Mechanism of small heat shock protein client sequestration and induced polydispersity

Miller, Adam P.; Reichow, Steve L.

doi:10.1038/s41467-025-58964-3

Download PDF

Article
Open access
Published: 16 April 2025

Mechanism of small heat shock protein client sequestration and induced polydispersity

Nature Communications volume 16, Article number: 3635 (2025) Cite this article

5752 Accesses
6 Citations
1 Altmetric
Metrics details

Subjects

Abstract

Small heat shock proteins (sHSPs) act as first responders during cellular stress, sequestering destabilized proteins (clients) to prevent aggregation and facilitate refolding or degradation. This critical function, conserved across all life, is linked to proteostasis and protein misfolding diseases. However, the extreme molecular plasticity of sHSP/client complexes has limited mechanistic understanding. Here, we present high-resolution cryo-EM structures of Methanocaldococcus jannaschii sHSP (mjHSP16.5) in apo and multiple client-bound states. The ensemble reveals molecular mechanisms of client sequestration, highlighting cooperative chaperone-client interactions. Client engagement polarizes scaffold stability, promoting higher-order assembly and enhanced sequestration. Higher-order states suggest multiple sHSP/client assembly pathways, including subunit insertion at destabilized geometrical features. These findings provide critical insights into sHSP chaperone function and the interplay between polydispersity and client handling under stress.

Asymmetric apical domain states of mitochondrial Hsp60 coordinate substrate engagement and chaperonin assembly

Article Open access 01 July 2024

Open architecture of archaea MCM and dsDNA complexes resolved using monodispersed streptavidin affinity CryoEM

Article Open access 27 November 2024

Tardigrade small heat shock proteins can limit desiccation-induced protein aggregation

Article Open access 30 January 2023

Introduction

Protein aggregation, driven by stressors such as heat, pH changes, or oxidative damage, underlies many age-related and neurodegenerative diseases, including Alzheimer’s, Parkinson’s, and Huntington’s disease, cataracts, and certain cancers^1,2,3. Cellular defense against irreversible aggregation relies on a proteostasis network of chaperones, proteases, and regulatory proteins⁴. Among these, small heat shock proteins (sHSPs) serve as a front line of defense, preventing proteotoxic aggregation by sequestering unfolded, aggregation-prone proteins (clients) through an ATP-independent ‘holdase’ mechanism^5,6,7,8,9. During stress recovery, sHSP/client complexes deliver clients to ATP-dependent chaperones, such as the HSP70 system for refolding or degradation^10,11,12,13. Mutations in human sHSPs link them to various aggregation diseases, making them attractive pharmacological targets^{14,15,16,17,18,19,20}. However, the structural plasticity and client-induced polydispersity of sHSP/client complexes have obscured mechanistic insights into their function^8,21,22.

sHSPs are small (~12–43 kDa) proteins with a conserved tripartite domain organization: a central alpha-crystallin domain (ACD), a variable N-terminal domain (NTD) involved in client interactions, and a flexible C-terminal domain (CTD) that mediates oligomerization. These proteins form dimers that often assemble into diverse oligomeric structures (200–800+ kDa), ranging from cage-like assemblies to more rare fibrillar and low-order forms^{23,24,25,26,27,28,29,30,31,32,33,34,35}. This structural diversity, conserved across all kingdoms of life, underpins sHSP function by enabling subunit exchange and oligomeric plasticity in response to stress.

The first high-resolution structure of a sHSP oligomer, mjHSP16.5 from Methanocaldococcus jannaschii, revealed an octahedrally symmetric cage (~12 nm) comprising 12 ACD dimers tethered by CTD interactions²³. The flexible NTD implicated in oligomerization and client binding was unresolved, though subsequent studies proposed multiple conformations, reflecting the structural plasticity essential for sHSP function^36,37. This adaptability is coupled to dynamic oligomeric assembly to meet functional demands during stress responses^31,38,39.

Using single-particle cryo-EM, we elucidated the structural basis of mjHSP16.5 chaperone function, capturing the underlying molecular plasticity of the apo-state and an ensemble of client-bound states formed with heat-destabilized client, lysozyme. The structures reveal how conserved hydrophobic regions of the NTD and flexible CTD tethering drive assembly and sequestration. Client engagement induces localized destabilization of the sHSP scaffold, enabling cooperative assembly through sequential recruitment of dimer/client units, unveiling a ‘polarized assembly’ mechanism for enhancing sequestration capacity. Structural and mutational analyses, combined with biophysical and functional studies, highlight the multimodal interactions critical for sHSP plasticity and client handling. These findings provide a comprehensive framework for understanding sHSP-mediated proteostasis under stress conditions, offering insights into potential therapeutic strategies for diseases associated with protein misfolding.

Results

Structural plasticity of the mjHSP16.5 apo-state is coupled to the NTD

mjHSP16.5 was recombinantly expressed and purified without genetic tags for structural and functional characterization (Supplementary Fig. 1). To assess temperature effects on the apo-state, oligomers were incubated at 37 °C (apo-37) or 75 °C (apo-75) before vitrification for cryo-EM studies. Both conditions yielded ~12 nm cage-like 24-mer assemblies, consistent with prior studies^23,36,40,41 (Fig. 1a–e, Supplementary Figs. 2–4).

**Fig. 1: Single-particle cryo-EM analysis of mjHSP16.5 in the absence and presence of destabilized client.**

The 12 dimeric building blocks displayed canonical β5–β7 loop exchange and ACD-CTD tethering, maintaining octahedral symmetry. Asymmetric reconstructions revealed disordered and helical densities within the cage, attributed to the NTD (Supplementary Fig. 3). 3D classification and variability analysis (3DVA) identified two apo-37 cage morphologies: a contracted state at 2.50 Å resolution and an expanded state at 2.35 Å (Fig. 1a, b, Supplementary Fig. 3; Supplementary Movies 1 and 2). Both resolved NTD density inside the cage, lining the inner ACD surface. An apo-75 consensus structure, resolved at 2.86 Å, showed similar features but with enhanced NTD flexibility (Supplementary Fig. 4).

Atomic models of the apo-37 and apo-75 states encompassed the complete ACD, CTD, and most of the NTD (residues 11–32) (Fig. 1b–e, Supplementary Fig. 4e and Supplementary Table 1). The NTD forms a helix-turn-helix-turn-β-sheet topology, with stabilizing inter- and intra-domain interactions. Dimeric subunits exhibited domain-swapped α1 helices, stabilized by Glu12-Lys16 salt bridges and Thr21-mediated hydrogen bonds (Fig. 1c, d). The α1 region interacts with an “ACD/NTD pocket,” formed by residues Phe42, Pro44, and Pro100 at the ACD dimer interface (Fig. 1e). Additional Van der Waals contacts between Met28 and Phe19 contribute to stability (Supplementary Fig. 3h).

Comparisons of apo-37 expanded and contracted states showed minimal differences in ACD/CTD domains (Cα r.m.s.d. = 0.3 Å) but significant NTD variability (Cα r.m.s.d. up to 1.5 Å) (Fig. 2a, b). The apo-75 state displayed distinct NTD arrangements, with weaker α1 density near Phe18/Phe19, likely reflecting increased flexibility (Fig. 2c, d, Supplementary Fig. 4f). Asymmetric reconstructions extracted from 3DVA results revealed transient interactions between unmodeled NTD residues (1–10) are dynamically remodeled during oligomer expansion and contraction, consistent with supporting dimer stability and higher-order assembly²⁴ (Supplementary Fig. 3).

**Fig. 2: Conserved Phe residues within the NTD are critical to sHSP assembly, stability and chaperone function.**

Overall, temperature-dependent activation at 75 °C induced only minor structural changes compared to apo-37 models, unlike other sHSPs that undergo gross morphological rearrangements^39,42,43. This suggests mjHSP16.5 activation primarily relies on enhanced subunit exchange dynamics⁴⁴, which were too transient to capture. These findings reinforce the role of dynamic NTD interactions in supporting sHSP oligomer plasticity and stability.

