer, intracellular polymer, and NE inhibitory activity for AAT variants.**

To characterize the impact of a variant on the status of the AAT protein fold responsible for its monomeric or polymeric states based on its folding, stability and trafficking itinerary through the secretory pathway (Fig. 1c), we utilize two conformation-specific antibodies, a monomer-specific antibody 16f8 (Supplementary Fig. 1a) and a polymer-specific antibody 2C1⁵². These antibodies were used in enzyme-linked immunosorbent assays (ELISA) (Supplementary Fig. 1a, b) to measure the level of AAT monomer or polymer in intracellular and extracellular environments (Fig. 1c) using high-throughput formats (see “Methods”). To measure the inhibitory activity of secreted AAT variants to neutrophil elastase (NE), its natural substrate in the lung, we used a sensitive fluorogenic NE substrate (Z-AAAA)2Rh110⁵³ (Supplementary Fig. 1b) (see “Methods”). A high level of NE inhibitory activity prevents the digestion of the fluorogenic NE substrate resulting in a low fluorescence signal (Supplementary Fig. 1c). Combined, use of the conformation-specific ELISA and NE inhibitory activity assays allows us to assess the different folding and functional states of AAT variants across its biological itinerary from the ER in the liver to secretion out of the cell for delivery to downstream tissues (Fig. 1c).

Defining the folding and functional properties of AAT variants

To generate a comprehensive understanding of the functional phenotypic properties of AAT in response to its physiologic secretion itinerary in the liver (Fig. 1c), WT and 75 variants were transfected in a hepatocyte-derived cell line Huh7.5 with endogenous AAT silenced by CRISPR/Cas9 (Huh7.5^null)^54,55,56. We also transfected these variants in a human bronchial epithelial lung cell line IB3 that has no detectable endogenous AAT⁵⁷ to capture epithelial pathways potentially contributing to lung function. The level of intracellular or extracellular AAT variants in either monomer or polymer form, and the NE inhibitory activity of the extracellular (secreted) pool of AAT variants were measured (Fig. 1d, Huh7.5^null cells; Supplementary Fig. 2, IB3 cells). The measured activity values for secreted monomer, intracellular polymer and NE inhibitory activity are highly correlated between the liver cell line Huh7.5 and the lung IB3 cell line (Supplementary Fig. 3a). These results demonstrate that the basic folding and functional features of AAT variants are conserved across different cellular environments. Furthermore, for each variant the intracellular level of monomer and polymer conformations are highly correlated with extracellular monomer and polymer (Supplementary Fig. 3b). These results demonstrate that extracellular levels of monomer and polymer generally reflect the intracellular monomer and polymer, respectively. Given the similarity of the conformational features associated with each AAT variant between the intracellular and extracellular space, we use the measured levels of intracellular polymer, secreted monomer and the NE inhibitory activity from Huh7.5^null cells henceforth to understand how, on a residue-by-residue basis, the folding of AAT is coupled to function to drive different disease states found in the AATD population.

We first compared AAT-WT to AAT-Z. As expected, AAT-Z shows low monomer secretion (~37% of WT) and low NE inhibitory activity (~36% of WT), while AAT-WT shows low intracellular polymer (~20%) when compared with AAT-Z (Fig. 1d, red rectangle vs green rectangle). These results are consistent with the severe liver and lung-disease phenotypes for patients with homozygous Z triggered by the combined gain of toxicity in the liver due to accumulation of AAT-Z intracellular polymer and loss-of-function in the lung due to AAT-Z deficiency³⁰. In contrast, the S allele (E288V) shows a median level of secreted monomer, intracellular polymer and NE inhibitory activity (Fig. 1d, yellow rectangle vs green rectangle), consistent with the mild disease phenotypes of patients with homozygous S allele. We found that our cell-based measurements of secreted monomer levels are significantly correlated with reported AAT serum levels from patients with different variants, indicating the utility of our cell-based assay to capture the clinical features of disease in the patient population (Fig. 1e, f). Moreover, the secreted monomer levels strongly correlate with their NE inhibitory activity (Fig. 1g, Pearson’s r = 0.77, P = 4.3 × 10⁻⁶), indicating that the monomer conformation we measure corresponds to the functional form of AAT. In contrast, the intracellular polymer levels of the variants are only moderately anti-correlated with the extracellular NE inhibitory activity values (Fig. 1h, Pearson’s r = −0.34). While variants with high intracellular polymer (Fig. 1h, polymer ≥1 (levels equal or above of AAT-Z)) generally show extracellular NE inhibitory activity lower than the mean of measured NE inhibitory activity of all variants (~73% of WT activity), variants with low intracellular polymer (Fig. 1h, polymer <0.2) have very diverse NE inhibitory activities ranging from ~26% of WT to 114% of WT (Fig. 1h), suggesting a low polymerization propensity does not necessarily confer high functional activity.

We categorize AAT variants into three classes based on their NE inhibitory activity. We choose a 75% cutoff threshold for loss-of-function given that of the 40 AAT variants that have been reported to cause significant clinical disease phenotypes, 33 of them (>80%) have NE inhibitory activity lower than 75% of WT activity and the mean of the predicted NE inhibitory activity for all variants is around 73%. We designate class I variants are loss-of-function defined as a NE inhibitory activity <75% of WT AAT; class II contains the function-impaired variants with NE inhibitory activity between 75% and 90% of WT AAT; class III variants show NE inhibitory activity comparable to WT AAT (>90% of WT activity) (Table 1). For the variants that have lower NE inhibitory than WT AAT, we further divide them into variants with high polymer (>50% of AAT-Z polymer) and low polymer (<50% of AAT-Z polymer). In the class I category comprising the function loss variants, more than a half do not show high polymer, suggesting that polymerization is not the major driving force for the functional deficiency of these variants. These results suggest a more complicated set of relationships between folding, stability and the functional properties of AAT variants contribute to changes in the protein fold responsible for disease that need to be managed by proteostasis in the AATD population.

Table 1 Classification of AAT variants based on the level of NE inhibitory activity and intracellular polymer

Full size table

GRP94 ATPase modulation rescues AATD phenotypes for AAT-Z

To test the potential role of GRP94 on the phenotypic diversity imposed by AAT variants, we used PU-WS13 (Fig. 2a), a GRP94-specific inhibitor that blocks the ATPase activity of GRP94^22,23,24. The IC50 of PU-WS13 for GRP94 inhibition is 0.22 μM, while the IC50 of PU-WS13 for inhibiting HSP90α, HSP90β, and TRAP1 are 27.3 μM, 41.8 μM, and 7.3 μM, respectively²³. We tested PU-WS13 in different doses in Huh7.5 cells (see below Supplementary Fig. 10a–e) revealing 1 μM PU-WS13 treatment achieved the maximum correction for Z-AAT NE activity, monomer secretion, and reducing polymer burden. In total, 1 μM PU-WS13 treatment should block most GRP94 ATPase activity while avoiding potential off-target effect on other Hsp90 family members.

**Fig. 2: GRP94 ATPase inhibitor PU-WS13 rescues the NE inhibitory activity, and monomer secretion of AAT variants.**

We first tested the effect of PU-WS13 on the NE inhibitory activity using the fluorogenic NE substrate (Z-AAAA)2Rh110⁵³ for wild-type (WT) AAT (AAT-WT) and an additional 75 AAT variants transiently transfected in a liver-derived Huh7.5 cell line in which AAT gene has been knocked-out (Huh7.5^null)^54,55. Intriguingly, we found that PU-WS13 treatment significantly increased the NE inhibitory activity for 37 of all 76 (~49%) tested variants (Fig. 2b). For the 37 variants with improved NE inhibitory activity in response to PU-WS13 treatment, 30 (~81%) of them have less than 75% of AAT-WT NE inhibitory activity in the vehicle control state (Fig. 2b), indicating a potential key role for GRP94 in managing the function of defective AAT variants.

As the NE inhibitory activity strongly correlates the monomer secretion in the basal state in the absence of drug treatment, we tested the impact of PU-WS13 on monomer secretion for 76 AAT variants using the conformation-dependent ELISA assay. PU-WS13 treatment significantly increases the monomer secretion for 40 of all 76 (~53%) tested variants (Fig. 2c). For the 40 variants with improved monomer secretion in response to PU-WS13 treatment, 38 (~95%) of them have monomer secretion levels less than 75% WT in the vehicle basal state (Fig. 2c). These results suggest that GRP94 impacts the features of AAT fold features that contribute to both its monomer secretion and NE inhibitory activity, including the prominent AAT-Z and AAT-S alleles contributing to clinical disease (Fig. 2b, c).

To validate the correction effect of PU-WS13 on the NE inhibitory activity and monomer secretion of AAT variants (Fig. 2b, c) with an endogenous promotor in a more physiological environment (Supplementary Fig. 4a), we tested the impact of PU-WS13 in hepatic lineages prepared from iPSCs derived from a homozygous AAT-ZZ patient expressing AAT under its endogenous promoter (iHepZZ) (Fig. 2d–f and Supplementary Fig. 4b, c). Compared to secretion levels observed iHepZZ cells under the endogenous promoter (Supplementary Fig. 4a), transient transfection of AAT-Z in our experimental cell Huh7.5^null cell line generates an ~1.2-fold increase of monomer secretion and ~1.3-fold of polymer secretion (Supplementary Fig. 4a). Treatment of iHepZZ cells by PU-WS13 at either 0.5 µM or 1 µM substantially increases both the total secretion of AAT (Supplementary Fig. 4b, c; 1.5-fold and 1.7-fold, respectively) and the NE inhibitory activity (Fig. 2d, ~1.7-fold), suggesting that the functional improvement by PU-WS13 treatment is applicable for AAT variants expressed under the endogenous promotor found in the patient-derived cellular environment. Notably, PU-WS13 increased NE inhibitory activity of AAT-ZZ in iHepZZ cells (Fig. 2d) is a value close to that of AAT-MZ individuals who are largely healthy^38,58. Whereas we found that the secretion of AAT-Z monomer is significantly improved (Fig. 2e), the polymer secretion was decreased by PU-WS13 treatment when measured by ELISA using polymer-specific antibody 2C1 (Fig. 2f)⁵². Consistent with these observations, native gel analysis of the secreted AAT-Z from Huh7.5^null cells shows that the monomer fraction is increased while the polymer fraction is decreased by PU-WS13 treatment (Supplementary Fig. 4d–f). The improvement of total secretion and NE inhibitory activity of AAT-Z, the increase of both intracellular and secreted AAT-Z monomer pools, and the decrease of both intracellular and secreted AAT-Z polymer by PU-WS13 treatment are prevented by reducing of GRP94 expression using siRNA (Supplementary Fig. 5), suggesting that the effect of PU-WS13 is specific for GRP94. These results indicate that the inhibition of GRP94 ATPase activity by PU-WS13 not only increases the level of AAT-Z secretion, but also improves the quality of secreted AAT-Z by decreasing the secretion of polymer, resulting in an improvement of NE inhibitory activity.

Using SCV to define functional rescue across the AAT sequence

While many AAT variants are rescued by PU-WS13 when expressed in Huh7.5^null cells or iHepZZ cell, it is a variable response reflecting the differential impact of each residue on AAT folding and function managed by GRP94 (Fig. 2b, c). Compared to the basal state (Supplementary Fig. 6a; Pearson’s r = 0.77, P = 4.3 × 10¹⁶), PU-WS13 treatment significantly reduces the correlation between the level of secreted monomer and NE inhibitory activity (Supplementary Fig. 6b, Pearson’s r = 0.59, P = 2.6 × 10⁻⁸). This result suggests GRP94 ATPase inhibition by PU-WS13 can, surprisingly, differentially manage the relationship between folding and functional properties of AAT on a global residue-by-residue basis.

To begin to understand the mechanism by which GRP94 modulation corrects the AAT fold on a residue-by-residue basis at atomic resolution, we applied our Gaussian process (GP)-spatial covariance (SCV) (GP-SCV) principled machine learning approach through variation spatial profiling (VSP)^{1,17,18,19,20,21}. GP-SCV principled relationships generated through VSP is based on a statistical paradigm used to find value in complex physical landscapes (see “Methods”). VSP integrates sparse genotype information found in the worldwide population to predict the role of each amino acid residue contributing to a phenotype in the context of all other residues in the polypeptide sequence with defined uncertainty^1,34. Given the impact of secreted monomer and NE inhibitory activity on lung disease, we plotted the normalized residue positions for all the measured missense variants (Fig. 3a, x axis, full-length polypeptide chain noted “1”) relative to their secreted monomer levels (Fig. 3a, y axis) to understand the impact of GRP94 on NE inhibitory activity for the entire protein sequence (Fig. 3a, z axis, color scale) in the absence (Fig. 3a, left panel) or presence of PU-WS13 (Fig. 3a, right panel).

