Allosteric Modulators of HSP90 and HSP70: Dynamics Meets Function

Perspective Cite This: J. Med. Chem. 2019, 62, 60−87

pubs.acs.org/jmc

Allosteric Modulators of HSP90 and HSP70: Dynamics Meets Function through Structure-Based Drug Design Mariarosaria Ferraro,† Ilda D’Annessa,† Elisabetta Moroni,‡ Giulia Morra,† Antonella Paladino,† Silvia Rinaldi,† Federica Compostella,§ and Giorgio Colombo*,†,∥ †

Istituto di Chimica del Riconoscimento Molecolare, CNR, Via Mario Bianco 9, 20131 Milano, Italy IRCCS MultiMedica, Via Fantoli 16/15, 20138 Milano, Italy § Dipartimento di Biotecnologie Mediche e Medicina Traslazionale, Università degli Studi di Milano, Via Saldini, 50, 20133 Milano, Italy ∥ Dipartimento di Chimica, Università di Pavia, V.le Taramelli 12, 27100 Pavia, Italy

J. Med. Chem. 2019.62:60-87. Downloaded from pubs.acs.org by UNIV OF GOTHENBURG on 01/18/19. For personal use only.

‡

ABSTRACT: Molecular chaperones HSP90 and HSP70 are essential regulators of the folding and activation of a disparate ensemble of client proteins. They function through ATP hydrolysis and the assembly of multiprotein complexes with cochaperones and clients. While their therapeutic relevance is recognized, important details underlying the links between ATP-dependent conformational dynamics and clients/cochaperones recruitment remain elusive. Allosteric modulators represent fundamental tools to obtain molecular insights into functional regulation. By selective perturbation of different aspects of HSP90/HSP70 activities, allosteric drugs can tune rather than completely inhibit signaling cascades, providing information on the relationships between structure-dynamics and function. Herein, we review advances in the design of HSP90 and HSP70 allosteric modulators. We consider inhibitors and activators in different biochemical and disease models. We discuss these compounds as probes to decipher the complexity of the chaperone machinery and that at the same time represent starting leads for the development of drugs against cancer and neurodegeneration.

■

INTRODUCTION Molecular chaperones are essential nodal proteins that integrate multiple biochemical networks fundamental for cell survival, proliferation, adaptation, and migration. In this framework, they act as the main regulators of proteostasis, supervising protein folding, activation, aggregation suppression, and degradation. All these tasks depend on finely tuned structural dynamics, which underlies enzymatic hydrolysis of ATP, and formation of multimolecular complexes with cochaperones and cofactors. Dysregulation of these functions1−3 has been associated with diverse kinds of human diseases, such as cancer, neurodegeneration, and inflammatory and infectious diseases. Because of their roles at the crossroads of multiple cellular functions and the likelihood that manipulation of their activities may impact on the treatment of diverse pathologies, efforts to target chaperone pathways with small molecules have been actively pursued for both theoretical and practical reasons. From the fundamental point of view, chemical tools that probe the intricate mechanistic aspects of chaperone function would represent important instruments to improve our knowledge of how complex biological systems are regulated. From the practical point of view, active small molecules could be engineered into new drug candidates © 2018 American Chemical Society

targeting a number of diseases with novel mechanisms of action. Indeed, some designed lead compounds targeting chaperones recently moved to clinical testing.4 HSPs include biomolecules such as HSP90, HSP70, HSP60, HSP47, HSP40, and HSP27, classified according to their respective molecular weights.5,6 These proteins are highly abundant in the cell and amount to ∼10% of the cellular mass. HSP90s and HSP70s alone account for more than half of that mass.7 In general, HSPs show a multipartite structural organization that consists of different domains endowed with distinct functions:2 in a simple schematic representation, a specific domain is dedicated to binding and processing the nucleotide, while different domains are dedicated to binding and processing substrate proteins, also known as “clients”. Conformational signals encoded by the nucleotide underlie communication among domains and the regulation of functionally oriented structural dynamics. In most cases, various cochaperones and effectors contribute additional levels of functional modulation by selecting different pools of chaperone conformations, binding and/or covalently modifySpecial Issue: Allosteric Modulators Received: May 24, 2018 Published: July 26, 2018 60

DOI: 10.1021/acs.jmedchem.8b00825 J. Med. Chem. 2019, 62, 60−87

Journal of Medicinal Chemistry

Perspective

Figure 1. Simplified representation of the client processing by the HSP90 and HSP70 chaperone systems. (A) Reaction scheme of the main steps in the coupled mechanisms of HSP70 and HSP90 with the main cochaperones and nucleotide states involved. (B) Three-dimensional structure of the complex HSP90-Cdc37-Cdk4 complex by Verba et al.176 The chaperone is depicted in blue, the cochaperone Cdc37 in gray. and the client Cdk4 in yellow. (C) Model obtained by manually fitting coordinates to the EM surface of the complex described in ref 20. The model has been kindly provided by David Agard. (D) 3D structure of the complex between HSP70 and DnaJ described by Kityk et al.218 The chaperone is depicted in blue, the cochaperone DnaJ in gray.

ing them.8,9 From the points of view of chemical biology and drug design, the members of the HSP70 and HSP90 families have been the focus of most of the attention (Figure 1). The initial interest in targeting these systems was based on the observation that, in cancer, key client oncoproteins depend on chaperone activity for functional stabilization. Indeed, HSP90 is a central member of the stress response machinery, and cancer cells have shown enhanced dependence on this protein for the efficient maintenance of intracellular proteostasis.10 Inhibition of HSP90 in cancer has proven to effectively exploit this biological vulnerability; by hitting key oncogenic networks at multiple points and blocking diverse client proteins, HSP90 inhibitors have also proved to minimize the development of drug resistance, e.g., through kinase mutation or pathway switching.11 In this context, it has also been observed that HSP70 overexpression12,13 correlated with

unfavorable prognosis and inhibition of cell apoptosis and senescence,14 facilitating the survival of transformed cells in otherwise adverse stress conditions.15 In neurodegenerative diseases,16 the HSP90-HSP70 system regulates the levels and cellular processing of the aberrant forms of the protein tau, whose aggregation has been linked to the formation of cytotoxic species in Alzheimer and other protein misfolding diseases.17 At the molecular level, HSP90 and HSP70 form multiprotein complexes18−20 with cochaperones to assist and facilitate the maturation of their client proteins (Figure 1). In this scenario, assemblies and networks of HSP90 and HSP70 (Figures 2 and 3, respectively) are ideal candidates to develop and test new chemical approaches and ideas aimed to mine structural and functional information on proteins and eventually translate them into novel effector molecules. 61