Conserved phenylalanine-rich regions of the NTD mediate oligomeric assembly and stability

The NTD of mjHSP16.5 contains an ‘aromatic region’ (residues 1–11) and the resolved α1 motif constitutes a ‘conserved region’, both rich in Phe residues (F2, F7, F11, F15, F18, F19). These Phe residues, conserved across diverse sHSPs (Fig. 2e), are proposed to facilitate transient hydrophobic interactions critical for structural plasticity and chaperone function⁴⁵. To probe their role, we generated variants replacing Phe with Ala: F15A (mj-1x), F15/18/19A (mj-3x), and F2/7/11/15/18/19A (mj-6x). mj-1x and mj-3x were expressed in high yields, while mj-6x exhibited reduced solubility and stability, limiting its analysis.

Dynamic light scattering (DLS, at 75 °C) revealed mj-1x and mj-3x formed oligomers similar in size to wild type, with hydrodynamic radii (R_h) of 6.3 ± 0.2 and 5.6 ± 0.2 nm, respectively, compared to 6.0 ± 0.04 nm for wild type (p > 0.1) (Fig. 2f ). In contrast, mj-6x showed a significantly larger R_h of 8.9 ± 0.6 nm (p < 0.01; Fig. 2g), measured at 25 °C due to its diminished heat stability (aggregation temperature ~60 °C) (Supplementary Fig. 1). Quantitative analysis by negative-stain EM (NS-EM) measuring single-particle Feret diameters (D_Feret) corroborated these findings and showed canonical cage-like morphologies for wild type, mj-1x, and mj-3x, while mj-6x displayed ring-like structures (~21.3 nm) and filamentous networks (~2 nm width), indicating severe disruption of the native quaternary structure (Fig. 2g–i, Supplementary Fig. 1 and Supplementary Table 2).

These results demonstrate conserved Phe residues in the ‘aromatic’ and ‘conserved’ regions of the NTD are essential for 24-mer assembly and oligomeric stability. These Phe residues may support oligomer condensation via the hydrophobic effect, while also allowing for oligomer plasticity through quasi-equivalent interactions. This aligns with previous studies showing truncation of the ‘aromatic region’ in mjHSP16.5 produces smaller oligomers⁴⁶ and mutation of the ‘conserved region’ in human α-crystallins alters oligomer size, stability, and chaperone activity^47,48.

Client-induced polydispersity of mjHSP16.5

To establish conditions to investigate mjHSP16.5 chaperone function under stress, binding assays were developed with hen egg lysozyme (14.3 kDa) as a model client. Lysozyme’s high melting temperature (T_m ~ 75 °C) aligns with mjHSP16.5 activation temperature (~60 °C), enabling subunit exchange dynamics essential for chaperone function. Control experiments confirmed that mjHSP16.5 does not protect against reduction-induced aggregation of lysozyme at 37 °C, and lysozyme does not fully aggregate by DLS up to ~80 °C (Supplementary Fig. 5). Binding assays at 75 °C were performed at 12:1 and 2:1 chaperone-to-client ratios (monomer ratios), with chaperone/client complexes monitored by DLS.

Wildtype mjHSP16.5 (mj-wt) formed client-bound complexes with increased R_h values, compared to the apo-state, becoming more pronounced at the 2:1 ratio (7.2 ± 0.4 nm). Similar trends were observed for mj-1x and mj-3x variants, with R_h values of 6.3 ± 0.2 and 6.8 ± 0.4 nm at 12:1, and 7.2 ± 0.4 and 8.5 ± 0.6 nm at 2:1, respectively (Fig. 2f). Differences between these variants were not significant (p > 0.05). Stable complexes were confirmed by SEC and SDS–PAGE, which also indicated increased size and polydispersity following lysozyme binding (Supplementary Fig. 5).

Negative-stain EM (NS-EM) provided detailed morphology and polydispersity analysis. For mj-wt, the primary D_Feret mode increased from 11.9 nm (apo-state) to 13.9 and 14.4 nm at 12:1 and 2:1 ratios, consistent with DLS R_h values. mj-1x and mj-3x exhibited similar increases, with D_Feret reaching 14.5 nm for mj-1x and 14.0–14.5 nm for mj-3x (Fig. 2h, i, Supplementary Fig. 6 and Table 2). All lysozyme-bound complexes showed significantly increased size distribution variability (polydispersity) compared to the apo-state (p < 0.0005, KS test). Morphologically, client-bound states displayed expansion, elongation, and clustering. Notably, clusters formed by mj-3x were larger and more amorphous, potentially reflecting destabilized NTD interactions.

Cryo-EM resolved an ensemble of chaperone/client complexes

To further characterize the structures of client-bound states, wildtype mjHSP16.5 was incubated with lysozyme at a 12:1 chaperone-to-client ratio (at the monomer level) at 75 °C and analyzed by cryo-EM. Single-particle analysis revealed a range of oligomeric states, and five distinct structures were resolved: a 24-mer (2.6 Å), 26-mer (3.7 Å), 32-mer (4.4 Å), 34-mer (4.7 Å), and a 36-mer (4.0 Å) (Fig. 1f, Supplementary Fig. 7 and Supplementary Movie 3). These structures demonstrated that client-induced oligomerization does not involve significant monomer insertion/removal, with dimers serving as the primary building blocks of the mjHSP16.5 chaperone.

These findings highlight the intrinsic plasticity of sHSPs to form diverse oligomeric assemblies through CTD-mediated coupling, enabling dynamic client sequestration. Higher-order assemblies exhibited increased internal cage volumes, reflecting enhanced client sequestration capacity (Fig. 1f). The internal density in all client-bound states showed lower local resolution, consistent with extensive client unfolding and/or transient binding interactions that prevented alignment. The NTD in the higher-order oligomers becomes unresolved, consistent with an order-to-disorder transition that accompanies client sequestration, likely due to complex forms of client interactions. The 24-mer closely resembled the apo-state with resolved NTDs, with a nearly filled internal cavity (~19 nm³ of non-NTD density), suggesting it likely represents unbound particles. In contrast, the 26-mer featured a cavity volume of ~100 nm³ (non-NTD), sufficient to sequester lysozyme (~22 nm³), with further increases observed in higher-order assemblies. The 36-mer, with an internal volume of ~140 nm³, could theoretically sequester up to six lysozyme molecules.

CTD flexibility enables client-induced sHSP polydispersity

Structural comparisons of client-bound assemblies revealed that CTD flexibility facilitates tethering of ACD dimers into diverse oligomeric states (Fig. 3a). While the ACD regions remained structurally consistent (minimal Cα r.m.s.d.), the CTD displayed a continuum of states with two primary conformations (upward and downward), with up to 10 Å deviations in the IXI motif (Fig. 3a, denoted by * and †). These conformations support the formation of novel structural features in higher-order oligomers, such as large 4-fold windows and 5-fold axes (Fig. 3b). For example, the 26-mer forms two large 4-fold windows (~3000 Å²) and two 3-fold axes, while the 36-mer exhibits three smaller 4-fold windows (~2000 Å²) and six 4-fold axes. The conformation of the CTD appears to be related to these features, where upward CTD conformation facilitates 4-fold windows, while the downward conformation supports 3-fold windows and 5-fold axes. The expanded windows increase internal cavity exposure, potentially enabling hydrophobic interactions between NTDs and clients, consistent with previous findings that larger complexes have enhanced client affinity^49,50.

**Fig. 3: Flexible CTD interactions facilitate multiple oligomeric states that correlate with adopted non-canonical geometrical features.**

Some client-bound structures exhibited features reflecting intrinsic dynamics. For instance, dimers connecting 3-fold axes of the 26-mer were of considerably low local resolution (∼10 Å), and the β5-β7 loop in the 32- and 34-mers showed higher Cα r.m.s.d. (Fig. 3a). The 26-, 32-, and 34-mer structures all exhibit weak or absent density corresponding to the CTD region of some monomers and were not modeled, presumably reflecting intrinsic dynamics at these sites (Fig. 3b, circled).

Despite these extensive reconfigurations, the inter-dimer and ACD/CTD tethering interfaces remained largely intact. The inter-dimer interface in the 24-mer contracted/expanded states (∼112–133 Å²) involves distal NTDs and the β5–β7 loop, while the ACD/CTD tethering interface (∼644–669 Å²) forms tongue-and-groove interactions between the CTD-IXI motif and the ACD β4/8-groove (Fig. 3c). Surprisingly, geometric reconfigurations in client-bound oligomers minimally disrupted these interfaces, with variations of only 12 Å² (9.5%) and 78 Å² (13%) for the inter-dimer and ACD/CTD tethering surfaces, respectively. These findings highlight the modularity of ACD dimers and CTD flexibility as key mechanisms enabling sHSPs to form polyhedron architectures and achieve high degrees of polydispersity while maintaining structural integrity.