**Fig. 3: GP-based phenotype landscapes in response to GRP94 ATPase inhibition.**

GP-SCV principled relationships first compute the separation distances between any pairwise combinations of the input variants based on their variant residue positions along the polypeptide sequence and the y axis feature, in this case secreted monomer levels (Fig. 3a, black lines). VSP subsequently generates the associated variance with NE inhibitory activity for each pairwise comparison (see “Methods”). These values are modeled by a variogram to illustrate how the NE inhibitory activity changes according to the differences of monomer secretion in relationship to its position in the polypeptide sequence (Fig. 3b). While we found a strong correlation between monomer secretion and the NE inhibitory activity across the entire polypeptide in the native basal state suggesting that a strong integration of folding and function features are required for secretion (Fig. 3b and Supplementary Fig. 6c, black line), PU-WS13 treatment dramatically shortened the distance range for covariance to 0.11 (i.e., covering ~46 residues) (Fig. 3b, blue line). These results suggest that reduced GRP94 ATPase activity relaxes the stringency for many sequence regions of the fold to allow a more diverse collection of monomer-activity covarying relationships to be secreted from ER^1,17,34. Furthermore, the plateau value of the variogram representing the global variance of NE inhibitory activity is reduced by PU-WS13 treatment (Fig. 3b and Supplementary Fig. 6c), consistent with a general restoration of NE inhibitory activity to WT level for most AAT variants (Fig. 2b).

Variogram modeling of SCV relationships between known variants provides a rigorous platform for GP to generate a multi-dimensional ‘phenotype landscape’ to map the GP-SCV principled relationships relating AAT sequence position to monomer secretion to NE inhibitory activity for every residue in the AAT polypeptide chain where input x- and y-coordinate values are used to predict the z-coordinate value for all residues¹ (Fig. 3c) (see “Methods”). As a probabilistic platform, GP-SCV based analysis also generates an associated uncertainty (i.e., a GP-based variance) for each prediction output based on variant input (Fig. 3d). A two-component Gaussian mixture model is used to fit the distribution of the associated uncertainty to separate the high-confidence vs low-confidence predictions (Fig. 3d). The mean of the distribution for the low uncertainty is illustrated in the landscape as the outermost bold contour and values lower than the mean are illustrated by the inner contours using the standard deviation of the distribution (Fig. 3c). The regions with uncertainty below the mean capture high-confidence predictions found in the phenotype landscape for every residue across the full sequence, generally encompassing 95% of all residues^{1,17,18,19,20,21} attesting to the strength of the approach for assigning value across the entire landscape (Fig. 3c). The variogram parameters required for phenotype landscape build can also be modeled using a Bayesian approach^59,60,61,62 to provide a comprehensive quantification of uncertainty in the modeling (Supplementary Fig. 11). We found that the predictions and associated uncertainties generated by a more extensive Bayesian approach were similar to GP-SCV predictions using our best-fit variogram parameters (Supplementary Fig. 11a, b; Pearson’s r = 0.99 for prediction; Pearson’s r = 0.98 for uncertainty) (Supplementary Fig. 11), indicating the value of a computationally tractable approach to address disease management based on SCV principled relationships.

The phenotype landscape predicting the NE inhibitory activity in the basal state (Fig. 3c, left panel, vehicle control) achieves significant prediction accuracy based on the leave-one-out cross-validation (Pearson’s r = 0.75, P = 5 × 10⁻¹⁴). Furthermore, the correlations between the predictions and experimental values are strong for the phenotype landscape at residues 1–139 (Pearson’s r = 0.78, P = 1.7 × 10⁻⁵), 140–278 (Pearson’s r = 0.7, P = 0.003) and 279–418 (Pearson’s r = 0.73, P = 9 × 10⁻⁷), illustrating the high value of the platform even for select regions distributed across the entire polypeptide chain. Strikingly, PU-WS13 treatment dramatically changes the NE inhibitory activity phenotype landscape, shifting large areas found in the AAT basal state landscape with defective NE inhibitory activity (Fig. 3c; left panel, red–orange–yellow) to normal activity (Fig. 3c, right panel, green), though there remain regions in the AAT polypeptide sequence that are insensitive to correction by PU-WS13 (Fig. 3c, right panel, red–orange). In the basal state landscape, most of the variants are well connected mechanistically with one another, triggering functional deficiency through high-confidence SCV relationships as illustrated by the broad contour lines (Fig. 3c, left panel, contours). PU-WS13 treatment largely shrinks and isolates these high-confidence regions (Fig. 3c, right panel, contours; Fig. 3d) and decreases the prediction accuracy (Pearson’s r = 0.45, P = 9 × 10⁻⁵), consistent with the short spatial correlation range observed in the variogram in the presence of PU-WS13 (Fig. 3b, blue line). These results suggest that reduced GRP94 ATPase activity in response to PU-WS13 alters the residue–residue SCV relationships leading to dysfunction in the AAT basal state to a more flexible folding state resulting in an increase in both monomer secretion and the NE inhibitory activity. Improving the general plasticity of SCV relationships could be a common mechanism by which proteostasis handles mutational load to restore function, in this case through the management of GRP94 ATPase activity that modulates GRP94 cycling. These results suggest a very dynamic role for GRP94 binding of client when transitioning the polypeptide chain set-point from unfolded to more functional folded states. We next mapped this conversion on a residue-by-residue basis by viewing each residue as part of a collective of states reflected in their SCV relationships to every residue in the polypeptide chain.

Phenotypic mapping GRP94 ATPase managed to function for every residue in AAT sequence

The phenotype landscapes report a range of relationships between NE inhibitory activity and secreted monomer for each residue in the AAT sequence (Fig. 4a, a 3D visualization of Fig. 3c). Taking residue 366 where the AAT-Z variant is located as an example (Fig. 4a, gray slice), we can plot the predicted relationships between NE inhibitory activity and secreted monomer for this residue as a curve in the presence (Fig. 4b, blue curve) or absence (Fig. 4b, black curve) of PU-WS13. The error bars represent the standard deviation of each prediction generated by GP-based variance to indicate the prediction confidence, and the high-confidence predictions defined by the Gaussian mixture modeling (Fig. 3d) are highlighted by darker error bars (Fig. 4b). The GP-based predictions in response to PU-WS13 on residue 366 (Fig. 4b, blue curve) are strikingly different from the basal state (Fig. 4b, black curve) in the range of monomer secretion from ~30% to ~60% of WT (Fig. 4b, 0.3<x < 0.6), where the NE inhibitory activity is greatly improved by PU-WS13 treatment, as shown by AAT-Z variant (Fig. 4b, arrows). In contrast, GRP94 ATPase inhibition does not generate a significant impact in the range of WT-like monomer secretion (Fig. 4b, x > 0.6), suggesting that GRP94 mainly manages the variants associated with misfolding and defective trafficking.

**Fig. 4: Residue-by-residue responses of NE inhibitory activity to PU-WS13.**

To understand the contribution of each residue in the correction of the pathogenic AAT fold by PU-WS13, we used inverse variance weighting (IVW) to average the predicted NE inhibitory activity in the context of monomer secretion across the entire landscape for each residue²¹. IVW allows us to assign the likelihood of function for a given residue position for the entire sequence that will be disrupted by variants with diverse chemical properties²¹, emphasizing that it is the probabilistic value of function derived from the collective of all variants that can be used to generate the SCV maps that rigorously address the role a given residue in the AAT wild-type protein sequence fold function. As expected, this function can be disrupted to different degrees by the different chemical properties of a particular variant residue found in the population impacting severity of disease. By using the reciprocal of prediction uncertainty (i.e., GP-based variance) as weight, IVW prioritizes the prediction with high confidence (i.e., low uncertainty/GP-based variance) over the prediction with low confidence (i.e., high uncertainty/GP-based variance) during the averaging (Fig. 4b). Through GP analyses (Fig. 4a, b), the IVW averaged value reports the most likely role of each residue contributing to the NE inhibitory activity based on SCV relationships. We also applied IVW to the phenotype landscape generated by Bayesian analysis, which yielded almost identical residue-based NE inhibitory activity for each residue across the AAT polypeptide (Supplementary Fig. 11f, g; Pearson’s r = 0.98), suggesting the robustness of this approach to estimating the most likely contribution of each residue to the function. The mean values for each residue are plotted as linear barcodes from the N-terminal to C-terminal of AAT polypeptide chain reflecting the NE inhibitory activity at the basal state (Fig. 4c, upper barcode (DMSO vehicle)) and after the treatment of PU-WS13 (Fig. 4c, middle barcode), as well as the delta (Δ) value between them (Fig. 4c, lower barcode). At the basal state with the DMSO vehicle control (Fig. 4c, upper barcode), the IVW barcode map defines three sequence regions at the N-terminal (N1), middle (M2) and carboxyl-terminal (C3) regions along the AAT polypeptide that are clustered with variants disrupting the NE inhibitory activity (Fig. 4c; labeled brackets, N1, M2, and C3) reflecting potential hotspots for managing protein misfolding.

These unanticipated results based on our sequence analysis and published structural efforts to date^63,64,65 suggest that AAT has evolved both short- and long-range cooperative interactions to manage NE inhibitory activity that could be sensitive to proteostasis management. Strikingly, all these regions are largely rescued by PU-WS13 treatment (Fig. 4c). Specifically, hA-s6B-hB and hD in the N1 region, a region around the S variant in M2 and most of C3 region, are highly rescued by PU-WS13 (Fig. 4c). These results suggest that GRP94, from a SCV perspective, can act at specific hotspots that involve different features of the fold protein spanning the entire protein sequence, possibly driven by regional fold energetics⁶⁶, that as a collective globally respond to not just variation in the population but to proteostasis management to improve activity. We next explored how this conversion occurs.

GRP94 ATPase activity can manage long-range residue interactions to correct AAT activity

To understand the structural mechanism(s) of correction by PU-WS13 treatment, we mapped the NE inhibitory activity predicted by GP analysis for each residue to the AAT 3D structure (PDB:3NE4)⁶⁷ to generate a “functional-structure” before (Fig. 4d) or after PU-WS13 treatment (Fig. 4e), as well as the structure representing the delta (Δ) value in response to PU-WS13 treatment (Fig. 4f). Interestingly, though N1, M2, and C3 sequence regions that contribute the NE inhibitory activity at the basal state are separated in the primary sequence of the polypeptide (Fig. 4c, upper barcode), they interact with each other through the β-sheet B (Fig. 4c, s6B strand for N1, s1-3B strand for M2 and s4-5B strand for C3, red dotted circles; Fig. 4d). The integration of each of the core SCV regions through β-sheet B indicates that the long-range interactions are crucial to evolve the function-structure relationships defining the AAT architecture (Supplementary Fig. 6d–h).

PU-WS13 administration generates corrections on many residues in the AAT structure to restore WT-like NE inhibitory activity to the AAT fold (Fig. 4e vs Fig. 4d, red–orange–yellow to green). Intriguingly, we discovered a core region that is highly responsive to PU-WS13 management, which encompasses residue interactions between hA-s6B from region N1, hG from M2 and s5B from region C3 (Fig. 4f, g). Notably, the highly responsive region to PU-WS13 treatment also involves the interactions between the N-terminal β-strand s6B and C-terminal β-strand s5B that are critical for assembling β-sheet group B (Fig. 4g). These results reveal that the integrity of β-sheet group B is critical for GRP94 managed NE inhibitory activity. These observations are consistent with the view that GRP94, as a member of the Hsp90 family, is generally involved in later stages of protein folding or in post-translational folding^14,15,68,69. Here, reduced ATPase activity may generate a more favorable kinetic/thermodynamic window for the N-terminal to C-terminal cooperative folding events disrupted by variants in the population, thereby significantly improving NE inhibitory activity correction (Fig. 4f, g). These results provide a rationale for the dynamic regulation of Hsp90 family members by co-chaperones^27,28,29 that could ensure coordination of folding with variation in sequence in the population to achieve an optimal folded state for function. Interestingly, non-responding or negatively impacted residues are clustered in the “bottom” of the AAT structure, comprising hC, s1A-hF, and the loop following hI to the beginning of s5A (Fig. 4c, lower barcode, Fig. 4f, h). These residue–residue relationships could potentially manage post-ER features involved in NE inhibitory activity. Overall, GP-principled SCV relationships provide a residue-by-residue roadmap of how GRP94 ATPase inhibition can differentially manage the AAT fold in response to variation to reshape/reinvent sequence-to-function-to-structure relationships.

