The Dynamic Triage Interplay of Hsp90 with its Chaperone Cycle and Client Binding

doi:10.21203/rs.3.rs-4704369/v1

Download PDF

Article

The Dynamic Triage Interplay of Hsp90 with its Chaperone Cycle and Client Binding

https://doi.org/10.21203/rs.3.rs-4704369/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Hsp90, a crucial molecular chaperone, significantly influences both normal biology and disease through its regulation of diverse client proteins. Central to its function is its remarkable conformational plasticity, driven by both ATPase activity and client interactions. However, comprehensive insights into Hsp90's dynamic transitions at the molecular level remain elusive. Using solution NMR spectroscopy, here we reveal how ATP binding, hydrolysis, and client engagement drive significant conformational and dynamic shifts in E. coli Hsp90, HtpG, throughout its chaperone cycle. We observe pronounced conformational fluctuations across the entire chaperone molecule, particularly in regions crucial for nucleotide binding and conformational transitions. ATP binding induces an ensemble of slow-exchanging conformations, representing discrete on-path transition states from open to closed forms. Subsequent ATP hydrolysis resolves this structural heterogeneity, shifting HtpG into a distinct, compact conformation. Furthermore, HtpG exhibits stage-specific mobility across the chaperone cycle, potentially enhancing client processing. Client binding acts as an allosteric switch, dynamically priming HtpG for elevated chaperone activity and, therefore, its efficient remodeling. These findings provide atomic-level insights into the functional interplay between Hsp90's conformation, dynamics, nucleotide, and client interactions, culminating in a refined model that elucidates the intricate machinery of Hsp90.

Biological sciences/Biochemistry/Protein folding/Chaperones

Biological sciences/Biochemistry/Proteins/Chaperones

Biological sciences/Structural biology/NMR spectroscopy/Solution-state NMR

The heat shock protein 90 (Hsp90), a highly abundant and evolutionary conserved molecular chaperone, plays a crucial role in facilitating the maturation and activation of a diverse array of structurally and functionally distinct client proteins^1–4. Its broad influence is underscored by its regulation of approximately 10% of the eukaryotic proteome, including critical components such as 60% of the human kinome, 30% of E3 ligases, and 7% of transcription factors³, positioning it as a central hub of protein homeostasis. Many of these client proteins are essential for signal transduction, establishing Hsp90 as a prime target for anticancer strategies^5–7.

Operating as an ATP-dependent homodimer, Hsp90 comprises three highly conserved domains in each protomer: the N-terminal domain (NTD), responsible for nucleotide binding; the middle domain (MD), primarily involved in client binding; and the C-terminal domain (CTD), mediating dimerization (Fig. 1a). Structurally, Hsp90s belong to the GHKL (grase, Hsp90, histidine kinase, MutL) family of ATPases, classified as a split ATPase with residues essential for ATP hydrolysis located in the MD, including a conserved arginine residue that interacts with the γ-phosphate of ATP.

Previous structural studies of Hsp90 in different nucleotide states have captured a series of snapshots during the ATPase-driven chaperone cycle, suggesting dramatic conformational changes (Fig. 1b). In the nucleotide-free state, the Hsp90 dimer exhibits a highly dynamic V-shaped conformation, encompassing a wide range of open extended states. Upon ATP binding, substantial structural rearrangements occur in the NTD, including the closure of a crucial structural element known as the 'lid'. This closure facilitates the trapping of bound ATP while exposing a hydrophobic surface that promotes N-terminal dimerization and the swapping of the N-terminal β-strands between interacting NTDs. These local conformational transitions further trigger a domain-level rotation of the NTD, association between the NTD and MD crucial for ATP hydrolysis, and arm-arm proximity. At this stage, Hsp90 adopts an interdigitated closed arrangement⁸. Subsequent ATP hydrolysis results in a compact state of ADP-bound Hsp90, followed by nucleotide release and the chaperone’s return to its open conformation^9,10. Cycling through this well-defined sequence of events, Hsp90 physically associates with its clients and eventually release them in a matured conformation^11–13.

While the described scheme offers a foundational understanding of Hsp90's functional machinery, ongoing debates persist around several key aspects, particularly the specific conformations of Hsp90 in its ATP-bound and ADP-bound states and the atomic-level depiction of its dynamic landscape across the chaperone cycle. Although there is some consensus suggesting that ATP-bound Hsp90 adopts a closed conformation⁸, other studies propose a spectrum of conformations^10,14,15, reflecting the protein's complex regulatory mechanisms. Conflicting results have further complicated our understanding of the ATPase-driven conformational closure of Hsp90. While Förster resonance energy transfer (FRET) data have suggested a sequential transition^16,17, recent PET fluorescence studies indicate that these changes occur in a cooperative manner¹⁸ (Fig. 1b).

The conformation of Hsp90 in its ADP-bound state also remains controversial. Structural investigations across various Hsp90 species have yielded divergent results: crystallography studies of E. coli Hsp90 (HtpG) and the endoplasmic reticulum-resident Hsp90 (Grp94) suggest configurations ranging from a parallel-like⁹ to a "twisted V" conformation. Conversely, electron microscopy (EM) and small-angle X-ray scattering (SAXS) of HtpG, along with FRET analysis of yeast Hsp90, reveal a mixture of open and more compact structures^9,10,15,19. Another set of FRET studies points to a predominantly open conformation for the ADP-bound yeast Hsp90^20,21. Furthermore, the transitions in Hsp90’s dynamics across the nucleotide-driven cycle, crucial for its function, are poorly defined (Fig. 1b). The mechanism by which ATPase activity is modulated by client protein engagement, though proposed as a general mechanism for Hsp90 action^17,22–25, remains incompletely understood. These highlight significant gaps in our understanding of Hsp90’s regulation and function, underscoring the need for further research to unravel these complex molecular interactions.