Perspective

Figure 2. Structures and conformational cycle of HSP90. (A) Three-dimensional structures of the mitochondrial asymmetric (TRAP1, PDB code 4IPE) and cytosolic symmetric (HSP90, PDB code 2CG9) representatives of the HSP90 family of chaperones. The circle indicates the client binding site. (B) Conformational cycle of HSP90, depicted through the 3D crystal structures of the various states solved so far. The representative PDB code is written near the respective structure.

cancer, for instance, it is now well appreciated that many mutated proteins or deregulated metabolic pathways are the leading cause of increasing costs and low success rate (typically 0.0001%) of “targetcentric” approach. Moreover, classical strategies aimed at targeting the active sites of enzymes appear to be approaching their limits. Due to the evolutionary and structural conservation of the latter across the proteome, many instances have shown that the efficacy of orthosteric drugs is hampered by selectivity issues, off-target effects, and development of drug resistance mechanisms.23,24 Targeting HSP90 and HSP70 pathways represents fresh opportunity for testing broad ideas on mechanistic regulation, protein interactions, and their effects on biological functions. HSP90 and HSP70 pathways may in fact be viewed as modular

At a more general level, this process would in principle overcome current limitations in drug development by taking into account the molecular complexities associated with the structural organization of biochemical networks: in postgenomic chemical biology and medicinal chemistry age, what has become clear is that, with the exception of a few monogenic pathologies driven by one single factor, in most cases the problem of novel drug discovery is more complex than the identification of a unique target and the consequent design of a molecule that may individually block its activity. This has often proved insufficient for an effective treatment. It has been proposed that the limitations of the “one target one drug” rule may indeed explain the low number of newly discovered active drugs in recent years.21,22 In the case of 62



Perspective

Figure 3. Structures and conformational cycle of HSP70. (A) Three-dimensional structures of HSP70 in the ATP (PDB code 3C7N) and ADP (PDB code 2KHO) states. (B) Simplified scheme of the conformational changes induced upon nucleotide exchange for HSP70, together with the cochaperones involved in the mechanism. (C) Conformational cycle of HSP70, depicted through the 3D crystal structures of the various states solved so far. The representative PDB code is written near the respective structure.

interaction networks,25 in which the chaperones act as hubs controlling the assembly of multiprotein complexes that eventually direct cells toward their final phenotypes. In this context, from the therapeutic perspective, it has recently been shown for instance that even modest HSP90 inhibition significantly impairs the evolution of resistance mechanisms to antiestrogens in breast cancer models, supporting the perspective of clinical testing of combined hormone antagonist and low-levels of HSP90 inhibitors.22 This approach can also be extended to combinations with kinase inhibitors.26−28

As noted by Gestwicki and co-workers, functional protein complexes in general, and chaperone complexes in particular, are in most cases assembled in a combinatorial manner,29,30 typically involving one enzyme combined with multiple nonenzymes. The composition of the resulting systems ultimately defines the type, location, and duration of cellular activities: depending on the cell or cell-compartment types, different proteins are combined with different mechanisms, favoring adaptation to disparate functional requirements. In addition to these aspects, it has been shown that client binding has a direct influence on the enzymatic activities of 63



Perspective

play a role in the design of such molecules, by shedding light on the detailed molecular mechanisms of HSP90 and HSP70 regulation.

both HSP90 and HSP70: recruitment of the client protein has in fact been shown to stimulate ATPase for HSP90,31−33 while in HSP70 ATP binding to the N-terminal nucleotide-binding domain (NBD) alters substrate affinity to the C-terminal substrate-binding domain (SBD) and substrate binding enhances ATP hydrolysis.34−36 Inhibition of the enzymatic component may result in the complete, indistinct shutdown of the functions of the system, while developing chemicals that mimic and/or enable the controlled modulation of fine-tuned mechanisms in the endogenous cellular environment7,19 may generate new opportunities in therapy and chemical biology. Allosteric modulators have represented an area of intense research in the past 2 decades,37−42 whereby the integration of molecular and physical biology with structure-based advances in the design and synthesis of new ligands has had a strong impact on disparate targets. In this frame of thought, three main realizations have encouraged the search for new allosteric modulators. First, allosteric sites tend to be under lower sequence and structural conservation pressure than active sites, ultimately facilitating the design of target-specific ligands with reduced risks of toxicity or side effects.43,44 Second, the role of protein dynamics has been increasingly recognized as fundamental for biological function: proteins carry out their tasks and participate in biochemical interaction networks by switching among a certain number of structural and functional substates, which favor adaptation to different partners and allow biomolecules to fine-tune their activity in response to varying conditions. Such conformational changes are induced by several biochemical factors, including covalent modification, redox-state changes, and most importantly, ligand binding.43,45−47 Third, allosteric ligands can be exploited to reshape protein−protein interaction surfaces by binding to sites far from the actual interface and thus influence the selection of one binding partner over another at a shared interface by favoring conformational states with specific recognition profiles. Allosteric events are the prime mechanisms by which fine regulation of enzymatic activities and recognition of binding partners in the assembly of large complexes, such as those formed by HSP90 and HSP70, can be achieved. In a simplified view, allostery defines how modifications at one site are propagated through the structure to a distal region.48,49 In this framework, modulation of ligand affinity toward the primary active site can be achieved by binding of an “effector” at an allosteric site.50 The end result is a shift in the structural population, whereby the activation of particular dynamic conformations may tune specific enzymatic activities and molecular recognition events.37,42,51 These physical modifications affect the ways in which chaperones coordinate and organize their networks and consequently their effects on clients and on the cellular pathways the latter belong to. Within this framework, our review will focus on the development of allosteric modulators of HSP90 and HSP70, critically assessing the recent literature in terms of the progress made in delivering novel chemical entities with potential prospects in chemical biology and in applicative therapies. We endeavor to address the questions of how to regulate the fundamental control points at the crossroads of different biological pathways and design/discover chaperone allosteric modulators displaying interesting properties. We will also consider how structure-based computational approaches may