Residue-by-residue-based correction of AAT monomer secretion

As GRP94 ATPase inhibition changes NE inhibitory activity and its relationship to monomer secretion for a number of variants, we wanted to understand the residue–residue basis for the correction of monomer secretion. For this purpose, we used NE inhibitory activity as the input y axis feature to predict as output the role of each residue in monomer secretion as the z axis feature of the phenotype landscape (Supplementary Fig. 7). PU-WS13 treatment reduced both the correlation range and the plateau value in the variogram modeling and decreased the leave-one-out cross-validation (from Pearson’s r = 0.65, P = 5 × 10⁻¹⁰ to Pearson’s r = 0.32, P = 0.007) (Supplementary Fig. 7b–d). These results are consistent with the interpretation that PU-WS13 treatment decreases the more rigid coupling between NE inhibitory activity and monomer secretion for AAT variants compared to the basal state without PU-WS13 treatment (Supplementary Fig. 6a, b).

To generate a residue-by-residue description of monomer secretion response to PU-WS13 treatment, we averaged the monomer secretion phenotype landscapes for each residue (Supplementary Fig. 7d) using IVW to generate residue-based monomer secretion barcodes (Fig. 5a). These mean values were mapped onto the AAT structure (Fig. 5b–f). We observed an increase of monomer secretion in the core cluster we identified above as a high response cluster for improving NE inhibitory activity (Figs. 4g and 5e). This core region comprises the interactions between hA-s6B (N1), hG (M2) and s5B (C3), suggesting that the high response of NE inhibitory activity in this region to PU-WS13 treatment (Fig. 4g) is at least partially due to the improvement of monomer secretion (Fig. 5e). However, the improvement of monomer secretion is much weaker (Fig. 5e) when compared to the NE inhibitory activity response (Fig. 4g), suggesting that there are additional effects generated by PU-WS13 treatment that increases the NE inhibitory activity independent of the level of monomer secretion. Indeed, we found that the top responding residues and structural regions for monomer secretion enhancement by PU-WS31 treatment (Fig. 5a, lower barcode; Fig. 5d) are in different sequence positions compared to those rescuing NE inhibitory activity (Fig. 4c, lower barcode; Fig. 4f). For example, the highly responsive region s2A-hE-s1A leading to improved monomer secretion (Fig. 5a, lower barcode; Fig. 5d, f) does not result in a reciprocal improvement of NE inhibitory activity (Fig. 4c, lower barcode; Fig. 4f). These results further suggest that GRP94 ATPase inhibition increases the tolerance of the ER for AAT folding in a manner that can differentially impact the monomer secretion and the NE inhibitory activity on a residue-by-residue basis. We next explored the impact of these changes on intracellular polymer accumulation- the hallmark of liver disease^32,70,71,72.

**Fig. 5: Residue-by-residue responses of monomer secretion to PU-WS13.**

Managing the N-terminal to C-terminal covariance folding of AAT to reduce polymer load

Given that the AAT-ZZ variant results in both intracellular and secreted polymer pools that can be corrected by PU-WS13 treatment (Fig. 2f and Supplementary Figs. 4d–f and 5f, h), we wanted to understand the molecular basis for these events across the entire AAT sequence. Therefore, we measured both the intracellular (Fig. 6a) and secreted polymer (Supplementary Fig. 8a) for all AAT variants in the absence or presence of PU-WS13 using an ELISA assay based on the polymer-specific antibody 2C1 (see ref. ⁵²). Among all measured 76 variants, 27 of them have intracellular polymer level at levels that are at least 50% of the severe intracellular AAT-Z polymer pool (Fig. 6a). In all, 13 of the 27 variants with high intracellular polymer (~48%) showed significant reduction in the intracellular polymer pool in the presence of PU-WS13 (Fig. 6a). Similarly, among all measured 76 variants, 20 of them have secreted polymer above 50% of the secreted AAT-Z polymer level (Supplementary Fig. 8a). PU-WS13 treatment significantly reduced the secreted polymer for 11 of the 20 variants with high secreted polymer (~55%) (Supplementary Fig. 8a). These results suggest that PU-WS13 has a broad impact on the AAT fold to decrease the polymer pools for around a half of the polymerogenic AAT variants, a metric highly relevant to monitoring liver disease in the patient population^32,70,71,72.

**Fig. 6: Residue-by-residue responses of intracellular polymer to PU-WS13.**

Consistent with the clinical observations that serum polymer level can be used as a biomarker for liver disease^71,72, the measured secreted polymer is strongly correlated with intracellular polymer for all the variants in both vehicle DMSO control (Supplementary Fig. 8b, left panel; Pearson’s r = 0.92, P = 4.3 × 10⁻³¹) and PU-WS13 treatment condition (Supplementary Fig. 8b, right panel; Pearson’s r = 0.85, P = 2 × 10⁻²²). This result indicates that the reduction of secreted polymer for AAT variants by PU-WS13 treatment is likely due to the reduced level of the intracellular polymer as suggested before^73,74,75.

To generate a residue-by-residue response map to PU-WS13 treatment for the intracellular polymer triggering liver disease, we built phenotype landscapes using secreted polymer as the y-coordinate to predict the intracellular polymer (z-coordinate) for every residue comprising the AAT polypeptide sequence (Supplementary Fig. 8c–f). We then used IVW to generate the residue-based barcodes (Fig. 6b) and related functional structure of AAT (Fig. 6c–f) to annotate the impact of PU-WS13 on intracellular AAT polymer pools. We also mapped the response of intracellular polymer pools to PU-WS13 to the more native AAT-Z polymer structure recently obtained using protein isolated from a ZZ-homozygous patient liver that revealed a “C-terminal polymerization” cryo-EM model^63,64. In the ‘C-terminal polymerization’ model, not only the RCL inserts as s4A in β-sheet A by “loop-sheet” insertion mechanism but also s4B and s5B insert into the β-sheet B of another molecule^63,64 (Fig. 6g-i). At the AAT basal state (Fig. 6c, g), the GP-based covariance prediction for polymer assembly is consistent with the C-terminal polymerization model (Fig. 6g) that highlights the critical role of assembly of β-sheet B in the polymer formation triggered by variation in the AATD population (Fig. 6c, g).

Strikingly, the residue interactions between s6B from the N1 region and s5B from the C3 region are clearly the highly responsive regions to PU-WS13 leading to significant polymer reduction (Fig. 6b, lower barcode; 6e, f, i). In contrast, the polymerogenic residues found in s5A that are critical for the “loop-sheet” insertion polymerization mechanism do not or only weakly respond to PU-WS13 (Fig. 6e, i). These results reveal that GRP94 ATPase inhibition by PU-WS13 primarily manages the N-terminal to C-terminal cooperative folding driving β-sheet B assembly to reduce the polymer accumulation in the liver-derived cells^63,64. Thus, by profiling the residue-by-residue response to GRP94 ATPase inhibition using GP-principled SCV relationships, we have uncovered an unanticipated mechanism used by the GRP94 chaperone to manage the polymerization state of AAT fold triggering aggregation disease.

Tracing the global management of function on a residue-by-residue basis by GRP94 ATPase activity

To generate a global view of how inhibition of GRP94 ATPase activity can differentially manage each residue of AAT fold for polymer accumulation in the ER, monomer secretion, and NE inhibitory activity in the hepatocyte in response to human variation, we overlaid the residue-based Δ value for each of these features in response to PU-WS13 treatment (Fig. 7a, gray (monomer), magenta (NE activity), cyan dots (intracellular polymer)). We observed a general improvement for all the three features on the N-terminal half of the N1 region encompassing hA-s6B-hB and the C3 region housing the s4B-s5B strands, suggesting overall a critical role of GRP94 managed N-terminal to C-terminal cooperative folding events to assemble β-sheet B to prevent polymer accumulation, leading to enhanced monomer secretion and NE inhibitory activity through manipulation of its ATPase activity (Fig. 7a). Interestingly, we also observed sequence regions that show improvement of monomer secretion and NE inhibitory activity without the correction of intracellular AAT polymer. For example, there is no change in the intracellular polymer pool for the sequence region spanning the C-terminal half of the N1 region to the beginning of the M2 region (Fig. 7a), residues that do not contribute to polymer formation in the basal state but instead contribute to monomer secretion and NE inhibitory activity. Notably, PU-WS13 treatment does generate a general and large improvement in both monomer secretion and NE inhibitory activity for selective sequence fragments in this region (Fig. 7a). These results further emphasize that the rescue of export from ER and restoration of activity by reduced GRP94 ATPase activity can be separated from its impact on polymer. A second example of the stabilizing activity of GRP94 in response PU-WS13 is found in the C-terminal half of M2 region (Fig. 7a). This region contains hG-hH where the AAT “S allele” is located (Fig. 7a). PU-WS13 treatment only weakly impacts its lower intracellular polymer pool when compared to the Z-variant pool, but does increase both monomer secretion and NE inhibitory activity (Fig. 7a). These results further emphasize that GRP94 ATPase inhibition by PU-WS13 can differentially manage the fold properties for specific residue regions to improve monomer secretion and NE inhibitory function in a manner that can be either dependent or independent on the AAT polymer status.

**Fig. 7: Differential responses of different phenotypes for each residue of AAT to GRP94 ATPase inhibition.**

To better quantitate these residue-specific differential responses, we calculated the delta (Δ) of the ratio between the NE inhibitory activity and secreted AAT monomer (referred to hereafter simply as “activity-to-monomer ratio”) for each residue in response to PU-WS13 treatment (Fig. 7b) and mapped these values onto AAT monomer structure (Fig. 7c), or the AAT-NE complex structure (PDB: 2D26)⁷⁶ (Fig. 7d). Interestingly, we found that the M2 region shows a general increase of the activity-to-monomer ratio (Fig. 7b–d). M2 includes s4C-s3C that comprises the previously suggested “gate” area^77,78, and s1B-s3B-s3B-hG-hH that forms “latch” interactions with the “gate”⁷⁷, both of which have been suggested to regulate the loop-sheet mechanism required for extracellular AAT function^77,78,79. Furthermore, s5A-RCL in C3 region that forms the loop-sheet insertion as s5A-s4A in the functional cycle of AAT with NE (Fig. 7d) also shows an increased activity-to-monomer ratio (Fig. 7b–d). These results suggest that for the region associated with the loop-sheet insertion mechanism, GRP94 ATPase inhibition potentially generates a more flexible AAT fold to improve the monomer ‘specific activity’ in its reaction with NE. In contrast to the improvement of the activity-to-monomer ratio, there are residues that show reduced responses (Fig. 7b–d). Interestingly, they are mostly clustered at the binding interface between AAT and NE (Fig. 7d, red residues), suggesting that the increased extracellular AAT monomer pool generated in response to PU-WS13 for these residues reflects a disrupted interaction with NE.

In summary, GP-principled SCV relationships generated in the absence or presence of PU-WS13 reveals at a residue-by-residue level an unanticipated pliable state of the protein fold and its sequence-to-function-to-structure relationships responsive to GRP94 ATPase activity—providing a plausible mechanistic view of Hsp90 family ATPase activity in shaping the protein fold in response to variation in the population.

Discussion

To evolve protein form and function requiring dynamic and metastable structural changes, a general computational framework is needed for understanding the balance in sequence-to-function-to-structure relationships dictating the genome to proteome transformation. A key feature in managing protein folding and function in response to human variation and the environment is proteostasis. In the case of AATD, past efforts focused on developing small molecules or peptides that block the aggregation of AAT-Z by interfering with the loop-sheet insertion mechanism responsible for NE inhibitory activity^{40,80,81,82,83,84}. While preventing polymer formation in the liver, they were of limited utility for lung disease in AATD given that the loop-sheet insertion is essential for AAT function. Another strategy to reduce AAT-Z polymer load in the ER was to activate the autophagy pathway to remove polymer burden^85,86,87, an approach that does not address the critical loss-of-function in the lung responsible for the vast majority of pathologies observed in the clinic.