The polyhedral architecture of mjHSP16.5, comprising ACD dimers (edges) and CTD interactions acting as nodes, is reminiscent of structural models proposed for higher-order eukaryotic sHSPs^51,52. Similarly, the dynamic NTD appears to play a conserved role in guiding polyhedral assembly and contributing to oligomeric polydispersity, as observed in other higher-order sHSP systems^{49,50,53,54,55}. These shared features suggest that the mechanistic principles gained from mjHSP16.5 are likely generalizable to a broad range of sHSP chaperones.

Client-induced sHSP polarized destabilization directs chaperone elongation

Local resolution analysis of cryo-EM maps obtained in the presence of lysozyme revealed further insight into the distinct stability variations across client-bound sHSP assemblies (Fig. 4a). In the 24-mer, the exterior scaffold composed of ACD and CTD regions showed uniform resolution (~2.5–3.0 Å), with lower-resolution internal density (~4.5 Å), consistent with apo-state features that include resolvable NTD motifs. In contrast, higher-order assemblies, including the 26-mer, 32-mer, and 34-mer, exhibited more pronounced resolution variation and a loss of defined NTD features. These maps showed high local resolution (~3–4 Å) centered around a single 4-fold axis, decreasing progressively along the elongating scaffold to ~9–10 Å. The 36-mer is defined by a smaller set of particles yielding an overall lower resolution (~7–8 Å), but more uniform, reflecting the internal symmetry of this state.

**Fig. 4: Polarized destabilization induced by client sequestration correlates with sites of recruited dimers to form higher-order sHSP oligomers.**

Based on these collective observations, we can attempt to infer aspects of the chaperone/client assembly pathway. The transition from the 24-mer to the 26-mer appears to serve as a critical intermediate in higher-order oligomer formation. Cryo-EM variability analysis showed flexible dynamics at newly formed 3-fold axes of the 26-mer (Supplementary Movie 4), suggesting regions of polarized destabilization may promote site-specific recruitment of additional dimers or dimer/client complexes. These destabilized regions may thus act as “hot spots” for cooperative subunit exchange, supporting further scaffold elongation (Fig. 4b, Supplementary Movie 4). This concept may be related to mechanisms proposed for the phospho-activation of HSP26⁵⁶. Such localized destabilization may be critical for the chaperone function, not only enhancing client capacity through cooperative subunit recruitment but also enabling reversible client release for downstream refolding machinery⁵⁷.

Interestingly, intermediates such as the 28-mer and 30-mer were not detected, implying these states are either intrinsically unstable or bypassed. However, a plausible transition from the 32-mer to the 34-mer was portrayed by 3DVA, involving dimer insertion into a large 4-fold window, generating new 3-fold windows and a 5-fold axis (Supplementary Movie 4). In contrast, a direct transition from the 34-mer to the 36-mer is not clear, as it would require significant rearrangement, possibly involving disassembly and reassembly of 5-fold axes to accommodate additional dimers. This suggests the 36-mer, and potentially other client-bound states, may form through alternative assembly pathways. Together, these findings highlight a dynamic mechanism of scaffold elongation, driven by localized destabilization that enables cooperative subunit recruitment and formation of diverse, high-order sHSP oligomers tailored to meet demands of enhanced client capacity.

Discussion

This study revealed previously unresolved features of sHSP plasticity and provided a detailed visualization of client-induced polydispersity-hallmarks of sHSP chaperone systems that have resisted structural characterization. The ensemble of states revealed how conserved hydrophobic regions and the dynamic properties of the NTD and CTD facilitate structural transitions critical for chaperone activity. Client engagement induced a striking polarization of stability within the sHSP oligomeric scaffold, pointing to a cooperative mechanism for client sequestration and subunit recruitment.

We propose two complementary models to explain these client-induced transitions: the ‘polarized assembly’ model (Fig. 4b), where localized destabilization drives directional stepwise dimer/client insertion, and the ‘seeded assembly’ model, in which initial client binding by exchanging sHSP subunits seed new oligomer formation by recruiting additional dimers or dimer/client complexes. This phenomenon is supported by recent single-molecule fluorescence studies of αB-crystallin and HSP27^58,59.

Notably, the ‘polarized assembly’ and ‘seeded assembly’ models are not mutually exclusive and may operate in tandem, contributing to the complexity of sHSP sequestration mechanisms. While precise pathways remain to be fully resolved, the formation of directionally elongated, cage-like assemblies appears to be an energetically favored outcome, enhancing the capacity for diverse client sequestration. Such client-induced elongation may parallel features described for human α-crystallins⁶⁰. However, the structural diversity of sHSPs and their ability to respond to varied stressors and clientele suggest the existence of additional, and yet-to-be-identified mechanisms. For example, monomeric HSP21 from A. thaliana has been shown to stabilize client proteins by interacting at a destabilized interface while retaining the client in a near-native structure⁶¹. Capturing transient intermediates will be crucial to further elucidating the molecular details of these processes and uncovering new insights into co-aggregation pathways relevant to sHSP-related diseases, including cataracts, neurodegenerative disorders, cardiomyopathies, and treatment-resistant cancers^{4,14,15,16,17,62}.

Methods

Construction of expression plasmids

The gene sequence of wildtype mjHSP16.5 (GENID: 1451140) was codon-optimized for bacterial expression and encoded into a pET23a(+) expression vector and sequences encoding variants within the N-terminal domain of mjHSP16.5: F15A (mj-1x), F15/18/19A (mj-3x), F2/5/11/15/18/19 (mj-6x), deletion of residues 1–32 (mj-Δ32, with Met1 at position 32), and deletion of residues 1–20 (mj-Δ20, with Met1 at position 20) were encoded into the pET21a(+) vector (Genscript). Protein expression constructs used in this study did not include solubility or purification tags. Full plasmid sequencing (Plasmidsaurus, Eugene, OR) confirmed the correct gene sequence, insertion site, and placement of mutations/truncations.

Expression and purification of mjHSP16.5 wildtype and variants

Wildtype (mj-wt) and variant constructs (mj-1x, mj-3x, mj-6x, mj-Δ32, and mj-Δ20) of mjHSP16.5 were expressed in bacteria and purified using the same protocols (modified from Quinlan et al.⁴⁰). Briefly, E. coli BL21(DE3) was used as an expression host for all constructs and the growth media (LB) was supplemented with ampicillin (0.5 mM). Cells were grown at 37 °C to an optical density (A.U. at 600 nm) of 0.6–1.0 and expression was induced with 1 mM isopropyl β-d-1-thiogalactosidase (IPTG). Cells were harvested by centrifugation (4000×g for 15 min at 4 °C) after 3–4 h post-induction at 37 °C. Pelleted cells were resuspended in 20 mM Tris–HCl (pH 8.0), aliquoted and frozen at −80 °C for further use.

For purification of each protein construct, cell suspensions were thawed and supplemented with 1,4-dithioreitol (DTT, 0.5 mM final concentration) and phenylmethylsulfonyl fluoride (PMSF, 0.1 mM final concentration), lysed by sonication on ice (70% amplitude, 6 rounds of 30 s on/off), supplemented with additional PMSF (0.2 mM final concentration), and cellular debris cleared by ultracentrifugation at 165,000 × g for 30 min at 4 °C. The supernatant was supplemented with NaCl (1 M final concentration) and 20 mM Tris–HCl (pH 8.0) and incubated in an 80 °C water bath for 30 min (V = 20 mL). The heated lysate was recovered on ice for 5 min and denatured protein was pelleted by ultracentrifugation at 165,000×g for 30 min at 4 °C. The supernatant containing the thermo-stable mjHSP16.5 was collected, DNase I (~400 units, Thermo Scientific) was added and incubated for 30 min on ice before being filtered at 0.22 µm prior to chromatography.