■

HSP90 The members of the 90 kDa heat shock protein (HSP90) family are homodimeric molecular chaperones that orchestrate different cellular pathways involved in cell development and maintenance, by regulating the late stage maturation, activation, and stability of a plethora of client proteins (updated list in http://www.picard.ch/downloads).52 Different cellular compartments host different paralogs, which oversee the assembly of specific networks of interactions. In humans, the main paralogs are HSP90 (with isoforms α and β) in the cytoplasm, Grp94 in the endoplasmic reticulum (ER), and TRAP1 in the mitochondria.8 Recently, important roles have also been invoked for extracellular HSP90.53,54 Microbial HSP90 has been linked to pathogenic infections in malaria and tuberculosis55,56(Figure 2). Structurally, the protein exists as a homodimer and each individual chain consists of three globular domains, namely, the N-terminal domain (NTD), the middle domain (MD), which is subdivided in large middle (LMD) and small middle (SMD) domains, and the C-terminal domain (CTD). HSP90, TRAP1, and Grp94 have a mutual sequence identity of about 30−40%, which reflects the high structural similarity of their individual domains.57−59 However, the preferential relative orientation of the domains in the crystal structures solved so far significantly varies depending on the isoform, cellular compartment, and organism.60 X-ray crystallography, smallangle X-ray scattering (SAXS) solution data, and kinetic measurements have led to the proposal of a general functional mechanism based on global conformational modulations triggered by ATP binding and hydrolysis, which integrates an array of structural information.59 In the absence of nucleotide, various, mostly open, conformations coexist. ATP binding shifts the chaperone to a partially closed and then into an asymmetric closed conformation (observed for TRAP1) that is significantly strained, leading to buckling of the MD:CTD interface (Figure 2). Interestingly, this region has a key role in client binding.31 Upon ATP hydrolysis, strain is relieved to yield a symmetric closed state reminiscent of the fully symmetric yeast HSP90 dimer.57 This model establishes a direct conformational communication between the ATP binding site and the client-remodeling site. In vitro experiments demonstrated that although the fundamental conformational states are well conserved among species (and/or paralogs), equilibria and kinetics are unique for every HSP90 homolog,60 suggesting adaptation to the specific pool of clients in a given cellular environment. Different HSP90 paralogs are associated with different molecular partners. The majority of clients identified for the cytosolic HSP90 are related to signal transduction, cell maintenance and growth, including many steroid hormone receptors and kinases.8 Grp94 clients include IgGs, some integrins, and the Toll-like receptor family. The validated list of TRAP1 clients is rather small and mostly related to mitochondrial protein homeostasis and respiration, involving mitochondrial ERK kinases and succinate dehydrogenase (SDH).61,62 Finally, interactions with cochaperones also appear to be paralog specific: while cytosolic HSP90 function is aided by numerous cochaperones, which organize complexes for client activation (e.g., Cdc37), enhance (e.g., in the case of 64



Perspective

Figure 4. Ligands that establish contacts with residues different from the ATPase catalytic residues: Grp94-selective purine based compounds and gambogic acid. The cartoon identifies their targeted domain.

ultimately exposing the patient to toxicity risks.70 Additional factors for the limited success of ATP-competitive inhibitors in clinic have recently been discussed by Neckers et al.28 One critical reason is linked to the pharmacodynamics markers of target engagement and to the cell models used to identify such markers, which are commonly represented by peripheral blood mononuclear cells (PBMCs) obtainable from patients before and after treatment. It has however been shown that the pharmacokinetics of HSP90 inhibitors differs significantly between tumor and normal tissues, that some tumor cells are more sensitive to N-terminal HSP90 inhibitors than normal ones,71 and that the HSP90 complex in cancer cells is biochemically distinct from that of normal cells. All these factors may affect the affinity for inhibitors, their distribution, and pharmacodynamics properties.19,72 Together with these limiting factors, the roles of HSP90 in mediating nuclear events beyond the stabilization of oncogenic kinases may also contribute to the low success rate of clinical trials.73 Finally, most preclinical models rely on immune compromised mice: thus, the impact of potential drugs on host immunity has not been fully characterized so far.74 Neckers and co-workers also provide a critical evaluation of prospects for the clinical evaluation of N-terminal directed compounds. We refer the reader to the original paper for more insights into these issues.28 One more critical factor that deserves mention is that, given the similarity of the catalytic domain in all HSP90 family members, most of these compounds lack paralog specificity and impact the whole HSP90 functional spectrum by indiscriminately blocking all HSP90-dependent processes. However, differences exist among paralogs in terms of expression levels, protein clientele, interactions with cochaperones, and consequently, molecular mechanisms at the basis of

Aha1) or slow (e.g., Sti1) the rate of ATPase activity, no cochaperones have been detected for the ER and mitochondrial HSP90s (see8). Pharmacological inhibition of the enzymatic activity has been pursued for HSP90 as a prominent strategy for drug development mainly focused on targeting the N-terminal ATPbinding site.10,63−66 The first candidates based on the natural product geldanamycin (i.e., 17-AAG, 17-DMAG, IPI504) showed interesting activities in clinical trials, but their development was discontinued due to high toxicity. Overall, about 20 ATP-competitive chemotypes have been tested or have undergone clinical trials against different types of cancers, such as HER2-positive breast cancer, myeloma, acute myelogenous leukemia (AML), prostate cancer, melanoma, ovarian cancer, non-small-cell lung cancer (NSCLC), gastrointestinal stromal tumors (GIST), rectal cancer, and melanoma.67 The compounds include (i) the purine scaffold series (BBIIB021, BIIB028, and PU-H71); (ii) the resorcylic pyrazoles and isoxazoles (VER52296); (iii) the resorcylic dihydroxybenzamides (AT13387 and KW-2478); and (iv) the 8-arylthiopurine CUDC-305, which is orally bioavailable and active in an orthotopic brain tumor model.64 The main limitation of HSP90 N-terminal inhibitors is that the drug concentration required to outcompete the abundant ATP and to induce client degradation is the same as that needed to induce the heat shock response, which results in the activation of the prosurvival factor heat shock factor 1 (HSF1) and induction of HSP27 and HSP70. HSP90 indeed sequesters HSF1 in unstressed cells, and disruption of this interaction is thought to determine release of HSF1 and initiation of the transcriptional response that leads to the heat shock response.68,69 Such drug-induced effect creates a vicious cycle in which higher doses at an increased frequency are required, 65