Nuclear magnetic resonance (NMR) spectroscopy stands out as the technique capable of providing atomic-resolution structural and dynamic insights into biological molecules within an aqueous solution, closely resembling physiological environments. Recent advancements in NMR, combined with sophisticated isotope labeling strategies, have extended its applicability to large protein systems ^26–33. In this study, we leverage these methodologies to explore the conformational dynamic of HtpG, the E. coli Hsp90, throughout its chaperone cycle. Our findings elucidate the intricate conformational changes and dynamic transitions of HtpG across various nucleotide states, driven notably by ATP binding and hydrolysis, thus revealing its dynamically polymorphic landscape at the atomic-level. Furthermore, we demonstrate how client binding induces widespread alterations in HtpG dynamics, effectively reducing both dynamic and conformational barriers, thereby facilitating chaperone activation. These results offer further insights into the dynamic machinery of Hsp90, illuminating its sophisticated conformational transition and dynamic priming essential for chaperone function.

The dynamic view of HtpG in its nucleotide-free, open state

We have recently achieved NMR assignment of the full-length HtpG and established that the dimeric chaperone in solution adopts an open conformation in absence of nucleotide, with a single set of NMR signals³⁴. To explore the dynamic properties of HtpG in its nucleotide-free state, we first investigated its fast-exchange (pico- to nanosecond, ps-ns) side-chain dynamics using the triple-quantum based NMR experiments³⁵, and obtained the relaxation-derived order parameter S², which quantifies the amplitude of motion (S² = 0: no restriction; S² = 1: complete restriction). These data reveal that the mobile residues are clustered at (sub)domain junctions and peripheral regions (Fig. 1c). Notably, the N-helix (residues 15 to 25 in HtpG), which undergoes critical position shifting for HtpG dimer closure (Fig. 1c)³⁶, displayed significant structural fluctuations. Elements critical for ATP catalysis, including the nucleotide-binding site, the “lid” segment that swings substantially to ensure nucleotide docking, and the so-called catalytic loop (residues 327–342 in HtpG) comprising the absolutely conserved Arg residue (Arg³³⁶) in the middle domain, whose significant structure remodeling and subsequent reaching to ATP enable catalysis (Fig. 1c) ³⁷, demonstrate moderate to high mobility in absence of nucleotide. The significant dynamics at these regions may contribute to moderating nucleotide affinity by providing an entropic barrier. The amphipathic helix in CTD (CTD^amp, residues 543–565), known as the “client discriminator helix”¹⁴, demonstrated somewhat structural rigidity at the ps-ns timescale. Interestingly, the C-terminal dimer interface showed conformational mobility, suggesting a breathing motion ³⁸.

We next investigated the slow-exchange (micro- to millisecond, µs-ms) timescale dynamics in the full-length HtpG using Carr–Purcell–Meiboom–Gill (CPMG) relaxation dispersion experiments, and measured ¹H-¹³C multiple-quantum methyl-TROSY dispersion profiles³⁹. The presence of multiple interconverting states of a protein result in line broadening of the NMR signal, and this effect can be modulated by applying of a series of 180° radiofrequency (RF) pulses in the CPMG pulse sequence⁴⁰. We observed a significant number of non-flat relaxation dispersions for methyl groups throughout the structure, underscoring the dynamic nature of HtpG (Fig. 1c). In particular, regions crucial for ATPase catalysis, including the N-helix, lid segment, nucleotide binding pocket, and the catalytic loop, exhibited apparent conformational plasticity on the µs‐ms timescale. Other regions undergoing slow conformational exchanges were predominantly distributed at the (sub)domain interfaces, as well as in the vicinity of CTD^amp. All dispersion curves could be fit well to a simple two-site exchange process, whereas in most cases the fast-exchanging rates (k_ex > 3000 S^− 1) preclude accurate estimation of populations of exchanging states or chemical shift differences. Notably, the conformational exchange rates of lid segment, as well as other regions involved in nucleotide binding and catalysis, appear significantly higher than that of the nucleotide association/disassociation¹⁹, suggesting the nucleotide binding process is largely thermally driven, as previously proposed¹⁹.

Taken together, these results the mobility profile of HtpG over a broad spectrum of time scales. The highly dynamic nature of HtpG may contribute for its remarkable capacity to accommodate structurally diverse client proteins, and facilitate its functional structure reorientation.

ATP binding shifts HtpG’s conformational equilibrium along the sequential open-to-closed transition pathway

As the functional roles of Hsp90 rely on in its unique, nucleotide-dependent conformational plasticity, we next assessed the conformation of HtpG across various nucleotide states. First, we acquired NMR spectra of HtpG in presence of AMP-PNP, a “nonhydrolysable” ATP analogue that has been shown to exert effects identical to ATP³⁶. Remarkably, upon binding to AMP-PNP, a significant number of methyl resonances throughout the entire HtpG molecule displayed signal splitting, accompanied by extensive chemical shift perturbations (Fig. 2a,b). Additional sets of NMR signals often indicate the presence of multiple conformational states undergoing slow exchange (k_ex < 1000 s^− 1). This signal splitting cannot be attributed to insufficient AMP-PNP supply, as the residues around the nucleotide binding site (e.g. L94, V133 and V137) shift completely from their original positions (Fig. 2a). Moreover, the evaluation of NMR signal integration revealed that different residues displayed varied splitting profiles concerning the number of split resonances and populations of conformational states (Fig. 2d). These observations ruled out the possibility that the NMR resonance multiplet was caused by an asymmetric configuration for each Hsp90 protomer, which would otherwise exhibit a uniformly distributed population profile. For instance, while the residues of L76, V83, L233 and L278 exhibit characteristic of triplet splitting, the population distribution ratio for the three states along the conformation transition pathway (states a, b and c, respectively) ranges from 5:1:4 to 2:1:1 (Fig. 2d). Residues displaying a doublet splitting profile, such as M20, I222, I306 and T548, exhibited populations ratio of states a and b ranges from 8:1 to 1:1 (Fig. 2d). These results thus provide a multifaceted and highly polymorphic perspective of HtpG in its AMP-PNP-bound state.