By profiling the genetic and phenotypic diversity associated with AATD through GP-SCV principled relationships in response to the ATPase inhibitor PU-WS13, we uncovered unexpected and precision residue–residue coupled relationships that simultaneously correct the defects in polymer formation, monomer folding, secretion, and NE inhibitory activity of most pathogenic AAT variants. This detailed phenotypic information cannot be captured in a conservation score derived from sequence alignment across different species (Supplementary Fig. 9), suggesting that GP-SCV principled relationships capture complex evolutionary information driving function revealed by the natural process of genetic diversity in the human population. Whereas past studies have focused on chemical kinetic stabilization of the fold to prevent diseases of misfolding, including transthyretin^88,89 and immunoglobulin light chain amyloidosis⁹⁰, Gaucher’s^91,92 and cystic fibrosis^21,93, GP-SCV provides a fresh approach to understand more completely the complex features of cooperative folding and fold architecture to develop insights into challenging sequence-to-function-to-structure relationships that could prevent misfolding while maintaining activity.

The improvement in NE inhibitory activity in response to PU-WS13 is largely correlated with improved monomer secretion along with decreasing both intracellular and secreted polymer for most AAT variants, including AAT-Z variant, shifting the folding balance from disease to a healthier state of the fold (Fig. 8). Specifically, a reduction of GRP94 ATPase activity revealed its role in managing the N-terminal (s6B) to C-terminal (s5B) cooperative folding of AAT essential to assemble β-sheet B contributing to NE inhibitory activity. This conclusion is supported by the ‘C-terminal polymerization’ model based on the cryo-EM structure study of AAT-Z polymer from patient liver⁶³, which indicates an important role of the assembly or disassembly of β-sheet B in polymerization of AAT variants leading to aggregation in the ER⁶³. Furthermore, the interaction region between s6B and s5B is distant to the reaction loop (RCL), therefore, unlike the loop-sheet mechanism, managing s6B-s5B interactions would not be expected to impact the extracellular NE inhibitory activity, and may even increase the monomer-specific activity as suggested by PU-WS13 treatment on AAT-Z and other variants (Fig. 8).

**Fig. 8: GRP94 manages the N- to C-terminal cooperative folding of AAT to shape the balance between AAT aggregation monomer secretion for function.**

It is interesting that at a concentration of 1 µM PU-WS13 used in all the assays (Supplementary Fig. 10a–e), we did not observe up-regulation of BiP⁹⁴ (Supplementary Fig. 10f). This is in contrast to siGRP94 which up-regulates the level of BiP in macrophages⁹⁵. These results suggest that GRP94 in the PU-WS13 bound form may be required for the rescue effect, for example, to facilitate the release of AAT variant to ERGIC-53 for secretion (Supplementary Fig. 10g, h). Moreover, whereas inhibition of GRP94 through PU-WS13 has been shown to reduce the stability of HER2 in a tumor-specific manner where siGRP94 mimics the effect of PU-WS13 to destabilize HER2²⁴, in this study we found that whereas PU-WS13 rescues the folding, secretion, and function of AAT-Z, silencing through siGRP94 does not rescue AAT-Z defects. Thus, SCV relationships capture residue-specific features of GRP94 ATPase function affecting its role in promoting uniquely AAT-Z stability and function. These results suggest that the correction effects we observed are part of a larger programmable system in which GRP94 ATPase activity (and potentially other factors⁹⁵) may be normally reprogrammed by co-chaperones to achieve optimal function in response to variation and the environment as has been observed for the Hsp70 family of chaperone/co-chaperone relationships^{89,90,96,97,98,99}.

The residue-by-residue responses to GRP94 ATPase inhibition is different from that we recently found for ATF6 activators that more globally control the differential activities of multiple Hsp70 and Hsp90 chaperones and their respective co-chaperones in the ER in response to misfolding stress and the environment³⁴. For example, in terms of NE inhibitory activity, s6B-s5B interactions only moderately respond to ATF6 activators³⁴. In contrast, the “gate” area that is strongly impacted by ATF6 activators³⁴ is only moderately responsive to GRP94 ATPase inhibitor. These observations highlight the fact that different proteostasis components manage different parts of the protein fold to allow for what we refer to as “functional conversion”—not unlike the PrPc/PrPs transitions leading to prion disease^100,101. These results raise the possibility that conversion is a common principle in biology³⁴. Moreover, whereas Hsp90 family members are generally thought to clamp the clients between the pocket created by the dimer interface in cryo-EM snapshots^102,103, the differing SCV relationships we have now defined for every residue using GP logistics raises the possibility of a potential sequence scanning role for Hsp90 family members for management of the fold trajectory in real time.

We now suggest that understanding GP-SCV principled relationships provides a framework for making use of a naturally occurring programmable system in which the Hsp90 ATPase activity is constantly managed by co-chaperones in response to the environment^{16,27,28,29,69} to achieve optimal function for fitness through variation (see below). As GRP94 has been shown to be involved in the ERAD of AAT variants^45,46,47, a process that is controlled by ER mannosidase I (ERManI/Man1b1)^{104,105,106,107,108,109}, particularly in infants of AATD patient population¹⁰⁸, it is likely that GRP94 establishes a variable level of stringency we have referred to as a plasticity set-point^1,20 that is defined by the activity of the resident co-chaperone regulators in each cell type, and in response to the environment and aging (Fig. 8). Our SCV relationships indicate that GRP94 ATPase inhibition can release this SCV based set-point stringency to allow the secretion of AAT variants with improved monomer-activity relationships- even leading to improved specific activity for some variants despite the presence of the variant residue in the secreted monomer³⁴. These results suggest that GP-SCV principled management strategies impacting the balance between folding for function and targeting to degradation, including autophagy pathways^85,86,87, will be of high value for disease management in the context of a covariant understanding delineating the probabilistic uncertainty inherent in human biology impact by the sequence features contributing to the clinical functional outcomes impacting health along the coupled liver–lung axis.

Our results support the view that the role of proteostasis is but one part of an integrated “quality system (QS)”²⁰ that is responsive to genetics, development, aging, and the environment to continually reshape sequence-to-function-to-structure relationships driving protein fold dynamics contributing to health and disease^{1,17,20,34,110} that is not captured in static structures^21,34. As it is well established by the pioneering efforts of Lindquist and others^{68,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127} that Hsp90 serves as a capacitor for natural selection and evolution, SCV principled relationships now suggest that Hsp90’s role in capacitance management of variation occurs on a residue-by-residue basis- but as a collective. The response of Hsp90 family members to ATPase inhibitors suggests their role in cellular QS²⁰ is acutely tuned by the use of different co-chaperone partners to facilitate fitness^{6,14,15,44,128}. Understanding the residue-based conversion of the protein fold and function^{1,17,18,19,20,21,32,34} through an SCV principled standard model¹²⁹ could provide a common framework from a machine learning perspective to assess the role of each of the over 2500 proteostasis components in the management of information flow from the genome to the functional proteome^130,131. Knowledge of how information flow is managed by proteostasis from a covariance principled standard model could be used to implement precision interventions in a broad range of inherited and somatic human diseases including neurodegeneration^{4,5,9,10,132,133} and cancer^{43,134,135,136} where management of variation is key to restoration of healthspan (Fig. 8).

Methods

Reagents

DMEM, LHC-8 medium, F-12 medium, fetal bovine serum (FBS), penicillin–streptomycin (P/S), and geneticin (G418) were purchased from Invitrogen Life Technologies Corporation (Carlsbad, CA). FuGENE6 Transfection Reagent Kit was purchased from Promega (Madison, WI). Plasmid DNA purification kit was purchased from QIAGEN Inc. (Valencia, CA). PU-WS13 was initially provided by Dr. Gabriela Chiosis (Memorial Sloan Kettering Cancer Center, NY), and then purchased from Cayman Chemical (Ann Arbor, MI). PU-WS13 was stocked at 10 mM in DMSO and stored at −20 °C. The goat anti-human AAT polyclonal antibody 80A was purchase from ICL Inc (Anaheim, CA) (Cat # GCYT-80A, 1:1000 dilution); The mice anti-human monomeric AAT antibody 16f8 were generated in-house (1:2000 dilution); The mice anti-human polymeric AAT antibody 2C1 and other antibodies used in the study were purchased from Hycult biotech (Wayne, PA) (Cat # HM2289, 1:1000 dilution). The ERGIC-53 antibody (Cat # 13364-1-AP, 1:20,000 dilution) and 6x-His Tag Monoclonal Antibody (HIS.H8) (Cat # 50-173-6447, 1:5000 dilution) were purchase from Proteintech Inc. GRP78(BiP) antibody (Cat # ab108615, 1:2000 dilution), GRP94 antibody (Cat # ab238126, 1:1000 dilution) and GAPDH antibody (Cat # ab8245, 1:5000 dilution) were purchased from Abcam (Waltham, MA). Secondary antibody Goat anti-Mouse HRP antibody (Cat # 32230; 1:5000 dilution), Goat anti-Rabbit HRP antibody (Cat# 32260, 1:5000 dilution) and Mouse anti-Goat HRP antibody (Cat#31400, 1:10,000 dilution) were all purchased from Thermo Fisher Science (Waltham, MA). GRP94 siRNA was purchased from Thermo Fisher Scientific (ID:119656). Human neutrophil elastase was purchased from Innovative Research (Novi, MI). NE fluorescence assay substrate 2Rh110 (Z-Ala-Ala-Ala-Ala) was purchased from Cayman Chemical (Ann Arbor, MI). All other chemicals were purchased from Sigma Chemical (St. Louis, MO).

Cell culture

IB3 cells stably transfected with AAT-Z, which were provided by T. Flotte, University of Massachusetts Medical School, Worcester, MA, were cultured in LHC-8 medium containing 10% (v/v) fetal bovine serum (FBS) and 100 µg/ml penicillin–streptomycin (P/S). Huh7.5 AAT knockout (KO) cells (Huh7.5^null), which were provided by Mark Brantly, University of Florida College of Medicine, Gainesville, FL., and Huh7 cells with His-tagged AAT-Z stably expressed (provided by T. Flotte, University of Massachusetts Medical School, Worcester, MA), were cultured in the DMEM/F-12 medium containing 10% FBS and primocin (100 µg/ml). Patient AAT-ZZ hiPSC cells were purchased from DefiniGEN (Cambridge, UK). Directed differentiation of hiPSC into hepatic lineage was carried out following the protocol provided by DefiniGEN.

AAT variant DNA constructs

AAT variant DNA constructs were in pcDNA3.1 (+) plasmid vector generated by Quintara Biosciences (Cambridge, MA). All plasmid sequences were validated by Genewiz, Inc (San Diego, CA)).

AAT-specific conformation (monomer or polymer) ELISA assays

Huh7.5 AAT^null cells were cultured in 96-well plates with the density 2 × 10⁴/well and transfected with AAT variant plasmids at 0.2 µg/well. After 24-h transfection and another 24-h PU-WS13 treatment, cells were washed with PBS and incubated with 100 µl/well FBS-free culture medium. After 3-h incubation, FBS-free culture medium was collected, and cell lysates were harvested. In total, 20 µl medium or cell lysates from each well were added into the capture antibody (goat anti-human AAT polyclonal antibody 80A) pre-coated and pre-blocked (with BSA) 96-well plate for ELISA assay. After overnight incubation at 4 °C, the medium or cells lysate were washed away by PBST. Conformation-specific antibody 16f8 or 2C1 (mouse anti-human) was added into plate and incubated for 2 h. After washing the conformation antibody by PBST, secondary HRP-conjugated goat anti-mouse antibody was incubated in the plate for 2 h. After washing the secondary antibody, TMB reagent was added into each well for 10 min reaction and then stopped by 2 M H₂SO₄. Plates were read by BioTek Synergy H1 Hybrid Reader (Santa Clara, CA) at 454 nm for absorbance. The reading values were normalized by total protein levels of each well sample (measured by Bradford assay). Monomer levels of AAT variant were normalized by WT monomers. Polymer levels of AAT variant were normalized by AAT-Z polymers. Commercial purchased AAT proteins were subject to serial dilution and prepared for ELISA standardized curve.