The clarified lysate was applied to a gel filtration chromatography column (S300 resin; Pharmacia) equilibrated with 20 mM Tris–HCl (pH 8.0), 1 mM EDTA and 0.5 mM DTT. Fractions from gel filtration were assessed via SDS–PAGE and fractions containing mjHSP16.5 (wildtype or mutants) were pooled and supplemented with DTT (0.5 mM final concentration). The pooled fractions were loaded onto a MonoQ anion exchange column (GE Healthcare) equilibrated with buffer A (20 mM Tris–HCl (pH 8.0), 0.16 mM EDTA, and 1 mM EGTA) and eluted with a NaCl gradient (buffer A with 1 M NaCl). Eluted fractions containing the target mjHSP16.5 construct were pooled, concentrated to a volume of ~2 mL with a 100,000 kDa cutoff spin concentrator (Vivaspin), and loaded onto a Superose 6 size-exclusion chromatography (SEC) column equilibrated with 20 mM HEPES (pH 7.4), 2 mM EDTA, 2 mM EGTA, and 100 mM NaCl. Fractions containing purified mjHSP16.5 constructs were pooled, aliquoted, and either flash-frozen in liquid nitrogen and stored at −80 °C for later use or incubated at 37 °C (wildtype) or 4 °C (mutants) for immediate use. Nucleic acid contamination was assessed by monitoring a UV absorbance ratio of 280/260 nm with all purified proteins having ratios >1.5, indicating minimal co-purification of nucleic acids. The concentration of purified protein was determined by UV absorbance at 280 nm using the extinction coefficient 8,257 M⁻¹ cm⁻¹ ⁶³. The variants mj-Δ20 and mj-Δ32 resulted in insoluble protein and were not purified for downstream analyses.

Aggregation assays of reduced lysozyme at 37 °C

All aggregation assays were performed in a reaction buffer containing 20 mM HEPES (pH 7.4), 2 mM EDTA, 2 mM EGTA, and 100 mM NaCl. Aggregation of hen egg-white lysozyme (Fisher, MS grade) at 37 °C was induced with the addition of 2 mM tris(2-carboxyethyl)phosphine (TCEP). Lysozyme (10 µM) aggregation by TCEP was monitored in the presence of 120 µM (monomeric ratio) and 20 µM (2:1 monomeric ratio)) or absence of mj-wt. Chaperone/client mixtures were allowed to equilibrate at 37 °C for 15 min prior to the addition of TCEP (2 mM final concentration). Turbidity measurements were monitored by absorption at 360 nm, collected in 384-well plates (Nucleon, flat black) on a Tecan Infinite M NANO+ for 2 h at 37 °C.

Heat-induced aggregation and binding assays with lysozyme at 75 °C

Dynamic light scattering (DLS) measurements were performed in an Aurora 384 well plate on a Wyatt DynaPro plate reader III (Wyatt Technology, Santa Barabara, USA) operating with an 830 nm laser and 150° DLS detector angle. All measurements were acquired with five reads and 10 s acquisition time in the Dynamics software v7.10.1 (Wyatt). To determine the aggregation temperature of lysozyme, the hydrodynamic radius in solution was monitored using DLS while ramping temperature from 25 to 85 °C at 0.91 °C min⁻¹ (n = 3). Due to the small size of the lysozyme (~2 nm radius) the working concentration of 10 µM was not detectable and 100 µM was used to determine the aggregation temperature. Likewise, lysozyme at 50 and 100 µM was monitored by DLS at a constant 75 °C for 2 h to show no immediate (<1 h) aggregation at these temperatures.

For binding assays, lysozyme (10 µM) was incubated in the presence and absence of mjHSP16.5 wildtype, mj-1x, and mj-3x at 120 µM (12:1 monomeric ratio) and 20 µM (2:1 monomeric ratio) at 75 °C for 2 h in 20 mM HEPES (pH 7.4), 2 mM EDTA, 2 mM EGTA, and 100 mM NaCl. Prior to incubation at 75 °C the mixed samples were incubated at 25 °C for 30 min. Additionally, samples of mjHSP16.5 were measured by DLS in the absence of lysozyme at 25 °C (mj-6x), or 37, 75 °C, or through a temperature ramp from 25–85 °C at a rate of 0.49 °C min⁻¹ (mj-wt, mj-1x, mj-3x: 120 µM). The aggregation temperature of mj-6x was determined using a heat ramp of 0.3°C min^-1 from 25–85°C. Replicates of DLS readings were pooled for downstream analyses (SEC/SDS–PAGE, NS-EM, cryo-EM). The average ± s.e.m. was calculated for the hydrodynamic radius of (n = 2–3) technical replicates across (n = 3–5) independent experiments for each sample. Statistical significance was assessed by completing an F-test for variability followed by a Student’s two-sided T-test (equal/unequal variance depending on F-test results).

Size-exclusion chromatography

Following binding assays performed at 75 °C, pooled DLS replicates obtained from binding reactions were loaded (125 µL injection) onto a Superose 6 SEC column equilibrated with 20 mM HEPES (pH 7.4), 2 mM EDTA, 2 mM EGTA, and 100 mM NaCl. Elution peaks were monitored by SDS–PAGE (17.5% acrylamide) and protein bands were visualized by silver staining.

Negative stain EM and single-particle morphology analysis

Negative stain EM was performed on purified apo-state mjHSP16.5 that was incubated at 37 °C for ~16 h and then diluted to ~3 µM in dilution buffer containing 20 mM HEPES (pH 7.4), 100 mM NaCl, 2 mM EDTA, and 2 mM EGTA. Chaperone assay reaction products of mj-wt, mj-1x, and mj-3x in the absence (apo) and presence of lysozyme (12:1 and 2:1 chaperone:client ratios) prepared at 75 °C were recovered on ice and diluted to ~3 µM (mjHSP16.5 concentration) with dilution buffer. Carbon-coated 400-mesh copper EM grid (Ted Pella) was glow discharged at 15 mA for 1 min prior to sample application. For each condition, 3 µL of the sample was applied to the grid, and excess protein/buffer was blotted with filter paper, washed twice with ultrapure water, stained with freshly prepared (0.75% wt vol⁻¹) uranyl formate (SPI-Chem), blotted with filter paper, and dried with laminar airflow. Grids of chaperone reactions were set within 30 min following ice recovery to quench subunit exchange dynamics following complex formation at 75 °C. Negatively stained specimens were imaged on a 120 kV TEM (Tecnai T12, FEI) equipped with either a 2K × 2K CCD camera (Eagle, FEI) at a nominal magnification of 49,000 and a calibrated pixel size of 4.4 Å pixel⁻¹ (mj-wt apo, 12:1, 2:1) or an AMT camera (model XR16) using the AMT Image Capture Engine (v602.591j) at a nominal magnification of 30,000 with a calibrated pixel size of 4.0 Å pixel⁻¹. Micrographs were collected with a defocus range from 1.5 to 2.2 µm.

Single-particle morphology analysis was performed as previously described⁶⁰. Briefly, unprocessed micrographs were imported into FIJI⁶⁴ and the scale was set based on the calibrated pixel size of the micrograph. Micrographs were processed using the fast-Fourier transform-based bandpass filter with default settings (filter large structures at 40 pixels, filter small structures at 3 pixels, 5% tolerance, auto-scale after filtering, saturate image when autoscaling) followed by a maximum filter (radius of 2 pixels) and background subtraction (rolling ball radius of 25–50 pixels). The filtered and background subtracted micrographs were binarized (dark background) and segmentation was optimized using the Remove Outliers tool and erosion/dilation of binary segments tools. Processed micrographs were compared to the raw micrograph during the optimization of binary segments. The Analyze Particles tool was used for the automated determination of the Feret diameter of each segment within a minimum particle area of 50 nm². Feret diameters are presented as raincloud plots generated in R Studio (v4.0.5). Statistical analysis was done in Excel (average ± s.e.m.) and Scipy⁶⁵ (Kolmogorov-Smirnov test).