Perspective

Figure 5. Chemical structures of the C-terminal inhibitors novobiocin and chlorbiocin.

membrane. In contrast, no major requirement for Grp94 was observed in cells with low HER2 plasma expression, where inhibition of HSP90 was sufficient to destabilize HER2. Overall, these results indicate that chaperone conformational dynamics may be aptly exploited to develop chemical tools that can be used in different cellular contexts, allowing precision interventions that take into account protein expression levels. This series of compounds was further developed by extensive SAR investigations leading to the design of 18c, a derivative with good potency for Grp94 (IC50 = 0.22 μM) and selectivity over other paralogs (>100- and 33-fold for HSP90α/β and TRAP1, respectively). The activity of 18c was explored in different cellular models of inflammation and cancer80 (Figure 4). An example of ligand that targets the NTD at a different location than the ATP-binding site is represented by gambogic acid.81 The molecule was identified via HTS screening of natural product libraries and shown to inhibit cell proliferation, determine degradation of stringent client proteins, disrupt the interaction of HSP90, HSP70, and Cdc37, and induce expression of HSP70, a typical indicator of HSP90 N-terminal inhibition. Binding to this domain with low micromolar Kd was confirmed by surface plasmon resonance, competition experiments with geldanamycin, and molecular docking experiments (Figure 4).

their functions. TRAP1, for instance, is less expressed in some cancers than in normal tissues.75 Indeed, recent data by the Neckers group indicate that TRAP1 expression is inversely correlated with tumor grade in several cancers, suggesting that, in some settings, this chaperone may even act as a tumor suppressor.76−78 In this context, the targeting of pockets, different from the ATP one, shows a first example of how to interfere with HSP90 mechanisms in an isoform selective manner exploiting the dynamics of ligand−protein cross-talk. Chiosis and co-workers79,80 combined library screening and structural and computational analyses to unveil paralog-specific chemical tools able to differentially modulate Grp94 versus cytosolic HSP90. Screening identified purine based compounds (Figure 4) with a preference for Grp94 greater than 100-fold over HSP90 and from 10- to 100-fold over TRAP1. Structural analysis indicated that specific rearrangements in and around the Grp94 NTD binding site open specific nonpolar pockets that can efficaciously be engaged by aromatic modifications on the purine scaffold. This work shows that the overall protein structure and conformational flexibility play important roles in configuring the binding sites of different members of this chaperone family and that the consideration of ligand−protein cross-talk may provide novel opportunities for the development of selective compounds. The availability of paralog-specific chemical tools permitted characterization of different chaperoning mechanisms for the client protein HER2 used by Grp94 or cytosolic HSP90. In particular, results suggested that Grp94 is specifically expressed on malignant HER2+ breast cells surface and may translocate to other membrane locations, such as those needed by the altered function of its oncogenic client protein, HER2. In this case, inhibition of Grp94 is sufficient to destabilize membrane HER2 and inhibit its signaling properties, implicating Grp94 in regulating oncogenic signal transduction at the plasma

■

ALLOSTERIC INHIBITORS OF HSP90 Fundamental for the development of allosteric modulators was the discovery of a second druggable site in the CTD of HSP90.82 Early work by the Neckers group started from the investigation of the interactions between HSP90 and coumarin antibiotics known to affect the activity of the homologous bacterial protein DNA gyrase.82,83 In particular, they observed 66



Perspective

Figure 6. Chemical structures of the C-terminal inhibitors developed by the Blagg group from novobiocin. The circles indicate pharmacophoric points that are important for activity.

expectedly have distinct impacts on the population of the proteins conformational states involved in the process of HSF1 recognition. HSP90 binding to HSF1 is favored for the ATPdependent, NTD dimerized chaperone, a conformation only rarely sampled by mammalian HSP90. Kijima et al.84 have used mutations that lock HSP90 in this closed state, proving that this is indeed the one relevant for HSF1 binding. The authors also succeeded in mapping out HSP90 binding site on HSF1, showing that the binding motif comprises the heptad repeat (HR)-A/B trimerization domain adjacent to the regulatory domain, the molecular region dedicated to sensing heat and initiating the HSR. ATP-competitive inhibitors are opposed to dimer closure,85 thus disfavoring the conformation that is required for stable HSF1 binding. In contrast, allosteric ligands targeting regions distal from the NTD (derived from novobiocin or different chemotypes; vide infra) can expectedly induce a different conformational landscape for the chaperone, in which structural ensembles suitable for HSF1 binding are significantly populated. Indeed, KU32, a novobiocin mimic developed by the Blagg group (vide infra), was shown not to disrupt the interaction between HSF1 and closed HSP90.84 HSP90 dynamics play a key role in modulating the HSR, and chemical tools that perturb such dynamics provide a chance to investigate their roles in signaling pathways. While no cocrystal structure for novobiocin or its derivatives bound to the C-terminal of the chaperone has been solved yet, intense design, synthesis, and SAR studies have been carried out to shed light on the main determinants of activity and to develop novel drug candidates with applications in cancer and neurodegeneration. Important steps in this direction were taken by the Blagg group,86,87 who first developed a focused library of compounds to explore the functional relevance of the benzamide side chain, coumarin core, and noviose sugar (Figure 6). The library was tested for efficacy on SK-BR-3 cancer cell lines at

that the coumarin antibiotics novobiocin and chlorbiocin (Figure 5) lead to a marked reduction in the cellular levels of HSP90 client proteins such as the protein tyrosine kinases (PTKs) p185(HER2) and p60(v-Src), the protein serine/ threonine kinase (PSTK) Raf-1, and the mutated p53 protein, using SK-BR-3 and MCF-7 human breast carcinoma cells and v-Src transformed mouse fibroblasts. Novobiocin showed an IC50 of around 700 μM. Immobilized novobiocin was used to isolate either full length endogenous HSP90 from cell lysates or HSP90 deletion fragments translated in vitro. The results showed that HSP90 binding to immobilized novobiocin was competed by soluble coumarins and ATP but not by Nterminally directed ligands like geldanamycin or radicicol. Importantly, the authors showed that a C-terminal HSP90 fragment could bind immobilized novobiocin but not immobilized geldanamycin. In contrast, a geldanamycinbinding N-terminal region did not bind novobiocin. These results showed for the first time the existence of an additional, alternative binding site on HSP90 that could be targeted by chemotypes different from ATP mimics or geldanamycin- and radicicol-based molecules. Importantly, this work showed that binding to a site distal from the nucleotide orthosteric pocket resulted in the modulation of the functional properties of the HSP90, leading to impairment of its chaperone functions and opening a whole new range of possibilities for pharmacological interference with HSP90 functions. Importantly, novobiocin and related natural products were not shown to induce the heat shock response (HSR) that instead represents one of the main drawbacks in the clinical development of HSP90 N-terminal inhibitors. This important property is shared by many of the allosteric ligands inspired by the discovery of novobiocin activity. The reasons for this differential behavior can aptly be explained hypothesizing that HSP90 binding to HSF1 is correlated to the chaperone conformation. Ligands targeting different HSP90 domains can 67