AAT-NE inhibitory activity assay

Cell culture medium was collected as described above. In all, 20 µl culture medium was added into the capture antibody (goat anti-human AAT polyclonal antibody 80A) pre-coated and pre-block (with BSA) plate for overnight incubation. Medium was washed away and then the plate was incubated with Human Neutrophil Elastase 5 ng/well for 2 h at 37 °C. Overall, 25 pmol/well Neutrophil Elastase substrate 2RH110 was added into the plate and incubated for another 1.5 h. The plate was read at BioTek Synergy H1 Hybrid Reader (Santa Clara, CA) for excitation 485 nm and emission at 525 nm. The reading values were normalized by total protein levels for each sample. The anti-elastase activity of AAT variant was normalized by WT-AAT activity.

Immunoblotting

Plasmids of AAT variants were transiently transfected in Huh7.5 AAT knockout cells. After 24-h transfection and another 24-h PU-WS13 treatment, cells were washed by PBS and then switched to FBS-free culture medium for 3-h incubation. Culture medium was harvested. Cells were washed twice with 1× PBS and then lysed with 50 µl/well of 1× cell lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1% (v/v), Triton X-100, and protease inhibitors at 2 mg/ml) on ice for 30 min. Samples were collected and then centrifuged at 20,817×g at 4 °C for 20 min. The supernatant was collected. The protein concentration levels of culture medium and cell lysate were determined by the Bradford assay (Bio-Rad, Hercules, CA). Culture medium samples or cell lysate samples were resuspended in 1× SDS sample buffer containing β-mercaptoethanol and incubated at 95 °C for 5 min. Samples containing 20 µg of total protein were separated on a 10% (v/v) SDS-PAGE, transferred to nitrocellulose, and immunoblotted with goat anti-human AAT antibody (80 A). Detection was performed using chemiluminescence and the appropriate horseradish peroxidase-conjugated secondary antibodies. GAPDH was used in general as loading control.

siRNA-mediated silencing

IB3 cells stably transfected by AAT-Z were plated in 12-well tissue culture dishes and grown to 60% confluency. Silencing of GRP94 was performed by transfecting with RNAiMax (Invitrogen) and a final concentration of 50 nM siRNA (Ambion; siRNA targeting HSP90B1 gene) into cells according to the manufacturer’s protocol (Invitrogen). After 48 h of RNA transfected, PU-WS13 was added into the culture medium. AAT proteins were harvested as described above and subjected into immunoblotting analysis.

Immunoprecipitation assay

Huh7 cells, which stably expressed the His-tagged AAT-Z, were seeded and cultured in 12-well plates (with cell population density 1.2 × 10⁵/well) for 24-h and then treated with PU-WS13 at different concentrations for another 24 h. Cell lysates were collected as described above and diluted in the equilibration buffer. Samples with equal amount total proteins and equal volumes were incubated with HisPur™ Ni-NTA Magnetic Beads from Invitrogen (Carlsbad, CA) for 1 h in an end-over-end rotator. Magnetic beads were incubated and vortexed with wash buffer to remove the unbound proteins in the supernatant. Elution buffer was added and incubated with magnetic beads for 15 min on the rotator. His-tagged Z-AAT proteins in the supernatant were removed and used for immunoblot analysis.

Variation spatial profiling (VSP) to build phenotype landscape

The VSP analysis^1,17,21,34 was performed using the gstat package (V2.0)^137,138 in R. VSP is based on a geostatistical prediction technique, Kriging¹³⁹, which is widely used to model the spatial dependency as a Gaussian process to interpolate the unmeasured value to construct the phenotype landscape for AAT.

Briefly, AAT variants were positioned by their sequence positions in the polypeptide chain on the “x” axis coordinate and their impact on a phenotype on the “y” axis coordinate to the impact on another phenotype on the ‘z’ axis coordinate. Ordinary Kriging¹³⁹ was applied to model z_{(x, y)} as:

$${z}_{(x,y)}={\mu }_{(x,y)}+{W}_{(x,y)}+{\epsilon }_{(x,y)}$$

(1)

Where ${\mu }_{(x,y)}$ is a constant local mean. ${W}_{(x,y)}$ is the spatially correlated stochastic part of variation. ${\epsilon }_{(x,y)}$ is a white noise that captures measurement error and microscale variation. The spatial structure of the data can be examined by variogram analysis. Suppose the ith observation at coordinates (x_i) and (y_i) in a dataset consists of a value ${z}_{({x}_{i},{y}_{i})}$. And jth observation at coordinates (${x}_{j}$) and (${y}_{j}$) in a dataset consists of a value ${z}_{({x}_{j},{y}_{j})}$. They are separated by a vector h.

The distance $d(h)$ between the i^th and j^th observations is calculated by:

$$d(h)=\sqrt{{({x}_{i}-{x}_{j})}^{2}+{({y}_{i}-{y}_{j})}^{2}}$$

(2)

The γ(h)-variance for a given distance $d(h)$ is defined by:

$$\gamma \left(h\right)=\frac{1}{2}{({z}_{({x}_{i},{y}_{i})}-{z}_{({x}_{j},{y}_{j})})}^{2}$$

(3)

where γ(h)-variance is the semivariance (the degree of dissimilarity) of the z value between the two observations, which is also the whole variance of z value for one observation at the given separation distance h, referred to as spatial variance here. The distance (h) and spatial variance (γ(h)) for all the data pairs are generated by the Eqs. (2) and (3). Then, the average values of spatial variance for each distance interval are calculated to plot the averaged spatial variance versus distance. The fitting of variograms were determined using GS+ Version 10 (Gamma Design Software) by both minimizing the residual sum of squares (RSS) and maximizing the leave-one-out cross-validation result (see below). The variogram enables us to compute the spatial covariance (SCV) matrices for any possible separation vector. The SCV at a distance (h) is calculated by C(h)= C(0) − γ(h), where C(0) is the covariance at zero distance representing the global variance of the data points under consideration (the plateau of the variogram). The approach aims to generate the prediction that has minimized estimation error (error variance) which is generated according to the expression:

$${\sigma }_{u}^{2}=E\left[{\left({{z}_{u}}^{*}-{z}_{u}\right)}^{2}\right]=\mathop{\sum }\limits_{i=1}^{n}\mathop{\sum }\limits_{j=1}^{n}{\omega }_{i}{\omega }_{j}{C}_{i,j}-2\mathop{\sum }\limits_{i=1}^{n}{\omega }_{i}{C}_{i,u}+{C}_{u,u}$$

(4)

where$\,{{z}_{u}}^{*}$ is the prediction value while z_u is the true but unknown value, C_i,j and C_i,u are SCV between data points i and j, and data points i and u, respectively, and C_u,u is the SCV within location u. ω_i is the weight for data point i. The SCV is obtained from the above molecular variogram analysis and the weight (ω_i) solved from Eq. (4) is used for following prediction. To ensure an unbiased result, the sum of weight is set as one:

$$\mathop{\sum }\limits_{i=1}^{n}{\omega }_{i}=1$$

(5)

Equations (4) and (5) not only solved the set of weights associated with input observations, but also provide the minimized “molecular variance” at location u which can be expressed as:

$$\,{\sigma }_{u}^{2}={C}_{u,u}-\left(\mathop{\sum }\limits_{i=1}^{n}{\omega }_{i}{C}_{i,u}+{{{{{\rm{\mu }}}}}}\right)$$

(6)

where C_u,u is the SCV within location u, ω_i is the weight for data point i, and C_i,u are SCV between data points i and u. μ is the Lagrange Parameter that is used to convert the constrained minimization problem in Eq. (6) into an unconstrained one. The resulting minimized molecular variance assessing the prediction uncertainty presents the confidence level of the prediction. With the solved weights, we can calculate the prediction of all unknown values to generate the complete fitness landscape by the equation:

$${{{{{{{\rm{z}}}}}}}_{u}}^{*}=\mathop{\sum }\limits_{i=1}^{n}{{{{{{\rm{\omega }}}}}}}_{i}{{{{{{\rm{z}}}}}}}_{i}$$

(7)

where ${{{{{{{\rm{z}}}}}}}_{u}}^{*}$ is the prediction value for the unknown data point u, ω_i is the weight for the known data point, and z_i is the measured value for data point i. Leave-one-out cross-validation (LOOCV) is used to validate the computational model because of small sample size modeling¹⁴⁰. In the LOOCV, we remove each data point, one at a time and use the rest of the data points to predict the missing value. We repeat the prediction for all data points and compare the prediction results to the measured value to generate the Pearson’s r value and its associated P value (ANOVA test performed in Originpro version 2020b (OriginLab)).

Defining the high-confidence prediction using probabilistic clustering

Each prediction on the phenotype landscape is associated with a GP-generated variance to indicate the prediction confidence. To separate the high-confidence vs low-confidence predictions in the GP-generated phenotype landscape, we fit a Gaussian mixture model with two components over the GP-generated variance by using the mclust package (V.5.4.10) in R. Gaussian mixture model is a probabilistic clustering tool that not only separates the predictions with low variance vs high variance, but also outputs the probability distributions for them. We use the mean of the distribution for the low variance as a cutoff and define the predictions with a lower variance than the mean of the distribution as high-confidence prediction. The mean and the standard deviation of the distribution for the low variance (i.e., high confidence) are illustrated as contours in the phenotype landscape.

Inverse variance weighting (IVW) to build the residue-by-residue phenotype barcode

Phenotype landscapes built based on a sparse collection of input variants map the full range of values describing function (based on the y- and z axis metrics) for the entire polypeptide sequence on a residue-by-residue basis (x axis). To get an averaged value of predicted phenotype for each residue, we use the reciprocal of GP-generated variance for the high-confidence predictions as weights to aggregate the phenotype values by using the following equation:

$$\hat{z}=\frac{{\sum }_{i}\frac{{z}_{i}}{{\sigma }_{i}^{2}}}{{\sum }_{i}\frac{1}{{\sigma }_{i}^{2}}}$$

(7)

where $\hat{z}$ is the weighted mean value for each residue, ${z}_{i}$ is the predicted phenotype value at z axis for every value on the y axis, ${\sigma }_{i}^{2}$ is the GP-generated variance for each prediction. We repeat this process for all the residues. The IVW averaged mean values for all the residues then are mapped as barcodes or mapped on AAT structures (PDB:3NE4⁶⁷ for AAT monomer, PDB:2D26⁷⁶ for the AAT-elastase complex and PDB:3T1P⁶⁴ for polymer). All the atomic resolution structure presentations were produced with the software of PyMOL.

As Ordinary Kriging used in the above process does not quantify the uncertainty in the model parameters, we also used Bayesian Kriging to take account for the error introduced by the estimation of the variogram model. Specifically, we used the bayesGeostatExtact() function in the spBayes package^59,60,61,62. We found that the phenotype landscape predictions generated by Ordinary Kriging were highly correlated with those generated by Bayesian Kriging (Supplementary Fig. 11a, Pearson’s r = 0.99). The uncertainty values assessed by Bayesian Kriging are generally larger than those generated by Ordinary Kriging (Supplementary Fig. 11b), but they are also highly correlated (Supplementary Fig. 11b, Pearson’s r = 0.98). The resulting landscapes between Ordinary Kriging and Bayesian Kriging are almost identical (Supplementary Fig. 11c, d), with minor differences observed for the uncertainty (Supplementary Fig. 11c, d, contour lines, Supplementary Fig. 11e). After the IVW, the weighted average function value for each residue is again highly correlated (Supplementary Fig. 11f, Pearson’s r = 0.98), which generates almost identical functional structures (Supplementary Fig. 11g). As Bayesian Kriging is much more computationally expensive than Ordinary Kriging, we used Ordinary Kriging in this study, but these results suggest that more advanced techniques can be applied to VSP to understand the sequence-to-function-to-structure relationships.