Cryo-electron microscopy data collection

Prior to vitrification, samples were incubated for >16 h at 37 °C (apo-37C) or for 2 h at 75 °C (reaction products from DLS experiments) in the absence and presence of lysozyme (apo-75C and mj:lyso-75C at a 12:1 ratio). 3 µL of each sample (~1 mg mL⁻¹) was applied to a fresh glow discharged (15 mA, 1 min) holey carbon copper grid (apo-37C: Cflat (EMS) R1.2/1.3, apo-75C and mj:lyso-75C samples: Quantifoil R2/1, 400 mesh). Grids were blotted (1.0–1.5 s) at room temperature and 90% humidity and the plunge froze into liquid ethane on a Vitrobot Mark IV (FEI, Thermo Fisher Scientific). Image datasets were collected at the Pacific Northwest Cryo-EM Center (OHSU, Portland, OR) on a 300 kV Titan Krios equipped with a K3 detector (Gatan) using SerialEM⁶⁶. Movies were collected in super-resolution mode at a calibrated physical/super-resolution pixel size of 0.788/0.394 Å pixel⁻¹ (apo-37C sample), 1.013/0.506 Å pixel⁻¹ (apo-75C sample), and 1.066/0.533 Å pixel⁻¹ (mj:lyso-75C) with a total dose rate of ~40 e⁻ per Å² over 70 frames for the apo-37C sample and ~50 e⁻ per Å² over 50 frames for apo-75C and mj:lyso-75C. Movies were collected over a defocus range of 1.0–2.5 µm. The apo-75C and mj:lyso-75C samples were collected using a GIF energy filter (Gatan) with a 10 eV slit width.

Cryo-EM image processing of apo-state mjHSP16.5 (37 °C)

All steps of cryo-EM image processing were performed in CryoSPARC v3.3.1⁶⁷. A dataset of 16,214 micrographs for mjHSP16.5 (apo-37C) was preprocessed with Patch Motion Correction (micrographs binned 2×, 0.788 Å pixel⁻¹) and Patch CTF estimation. Low-quality micrographs were removed based on CTF resolution fit. A subset of 100 micrographs was subjected to blob picking (120–160 Å diameter) to yield a particle set of ~3.4 million particles extracted with binning (2.46 Å pixel⁻¹). Noisy particles and low occupancy classes were removed by 2D classification to give a set of 1,170,772 particles used for multi-class ab initio model generation with 4 classes and a maximum resolution of 6 Å. Multi-class ab initio generation yielded 2 good classes (1,060,133 total particles) corresponding to 24-mer caged assemblies with slightly different diameters. Further rounds of 2D classification yielded 968,458 particles that were again subjected to multi-class ab initio reconstruction (3 classes), yielding two distinct classes of 24-meric cages with 441,253 particles in class 1 (expanded state) and 421,283 particles in class 2 (contracted state). A consensus refinement of re-extracted particles (1.05 Å pixel⁻¹) of the combined classes (862,436 particles) without symmetry (C1) yielded a consensus reconstruction at 2.99 Å resolution. 2D classification and heterogeneous refinement (C1, 6 classes) and removal of low occupancy classes yielded 630,757 particles which refined with octahedral (O) symmetry to 2.44 Å.

The C1 consensus refinement of the 630,757 particle stack was used as input for 3D Variability analysis with three orthogonal principal modes and a filter resolution of 5 Å⁶⁸. Additionally, this particle set was expanded with octahedral (O) symmetry (15,136,968 expanded particles) and used for 3D Variability analysis with three orthogonal principal modes and a filter resolution of 5 Å. These particles were subjected to heterogeneous refinement with four classes which gave two high occupancy classes at ~2.7 Å resolution (expanded state: 257,168 particles; contracted state: 205,181 particles). Separate cleanup of the two particle sets was done by 2D classification yielding final non-uniform refinements (O symmetry) of 2.35 Å for the expanded state (256,929 particles) and 2.50 Å for the contracted state (186,720 particles).

Cryo-EM image processing of apo-state mjHSP16.5 (75 °C)

All steps of cryo-EM image processing were performed in CryoSPARC 4.4.1⁶⁷. The full dataset of 6,460 micrographs for mjHSP16.5 apo-75C was preprocessed with Patch Motion Correction (micrographs binned by 2×, 1.0125 Å pixel⁻¹) and Patch CTF estimation. The resulting micrographs were culled based on CTF estimation resolution, relative ice thickness, and average intensity to yield 6186 micrographs carried forward for particle picking. Blob particle picking on the full micrograph stack generated ~2.4 million picks that were extracted at 2.373 Å pixel⁻¹ and cleaned up by two rounds of 2D classification to yield 185,806 particles for further analysis. Results from multi-class ab initio generation (4 classes) were input into a heterogeneous refinement (C1 symmetry, 4 classes) which yielded two cage-like maps at 5.69 Å (54,739 particles, set 1) and 4.92 Å (99,715 particles, set 2) resolution. Particles from these two classes were combined and re-extracted at 1.187 Å pixel⁻¹ and the pooled particle set (153,807 particles) was reconstructed with O symmetry to 2.86 Å. Results from a C1 consensus non-uniform refinement were input into a 3D variability analysis with three orthogonal principal modes and a filter resolution of 5 Å⁶⁸.

Cryo-EM image processing of mjHSP16.5/lysozyme complexes (12:1 ratio)

All steps of cryo-EM image processing were performed in CryoSPARC v4.2.1-4.4.1⁶⁷. The full dataset of 13,276 movies for mj:lyso-75C was preprocessed with Patch Motion correction (micrographs binned by 2×, 1.0655 Å pixel⁻¹) and Patch CTF estimation. The micrographs were culled based on CTF estimation resolution, relative ice thickness, and average intensity to yield 12,704 micrographs carried forward for particle picking. Blob picking (120–220 Å diameter) on a subset of 500 micrographs to yield 197,163 particles. Particles were extracted at 2.5 Å pixel⁻¹, and submitted to 2D classification, and the resulting good classes were used as 2D templates for particle picking. Inspection of template-based picks resulted in ~8.4 million particles that were subjected to two rounds of 2D classification to yield ~3 million good particles which were then extracted at 3.33 Å pixel⁻¹. This particle set was used as input for a multi-class ab initio job (8 classes, initial resolution 80 Å). The resulting eight ab initio models along with the full good particle stack were input into a heterogeneous refinement job which gave six good classes (2,826,578 total particles) and two noisy classes. A second round of heterogeneous refinement was performed using the six good maps/particles and the two noisy maps (to assist in removing noisy particles) which generated four good classes identified as a 24-mer (985,163 particles), 26-mer (450,391 particles), 32-mer (653,689 particles), and a 36-mer oligomeric states (242,970 particles) that were used for further analysis.

The initial particle set pertaining to the 24-mer oligomeric state (985,163 particles) was cleaned up by 2D classification to produce a particle set of 960,040 that was extracted at 1.25 Å pixel⁻¹ and input into 3D variability analysis with three orthogonal principal modes and filter resolution of 5 Å⁶⁸. Intermediate mode analysis of the first component was done with five intermediate maps and particles were used as inputs for heterogeneous refinement which produced two classes below 4 Å resolution. These two classes were pooled and refined (non-uniform refinement) without applied symmetry to 2.60 Å resolution.

The particle set associated with the 26-mer oligomeric state class (450,391 particles) was extracted at ~1.2× binning (1.25 Å pixel⁻¹) followed by global CTF refinement and non-uniform refinement without symmetry (C1) to 3.65 Å. 3D variability analysis was performed with three orthogonal principal modes and intermediates analysis of the three components generated five intermediates states that refined to ~4–8.5 Å resolution without imposed symmetry (C1).

The 32-mer oligomeric state particle set (653,689 particles) was extracted at 1.25 Å pixel⁻¹ and cleaned up by 2D classification to give 645,384 particles that were refined without applied symmetry to 4.80 Å resolution. 3D variability analysis was performed with three orthogonal principal modes at a filter resolution of 5 Å, followed by intermediates analysis to generate five intermediates states that were used as input for heterogeneous refinement⁶⁸. Heterogeneous refinement produced two classes below 7 Å resolution that corresponded to a 32-mer state (244,887) and a 34-mer state (202,648) that refined (non-uniform refinement) without symmetry (C1) to 4.37 and 4.79 Å, respectively. The 34-mer particles went through global CTF refinement and a final non-uniform refinement (C1) to yield a 4.71 Å final map.