In the sequential transition model, Hsp90’s conformational change from open to closed involves two principal stages: (1) a ~ 90° rotation of the NTD accompanied by local structural alterations and MD association; and (2) arm-arm approximation (Fig. 2c)^14,17. Our NMR data of HtpG-AMPPNP provide an allosteric fingerprint of these essential conformational stages, thus strongly supporting the sequential model. We propose that ATP-bound HtpG exists in an equilibrium of multiple conformational states, spanning all three on-path transitional states: the open state (state a), the intermediate NTD-rotated state (state b), and the closed state (state c). This is evidenced by multiple lines of evidence. Firstly, NMR spectra of the AMPPNP-bound HtpG^NM, capable of NTD-rotation but not arm closure due to lacking the dimerization-pivoting CTD, showed conformation transition halting at state b, the NTD-rotated intermediate state (Fig. 2e and Extended Data Fig. 1). Additionally, signal splitting in residues I306 and M314, corresponding to residues I350 and I358 in yeast Hsp82, has been shown to reflect the conformational transition to the NTD-rotated phase¹⁴. The linear progression of triplet-split signals provides further support for a sequential transition process (Fig. 2d,e). The inability of residues to transition beyond state b suggests a unique chemical environment preserved across these states, resulting in unchanged chemical shifts or signal degeneration due to rapid averaging. Notably, the arm closure-impaired HtpG^NM construct in its AMPPNP-bound state shows an increased relative population of state b, aligning with expectations for a system where progression to state c is obstructed (Fig. 2f). Lastly, negative-staining EM analysis confirm that, even in the presence of AMPPNP, a substantial portion of HtpG retains the V-shaped open state (Fig. 2h), in alignment with previous SAXS studies that have shown the co-existence of open and closed Hsp90 conformations with ATP bound¹⁵.

These findings affirm the sequential nature of HtpG’s conformational transitions upon ATP binding and highlight the sophisticated regulation of these transitions at the molecular level. Further comprehensive spectral analysis through signal integration has quantified the populations of the AMPPNP-bound HtpG conformations: 48 ± 7% in the open state (state a), 23 ± 11% in the intermediate NTD-rotated state (state b), and 29 ± 10% in the closed state (state c). These findings suggest that the conformations of AMPPNP-bound HtpG are best represented by a rugged energy landscape, as illustrated in Fig. 2i.

Conformational Closure of HtpG is Not Required for ATP hydrolysis

Prior research has established the conserved arginine residue in the α2 (R33 in HtpG) as a critical conformational switch within the interaction network between the NTD and MD of Hsp90 and ATP^14,41 (Fig. 2i). Supporting the transition scenario mentioned above (Fig. 2c), mutation of R33 impedes the transition from the open to the closed state, as evidenced by the absence of resonance splitting in the NMR spectra of the HtpG^R33A mutant bound to AMP-PNP (Fig. 2j and Extended Data Fig. 2), which indicates a halt at the open state and an inability to progress along the transition pathway. Furthermore, the R33 mutation induces extensive chemical shift perturbations beyond the NTD, underscoring its central role in the conformational modulation of Hsp90 (Extended Data Fig. 2).

The necessity of conformational closure for ATP hydrolysis by Hsp90 has remained uncertain. To address this, we measured the ATPase activity of the conformation-transition impaired mutant, HtpGR^33A, finding it to exhibit substantial catalytic activity—approximately 80% greater than that of the wild-type protein (Extended Data Fig. 3a). This mutant, despite its inability to achieve dimer closure, efficiently catalyzes ATP hydrolysis. Further enzymatic kinetics analysis revealed that the mutation leads to an approximately 2.7-fold increase in turnover rates, while the Michaelis-Menten constant (K_m) remains largely unchanged (Fig. 2k and Extended Data Fig. 3b,c). These findings align with hybrid quantum/classical (QM/MM) free-energy calculations⁴¹, which suggest that mutation of the conserved arginine residue significantly lowers the reaction energy barrier, thereby enhancing the catalysis rate.

ATP Hydrolysis Induces a Transition from Structural Heterogeneity to a 'Compact' Conformation in HtpG

We next sought to use NMR to elucidate the conformational state of the ADP-bound HtpG in solution. In sharp contrast to the split spectra observed for AMPPNP-HtpG, the ADP-bound chaperone displays a singular methyl-TROSY resonance set (Fig. 3a), suggesting both HtpG protomers are uniformly sampling the same conformational space within the fast-exchange NMR timescale, specifically under 1 ms, without slow conformation exchanging events occurring. This observation precludes slow conformational exchange events and indicates a homogenous conformation as opposed to a heterogeneous mixture, challenging previous assertions based on negative stain EM and SAXS studies¹⁰. Furthermore, ADP binding induced substantial chemical shift perturbations (CSPs), mapping not only to the NTD containing the nucleotide pocket, but throughout all three HtpG domains and the inter-domain linkers (Fig. 3b). These widespread CSPs suggest ADP binding triggers global conformational rearrangements in HtpG. The controversy surrounding whether ADP-bound HtpG assumes a parallel⁹ or a more 'compact' conformation¹⁰, with the NTD bending toward the MD, persists (Fig. 3c). However, NMR's high conformational sensitivity enabled the distinction between these two disparate structures. Specifically, the 'compact' conformation, characterized by the NTD's extensive interaction with the MD, was evidenced by significant CSPs at the interface, notably at residues I30, V193, I196, A254, A283, and M288 (Fig. 3d). These perturbations are consistent with the 'compact' model and cannot be readily explained by a parallel orientation. Thus, our NMR findings support the conclusion that HtpG predominantly adopts a 'compact' conformation in solution, which is further corroborated by our negative staining data (Fig. 3e and Extended Data Fig. 4).