Perspective

Figure 7. (A) Coumarin structure−activity relationships (SARs) and structure of KU-135. (B) Biphenyl derivatives and modifications of the biphenyl scaffold in the study of SARs discussed in ref 96. (C) Biphenyl compounds with noviose substitutions.

100 μM concentration, followed by Western blot analysis of cell lysates to check for the degradation of HSP90-dependent client proteins. Compound A4 turned out to be the most active

one in this assay and was subsequently tested in a mutated androgen receptor-dependent prostate cancer cell line (LNCaP) and a wild-type androgen receptor prostate cancer 68



Perspective

levels constant without causing HSP70 overexpression, the latter being a hallmark of cytotoxic HSP90 C-terminal inhibitors. A different set of molecules from this series (both coumarinand biphenyl-based derivatives) was shown to modulate HSP90 functional responses in a different manner. Some of these compounds (e.g., KU-32, KU-596, Figure 6) raised interest for their cytoprotective effects,89,97 since they could induce the prosurvival HSR yet at significantly lower concentrations than those needed for client protein degradation.89,90,97−99 Such properties made them interesting as agents able to protect neural cells from degeneration. Indeed, KU-32 and KU-596 (Figure 6) turned out to be efficacious in cellular models of Alzheimer disease: their modulation of chaperone functions resulted in an induction of HSP70, which could refold denatured proteins, decrease the levels of abnormal proteins, and prevent protein aggregation, ultimately alleviating the phenotypes associated with Alzheimer and other neurodegenerative diseases. Further studies showed the possibility to reverse clinical symptoms of biabetic peripheral neuropathy in vivo and protect against neuronal glucotoxicity.100,101 KU-32 specifically protected against glucose-induced death of embryonic DRG (dorsal root ganglia) neurons cultured for 3 days in vitro. Similarly, KU-32 significantly decreased neuregulin-1-induced degeneration of myelinated Schwann cell DRG neuron cocultures prepared from WT (wild-type) mice.100,101 KU-596 has recently been examined for its potential in decreasing c-jun expression and improving motor function in an inducible transgenic model of a Schwann cell-specific demyelinating neuropathy (MPZ-Raf mice).102 KU-596 therapy reduced c-jun expression, significantly improved motor performance, and ameliorated the extent of peripheral nerve demyelination in both prevention and intervention studies. These effects were reconnected to KU-596 controlled induction of HSP70 expression: indeed the drug’s neuroprotective efficacy was absent in HSP70 knockout mice. These data are very interesting as they show that modulation of chemical pathways, namely, HSP70 expression levels, via designed chemicals provides the basis for novel therapies to improve conditions in demyelinating neuropathies in humans. The accessibility and usability of this class of HSP90 Cterminal allosteric modulators were further expanded by substituting the synthetically complex noviose sugar with simple cyclohexyl moieties while maintaining biological efficacy90 (Figure 7C). Some of the derivatives once more proved to induce HSP70 levels. Activities were further probed by the c-jun assay and by measuring mitochondrial bioenergetics. Reduction of c-jun and increased mitochondrial respiratory capacity (MRC) were observed for selected noviomimetics, lending support to the notion that designed modulators of HSP90 can also impact mechanistic features that are associated with neuroprotective efficacy. Finally, it is worth mentioning that other natural product inspired ligands have been used as a source of inspiration for the development of mimics able to target HSP90 CTD and allosterically inhibit it. Among these, the most relevant ones are epigallocatechin 3-gallate (EGCG) and coumermycin A1. EGCG103 was demonstrated to bind the CTD of HSP90 by affinity chromatography. The molecule was successively developed into a series of derivatives with nanomolar IC50’s against SK-BR-3 and MCF-7 breast cancer cell lines as low as 4 μM.104 Coumermycin A1 was identified in the same context of

cell line (LAPC-4) and showed a relevant effect on the concentrations of the mutant androgen receptor, AKT, and HIF-1 at ∼1 μM in the LNCaP cell line. The compound reduced levels of the androgen receptor at lower concentrations in the LAPC-4 cell line.86,87 The results highlighted that attachment of noviose to the 7position and an amide linker at the 3-position of the coumarin ring are critical for HSP90 inhibitory activity (Figure 6, compounds DHN1, DHN2). In SK-BR-3 breast cancer cells DHN1 and DHN2 induced degradation of HSP90-dependent HER2 and p53 between 5 and 10 μM and between 0.1 and 1.0 μM, respectively. It was also found that activity increased in the presence of a diol on the sugar ring. Interestingly, in contrast to N-terminal inhibitors, such derivatives of novobiocin perturbed HSP90 at concentrations 1000- to 10000-fold lower than that required for client protein degradation, a phenomenon that showed how allosteric compounds could be active with lower toxicity: for these reasons, these compounds were tested as neuroprotective agents, with promising efficacy in a culture model for Alzheimer disease (EC50 of 6 nM).88−90 These results supported the investigation of different modifications on the scaffold based on the identified pharmacophores. In this context, the sugar rings were substituted by more synthetically accessible cyclic or acyclic moieties carrying hydrogen-bonding functions, identifying the piperidine ring as the best substitution for the carbohydrate. Moreover, the chain on the amide side was extended with biphenyl moieties (Figure 7A).91−94 A relevant example of these compounds, KU135, inhibited Jurkat T-lymphoblastoid leukemia cells proliferation with an IC50 of 416 nM after 48 h of treatment.93 In this test, KU-135 turned out to be ∼3, ∼10, and ∼600 times more potent than etoposide, 17-AAG, and novobiocin, respectively. Furthermore, KU135 proved to inhibit cell division and induce apoptosis in the leukemic cells, together with the degradation of HSP90 client proteins AKT and HIF1 (α subunit) without inducing the overexpression of HSP70, indicative of an adverse HSR. Surface plasmon resonance (SPR) experiments showed that KU135 could bind the isoform β of HSP90 Kd of 1−2 μM. A further interesting evolution of this class of allosteric modulators/inhibitors of HSP90 was represented by the introduction of a biphenyl scaffold in lieu of the coumarin ring system found in novobiocin. Initial structure−activity studies showed that the distance between the nitrogen atoms on the piperidine ring and the amide was important for HSP90 C-terminal inhibition.95 Computational docking studies identified the biphenyl scaffold in Figure 6 containing the para−meta substitution pattern as the optimal core surrogate to mimic the stereochemical relationships among the pharmacophores presented by the coumarin rings. Initial tests in SK-BR-3 and MCL-7 breast cancer cells yielded low micromolar IC50 values (between ∼0.5 and ∼5 μM) and degradation of oncogenic HSP90 clients HER2, Raf-1, and AKT.95 Extensive SAR analyses were then performed on these compounds, exploring modifications of the core-biphenyl scaffold, the amide chains, and different types of amine substituents on the other side of the molecule (Figure 7B). Many of these derivatives manifested interesting antiproliferative activities, reaching nanomolar IC50’s against SK-BR-3 and MCF-7 breast cancer cell lines,96 the best compounds being as potent as ∼150 nm. The compounds led to degradation of several client proteins (HER2, ERa, and AKT) and left HSP90 69