Statistical analyses

Experimental data showed as the mean ± SD from three or more than three independent experiments. The differences between two independent groups were analyzed using Student’s t test. A P value < 0.05 was considered statistically significant. The P value of the Pearson’s r value is performed using ANOVA test. The null hypothesis for the P value of the Pearson correlation is that the correlation coefficient is not significantly different from 0. The P value represents the probability that the correlation occurred by chance. We used P < 0.05 to reject the null hypothesis and conclude there is a significant Pearson correlation. To compare the Pearson’s correlation coefficients, we used Fisher Z-transformation to convert the Pearson’s correlation coefficients into approximate normally distributed variables. We then conducted a hypothesis test with a null hypothesis where the transformed coefficients are equal. We used P < 0.05 to reject the null hypothesis.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

Source datasets for each figure are provided in the Source Data file. Due to large data size, source data for Figs. 3c, d and 4a and Supplementary Figs. 7c, d, 8e, f, and 11 are included in the Mendeley Data (https://doi.org/10.17632/nt59rwz6r6.1). PDB files used in this study include 3NE4, 2D26 and 3T1P. Source data are provided with this paper.

Code availability

The input data, R-code scripts, and output files for the GP-based analysis have been deposited in the public Mendeley database with https://doi.org/10.17632/nt59rwz6r6.1. The files are under CC BY 4.0 license.