Transitions in the Dynamics of HtpG Throughout Its ATPase-Driven Chaperone Cycle

To ascertain whether the dynamics of HtpG, in addition to its conformations, are modulated during its chaperone cycle, we further employed NMR to characterize the protein’s dynamics in complex with AMP-PNP or ADP. For AMP-PNP-bound HtpG, in most cases, our dynamic analysis primarily centered on the strongest signals, those corresponding to the open conformation (state a). Both nucleotides were found to significantly influence the order parameters (S²) across the entire HtpG molecule, including critical regions such as the nucleotide binding pocket (e.g., residue L32), the lid segment (e.g., residue L119), NTD-MD interface (e.g., residues L339 and L352), MD inter-subdomain interface (e.g., residue A397), and MD-CTD joint region (e.g., residue V539), with each nucleotide state displaying unique dynamic signatures (Fig. 4a). This data indicates that nucleotide engagement at the NTD has a far-reaching effect, allosterically modulating HtpG’s movements across the ps-ns timescale. In particular, regions proximal to the nucleotide-binding site exhibited decreased mobility, suggesting an entropically favorable environment for nucleotide accommodation (Extended Data Fig. 5a). Of note, AMP-PNP binding, compared to the apo or ADP-bound state, increases the order parameter for the catalytic loop residue L339 from ~ 0.2 to 0.4 (Fig. 4a). This change implies that ATP's gamma-phosphate rigidified the catalytic loop "breathing motion," corroborating its role in reinforcing the NTD-MD contacts ⁴².

We recently elucidated solution complex structure of HtpG with a disordered client (Δ131Δ)³⁴. Here we explored the dynamic transitions in the client-engagement regions during the chaperone cycle. Our results reveal a complex dynamic transition profile, with residues proximal to or within the client-binding grooves displaying stage-specific mobility patterns (Fig. 4b). These subtly refined dynamic environments may exert distinct chaperoning influences on client at each stage of the cycle, potentially facilitating their processing.

CPMG experiments have provided further insights into the µs-ms timescale dynamics of HtpG during the distinct stages of the chaperone cycle. In particular, residues M85, I91, L109, V125, and V164 near the nucleotide pocket and A254 at the NTD-MD interface exhibited individualized relaxation dispersion profiles according to the nucleotide state (Fig. 4c). While AMP-PNP binding appears to stabilize the nucleotide-binding pocket and surrounding residues—evidenced by the decreased conformational exchange rates for residues such as I91, V125, and V164—this is not uniform across all residues (Extended Data Fig. 5b, c). Conversely, the ADP state revealed a varied response, with some residues displaying increased exchange rates and others showing decreased or unchanged rates. These differences in the conformational exchange rates, notably higher in the ADP-bound state compared to AMP-PNP, align with the more rapid kinetics of ADP binding and release, shedding light on nucleotide turnover rates. Notably, the µs-ms dynamics of the catalytic loop residue L339 remained largely unchanged (Extended Data Fig. 5b), exhibiting an exchange rate (k_ex ≈ 3000 s⁻¹) that greatly surpasses the catalytic rate (k_cat≈0.005 s⁻¹). This suggests that movements of the catalytic loop are not the rate-limiting step for ATP hydrolysis, in agreement with prior observations⁴².

Furthermore, each conformational state of AMPPNP-bound HtpG, as illustrated by residue M20, presented distinct CPMG profiles within the intermediate to slow exchange time regime (Fig. 4d). This observation introduces an additional dimension of dynamic intricacy, underscoring the complex conformational landscape that HtpG navigates during its functional cycle.

Collectively, our findings assemble a detailed picture of the dynamic reconfiguration of HtpG throughout its chaperone cycle, underscoring the nuanced allosteric regulation and conformational adaptability essential to its function.

The dynamic response of HtpG to client binding

The modulation of Hsp90 ATPase activity by client binding has been proposed as a general mechanism of the action for the chaperone, however, the underlying mechanism remains unclear. To investigate this aspect in terms of protein dynamics, we conducted NMR analyses on the ¹³CH₃ methyl moieties of the HtpG-Δ131Δ complex. Surprisingly, we observed an overall increase in ps-ns timescale motion (ΔS² < 0) in HtpG upon client binding, particularly around the nucleotide binding pocket in the NTD (Fig. 5a). Notably, A114 and A551, situated in the lid and "client discriminator helix", respectively, gained a large amplitude of structural mobility in response to Δ131Δ binding, with marked decrease in S² values of ∼0.5 and ∼0.9, respectively (Fig. Extended Data Fig. 6a). However, the effects of client binding on HtpG’s internal motions are non-uniform; certain residues, like L339 that at the NTD-MD interface, displayed decreased conformational entropy (Extended Data Fig. 6a).

Consistent with these findings, relaxation dispersion experiments indicated that client binding to the MD caused an overall increase in the fluctuation rates of HtpG at the µs-ms timescale, with many of these residues being remotely located in the NTD, indicating a reduced activation energy barrier at the catalysis center (Fig. 5b). Additionally, while some residues, such as L360, A550 and M550, exhibited distinct conformation transition profiles in the client-free form, Δ131Δ binding eliminated any alternative conformational states occurring on the µs-ms timescale (the detection limit of the relaxation dispersion experiments is ∼0.5%) (Extended Data Fig. 6b).

Subsequently, we acquired methyl-TROSY NMR spectra of the HtpG-Δ131Δ complex in presence of AMP-PNP, displaying resonance patterns akin to the client-free state (Extended Data Fig. 7). These data indicate that in the AMP-PNP state, the HtpG-Δ131Δ complex, resembling client-free HtpG, adopts multiple conformations in equilibrium. Nevertheless, conformational populations of HtpG-Δ131Δ complex in the AMP-PNP state differ significantly from those of client-free Hsp90 (Fig. 5c). For instance, residues M20 and M314 exhibited increased populations of state b, corresponding to the transitioned state upon AMP-PNP binding, by 2- and 3-fold, respectively (Fig. 5d). Furthermore, residues L76 and L233 displayed a substantial decrease in the intermediate state (state b) due to Δ131Δ binding, shifting the conformation equilibrium towards the closed state (state c) by around 10% and 15%, respectively (Fig. 5e).