Perspective

Figure 8. Chemical structures of the deguelin analog L80, of the C-terminal-targeted compound C9, and of the cyclopeptide SanA. Their binding on HSP90 domains is schematically presented.

novobiocin82,83 and subsequently used for the development of novel derivatives and SAR studies.105 The best derivative showed an IC50 of about 2 μM in SK-BR-3 and MCF-7 breast cancer cell lines. One important aspect of the results described above is that target engagement was demonstrated in all cases by affinity purification with isolated HSP90 domains, proteolytic fingerprinting, SPR, mutation, or competition assays. The importance of target engagement issues in HSP90 drug development is thoroughly discussed by Neckers and coworkers.28 An additional example of a small molecule modulator targeting the HSP90 CTD is represented by a natural rotenoid called deguelin and related compounds.106 In this context, Lee et al.106 introduced compound L80 (Figure 8), a derivative that, according to biochemical and biophysical analyses, was shown to target the CTD of HSP90. L80 not only decreased the expression of various HSP90 client proteins and disrupted the interaction between the α subunit of HIF-1 and HSP90 but also led to downregulation of the former and its target genes, including vascular endothelial growth factor (VEGF) and insulin-like growth factor 2 (IGF2). This translated into inhibition of viability, colony formation, angiogenesis-stimulating activity, migration, and invasion in a panel of NSCLC cell lines and their paclitaxel-resistant sublines.

Regan and co-workers reported on the discovery of modulators of the chaperone’s PPIs by targeting the interaction of HSP90 with the cochaperone named HSP organizing protein (HOP).107−111 HOP acts as a linker protein that holds together HSP90 and HSP70 by forming complexes with its CTD and modulates the activities of these chaperones. The authors used the AlphaScreen112 high throughput technology to identify compounds that inhibited this interaction. The discovered compounds showed activities in vivo against cancer cell lines BT-474 and SK-BR-3, inducing decreased levels of client protein HER2, with associated cell death.107 One compound was thoroughly characterized, namely, 1,6-dimethyl-3-propylpyrimido[5,4-e][1,2,4]triazine5,7-dione (C9, Figure 8).110 In vitro experiments using purified protein components were used to prove that C9 antagonizes HSP90/HOP interaction. In vivo, C9 was cytotoxic also against the highly metastatic MDA-MB-231 breast cancer cell line and did not induce the transcriptional upregulation of HSP70, a hallmark of HSP90 C-terminal inhibitors’ activity. Interestingly, the authors treated cells with a combination of C9 and either 17-AAG or NVP-AUY922, two active N-terminal directed ATP competitive inhibitors: their results indicated that the overexpression of HSP70, typically induced by the NTD ligands, could be significantly counteracted by C9. Moreover, in this setting, the lethal dose of C9 was also decreased compared to the use of C9 alone. 70



Perspective

Figure 9. (a) Chemical structure of the activator goniothalamin. (b) Chemical structures of the artificial accelerators discovered via FRET screening by the Buchner group. (c) Rationally designed stimulators of the ATPase activity of HSP90, together with the structure of the complexes on which the design was successfully based. The black square indicates the allosteric binding site for the ligand, partially overlapping with the clientbinding region.

■

These data support the concept that combinations of ATPcompetitive inhibitors and allosteric modulators may synergize to induce cancer cell death via a combination of different molecular mechanisms while potentially minimizing undesired effects. The McAlpine group characterized allosteric modulators derived from the natural compound sansalvamide A, which were shown to bind at the interface between NTD and MD of HSP90 and to interfere with the recruitment of HOP, inositol hexakisphosphate kinase 2 (IP6K2), and FK506-binding protein 52 (FKBP52) at the CTD.113−115 This group also introduced peptide-based macrocycles that, targeting the CTD, blocked the recruitment of cochaperones and clients without inducing the HSR116−118 (Figure 8). In the case of large peptides and macrocycles, as well as for natural products and their mimics, considering (i) their potential points of metabolic liability, (ii) the large molecular weight, and (iii) aggregation tendencies, caution should be given to the issue of ruling out off-target effects, problems with adsorption and distribution, and PAINS (pan assay interference compounds).119 Overall, the synthetic development of C-terminal allosteric modulators, started from the initial discovery of a secondary binding site in HSP90 able to host the natural compound novobiocin,82,83 shows that different cellular functions can be modulated by modifying selected scaffolds. It is therefore possible to develop chemical tools to direct cells toward different fates, regulating the types and levels of HSR and acting on the folding/unfolding of client proteins associated with different biochemical pathways.