References

Wang, C. & Balch, W. E. Bridging genomics to phenomics at atomic resolution through variation spatial profiling. Cell Rep. 24, 2013–2028 e6 (2018).
Gershenson, A., Gierasch, L. M., Pastore, A. & Radford, S. E. Energy landscapes of functional proteins are inherently risky. Nat. Chem. Biol. 10, 884–891 (2014).
Article CAS PubMed PubMed Central Google Scholar
Balch, W. E., Morimoto, R. I., Dillin, A. & Kelly, J. W. Adapting proteostasis for disease intervention. Science 319, 916–919 (2008).
Article ADS CAS PubMed Google Scholar
Sinnige, T., Yu, A. & Morimoto, R. I. Challenging proteostasis: role of the chaperone network to control aggregation-prone proteins in human disease. Adv. Exp. Med. Biol. 1243, 53–68 (2020).
Article CAS PubMed PubMed Central Google Scholar
Jayaraj, G. G., Hipp, M. S. & Hartl, F. U. Functional modules of the proteostasis network. Cold Spring Harb. Perspect. Biol. 12, a033951 (2020).
Article CAS PubMed PubMed Central Google Scholar
Powers, E. T. & Balch, W. E. Diversity in the origins of proteostasis networks—a driver for protein function in evolution. Nat. Rev. Mol. Cell Biol. 14, 237–248 (2013).
Article CAS PubMed PubMed Central Google Scholar
Sala, A. J. & Morimoto, R. I. Protecting the future: balancing proteostasis for reproduction. Trends Cell Biol. 32, 202–215 (2022).
Article CAS PubMed Google Scholar
Morimoto, R. I. Cell-nonautonomous regulation of proteostasis in aging and disease. Cold Spring Harb. Perspect. Biol. 12, a034074 (2020).
Article CAS PubMed PubMed Central Google Scholar
Hipp, M. S., Kasturi, P. & Hartl, F. U. The proteostasis network and its decline in ageing. Nat. Rev. Mol. Cell Biol. 20, 421–435 (2019).
Article CAS PubMed Google Scholar
Kelly, J. W. Pharmacologic approaches for adapting proteostasis in the secretory pathway to ameliorate protein conformational diseases. Cold Spring Harb. Perspect. Biol. 12, a034108 (2020).
Article CAS PubMed PubMed Central Google Scholar
Wiseman, R. L., Mesgarzadeh, J. S. & Hendershot, L. M. Reshaping endoplasmic reticulum quality control through the unfolded protein response. Mol. Cell 82, 1477–1491 (2022).
Article CAS PubMed PubMed Central Google Scholar
Grandjean, J. M. D. & Wiseman, R. L. Small molecule strategies to harness the unfolded protein response: where do we go from here? J. Biol. Chem. 295, 15692–15711 (2020).
Article CAS PubMed PubMed Central Google Scholar
Marciniak, S. J., Chambers, J. E. & Ron, D. Pharmacological targeting of endoplasmic reticulum stress in disease. Nat. Rev. Drug Discov. 21, 115–140 (2022).
Article CAS PubMed Google Scholar
Biebl, M. M. & Buchner, J. Structure, function, and regulation of the Hsp90 machinery. Cold Spring Harb. Perspect. Biol. 11, a034017 (2019).
Article CAS PubMed PubMed Central Google Scholar
Pugh, K. W., Alnaed, M., Brackett, C. M. & Blagg, B. S. J. The biology and inhibition of glucose-regulated protein 94/gp96. Med. Res. Rev. 42, 2007–2024 (2022).
Article CAS PubMed PubMed Central Google Scholar
Schopf, F. H., Biebl, M. M. & Buchner, J. The HSP90 chaperone machinery. Nat. Rev. Mol. Cell Biol. 18, 345–360 (2017).
Article CAS PubMed Google Scholar
Wang, C. et al. Quantitating the epigenetic transformation contributing to cholesterol homeostasis using Gaussian process. Nat. Commun. 10, 5052 (2019).
Article ADS PubMed PubMed Central Google Scholar
Wang, C. et al. Individualized management of genetic diversity in Niemann-Pick C1 through modulation of the Hsp70 chaperone system. Hum. Mol. Genet. 29, 1–19 (2020).
Article CAS PubMed Google Scholar
Wang, C., Elghobashi-Meinhardt, N. & Balch, W. E. Covariant fitness clusters reveal structural evolution of SARS-CoV-2 polymerase across the human population. Preprint https://www.biorxiv.org/content/10.1101/2022.01.07.475295v1 (2022).
Angles, F., Wang, C. & Balch, W. E. Spatial covariance analysis reveals the residue-by-residue thermodynamic contribution of variation to the CFTR fold. Commun. Biol. 5, 356 (2022).
Article PubMed PubMed Central Google Scholar
Wang, C., Angles, F. & Balch, W. E. Triangulating variation in the population to define mechanisms for precision management of genetic disease. Structure 30, 1190–1207 (2022).
Article CAS PubMed PubMed Central Google Scholar
Yan, P. et al. Molecular stressors engender protein connectivity dysfunction through aberrant N-glycosylation of a chaperone. Cell Rep. 31, 107840 (2020).
Article CAS PubMed PubMed Central Google Scholar
Patel, H. J. et al. Structure-activity relationship in a purine-scaffold compound series with selectivity for the endoplasmic reticulum Hsp90 paralog Grp94. J. Med. Chem. 58, 3922–3943 (2015).
Article CAS PubMed PubMed Central Google Scholar
Patel, P. D. et al. Paralog-selective Hsp90 inhibitors define tumor-specific regulation of HER2. Nat. Chem. Biol. 9, 677–684 (2013).
Article CAS PubMed PubMed Central Google Scholar
Que, N. L. S. et al. Structure based design of a Grp94-selective inhibitor: exploiting a key residue in Grp94 to optimize paralog-selective binding. J. Med. Chem. 61, 2793–2805 (2018).
Article CAS PubMed PubMed Central Google Scholar
Huck, J. D. et al. NECA derivatives exploit the paralog-specific properties of the site 3 side pocket of Grp94, the endoplasmic reticulum Hsp90. J. Biol. Chem. 294, 16010–16019 (2019).
Article CAS PubMed PubMed Central Google Scholar
Huang, B. et al. The endoplasmic reticulum chaperone BiP is a closure-accelerating cochaperone of Grp94. Proc. Natl. Acad. Sci. USA 119, e2118793119 (2022).
Article CAS PubMed PubMed Central Google Scholar
Rosenbaum, M. et al. MZB1 is a GRP94 cochaperone that enables proper immunoglobulin heavy chain biosynthesis upon ER stress. Genes Dev. 28, 1165–1178 (2014).
Article CAS PubMed PubMed Central Google Scholar
Liu, B. et al. Folding of Toll-like receptors by the HSP90 paralogue gp96 requires a substrate-specific cochaperone. Nat. Commun. 1, 79 (2010).
Article ADS PubMed Google Scholar
Strnad, P., McElvaney, N. G. & Lomas, D. A. Alpha1-antitrypsin deficiency. New Engl. J. Med. 382, 1443–1455 (2020).
Article CAS PubMed Google Scholar
Wang, C. & Balch, W. E. Managing the adaptive proteostatic landscape: restoring resilience in alpha-1 antitrypsin deficiency. In Alpha-1 Antitrypsin: Role in Health and Disease (eds. Wanner, A. & Sandhaus, R. A.) 53–83 (Springer International Publishing, Cham, 2016).
Wang, C., Zhao, P., Sun, S., Teckman, J. & Balch, W. E. Leveraging population genomics for individualized correction of the hallmarks of alpha-1 antitrypsin deficiency. Chronic Obstr. Pulm. Dis. 7, 224–246 (2020).
PubMed PubMed Central Google Scholar
Giacopuzzi, E. et al. Real-world clinical applicability of pathogenicity predictors assessed on SERPINA1 mutations in alpha-1-antitrypsin deficiency. Hum. Mutat. 39, 1203–1213 (2018).
Article CAS PubMed Google Scholar
Sun, S. et al. Capturing the conversion of the pathogenic alpha-1-antitrypsin fold by ATF6 enhanced proteostasis. Cell Chem. Biol. 30, 22–42.e5 (2023).
Article CAS PubMed PubMed Central Google Scholar
Patel, D., McAllister, S. L. & Teckman, J. H. Alpha-1 antitrypsin deficiency liver disease. Transl. Gastroenterol. Hepatol. 6, 23 (2021).
Article PubMed PubMed Central Google Scholar
Suri, A., Patel, D. & Teckman, J. Alpha-1-antitrypsin deficiency. Clin. Liver Dis. 19, 89–92 (2022).
Article Google Scholar
Perlmutter, D. H. Alpha1-antitrypsin deficiency: a misfolded secretory protein variant with unique effects on the endoplasmic reticulum. Cell Pathol. 3, 63–72 (2016).
CAS Google Scholar
Strange, C. Alpha-1 antitrypsin deficiency associated COPD. Clin. Chest Med. 41, 339–345 (2020).
Article PubMed Google Scholar
Brantly, M. L., Lascano, J. E. & Shahmohammadi, A. Intravenous alpha-1 antitrypsin therapy for alpha-1 antitrypsin deficiency: the current state of the evidence. Chronic Obstr. Pulm. Dis. 6, 100–114 (2018).
PubMed PubMed Central Google Scholar
Lomas, D. A. et al. Development of a small molecule that corrects misfolding and increases secretion of Z alpha1 -antitrypsin. EMBO Mol. Med. 13, e13167 (2021).
Article CAS PubMed PubMed Central Google Scholar
Lomas, D. A. New therapeutic targets for alpha-1 antitrypsin deficiency. Chronic Obstr. Pulm. Dis. 5, 233–243 (2018).
PubMed PubMed Central Google Scholar
Sanchez, J., Carter, T. R., Cohen, M. S. & Blagg, B. S. J. Old and new approaches to target the Hsp90 chaperone. Curr. Cancer Drug Targets 20, 253–270 (2020).
Article CAS PubMed PubMed Central Google Scholar
Duan, X. et al. Molecular chaperone GRP94/GP96 in cancers: oncogenesis and therapeutic target. Front. Oncol. 11, 629846 (2021).
Article CAS PubMed PubMed Central Google Scholar
Gewirth, D. T. Paralog specific Hsp90 inhibitors—a brief history and a bright future. Curr. Top. Med. Chem. 16, 2779–2791 (2016).
Article CAS PubMed PubMed Central Google Scholar
Zhong, Y. et al. Identification of ERAD components essential for dislocation of the null Hong Kong variant of alpha-1-antitrypsin (NHK). Biochem. Biophys. Res. Commun. 458, 424–428 (2015).
Article CAS PubMed Google Scholar
Schmidt, B. Z. & Perlmutter, D. H. Grp78, Grp94, and Grp170 interact with alpha1-antitrypsin mutants that are retained in the endoplasmic reticulum. Am. J. Physiol. Gastrointest. Liver Physiol. 289, G444–G455 (2005).
Article CAS PubMed Google Scholar
Christianson, J. C., Shaler, T. A., Tyler, R. E. & Kopito, R. R. OS-9 and GRP94 deliver mutant alpha1-antitrypsin to the Hrd1-SEL1L ubiquitin ligase complex for ERAD. Nat. Cell Biol. 10, 272–282 (2008).
Article CAS PubMed PubMed Central Google Scholar
Blanco, I. et al. Alpha-1 antitrypsin Pi*Z gene frequency and Pi*ZZ genotype numbers worldwide: an update. Int. J. Chron. Obstruct Pulmon Dis. 12, 561–569 (2017).
Article CAS PubMed PubMed Central Google Scholar
Zaimidou, S. et al. A1ATVar: a relational database of human SERPINA1 gene variants leading to alpha1-antitrypsin deficiency and application of the VariVis software. Hum. Mutat. 30, 308–313 (2009).
Article CAS PubMed Google Scholar
Landrum, M. J. et al. ClinVar: improving access to variant interpretations and supporting evidence. Nucleic Acids Res. 46, D1062–D1067 (2018).
Article CAS PubMed Google Scholar
Karczewski, K. J. et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 581, 434–443 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Miranda, E. et al. A novel monoclonal antibody to characterize pathogenic polymers in liver disease associated with alpha1-antitrypsin deficiency. Hepatology 52, 1078–1088 (2010).
Article CAS PubMed Google Scholar
Johnson, A. F., Struthers, M. D., Pierson, K. B., Mangel, W. F. & Smith, L. M. Nonisotopic DNA detection system employing elastase and fluorogenic rhodamine substrate. Anal. Chem. 65, 2352–2359 (1993).
Article CAS PubMed Google Scholar
Khodayari, N. et al. Modulation of calreticulin expression reveals a novel exosome-mediated mechanism of Z variant alpha1-antitrypsin disposal. J. Biol. Chem. 294, 6240–6252 (2019).
Article CAS PubMed PubMed Central Google Scholar
Khodayari, N. et al. Erdj3 has an essential role for Z variant alpha-1-antitrypsin degradation. J. Cell Biochem 118, 3090–3101 (2017).
Article CAS PubMed PubMed Central Google Scholar
Khodayari, N. et al. SVIP regulates Z variant alpha-1 antitrypsin retro-translocation by inhibiting ubiquitin ligase gp78. PLoS ONE 12, e0172983–e0172983 (2017).
Article PubMed PubMed Central Google Scholar
Bouchecareilh, M., Hutt, D. M., Szajner, P., Flotte, T. R. & Balch, W. E. Histone deacetylase inhibitor (HDACi) suberoylanilide hydroxamic acid (SAHA)-mediated correction of alpha1-antitrypsin deficiency. J. Biol. Chem. 287, 38265–38278 (2012).
Article CAS PubMed PubMed Central Google Scholar
Teckman, J. H. & Jain, A. Advances in alpha-1-antitrypsin deficiency liver disease. Curr. Gastroenterol. Rep. 16, 367 (2014).
Article PubMed Google Scholar
Banerjee, S., Carlin, B. & Gelfand, A. Hierarchical Modeling and Analysis for Spatial Data (Chapman and Hall/CRC, 2014).
Finley, A. O. & Banerjee, S. Bayesian spatially varying coefficient models in the spBayes R package. Environ. Model. Softw. 125, 104608 (2020).
Article Google Scholar
Finley, A. O., Banerjee, S. & Gelfand, A. E. spBayes for large univariate and multivariate point-referenced spatio-temporal data models. J. Stat. Softw. 63, 1–28 (2015).
Article Google Scholar
Finley, A. O., Banerjee, S. & Carlin, B. P. spBayes: an R package for univariate and multivariate hierarchical point-referenced spatial models. J. Stat. Softw. 19, 1–24 (2007).
Article PubMed PubMed Central Google Scholar
Faull, S. V. et al. The structural basis for Z alpha1-antitrypsin polymerization in the liver. Sci. Adv. 6, eabc1370 (2020).
Yamasaki, M., Sendall, T. J., Pearce, M. C., Whisstock, J. C. & Huntington, J. A. Molecular basis of alpha1-antitrypsin deficiency revealed by the structure of a domain-swapped trimer. EMBO Rep. 12, 1011–1017 (2011).
Article CAS PubMed PubMed Central Google Scholar
Huntington, J. A., Read, R. J. & Carrell, R. W. Structure of a serpin-protease complex shows inhibition by deformation. Nature 407, 923–926 (2000).
Article ADS CAS PubMed Google Scholar
Fersht, A. R., Matouschek, A. & Serrano, L. The folding of an enzyme. I. Theory of protein engineering analysis of stability and pathway of protein folding. J. Mol. Biol. 224, 771–782 (1992).
Article CAS PubMed Google Scholar
Patschull, A. O. et al. Therapeutic target-site variability in alpha1-antitrypsin characterized at high resolution. Acta Crystallogr Sect. F. Struct. Biol. Cryst. Commun. 67, 1492–1497 (2011).
Article CAS PubMed PubMed Central Google Scholar
Zabinsky, R. A., Mason, G. A., Queitsch, C. & Jarosz, D. F. It’s not magic—Hsp90 and its effects on genetic and epigenetic variation. Semin Cell Dev. Biol. 88, 21–35 (2019).
Article CAS PubMed Google Scholar
Singh, J. K. et al. Management of Hsp90-dependent protein folding by small molecules targeting the Aha1 co-chaperone. Cell Chem. Biol. 27, 292–305 e6 (2020).
Article CAS PubMed PubMed Central Google Scholar
Ordonez, A., Harding, H. P., Marciniak, S. J. & Ron, D. Cargo receptor-assisted endoplasmic reticulum export of pathogenic alpha1-antitrypsin polymers. Cell Rep. 35, 109144 (2021).
Article CAS PubMed PubMed Central Google Scholar
Núñez, A. et al. Association between circulating alpha-1 antitrypsin polymers and lung and liver disease. Respir. Res. 22, 244 (2021).
Article PubMed PubMed Central Google Scholar
Sark, A. D. et al. The relationship between plasma alpha-1-antitrypsin polymers and lung or liver function in ZZ alpha-1-antitrypsin-deficient patients. Biomolecules 12, 380 (2022).
Article CAS PubMed PubMed Central Google Scholar
Ordóñez, A., Harding, H. P., Marciniak, S. J. & Ron, D. Cargo receptor-assisted endoplasmic reticulum export of pathogenic α1-antitrypsin polymers. Cell Rep. 35, 109144 (2021).
Article PubMed PubMed Central Google Scholar
Tan, L. et al. Circulating polymers in α1-antitrypsin deficiency. Eur. Respir. J. 43, 1501–1504 (2014).
Article CAS PubMed Google Scholar
Fra, A. et al. Polymers of Z α1-antitrypsin are secreted in cell models of disease. Eur. Respir. J. 47, 1005–1009 (2016).
Article CAS PubMed Google Scholar
Dementiev, A., Dobo, J. & Gettins, P. G. Active site distortion is sufficient for proteinase inhibition by serpins: structure of the covalent complex of alpha1-proteinase inhibitor with porcine pancreatic elastase. J. Biol. Chem. 281, 3452–3457 (2006).
Article CAS PubMed Google Scholar
Fra, A. M. et al. Three new alpha1-antitrypsin deficiency variants help to define a C-terminal region regulating conformational change and polymerization. PLoS ONE 7, e38405 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Wardell, M. R. et al. Preparative induction and characterization of L-antithrombin: a structural homologue of latent plasminogen activator inhibitor-1. Biochemistry 36, 13133–13142 (1997).
Article CAS PubMed Google Scholar
Chang, W. S. et al. Importance of the release of strand 1C to the polymerization mechanism of inhibitory serpins. Protein Sci. 6, 89–98 (1997).
Article CAS PubMed PubMed Central Google Scholar
Zhou, A. et al. How small peptides block and reverse serpin polymerisation. J. Mol. Biol. 342, 931–941 (2004).
Article CAS PubMed Google Scholar
Ronzoni, R. et al. The molecular species responsible for alpha1-antitrypsin deficiency are suppressed by a small molecule chaperone. FEBS J. 288, 2222–2237 (2021).
Article CAS PubMed Google Scholar
Liddle, J. et al. The development of highly potent and selective small molecule correctors of Z alpha1-antitrypsin misfolding. Bioorg. Med. Chem. Lett. 41, 127973 (2021).
Article CAS PubMed Google Scholar
Mallya, M. et al. Small molecules block the polymerization of Z alpha1-antitrypsin and increase the clearance of intracellular aggregates. J. Med. Chem. 50, 5357–5363 (2007).
Article CAS PubMed PubMed Central Google Scholar
Zhang, X. et al. A novel small molecule inhibits intrahepatocellular accumulation of Z-variant alpha 1-antitrypsin in vitro and in vivo. Cells 8, 1586 (2019).
Article CAS PubMed PubMed Central Google Scholar
Hidvegi, T. et al. An autophagy-enhancing drug promotes degradation of mutant alpha1-antitrypsin Z and reduces hepatic fibrosis. Science 329, 229–232 (2010).
Article ADS CAS PubMed Google Scholar
Tang, Y. et al. Autophagy induced by exogenous bile acids is therapeutic in a model of alpha-1-AT deficiency liver disease. Am. J. Physiol. Gastrointest. Liver Physiol. 311, G156–G165 (2016).
Article PubMed Google Scholar
Tang, Y., Blomenkamp, K. S., Fickert, P., Trauner, M. & Teckman, J. H. NorUDCA promotes degradation of alpha1-antitrypsin mutant Z protein by inducing autophagy through AMPK/ULK1 pathway. PLoS ONE 13, e0200897 (2018).
Article PubMed PubMed Central Google Scholar
Burton, A. et al. Drug discovery and development in rare diseases: taking a closer look at the Tafamidis Story. Drug Des. Dev. Ther. 15, 1225–1243 (2021).
Article Google Scholar
Chiti, F. & Kelly, J. W. Small molecule protein binding to correct cellular folding or stabilize the native state against misfolding and aggregation. Curr. Opin. Struct. Biol. 72, 267–278 (2022).
Article CAS PubMed Google Scholar
Yan, N. L., Morgan, G. J., Petrassi, H. M., Wilson, I. A. & Kelly, J. W. Pharmacological stabilization of the native state of full-length immunoglobulin light chains to treat light chain amyloidosis. Curr. Opin. Chem. Biol. 75, 102319 (2023).
Article CAS PubMed Google Scholar
Platt, F. M., d’Azzo, A., Davidson, B. L., Neufeld, E. F. & Tifft, C. J. Lysosomal storage diseases. Nat. Rev. Dis. Prim. 4, 27 (2018).
Article PubMed Google Scholar
Martinez-Bailen, M., Clemente, F., Matassini, C. & Cardona, F. GCase enhancers: a potential therapeutic option for Gaucher disease and other neurological disorders. Pharmaceuticals 15, 823 (2022).
Article CAS PubMed PubMed Central Google Scholar
Bacalhau, M. et al. Elexacaftor-Tezacaftor-Ivacaftor: a life-changing triple combination of CFTR modulator drugs for cystic fibrosis. Pharmaceuticals 16, 410 (2023).
Article CAS PubMed PubMed Central Google Scholar
Chaumonnot, K. et al. The HSP GRP94 interacts with macrophage intracellular complement C3 and impacts M2 profile during ER stress. Cell Death Dis. 12, 114 (2021).
Article CAS PubMed PubMed Central Google Scholar
Eletto, D. et al. Limitation of individual folding resources in the ER leads to outcomes distinct from the unfolded protein response. J. Cell Sci. 125, 4865–4875 (2012).
CAS PubMed PubMed Central Google Scholar
Johnson, O. T. & Gestwicki, J. E. Multivalent protein-protein interactions are pivotal regulators of eukaryotic Hsp70 complexes. Cell Stress Chaperones 27, 397–415 (2022).
Article CAS PubMed PubMed Central Google Scholar
Zuiderweg, E. R., Hightower, L. E. & Gestwicki, J. E. The remarkable multivalency of the Hsp70 chaperones. Cell Stress Chaperones 22, 173–189 (2017).
Article CAS PubMed PubMed Central Google Scholar
Assimon, V. A., Gillies, A. T., Rauch, J. N. & Gestwicki, J. E. Hsp70 protein complexes as drug targets. Curr. Pharm. Des. 19, 404–417 (2013).
Article CAS PubMed PubMed Central Google Scholar
Li, X., Shao, H., Taylor, I. R. & Gestwicki, J. E. Targeting allosteric control mechanisms in heat shock protein 70 (Hsp70). Curr. Top. Med. Chem. 16, 2729–2740 (2016).
Article CAS PubMed PubMed Central Google Scholar
Ayers, J. I., Paras, N. A. & Prusiner, S. B. Expanding spectrum of prion diseases. Emerg. Top. Life Sci. 4, 155–167 (2020).
Article CAS PubMed Google Scholar
Condello, C., DeGrado, W. F. & Prusiner, S. B. Prion biology: implications for Alzheimer’s disease therapeutics. Lancet Neurol. 19, 802–803 (2020).
Article PubMed Google Scholar
Wang, R. Y. et al. Structure of Hsp90-Hsp70-Hop-GR reveals the Hsp90 client-loading mechanism. Nature 601, 460–464 (2022).
Article ADS CAS PubMed Google Scholar
Noddings, C. M., Wang, R. Y., Johnson, J. L. & Agard, D. A. Structure of Hsp90-p23-GR reveals the Hsp90 client-remodelling mechanism. Nature 601, 465–469 (2022).
Article ADS CAS PubMed Google Scholar
Sifers, R. N. Intracellular processing of alpha1-antitrypsin. Proc. Am. Thorac. Soc. 7, 376–380 (2010).
Article CAS PubMed PubMed Central Google Scholar
Pan, S. et al. Golgi localization of ERManI defines spatial separation of the mammalian glycoprotein quality control system. Mol. Biol. Cell 22, 2810–2822 (2011).
Article CAS PubMed PubMed Central Google Scholar
Pan, S., Cheng, X. & Sifers, R. N. Golgi-situated endoplasmic reticulum alpha-1, 2-mannosidase contributes to the retrieval of ERAD substrates through a direct interaction with gamma-COP. Mol. Biol. Cell 24, 1111–1121 (2013).
Article CAS PubMed PubMed Central Google Scholar
Iannotti, M. J., Figard, L., Sokac, A. M. & Sifers, R. N. A Golgi-localized mannosidase (MAN1B1) plays a non-enzymatic gatekeeper role in protein biosynthetic quality control. J. Biol. Chem. 289, 11844–11858 (2014).
Article CAS PubMed PubMed Central Google Scholar
Pan, S. et al. Single nucleotide polymorphism-mediated translational suppression of endoplasmic reticulum mannosidase I modifies the onset of end-stage liver disease in alpha1-antitrypsin deficiency. Hepatology 50, 275–281 (2009).
Article CAS PubMed Google Scholar
Sun, A. H., Collette, J. R. & Sifers, R. N. The cytoplasmic tail of human mannosidase Man1b1 contributes to catalysis-independent quality control of misfolded alpha1-antitrypsin. Proc. Natl. Acad. Sci. USA 117, 24825–24836 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Neal, S. et al. Quality Control: Maintaining molecular order and preventing cellular chaos. Mol. Cell 82, 1390–1397 (2022).
Article CAS PubMed Google Scholar
Karras, G. I. et al. HSP90 shapes the consequences of human genetic variation. Cell 168, 856–866 e12 (2017).
Article CAS PubMed PubMed Central Google Scholar
Jarosz, D. F. & Lindquist, S. Hsp90 and environmental stress transform the adaptive value of natural genetic variation. Science 330, 1820–1824 (2010).
Article ADS CAS PubMed PubMed Central Google Scholar
Jarosz, D. F., Taipale, M. & Lindquist, S. Protein homeostasis and the phenotypic manifestation of genetic diversity: principles and mechanisms. Annu Rev. Genet. 44, 189–216 (2010).
Article CAS PubMed Google Scholar
Queitsch, C., Sangster, T. A. & Lindquist, S. Hsp90 as a capacitor of phenotypic variation. Nature 417, 618–624 (2002).
Article ADS CAS PubMed Google Scholar
Rohner, N. et al. Cryptic variation in morphological evolution: HSP90 as a capacitor for loss of eyes in cavefish. Science 342, 1372–1375 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Rutherford, S. L. & Lindquist, S. Hsp90 as a capacitor for morphological evolution. Nature 396, 336–342 (1998).
Article ADS CAS PubMed Google Scholar
Geller, R., Pechmann, S., Acevedo, A., Andino, R. & Frydman, J. Hsp90 shapes protein and RNA evolution to balance trade-offs between protein stability and aggregation. Nat. Commun. 9, 1781 (2018).
Article ADS PubMed PubMed Central Google Scholar
Hummel, B. et al. The evolutionary capacitor HSP90 buffers the regulatory effects of mammalian endogenous retroviruses. Nat. Struct. Mol. Biol. 24, 234–242 (2017).
Article CAS PubMed Google Scholar
Lemus, T. et al. AGO1 and HSP90 buffer different genetic variants in Arabidopsis thaliana. Genetics 223, iyac163 (2023).
Article PubMed Google Scholar
Levis, N. A. & Pfennig, D. W. Phenotypic plasticity, canalization, and the origins of novelty: evidence and mechanisms from amphibians. Semin Cell Dev. Biol. 88, 80–90 (2019).
Article PubMed Google Scholar
Geiler-Samerotte, K., Sartori, F. M. O. & Siegal, M. L. Decanalizing thinking on genetic canalization. Semin Cell Dev. Biol. 88, 54–66 (2019).
Article CAS PubMed Google Scholar
Zheng, X. et al. Hsf1 phosphorylation generates cell-to-cell variation in Hsp90 levels and promotes phenotypic plasticity. Cell Rep. 22, 3099–3106 (2018).
Article CAS PubMed PubMed Central Google Scholar
Geller, R., Andino, R. & Frydman, J. Hsp90 inhibitors exhibit resistance-free antiviral activity against respiratory syncytial virus. PLoS ONE 8, e56762 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Lachowiec, J., Lemus, T., Borenstein, E. & Queitsch, C. Hsp90 promotes kinase evolution. Mol. Biol. Evol. 32, 91–99 (2015).
Article CAS PubMed Google Scholar
Geller, R., Taguwa, S. & Frydman, J. Broad action of Hsp90 as a host chaperone required for viral replication. Biochim. Biophys. Acta 1823, 698–706 (2012).
Article CAS PubMed Google Scholar
Aviner, R. & Frydman, J. Proteostasis in viral infection: unfolding the complex virus-chaperone interplay. Cold Spring Harb. Perspect. Biol. 12, a034090 (2020).
Article CAS PubMed PubMed Central Google Scholar
Takahashi, K. H., Ishimori, M. & Iwata, H. HSP90 as a global genetic modifier for male genital morphology in Drosophila melanogaster. Evolution 72, 2419–2434 (2018).
Article CAS PubMed Google Scholar
Serapian, S. A., Sanchez-Martin, C., Moroni, E., Rasola, A. & Colombo, G. Targeting the mitochondrial chaperone TRAP1: strategies and therapeutic perspectives. Trends Pharm. Sci. 42, 566–576 (2021).
Article CAS PubMed Google Scholar
Loguercio, S. et al. Undertanding the host-pathogen evolutionary balance through Gaussian process modeling of SARS-CoV-2. Patters 4, 100800 (2023).
Article Google Scholar
Consortium, P. et al. A comprehensive enumeration of the human proteostasis network. 1. Components of transaltion, protein folding, and organelle-specific systems. Preprint at https://www.biorxiv.org/content/10.1101/2022.08.30.505920v1.full.pdf (2022).
Elsasser, S. et al. A comprehensive enumeration of the human proteostasis network. 2. Components of the autophagy-lysosome pathway. Preprint at https://www.biorxiv.org/content/10.1101/2023.03.22.533675v1 (2023).
Hetz, C. Adapting the proteostasis capacity to sustain brain healthspan. Cell 184, 1545–1560 (2021).
Article CAS PubMed Google Scholar
Ginsberg, S. D., Sharma, S., Norton, L. & Chiosis, G. Targeting stressor-induced dysfunctions in protein-protein interaction networks via epichaperomes. Trends Pharm. Sci. 44, 20–33 (2023).
Article CAS PubMed Google Scholar
Taldone, T. et al. A chemical biology approach to the chaperome in cancer-HSP90 and beyond. Cold Spring Harb. Perspect. Biol. 12, a034116 (2020).
Article CAS PubMed PubMed Central Google Scholar
Yan, P., Wang, T., Guzman, M. L., Peter, R. I. & Chiosis, G. Chaperome networks— redundancy and implications for cancer treatment. Adv. Exp. Med. Biol. 1243, 87–99 (2020).
Article CAS PubMed PubMed Central Google Scholar
Gestwicki, J. E. & Shao, H. Inhibitors and chemical probes for molecular chaperone networks. J. Biol. Chem. 294, 2151–2161 (2019).
Article CAS PubMed Google Scholar
Pebesma, E. Multivariable geostatistics in S: the gstat package. Comput. Geosci. 30, 683–691 (2004).
Article ADS Google Scholar
Gräler, B., Pebesma, E. & Heuvelink, G. Spatio-temporal interpolation using gstat. R. J. 8, 204–218 (2016).
Article Google Scholar
Chiles, J.-P. & Delfiner, P. Geostatistics: Modeling Spatial Uncertainty (John Wiley & Sons, 2012).
Zhang, Y. & Yang, Y. Cross-validation for selecting a model selection procedure. J. Econ. 187, 95–112 (2015).
Article MathSciNet Google Scholar
Jagger, A. et al. Alpha1-antitrypsin: structure and dynamics in health, disease and drug development. In Alpha-1-antitrypsin Deficiency (eds Kalsheker, N. & Stockley, R.) 49–80 (Academic Press, 2017).