Taken together, our results reveal that despite the spatial separation of ATP hydrolysis and client binding sites in the NTD and MD, respectively, these distinct functional processes are dynamically communicated across the full-length HtpG. In this scenario, client binding reconfigures the global internal motions, effectively lowering the rate-limiting conformational barrier and shifting the conformational equilibrium of Hsp90 during the chaperone cycle, in line with recent yeast Hsp90 studies¹⁴.

In this study, we unveil the complex interplay within HtpG, the E. coli Hsp90, which is orchestrated by its ATPase-driven functional cycle that meticulously integrates conformational shifts with dynamic transitions. Our findings highlight the mechano-chemical coupling between ATP hydrolysis and the extensive structural rearrangements within HtpG. ATP binding initiates HtpG's transition towards a closed state, setting the stage for an equilibrium among the discrete conformations of open, NTD-rotated intermediate, and closed states. Crucially, the observation of all on-path transition states during this conformational shift, coupled with the demonstrative effects of specific truncation and mutations, robustly support a sequential transition mechanism for HtpG (Fig. 2c). Following ATP hydrolysis, this structural heterogeneity is resolved, and our solution NMR spectra suggest that the chaperone transitions into a distinct, singular 'compact' conformation in its ADP-bound state. This sequence of events yields insights into how the energies from nucleotide binding and hydrolysis are elegantly utilized to drive protein conformational changes.

A noteworthy parallel is drawn with eukaryotic Hsp90 through the R33A mutation, which similarly restricts HtpG to an open state, thereby impeding the transition to subsequent conformations along the pathway (Fig. 2i)^14,41. This similarity suggests a conserved conformational transition mechanism across prokaryotic and eukaryotic Hsp90s, orchestrated by an electrostatic network between the NTD and the catalytic loop within the MD^41,43. In this scenario, the allosteric signal of ATP binding is initially detected by the nucleotide-binding pocket and is then propagated throughout the entire protein, including the client-binding regions. This implies that HtpG’s conformational transition cycle is inexorably linked to its intrinsic ATPase activity.

Interestingly, this portrayal of E. coli Hsp90’s conformational change as being unilaterally linked to ATP hydrolysis appears to diverge from that observed in eukaryotic Hsp90s. FRET analyses suggest that for yeast Hsp90, thermal fluctuations rather than ATP binding or hydrolysis predominantly dictate conformational states, exhibiting minimal directionality in this transformation^19,21. Conversely, NMR studies show that ATP binding induces signal splitting in yeast Hsp90 into doublets—representative of the open and NTD-rotated states¹⁴—a stark contrast to the triplet signals for many residues in E. coli Hsp90, which span all three identified states (Fig. 2a,d). This discrepancy raises an intriguing question about the evolutionary advantages conferred by the distinct conformational states and dynamics between these species-specific Hsp90 homologs.

Additionally, we present distinct dynamic signatures for HtpG at each stage of the chaperone cycle, covering timescales from nanoseconds to milliseconds. These dynamics, critical for the chaperone’s function, have previously been vaguely depicted. Recent NMR studies on human Hsp90α NTD have identified transiently populated states, presumably representing the lid-closed state, within the isolated domain⁴⁴. Our findings extend this observation, demonstrating that such conformational exchanges are not confined to the NTD but also occur across multiple regions of the full-length E. coli Hsp90 molecule. While we cannot yet unambiguously assign these transient states, many are situated near the nucleotide-binding pocket or at (sub)domain interfaces, emphasizing the critical role of this structural plasticity in ATP catalysis and subsequent conformational rearrangements. Notably, the conformational exchange rates are remarkably similar across HtpG, human Hsp90α⁴⁴, and yeast Hsp90²⁰, suggesting a comparable level of structural plasticity conserved across different Hsp90 species.

The precise functional role of Hsp90 ATPase activity in its chaperoning machinery continues to be a subject of debate. While it has been suggested to regulate client interactions⁴⁵, our recent findings indicate that the engagement of HtpG with the disordered client Δ131Δ is largely nucleotide-independent³⁴, consistent with interaction studies of Hsp90 with Tau²⁷ and p53¹⁴. Building on the structure of the HtpG-Δ131Δ encounter complex, we have proposed that Hsp90 functions as an ATP-driven mechanochemical clamp, where two protomers synergistically capture discrete hydrophobic segments within client proteins³⁴. In this study, we extend our model by demonstrating that HtpG not only adopts distinct conformations but also exhibits unique dynamic signatures across various nucleotide states, indicating that the ATP-driven chaperone cycle intricately fine-tunes the dynamic properties of its client-engagement grooves. These findings have led us to propose a refined model of Hsp90's chaperoning mechanism (Fig. 6). In this model, the conformations of bound clients are synchronized with Hsp90’s open-to-closed conformational cycle, and the local dynamic properties of the client-binding grooves significantly influence transient, low-affinity (re)bindings of clients at multiple sites. This refined understanding suggests that ATP's role extends simply providing mechanical forces for macroscopic manipulation of clients, such as stretching and squeezing. More subtly, ATP is crucial for fine-tuning the dynamic environments that balance client hold-folding processes at numerous sites, thereby facilitating their conformational sampling and eventual maturation. This comprehensive view highlights ATP's dual function in driving the mechanical aspects of client processing while finely regulating the molecular dynamics essential for optimal chaperone activity.

Our study also presents compelling evidence that client binding instigates a widespread transformation in HtpG dynamics. This transformation effectively triggers chaperone activation by lowering the dynamic and conformational barriers required for the transition to the closed state. These findings enhance our understanding of how client binding intricately regulates Hsp90's ATPase activity. By dynamically adapting to the presence of a client, Hsp90 showcases its capacity to modulate its behavior in response to the cellular context. In this reciprocal interaction scenario, the client binding functions as a dynamic switch, priming Hsp90 with a refined platform that potentially facilitates both the chaperone process and the sampling of client conformations, thereby facilitating efficient client remodeling.