The 3x binned particle set for the 36-mer class (242,970) was refined with (D3) and without symmetry (C1) and the resulting maps were used as input for a heterogeneous refinement (C1) with two classes (40 Å initial lowpass filter) and generated one class displaying D3 symmetric features at 8.27 Å resolution with 111,271 particles. Low quality particles were removed by 2D classification resulting in 85,180 particles that refined to 4.50 Å resolution with D3 symmetry. This particle set was expanded with D3 symmetry to yield 511,080 particles that were subjected to local refinement (C1) resulting in a 4.30 Å reconstruction. Output from local refinement was used as input for 3D Variability with three orthogonal principal modes and a filter resolution of 5 Å⁶⁸. Intermediate mode analysis (five intermediates) of the principal components resulted in five intermediate maps with corresponding particle sets that were input for 3D classification without alignment (5 classes). A highly populated class containing 144,177 particles was used for local refinement (C1) resulting in a final reconstruction of the 36-subunit cage structure at 4.01 Å resolution. A reconstruction without symmetry (C1) of the expanded particles was refined to 6.39 Å and used for local resolution analysis. Local resolution estimation of the final refined maps of each oligomeric state was performed in CryoSPARC using default parameters and resolution-based coloring of each map was done in ChimeraX (v1.7)⁶⁹.

Atomic modeling, validation, and analysis

Atomic model building into the mjHSP16.5 apo-state (37 °C) in both contracted and expanded states was initiated using a dimer model from the previously published crystal structure of mjHSP16.5 (PDB ID: 1SHS²³). Final maps (O symmetric) corresponding to the contracted and expanded cages were sharpened using Phenix AutoSharpen⁷⁰. Dimers were initially fit as rigid bodies into each map using ChimeraX to produce a 24-meric model and refinement was done using phenix real space refinement with secondary structure and NCS restraints⁷¹. Atomic model building of NTD residues 11–32 for the contracted and expanded states were built in COOT as a polyalanine chain, refined, and side chains added before further refinement and side chain adjustment. Iterative manual and automatic model refinement was done in COOT and Phenix (real space refinement) using secondary structure and NCS restraints, and in Isolde without NCS restraints.

The final 24-mer map (C1) from the mjHSP16.5/lysozyme (75 °C) dataset was sharpened using Phenix AutoSharpen and model building was initiated by rigidly fitting the expanded model from the apo-37C dataset with the deletion of residues 11–26^70,71. Real-space refinement was performed in Phenix using a reference model and secondary structure restraints. The final 24-mer model contained residues 27–147. All subsequent model building was initiated using a dimer model from the 24-mer rigidly fit into the final 26-mer, 32-mer, 34-mer, and 36-mer maps. For each oligomeric state various deletions of NTD and CTD residues of the monomers were done to agree with resolved map density and iterative manual remodeling of CTDs to fit the map density was performed in COOT and ISOLDE along with real-space refinement in Phenix^71,72,73. Model building of the 26-mer state rigidly fit 13 dimers into the unsharpened map with truncations yielding 22 chains with residues 34–147 and 4 chains with residues 34–139. Sixteen dimers were refined in the sharpened (Phenix local anisotropic sharpening) 32-mer map, resulting in 31 chains with residues 31–147 and one chain with residues 35–143. For the 34-mer state, 17 dimers refined into a Phenix auto-sharpened map, resulting in 33 chains covering residues 33–147 and one chain with residues 35–143. The 36-mer state was built with 18 refined dimers fit into a Phenix AutoSharpened map with 36 chains covering residues 31–147. For all models, validation of model refinement and map-to-model fit was done using Phenix validation and the PDB validation server⁷⁰.

For visualization of unmodeled density, final maps were low-pass filtered at 7 Å resolution and density corresponding to the atomic models of each state were generated using the molmap (7 Å) function in ChimeraX v1.17.1. Density corresponding to the 7 Å molmaps was subtracted from the respective 7 Å low-pass filtered Cryo-EM maps to generate a subtracted map containing the unmodeled internal density. Cα r.m.s.d. calculations were generated in ChimeraX using only chains with full CTDs (through residue 147) resulting in 24 chains for the 24-mer, 22 chains for the 26-mer, 31 chains for the 32-mer, 33 chains for the 34-mer, and 36 chains for the 36-mer for this comparative analysis. Coloring based on Cα r.m.s.d. and local resolution was done in ChimeraX with the color by attributes and surface color utilities, respectively. Buried surface areas for the ACD-dimer interface, inter-dimer interface, and the canonical CTD/ACD-groove interface were calculated using the Interfaces function in ChimeraX with default settings except areaCutoff set to 100 Å². For visualization, modeling of NTD residues 1–10 (Fig. 2e) was done by extension of residues distally from residue 11 of the contracted model and subsequent addition of side chains and refinement in COOT. Measurement of the internal volume density (at ~2σ) was performed in ChimeraX (v1.7) by subtracting cage density using molmaps generated at 7 Å and the volume subtraction tool. PDB 1DPX was used for the measurement of lysozyme volume. Assessment of non-NTD internal density volume was performed by calculating the volume of residues 1–32 from a 7 Å molmap, multiplying by the number of subunits to determine the total volume occupied by NTDs, and subtracting this from the total internal volume density of the cryo-EM map (Supplementary Movie 4).

Figure preparation

Structural models and cryo-EM density maps were visualized and prepared for presentation using ChimeraX. Final figures were composed in Photoshop.

Reporting summary

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

Data availability

Cryo-EM density maps have been deposited to the Electron Microscopy Data Bank for the mjHSP16.5 apo-state datasets (apo37-contracted: EMDB-49828 [https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-49828], apo37-expanded: EMDB-49829 [https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-49829], apo-75: EMDB-49830 [https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-49830]) and the mjHSP16.5/lysozyme dataset (24mer: EMDB-49832 [https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-49832], 26mer: EMDB-49834 [https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-49834], 32mer: EMDB-49836 [https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-49836], 34mer: EMDB-49837 [https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-49837], 36mer: EMDB-49838 [https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-49838]. Coordinates for atomic models have been deposited to the Protein Data Bank for the mjHSP16.5 apo-state datasets (apo37-contracted: PDB 9NV4, apo37-expanded: PDB 9NV7, apo75: PDB 9NV8) and the mjHSP16.5/lysozyme dataset (24mer: PDB 9NVC, 26mer: PDB 9NVF, 32mer: PDB 9NVI, 34mer: PDB 9NVJ, 36mer: PDB 9NVK). The original multi-frame micrographs have been deposited to EMPIAR for the apo-state datasets (apo-37: EMPIAR-12645, apo-75: EMPIAR-12642) and the mjHSP16.5/lysozyme dataset (EMPIAR-12643 [https://www.ebi.ac.uk/empiar/EMPIAR-12643]). Plasmids used for protein expression are available upon request. The negative stain EM and DLS datasets generated in this study are provided in the Source Data file deposited to Zenodo [https://doi.org/10.5281/zenodo.15047210]. Previously published models used for initial model building can be found here: PDB: 1SHS. Source data are provided with this paper.