ALLOSTERIC ACTIVATORS OF HSP90

Allosteric modulators offer a chance to garner new mechanistic information on biochemical processes by enhancing protein functions at specific levels along their pathways, as opposed to completely abrogating their activity.120 This interesting opportunity applies to HSP90 as allosteric stimulators of its catalytic activity have recently appeared. Molecules with these properties represent new chemical tools useful to gain insights into the role of HSP90 conformational dynamics in determining its function as well as for the design of novel therapeutic strategies.121−124 In this context, Yokoyama and co-workers reported on the stimulatory effects of the natural lactone goniothalamin on the ATPase activity of the bacterial homolog of HSP90, HtpG. Goniothalamin induced a 2-fold increase in both kcat and kM. Competition assays with in vitro reconstituted truncated versions of the chaperone identified the NTD of HtpG as the domain hosting the goniothalamin binding site. Importantly, the stimulator had a negative impact on the client refolding activity of HtpG in cooperation with HSP70121 (Figure 9a). The Buchner group used a FRET-based (fluorescence resonance energy transfer) screening, reporting on the conformational rearrangements of HSP90, to identify allosteric activators of the chaperone closure kinetics.122 The authors reasoned that as the conformational changes are the ratelimiting steps of HSP90 ATPase,85,125 molecules that influence such kinetics should aptly influence the ATPase activity. Upon screening a library of 10 000 compounds, the authors identified 5 compounds accelerating the ATPase activity 2- to 5-fold (Figure 9b). The authors further mapped the binding site of a 71



Perspective

exploited to guide the design of a second generation of derivatives with improved ability to stimulate ATP hydrolysis. Therefore, we could deliver a series of ligands that permitted enhancement of ATPase turnover by up to 7-fold.123,124,130,131 Furthermore, decorating the best activator with a mitochondrial penetrating group determined cytotoxicity versus cancer cells in the order of 500 nM.123,124,130 The combination of experimental and MD-based investigations of binding and activation trends for different derivatives also facilitated the development of a quantitative SAR model that correlates HSP90 activation to the presence of a certain allosteric activator. Such models integrate the information on the dynamic cross-talk that allows protein conformations to adapt to the presence of the ligand, selecting substates that favor the activation process. We characterized the impact of compound-induced activation on HSP90 interactions in vitro and on the stability of a number of clients in cellular models: consistent with observations from other groups, administration of accelerators induced a marked decrease in levels and stability of stringent client proteins, supporting the concept that acceleration of conformational dynamics may in fact represent a new way of perturbing the chaperoning mechanisms that underlie cell viability: indeed, some of our compounds inhibit the proliferative potential of tumor cells including those resistant to HSP90 ATP-competitive inhibitors.123,124 Finally, it is worth noting that the detrimental effect of conformational closure and ATPase stimulation on cell viability was also observed through careful mutational and biophysical analysis.128,132 Very recently, Prodromou and co-workers133 showed that dihydropyridine derivatives could target the MD/CTD interface of HSP90 and induce activation of its ATPase activity. Interestingly, binding to this region of HSP90 was validated through mutagenesis. Indeed, dihydropyridines were shown to upregulate the HSR, inducing a neuroprotective effect in an APPxPS1 mouse model of Alzheimer disease. Importantly, such HSR could be observed only in diseased cells. On the basis of their data and starting from the observation that dihydropyridines could function as coinducers by upregulating chaperones and cochaperones only under pathological conditions, the authors suggested that allosteric ligand binding compromises HSP90 turnover rate, which consequently induces the HSR in diseased cells. Interestingly, this may represent the mechanism whereby reduced amyloid plaque and increased dendritic spine density in the APPxPS1 Alzheimer mouse model could be initiated, indicating the compounds as therapeutic leads for neurodegenerative diseases. At this point, it is tempting to speculate on the molecular reasons why both inhibition and acceleration of the HSP90 cycle would lead to client degradation and impact on cell viability. HSP90 functions depend on a delicate balance between ATP hydrolysis, large conformational changes, and assembly of multiprotein complexes that include clients, stimulator, and/or inhibitor cochaperones at different points in time. In such a highly regulated system, it is crucial for HSP90 to reach, populate, and spend the correct amount of time in the conformations required to correctly assemble the functional complexes and as a consequence process client proteins. Artificial accelerators would force the system to spend less time than that required in such conformations. Inhibitors, on the other hand, could potentially stall the cycle at specific

selected activator using NMR spectroscopy, identifying the NTD region close to helix α2 as the pocket engaged by the ligand. Targeting this region induces structural changes in the active site that reverberate on nucleotide hydrolysis rates. Next, the authors used a yeast model to test the effect of the stimulators on the folding of the stringent client glucocorticoid receptor (GR) in vivo. Administration of accelerators reduced GR activity while increasing HSP90 affinity for the client. Collectively, these results indicate that the presence of the stimulators negatively affects client processing in vivo. Our group reported for the first time on the rational design of activators by combining molecular simulations and theoretical characterization of allosteric communication mechanisms in HSP90.123,124 First, the allosteric site was identified through the analysis of the patterns of long-range coordination among residue-pairs: the sets of amino acids endowed with potential allosteric control properties were designated as the ones belonging to the M- and C-terminal domains and characterized by minimal values of the coordination propensity (CP) parameter. This parameter evaluates the coordination between any two residues (separated by a defined cutoff distance) as a function of the mean square fluctuation of their distance. Small CP values define (groups of) mechanically connected amino acids that move in a cooperative manner giving the highest contribution in modulating functional motions. The CP profile may change in the presence/absence of a ligand allowing connection between the internal dynamic and the observed ligand effects.47,126−128 With this approach, we identified residues at the CTD interface with the SMD that respond to the specific bound nucleotide and define a pocket with stereochemical properties apt to bind drug-like molecules. Interestingly, later independent experimental results reported consistency between the predicted allosteric site and the client binding regions31,129 (Figure 9c). Possible complementary interactions with the putative site were thus mapped with probe molecules, identifying conserved regions in ensembles of HSP90 conformations where the same interactions were consistently made with the majority of structures. Translating this information into pharmacophore models to screen drug databases allowed the identification of several HSP90 allosteric binders with interesting anticancer properties. It is worth noting that the use of pharmacophores derived from molecular dynamics and the avoidance of constraints on chemical scaffolds permitted expansion of the chemical diversity of the predicted ligands.126 Compound 1 (Figure 9c)126 turned out to be the molecule with the most promising anticancer properties, also showing the possibility for evolution to a number of different derivatives. Using structural and dynamic information obtained from various sets of molecular dynamics (MD) simulations of HSP90 in different conditions, we were successively able to evolve 1 into new chemical entities that enable controlled activation of HSP90 ATPase function. Designed derivatives showed (i) a structure-dependent ability to stimulate HSP90 ATPase; (ii) the capability of altering conformational dynamics favoring synergistic effects with the activating cochaperone Aha1; (iii) modulation of HSP90 direct interactions with the cochaperone Sba1 (also named p23); (iv) competition with the model client protein Δ131ΔM31. The putative binding site was also validated through mutational analysis.123,124 The analyses of protein responses to first-generation activators in computational models and in different sets of experiments were 72