Expression and Preparation of Proteins

Protein expression and preparation were carried out using the following constructs: The E. coli Hsp90 (residues 1-624), which was cloned into a pET-SUMO vector with an N-terminal 6×His tag and an Ulp1 cleavage site. Additionally, The HtpG^NM construct (1-495) was cloned into a pET-16a vector and fused to a 6×His tag, including a tobacco etch virus (TEV) protease cleavage site. Fusion construct between Hsp90 and Δ131Δ was prepared by fusing the Δ131Δ to the C terminus of Hsp90, with a linker of 65 residues to ensure no bias in the complex conformation. The HtpG mutant R33A was generated using PrimeSTAR Max DNA Polymerase (Takara).

All constructs were transformed into BL21(DE3) cells. Cells were grown in LB medium at 37°C with kanamycin (100 µg/ml). Protein expression was induced by adding 0.1 mM IPTG at OD₆₀₀ ~ 0.5, followed by 4 to 6 hours of incubation at 37°C.

Purification of proteins

Cells were harvested by centrifugation at 5,000g for 20 min at 4°C, resuspended in lysis buffer (25 mM Tris-HCl pH 8.0, 500 mM NaCl, 10 mM imidazole, and 1 mM PMSF), and stored at -80°C until use. After thawing, cells were lysed by a JN-Mini Pro Low-temperature Ultra-high pressure cell disrupter (JNBIO) at 4°C. The lysate was clarified by centrifugation at 50,000g for 30 min at 4°C, and the supernatant was passed through a Ni²⁺-chelating column (GE Healthcare) pre-equilibrated with lysis buffer.

The column was then washed with 25 mM Tris-HCl pH 8.0, 500 mM NaCl, and 25 mM imidazole to remove non-specifically bound proteins, and the target protein was eluted with 25 mM Tris-HCl pH 8.0, 150 mM NaCl, and 300 mM imidazole. The eluted protein was subjected to TEV protease or SUMO protease cleavage for removal of amino-end tags by dialysis. Briefly, the eluted protein was mixed with SUMO protease and dialyzed overnight against 50 mM Tris-HCl pH 8.0, 100 mM NaCl, and 1.5 mM β-mercaptoethanol (βme). The protein mixture was then subjected to another round of Ni²⁺-chelating column purification to subtract the amino-terminal tag and His-tagged TEV protease or His-tagged SUMO protease.

The flow-through was collected, concentrated, and applied to a HiLoad 16/600 Superdex 200 size exclusion column (GE Healthcare). HtpG samples were eluted with 50 mM phosphate pH 7.0, 100 mM NaCl, and 3 mM βme. The purity of each protein was assessed by SDS-PAGE analysis, and the concentration was determined using the Epoch 2 Microplate Spectrophotometer (BioTek) at 280 nm with the corresponding extinction coefficients.

Protein Isotope Labeling for NMR Studies

To produce methyl-protonated samples, we grew cells in 99.9% D₂O (CIL and Isotec) supplemented with D₇, ¹²C-glucose (2g/L), [¹³CH₃]-a-ketobutyric acid (46 mg/L), [¹³CH₃]-Met (50 mg/L), [²H₂, ¹³CH₃]-Ala (50 mg/L), [²H₂, ¹³CH₃]-Thr (50 mg/L), [¹³CH₃]-a-ketoisovaleric acid (85 mg/L), and [²H₅]-Gly (100 mg/L) to the culture 30 minutes before adding IPTG.

NMR Spectroscopy

Isotopically labeled samples were prepared in 50 mM phosphate pH 7.0, 100 mM NaCl, 0.05% NaN₃, and 7% D₂O. NMR experiments were conducted using Bruker Avance III 600MHz and 850MHz spectrometers that are equipped with cryogenic probes. NMR experiments of HtpG and its complex were conducted at 37°C. The protein sample concentrations ranged from 0.5 to 1.0 mM. Nucleotides were added to a final concentration of 10 mM. All NMR spectra were processed by NMRPipe ⁴⁶ and analyzed using the NMRViewJ software (http://www.onemoonscientific.com).

NMR Relaxation Experiments

To analyze side-chain dynamics in the fast timescale (ps-ns), we used an 850 MHz spectrometer equipped with a Bruker AVANCE III console. We prepared methyl-labeled samples of HtpG, HtpG-AMP-PNP, HtpG-ADP, and HtpG-Δ131Δ fusion in buffer PB7.0 and collected relaxation data at 37°C. The homogeneity of all samples was confirmed by dynamic light scattering, and the global rotational correlation time (τ_c) was determined. Spectra were collected using an interleaved acquisition scheme with relaxation delays of 2, 7, 12, 17, 22, 27, 32, 37, and 42 ms. The order parameters were extracted as described ⁴⁷.

Methyl relaxation dispersion experiments

Multiple-quantum ¹³C-CPMG experiments were carried out on both 600 and 850 MHz spectrometers equipped with Bruker AVANCE and AVANCE III consoles, respectively. Spectra were acquired in an interleaved manner with CPMG field strength of 100, 200, 300, 400, 500, 700, 1000, 2000, 3000, 4000, 5000, 6000, 8000 ,10000, and 12000Hz on 600MHz and 100, 200, 400, 600, 800, 1000, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 10000, and 12000Hz on 850MHz.The data were processed using NMRPipe, and peak intensities were extracted using NMRViewJ. Subsequently, the peak intensities were converted to transverse decay rates, R₂^eff, as previously described⁴⁸.