References

Dobson, C. M. Protein folding and misfolding. Nature 426, 884–890 (2003).
Article ADS CAS PubMed Google Scholar
Mathieu, C., Pappu, R. V. & Taylor, J. P. Beyond aggregation: pathological phase transitions in neurodegenerative disease. Science 370, 56–60 (2020).
Article ADS 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
Hartl, F. U., Bracher, A. & Hayer-Hartl, M. Molecular chaperones in protein folding and proteostasis. Nature 475, 324–332 (2011).
Article CAS PubMed Google Scholar
Caspers, G.-J., Leunissen, J. A. M. & de Jong, W. W. The expanding small heat-shock protein family, and structure predictions of the conserved “α-crystallin domain”. J. Mol. Evol. 40, 238–248 (1995).
Article ADS CAS PubMed Google Scholar
Jakob, U., Gaestel, M., Engel, K. & Buchner, J. Small heat shock proteins are molecular chaperones. J. Biol. Chem. 268, 1517–1520 (1993).
Article CAS PubMed Google Scholar
Horwitz, J. Alpha-crystallin can function as a molecular chaperone. Proc. Natl Acad. Sci. USA 89, 10449–10453 (1992).
Article ADS CAS PubMed PubMed Central Google Scholar
Haslbeck, M., Weinkauf, S. & Buchner, J. Small heat shock proteins: simplicity meets complexity. J. Biol. Chem. 294, 2121–2132 (2019).
Article CAS PubMed Google Scholar
Mymrikov, E. V., Daake, M., Richter, B., Haslbeck, M. & Buchner, J. The chaperone activity and substrate spectrum of human small heat shock proteins. J. Biol. Chem. 292, 672–684 (2017).
Article CAS PubMed Google Scholar
Ungelenk, S. et al. Small heat shock proteins sequester misfolding proteins in near-native conformation for cellular protection and efficient refolding. Nat. Commun. 7, 13673 (2016).
Article ADS PubMed PubMed Central Google Scholar
Goncalves, C. C., Sharon, I., Schmeing, T. M., Ramos, C. H. I. & Young, J. C. The chaperone HSPB1 prepares protein aggregates for resolubilization by HSP70. Sci. Rep. 11, 17139 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Żwirowski, S. et al. Hsp70 displaces small heat shock proteins from aggregates to initiate protein refolding. EMBO J. 36, 783–796 (2017).
Article PubMed PubMed Central Google Scholar
Tedesco, B., Vendredy, L., Timmerman, V. & Poletti, A. The chaperone-assisted selective autophagy complex dynamics and dysfunctions. Autophagy 19, 1619–1641 (2023).
Jong, W. W. de, Workum, F. P. A. van, Bosman, G. J. C. G., Vooter, C. E. M. & Renkawek, K. Expression of αB-crystallin in Alzheimer’s disease. Acta Neuropathol. 87, 155–160 (1994).
Vicart, P. et al. A missense mutation in the αB-crystallin chaperone gene causes a desmin-related myopathy. Nat. Genet. 20, 92–95 (1998).
Article CAS PubMed Google Scholar
Braak, H., Tredici, K., Sandmann-Keil, D., Rüb, U. & Schultz, C. Nerve cells expressing heat-shock proteins in Parkinson’s disease. Acta Neuropathol. 102, 449–454 (2001).
Article CAS PubMed Google Scholar
Clark, A. R., Lubsen, N. H. & Slingsby, C. sHSP in the eye lens: crystallin mutations, cataract and proteostasis. Int. J. Biochem. Cell Biol. 44, 1687–1697 (2012).
Article CAS PubMed Google Scholar
Kampinga, H. H. & Garrido, C. HSPBs: small proteins with big implications in human disease. Int J. Biochem Cell Biol. 44, 1706–1710 (2012).
Article CAS PubMed Google Scholar
Abati, E. et al. Charcot–Marie–Tooth disease type 2F associated with biallelic HSPB1 mutations. Ann. Clin. Transl. Neurol. 8, 1158–1164 (2021).
Article CAS PubMed PubMed Central Google Scholar
Liu, G.-S. et al. Regulation of BECN1-mediated autophagy by HSPB6: insights from a human HSPB6S10F mutant. Autophagy 14, 80–97 (2018).
Article ADS CAS PubMed PubMed Central Google Scholar
Siezen, R. J., Bindels, J. G. & Hoenders, H. J. The quaternary structure of bovine alpha-crystallin. Size and charge microheterogeneity: more than 1000 different hybrids? Eur. J. Biochem. 91, 387–396 (1978).
Article CAS PubMed Google Scholar
Janowska, M. K., Baughman, H. E. R., Woods, C. N. & Klevit, R. E. Mechanisms of small heat shock proteins. Cold Spring Harb. Perspect. Biol. 11, a034025 (2019).
Article CAS PubMed PubMed Central Google Scholar
Kim, K. K., Kim, R. & Kim, S.-H. Crystal structure of a small heat-shock protein. Nature 394, 595–599 (1998).
van Montfort, R. L. M., Basha, E., Friedrich, K. L., Slingsby, C. & Vierling, E. Crystal structure and assembly of a eukaryotic small heat shock protein. Nat. Struct. Biol. 8, 1025–1030 (2001).
Article PubMed Google Scholar
White, H. E. et al. Multiple distinct assemblies reveal conformational flexibility in the small heat shock protein Hsp26. Struct. Lond. Engl. 1993 14, 1197–1204 (2006).
CAS Google Scholar
Shi, X. et al. Small heat shock protein AgsA forms dynamic fibrils. FEBS Lett. 585, 3396–3402 (2011).
Article CAS PubMed Google Scholar
Braun, N. et al. Multiple molecular architectures of the eye lens chaperone αB-crystallin elucidated by a triple hybrid approach. PNAS 108, 20491–20496 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Hanazono, Y., Takeda, K., Yohda, M. & Miki, K. Structural studies on the oligomeric transition of a small heat shock protein, StHsp14.0. J. Mol. Biol. 422, 100–108 (2012).
Article CAS PubMed Google Scholar
Hanazono, Y. et al. Nonequivalence observed for the 16-meric structure of a small heat shock protein, SpHsp16.0, from Schizosaccharomyces pombe. Structure 21, 220–228 (2013).
Article CAS PubMed Google Scholar
Weeks, S. D. et al. Molecular structure and dynamics of the dimeric human small heat shock protein HSPB6. J. Struct. Biol. 185, 342–354 (2014).
Article CAS PubMed Google Scholar
Fleckenstein, T. et al. The chaperone activity of the developmental small heat shock protein Sip1 is regulated by pH-dependent conformational changes. Mol. Cell 58, 1067–1078 (2015).
Article CAS PubMed Google Scholar
Mani, N. et al. Multiple oligomeric structures of a bacterial small heat shock protein. Sci. Rep. 6, 24019 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Clark, A. R. et al. Terminal regions confer plasticity to the tetrameric assembly of human HspB2 and HspB3. J. Mol. Biol. 430, 3297–3310 (2018).
Article CAS PubMed PubMed Central Google Scholar
Kaiser, C. J. O. et al. The structure and oxidation of the eye lens chaperone αA-crystallin. Nat. Struct. Mol. Biol. 26, 1141–1150 (2019).
Article CAS PubMed PubMed Central Google Scholar
Biswas, S., Garg, P., Dutta, S. & Suguna, K. Multiple nanocages of a cyanophage small heat shock protein with icosahedral and octahedral symmetries. Sci. Rep. 11, 21023 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Lee, J., Ryu, B., Kim, T. & Kim, K. K. Cryo-EM structure of a 16.5-kDa small heat-shock protein from Methanocaldococcus jannaschii. Int. J. Biol. Macromol. 258, 128763 (2024).
Article CAS PubMed Google Scholar
Koteiche, H. A., Chiu, S., Majdoch, R. L., Stewart, P. L. & Mchaourab, H. S. Atomic models by Cryo-EM and site-directed spin labeling: application to the N-terminal region of Hsp16.5. Structure 13, 1165–1171 (2005).
Article CAS PubMed Google Scholar
Santhanagopalan, I. et al. It takes a dimer to tango: oligomeric small heat shock proteins dissociate to capture substrate. J. Biol. Chem. 293, 19511–19521 (2018).
Article CAS PubMed PubMed Central Google Scholar
Haslbeck, M. et al. Hsp26: a temperature-regulated chaperone. EMBO J. 18, 6744–6751 (1999).
Article CAS PubMed PubMed Central Google Scholar
Quinlan, R. A. et al. Changes in the quaternary structure and function of MjHSP16.5 attributable to deletion of the IXI motif and introduction of the substitution, R107G, in the α -crystallin domain. Philos. Trans. R. Soc. B Biol. Sci. 368, 20120327 (2013).
Article Google Scholar
Shi, J. et al. Cryoelectron microscopy analysis of small heat shock protein 16.5 (Hsp16.5) complexes with T4 lysozyme reveals the structural basis of multimode binding. J. Biol. Chem. 288, 4819–4830 (2013).
Article CAS PubMed Google Scholar
Stengel, F. et al. Quaternary dynamics and plasticity underlie small heat shock protein chaperone function. Proc. Natl Acad. Sci. USA 107, 2007–2012 (2010).
Article ADS CAS PubMed PubMed Central Google Scholar
Lelj-Garolla, B. & Mauk, A. G. Self-association and chaperone activity of Hsp27 are thermally activated *. J. Biol. Chem. 281, 8169–8174 (2006).
Article CAS PubMed Google Scholar
Bova, M. P., Huang, Q., Ding, L. & Horwitz, J. Subunit exchange, conformational stability, and chaperone-like function of the small heat shock protein 16.5 from Methanococcus jannaschii. J. Biol. Chem. 277, 38468–38475 (2002).
Article CAS PubMed Google Scholar
Greaves, R. B. & Warwicker, J. Mechanisms for stabilisation and the maintenance of solubility in proteins from thermophiles. BMC Struct. Biol. 7, 18 (2007).
Article PubMed PubMed Central Google Scholar
Kim, R. et al. On the mechanism of chaperone activity of the small heat-shock protein of Methanococcus jannaschii. Proc. Natl Acad. Sci. USA 100, 8151–8155 (2003).
Article ADS CAS PubMed PubMed Central Google Scholar
Pasta, S. Y., Raman, B., Ramakrishna, T. & Rao, C. M. Role of the conserved SRLFDQFFG region of alpha-crystallin, a small heat shock protein. Effect on oligomeric size, subunit exchange, and chaperone-like activity. J. Biol. Chem. 278, 51159–51166 (2003).
Article CAS PubMed Google Scholar
Haley, D. A., Horwitz, J. & Stewart, P. L. The small heat-shock protein, αb-crystallin, has a variable quaternary structure. J. Mol. Biol. 277, 27–35 (1998).
Article CAS PubMed Google Scholar
Mishra, S. et al. Engineering of a polydisperse small heat-shock protein reveals conserved motifs of oligomer plasticity. Structure 26, 1116–1126.e4 (2018).
Article CAS PubMed PubMed Central Google Scholar
Shi, J., Koteiche, H. A., McHaourab, H. S. & Stewart, P. L. Cryoelectron microscopy and EPR analysis of engineered symmetric and polydisperse Hsp16.5 assemblies reveals determinants of polydispersity and substrate binding. J. Biol. Chem. 281, 40420–40428 (2006).
Article CAS PubMed Google Scholar
Baldwin, A. J. et al. The polydispersity of αB-crystallin is rationalized by an interconverting polyhedral architecture. Structure 19, 1855–1863 (2011).
Article CAS PubMed Google Scholar
Shepherd, D. A., Marty, M. T., Giles, K., Baldwin, A. J. & Benesch, J. L. P. Combining tandem mass spectrometry with ion mobility separation to determine the architecture of polydisperse proteins. Int. J. Mass Spectrom. 377, 663–671 (2015).
Article CAS Google Scholar
Clouser, A. F. et al. Interplay of disordered and ordered regions of a human small heat shock protein yields an ensemble of ‘quasi-ordered’ states. eLife 8, e50259 (2019).
Article PubMed PubMed Central Google Scholar
Jehle, S. et al. N-terminal domain of alphaB-crystallin provides a conformational switch for multimerization and structural heterogeneity. Proc. Natl Acad. Sci. USA 108, 6409–6414 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Woods, C. N., Ulmer, L. D., Guttman, M., Bush, M. F. & Klevit, R. E. Disordered region encodes α-crystallin chaperone activity toward lens client γD-crystallin. Proc. Natl Acad. Sci. USA 120, e2213765120 (2023).
Article CAS PubMed PubMed Central Google Scholar
Mühlhofer, M. et al. Phosphorylation activates the yeast small heat shock protein Hsp26 by weakening domain contacts in the oligomer ensemble. Nat. Commun. 12, 6697 (2021).
Article ADS PubMed PubMed Central Google Scholar
Lee, G. J. & Vierling, E. A small heat shock protein cooperates with heat shock protein 70 systems to reactivate a heat-denatured protein. Plant Physiol. 122, 189–198 (2000).
Article CAS PubMed PubMed Central Google Scholar
Rice, L., Marzano, N., Cox, D., Oijen, A. van & Ecroyd, H. Single-molecule observations of human small heat shock proteins in complex with aggregation-prone client proteins. Preprint at https://doi.org/10.1101/2024.02.08.579576 (2024).
Johnston, C. L. et al. Single-molecule fluorescence-based approach reveals novel mechanistic insights into human small heat shock protein chaperone function. J. Biol. Chem. 296, 100161 (2021).
Article CAS PubMed Google Scholar
Miller, A. P., O’Neill, S. E., Lampi, K. J. & Reichow, S. L. The α-crystallin chaperones undergo a quasi-ordered co-aggregation process in response to saturating client interaction. J. Mol. Biol. 436, 168499 (2024).
Article CAS PubMed PubMed Central Google Scholar
Yu, C. et al. Structural basis of substrate recognition and thermal protection by a small heat shock protein. Nat. Commun. 12, 3007 (2021).
Article ADS PubMed PubMed Central Google Scholar
Martin, J. L., Mestril, R., Hilal-Dandan, R., Brunton, L. L. & Dillmann, W. H. Small heat shock proteins and protection against ischemic injury in cardiac myocytes. Circulation 96, 4343–4348 (1997).
Article CAS PubMed Google Scholar
Xi, D., Wei, P., Zhang, C. & Lai, L. The minimal α-crystallin domain of Mj Hsp16.5 is functional at non-heat-shock conditions. Proteins Struct. Funct. Bioinform. 82, 1156–1167 (2014).
Article CAS Google Scholar
Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).
Article CAS PubMed Google Scholar
Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat. Methods 17, 261–272 (2020).
Article CAS PubMed PubMed Central Google Scholar
Mastronarde, D. N. SerialEM: a program for automated tilt series acquisition on tecnai microscopes using prediction of specimen position. Microsc. Microanal. 9, 1182–1183 (2003).
Article ADS Google Scholar
Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017).
Article CAS PubMed Google Scholar
Punjani, A. & Fleet, D. J. 3D variability analysis: resolving continuous flexibility and discrete heterogeneity from single particle cryo-EM. J. Struct. Biol. 213, 107702 (2021).
Article CAS PubMed Google Scholar
Meng, E. C. et al. UCSF ChimeraX: tools for structure building and analysis. Protein Sci. Publ. Protein Soc. 32, e4792 (2023).
Article CAS Google Scholar
Afonine, P. V. et al. New tools for the analysis and validation of cryo-EM maps and atomic models. Acta Crystallogr. Sect. Struct. Biol. 74, 814–840 (2018).
Article ADS CAS Google Scholar
Afonine, P. V. et al. Real-space refinement in PHENIX for cryo-EM and crystallography. Acta Crystallogr. Sect. Struct. Biol. 74, 531–544 (2018).
Article ADS CAS Google Scholar
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of COOT. Acta Crystallogr. Sect. Struct. Biol. D. 66, 486–501 (2010).
Article ADS CAS Google Scholar
Croll, T. I. ISOLDE: a physically realistic environment for model building into low-resolution electron-density maps. Acta Crystallogr. Sect. Struct. Biol. 74, 519–530 (2018).
Article ADS CAS Google Scholar