Download references

Acknowledgements

Grant support provided by NIH HL095524, DK051870, AG070209, AG049665, HL141810, and HG10881 to W.E.B. We thank Dr. Gabriela Chiosis for initially providing PU-WS13 compound. P.Z. was supported by in part the Ara Parseghian Medical Research Foundation Fellowship. C.W. was in part supported by the Alpha-1 Foundation Fellowship. We thank Dr. Mark Brantly for providing the Huh7.5 AAT−/− knockout cell line. We thank Ms. Fang Xiao for assistance on the conservation scores for each residue of AAT and Mr. Liangxiao Li for the statistical comparison between correlations.

Author information

Chao Wang
Present address: Institute of Systems and Physical Biology, Shenzhen Bay Laboratory, Shenzhen, China
Shuhong Sun
Present address: Department of Nutrition and Food Hygiene, Center for Global Health, School of Public Health, Nanjing Medical University, Nanjing, China
Shuhong Sun
Present address: Institute for Brain Tumors, Collaborative Innovation Center for Cancer Personalized Medicine, and Center for Global Health, Nanjing Medical University, Nanjing, China
Xi Wang
Present address: National Laboratory of Biomacromolecules, Institute of Biophysics, Chinese Academy of Sciences, Beijing, China
These authors contributed equally: Pei Zhao, Chao Wang.

Authors and Affiliations

Department of Molecular Medicine, Scripps Research, La Jolla, CA, USA
Pei Zhao, Chao Wang, Shuhong Sun, Xi Wang & William E. Balch

Authors

Pei Zhao
View author publications
You can also search for this author in PubMed Google Scholar
Chao Wang
View author publications
You can also search for this author in PubMed Google Scholar
Shuhong Sun
View author publications
You can also search for this author in PubMed Google Scholar
Xi Wang
View author publications
You can also search for this author in PubMed Google Scholar
William E. Balch
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

P.Z., C.W., S.S., X.W., and W.E.B. contributed to the experimental design. P.Z., C.W., and S.S. performed the experiments; C.W. and P.Z. performed the computational analysis; C.W., P.Z., S.S., and W.E.B. generated the manuscript.

Corresponding authors

Correspondence to Chao Wang or William E. Balch.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Emidio Capriotti, Richard Sifers and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Zhao, P., Wang, C., Sun, S. et al. Tracing genetic diversity captures the molecular basis of misfolding disease. Nat Commun 15, 3333 (2024). https://doi.org/10.1038/s41467-024-47520-0

Download citation

Received: 04 April 2023
Accepted: 04 April 2024
Published: 18 April 2024
DOI: https://doi.org/10.1038/s41467-024-47520-0

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.