$\:{K}_{ex}$ and $\:\varnothing\:$ were measured by using the equation:

$$\:{R}_{2}^{eff}={R}_{0}^{2}+\frac{\varnothing\:}{{k}_{ex}}\:\left[\:1-\frac{4{v}_{CPMG}}{{k}_{ex}}tanh\left(\frac{{k}_{ex}}{4{v}_{CPMG}}\right)\right]\:$$

The software NESSY was used to fit the relaxation data⁴⁹. All dispersion curves could be well-fitted to a simple two-site exchange process. However, in most cases, the fast-exchanging rates (K_ex > 3000 S^− 1) hindered accurate estimation of populations of exchanging states or chemical shift differences.

NMR chemical shift perturbations

The interactions of HtpG with nucleotides were monitored by methyl-TROSY experiments. AMP-PNP (10mM) or ADP (10mM) was add to 250 µM (dimer concentration) of methyl-labelled Hsp90 in 50 mM phosphate pH7.0, 100 mM NaCl, 0.05% NaN₃, and 7% D₂O. CSPs were calculated in Hz at 850 MHz using the following Eq. 5⁰:

$$\:\varvec{\Delta\:}\varvec{v}={\left({\left({\varvec{\Delta\:}}^{1}\text{H}\text{*}850\right)}^{2}+{\left({\varvec{\Delta\:}}^{13}\text{C}\text{*}213.8\right)}^{2}\right)}^{1/2}$$

Negative stain electron microscopy

For conformation visualization of HtpG at different nucleotide states, we incubated 4 µL of HtpG (0.007 mg/mL) and HtpG (0.007 mg/mL) with AMP-PNP (1 mM) or ADP (1 mM) for 1 h with 20 mM sodium molybdate at room temperature in 25 mM Tris-HCl, pH 8.0, 25 mM KCl, and 10 mM MgCl₂. Subsequently, we negatively stained the samples with 1% (w/v) uranyl acetate on carbon-film 400 mesh copper grids and imaged them using a FEI TF20 operated at 200 keV with 62,000× nominal magnification.

ATPase activity measurements

Enzymatic assays were carried out using 5 µM HtpG or HtpG^R33A at 37°C in 25 mM Tris-HCl pH 8.0, 25 mM KCl, 10 mM MgCl₂, and 0.5 mM ATP for 1 hour. Steady-state kinetic analysis was performed with 2 µM enzyme under the same conditions, except with various concentrations of ATP from 0.05 to 10 mM. Steady-state kinetic parameters were determined by double reciprocal Lineweaver-Burk plot for reactions that followed Michaelis-Menten kinetics. All experiments were repeated independently at least three times.