Download references

Acknowledgements

We thank Dr. Kirsten Lampi for the helpful discussions. We are grateful for instrumentation access and training provided by the staff at the OHSU Multiscale Microscopy Core and Advanced Computing Center (supported by NIH grant: S10OD034224), and the Pacific Northwest Cryo-EM Center (supported by NIH grant R24GM154185) with the assistance of Dr. Janette Myers. The research was funded by NIH grants R01EY030987 and R35GM124779 (to S.L.R.) and fellowship F31EY033226 (to A.P.M.).

Author information

Authors and Affiliations

Department of Chemical Physiology and Biochemistry, Oregon Health and Science University, Portland, OR, 97239, USA
Adam P. Miller & Steve L. Reichow
Vollum Institute, Oregon Health and Science Institute, Portland, OR, 97239, USA
Adam P. Miller & Steve L. Reichow

Authors

Adam P. Miller
View author publications
Search author on:PubMed Google Scholar
Steve L. Reichow
View author publications
Search author on:PubMed Google Scholar

Contributions

A.P.M. performed the cryo-EM studies, biophysical characterizations, functional experiments, and analyzed these data; S.L.R. and A.P.M. conceived of the study and prepared the manuscript.

Corresponding author

Correspondence to Steve L. Reichow.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Hideki Shigematsu 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 (download PDF )

Description of Additional Supplementary Files (download PDF )

Supplementary Movie 1 (download MOV )

Supplementary Movie 2 (download MOV )

Supplementary Movie 3 (download MOV )

Supplementary Movie 4 (download MOV )

Reporting Summary (download PDF )

Transparent Peer Review file (download PDF )

Source data

Source data (download XLSX )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, 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 you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. 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-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Miller, A.P., Reichow, S.L. Mechanism of small heat shock protein client sequestration and induced polydispersity. Nat Commun 16, 3635 (2025). https://doi.org/10.1038/s41467-025-58964-3

Download citation

Received: 17 January 2025
Accepted: 07 April 2025
Published: 16 April 2025
Version of record: 16 April 2025
DOI: https://doi.org/10.1038/s41467-025-58964-3