Zhao R et al (2005) Navigating the chaperone network: an integrative map of physical and genetic interactions mediated by the hsp90 chaperone. Cell 120:715–727
Taipale M, Jarosz DF, Lindquist S (2010) HSP90 at the hub of protein homeostasis: emerging mechanistic insights. Nat Rev Mol Cell Biol 11:515–528
Taipale M et al (2012) Quantitative analysis of HSP90-client interactions reveals principles of substrate recognition. Cell 150:987–1001
Lackie RE et al (2017) The Hsp70/Hsp90 Chaperone Machinery in Neurodegenerative Diseases. Front Neurosci 11:254
Whitesell L, Lindquist SL (2005) HSP90 and the chaperoning of cancer. Nat Rev Cancer 5:761–772
Trepel J, Mollapour M, Giaccone G, Neckers L (2010) Targeting the dynamic HSP90 complex in cancer. Nat Rev Cancer 10:537–549
Jaeger AM, Whitesell L (2019) HSP90: Enabler of Cancer Adaptation. Annual Rev Cancer Biology 3:275–297
Ali MM et al (2006) Crystal structure of an Hsp90-nucleotide-p23/Sba1 closed chaperone complex. Nature 440:1013–1017
Shiau AK, Harris SF, Southworth DR, Agard DA (2006) Structural Analysis of E. coli hsp90 reveals dramatic nucleotide-dependent conformational rearrangements. Cell 127, 329 – 40
Southworth DR, Agard DA (2008) Species-dependent ensembles of conserved conformational states define the Hsp90 chaperone ATPase cycle. Mol Cell 32:631–640
Verba KA et al (2016) Atomic structure of Hsp90-Cdc37-Cdk4 reveals that Hsp90 traps and stabilizes an unfolded kinase. Science 352:1542–1547
Wang RY et al (2022) Structure of Hsp90-Hsp70-Hop-GR reveals the Hsp90 client-loading mechanism. Nature 601:460–464
Noddings CM, Wang RY, Johnson JL, Agard DA (2022) Structure of Hsp90-p23-GR reveals the Hsp90 client-remodelling mechanism. Nature 601:465–469
Lopez A et al (2021) Client binding shifts the populations of dynamic Hsp90 conformations through an allosteric network. Sci Adv 7:eabl7295
Krukenberg KA, Forster F, Rice LM, Sali A, Agard DA (2008) Multiple conformations of E. coli Hsp90 in solution: insights into the conformational dynamics of Hsp90. Structure 16, 755 – 65
Hessling M, Richter K, Buchner J (2009) Dissection of the ATP-induced conformational cycle of the molecular chaperone Hsp90. Nat Struct Mol Biol 16:287–293
Street TO et al (2012) Cross-monomer substrate contacts reposition the Hsp90 N-terminal domain and prime the chaperone activity. J Mol Biol 415:3–15
Schubert J, Schulze A, Prodromou C, Neuweiler H (2021) Two-colour single-molecule photoinduced electron transfer fluorescence imaging microscopy of chaperone dynamics. Nat Commun 12:6964
Ratzke C, Berkemeier F, Hugel T (2012) Heat shock protein 90's mechanochemical cycle is dominated by thermal fluctuations. Proc Natl Acad Sci U S A 109:161–166
Schulze A et al (2016) Cooperation of local motions in the Hsp90 molecular chaperone ATPase mechanism. Nat Chem Biol 12:628–635
Mickler M, Hessling M, Ratzke C, Buchner J, Hugel T (2009) The large conformational changes of Hsp90 are only weakly coupled to ATP hydrolysis. Nat Struct Mol Biol 16:281–286
McLaughlin SH, Smith HW, Jackson SE (2002) Stimulation of the weak ATPase activity of human hsp90 by a client protein. J Mol Biol 315:787–798
Motojima-Miyazaki Y, Yoshida M, Motojima F (2010) Ribosomal protein L2 associates with E. coli HtpG and activates its ATPase activity. Biochem Biophys Res Commun 400:241–245
Lorenz OR et al (2014) Modulation of the Hsp90 chaperone cycle by a stringent client protein. Mol Cell 53:941–953
Street TO, Lavery LA, Agard DA (2011) Substrate binding drives large-scale conformational changes in the Hsp90 molecular chaperone. Mol Cell 42:96–105
Saio T, Guan X, Rossi P, Economou A, Kalodimos CG (2014) Structural basis for protein antiaggregation activity of the trigger factor chaperone. Science 344:1250494
Karagoz GE et al (2014) Hsp90-Tau complex reveals molecular basis for specificity in chaperone action. Cell 156:963–974
Huang C, Rossi P, Saio T, Kalodimos CG (2016) Structural basis for the antifolding activity of a molecular chaperone. Nature 537:202–
Huang C, Kalodimos CG (2017) Structures of Large Protein Complexes Determined by Nuclear Magnetic Resonance Spectroscopy. Annual Rev Biophys 46:46 317–336
Jiang Y, Rossi P, Kalodimos CG (2019) Structural basis for client recognition and activity of Hsp40 chaperones. Science 365:1313–1319
Rosenzweig R, Kay LE (2014) Bringing dynamic molecular machines into focus by methyl-TROSY NMR
Burmann BM et al (2019) Regulation of α-synuclein by chaperones in mammalian cells. 577:127–132
Faust O et al (2020) HSP40 proteins use class-specific regulation to drive HSP70 functional diversity. 1–6
Qu X et al (2024) Structural basis for the dynamic chaperoning of disordered clients by Hsp90. Nat Struct Mol Biol
Sun H, Kay LE, Tugarinov V (2011) An optimized relaxation-based coherence transfer NMR experiment for the measurement of side-chain order in methyl-protonated, highly deuterated proteins. J Phys Chem B 115:14878–14884
Karagoz GE et al (2011) N-terminal domain of human Hsp90 triggers binding to the cochaperone p23. Proc Natl Acad Sci U S A 108:580–585
Pearl LH, Review (2016) The HSP90 molecular chaperone-an enigmatic ATPase. Biopolymers 105:594–607
Ratzke C, Mickler M, Hellenkamp B, Buchner J, Hugel T (2010) Dynamics of heat shock protein 90 C-terminal dimerization is an important part of its conformational cycle. Proc Natl Acad Sci U S A 107:16101–16106
Korzhnev DM, Kloiber K, Kanelis V, Tugarinov V, Kay LE (2004) Probing slow dynamics in high molecular weight proteins by methyl-TROSY NMR spectroscopy: application to a 723-residue enzyme. J Am Chem Soc 126:3964–3973
Palmer AG 3rd, Kroenke CD, Loria JP (2001) Nuclear magnetic resonance methods for quantifying microsecond-to-millisecond motions in biological macromolecules. Methods Enzymol 339:204–238
Mader SL et al (2020) Conformational dynamics modulate the catalytic activity of the molecular chaperone Hsp90. Nat Commun 11:1410
Cunningham CN, Southworth DR, Krukenberg KA, Agard DA (2012) The conserved arginine 380 of Hsp90 is not a catalytic residue, but stabilizes the closed conformation required for ATP hydrolysis. Protein Sci 21:1162–1171
Schmid S, Gotz M, Hugel T (2018) Effects of Inhibitors on Hsp90's Conformational Dynamics, Cochaperone and Client Interactions. ChemPhysChem 19:1716–1721
Henot F et al (2022) Visualizing the transiently populated closed-state of human HSP90 ATP binding domain. Nat Commun 13:7601
Kirschke E, Goswami D, Southworth D, Griffin PR, Agard DA (2014) Glucocorticoid receptor function regulated by coordinated action of the Hsp90 and Hsp70 chaperone cycles. Cell 157:1685–1697
Delaglio F et al (1995) NMRPipe: A multidimensional spectral processing system based on UNIX pipes. 6:277–293
Sprangers R, Kay (2007) L.E.J.N. Quantitative dynamics and binding studies of the 20S proteasome by NMR. 445:618–622
McDonald LR, Boyer JA, Lee AL (2012) J.S. Segmental motions, not a two-state concerted switch, underlie allostery in CheY. 20 8:1363–1373
Bieri M, Gooley PR (2011) J.B.B. Automated NMR relaxation dispersion data analysis using NESSY. 12:421–421
Conicella AE et al (2020) An intrinsically disordered motif regulates the interaction between the p47 adaptor and the p97 AAA + ATPase. 117:26226–26236

There is NO Competing Interest.

ExtendedDataFig.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

The Dynamic Triage Interplay of Hsp90 with its Chaperone Cycle and Client Binding

Status:

Version 1

Abstract

Figures

Introduction

Results

The dynamic view of HtpG in its nucleotide-free, open state

ATP binding shifts HtpG’s conformational equilibrium along the sequential open-to-closed transition pathway

Conformational Closure of HtpG is Not Required for ATP hydrolysis

ATP Hydrolysis Induces a Transition from Structural Heterogeneity to a 'Compact' Conformation in HtpG

Transitions in the Dynamics of HtpG Throughout Its ATPase-Driven Chaperone Cycle

The dynamic response of HtpG to client binding

Discussion

Materials and Methods

Expression and Preparation of Proteins

Purification of proteins

Protein Isotope Labeling for NMR Studies

NMR Spectroscopy

NMR Relaxation Experiments

Methyl relaxation dispersion experiments

NMR chemical shift perturbations

Negative stain electron microscopy

ATPase activity measurements

References

Additional Declarations

Supplementary Files

Status:

Version 1