Perspective

offering not only an alternative approach to antagonize HSP90 but also providing the possibility to pursue combination strategies with HSP90 inhibitors, eventually maximizing therapeutic efficacy. In neurodegeneration, HSP70 has been linked to several diseases such as Alzheimer disease, Pick’s disease, progressive supranuclear palsy, corticobasal degeneration, and others. All these pathologies share the common trait of an aberrant accumulation of hyperphosphorylated tau, called neurofibrillary tangles.144 The main accepted strategy against neurodegeneration focuses on eliminating toxic oligomers or preventing the misfolding/aggregation of the relevant proteins. In some cases, inhibition of HSP70 has been shown to correlate with a rapid increase of tau ubiquitination and proteasome-dependent tau degradation. However, HSP70 also participates in the clearance of tau misfolded aggregates145 via mechanisms that have not fully been understood yet. In this framework, it may be hypothesized that these disorders may also be approached with drugs that promote HSP70-dependent degradation. Drugs that modulate HSP70-dependent degradation may be combined with HSP90-targeted modulators to maximize effects on the elimination of aggregated client proteins. Interestingly, while HSP70 is known to interact with HSF1,146−148 its inhibition does not result in the induction of the HSR70,84,143,149,150 while leading to the degradation of many clients that are also common to HSP90. This may be ascribed to different possible reasons. HSP70 binds to other regions of HSF1 than HSP90,84 and small molecules may be unable to outcompete such interactions. Alternatively, small molecules may target a specific subset of HSP70 isoforms or populations that are sufficient to block chaperoning of client proteins but not repressor activity on HSF1: In the titration model of HSR,149,151,152 a subset of HSP70 in the cell binds HSF1, repressing its transcriptional activity. Upon different stresses, client proteins can start to unfold, causing an accumulation of misfolded isoforms. Such abnormal species may compete HSP70 away from HSF1, eventually activating transcriptional activity. In this context, it is possible to speculate that small molecules target the subpopulation/ isoform of HSP70 relevant to disrupt the formation of the chaperone complex with the client but not the HSP70 subpopulation/isoform that binds and represses HSF1. Finally, it should also be considered that HSP70 perturbation induces apoptosis in transformed cells:70,143,153 directing transformed cells toward this state may precede the observation of the adverse heat shock response displayed by HSP90 N-terminal targeted, ATP-competitive inhibitors. These facts vividly illustrate the complexity of mechanisms presiding HSP70 biological roles, a complexity that can be effectively disentangled via chemical tools that modulate chaperone activities at different levels.

conformations. Overall, the perturbation of HSP90 conformational landscape and timing of conformational transitions by both accelerators and inhibitors could potentially reverberate in the nonoptimal processing of client proteins, with effects on relative downstream pathways and ultimately on cell viability. In general, it appears that perturbing the system by artificial acceleration can provide novel modes to investigate the relationships between ATP processing and in vivo activities of the chaperone in the cell, in the absence of genetically induced perturbations to HSP90 levels. The available data underscore the fact that ATPase does not fully correlate with chaperoning activity. Rather, an ordered and coordinated sequence of conformational events is required for correct HSP90 function and client processing.123,132 The idea of perturbing HSP90 functions through accelerators may lead to the identification of new chemical tools that complement biochemical and molecular biology approaches in the investigation of the complex and multifaceted mechanisms of HSP90 at different stages of the chaperone cycle.134,135 Furthermore, observed cytotoxic activities and, in some cases, the lack of HSR induction indicate new potential therapeutic opportunities.

■

HSP70 The HSP70 family of molecular chaperones consists of ubiquitous and highly conserved proteins. The most abundant members are cytosolic constitutive (HSC70) and stressinducible (HSP70) (encoded by HSPA1A and HSPA1B genes)136 forms; ER and mitochondrial paralogs (HSPA5 also known as BiP or GRP78 in the ER; HSPA9, also known as mortalin, GRP75, or mtHSP70, in mitochondria) have also been characterized.136 HSP70 proteins share a highly conserved bipartite structure composed by two domains, connected by a flexible linker: a ∼44 kDa N-terminal nucleotide binding domain (NBD) with ATPase activity, and a ∼25 kDa substrate binding domain (SBD).35,36,137,138 The SBD and the NBD communicate via an allosteric cross-talk that, in turn, controls the conformational cycle: the ATPase driven interconversion between open (low substrate affinity) and closed (high substrate affinity) states regulates the folding machinery, which leads to binding or release of SBD substrates, typically linear hydrophobic polypeptides or partially folded proteins. The binding (or release) of these substrates can, in turn, enhance (or decrease) NBD ATPase activity. One additional level of regulation is provided by a number of cochaperones that modify HSP70 functions not only by modulating ATPase rates and substrate specificity (HSP40 or DnaJ),139 but also by exerting control on nucleotide exchange (BAG and HSp110, also known as nucleotide exchange factors (NEFs))140 and client substrates for ubiquitination. Indeed, cochaperones such as the J proteins and NEFs coordinate ATP cycling and client loading onto HSP70141 (Figures 1 and 3). HSP70 participates in mediating correct folding as well as buffering the toxicity of denatured/misfolded proteins, and its dysregulation has been linked to several diseases, ranging from cancer to neurodegenerative disorders. HSP70 forms complexes with the HSP90 chaperone machinery through the action of the cochaperone HOP to buffer the destabilizing effect of oncogenic mutations on the structural stability of client proteins, mainly kinases.142 In this context, HSP70 inhibition alone has proven to effectively disrupt the HSP90-centered folding machinery.143 This evidence supports HSP70 as an effective cancer drug target,

■

ALLOSTERIC MODULATORS OF HSP70 The allosteric mechanism of HSP70 interdomain communication and the interdependence between substrate affinity and nucleotide processing36,154 provide interesting opportunities for chemical modulation. In the context of HSP70 targeting, the classical drug-discovery approach focused on ATPcompeting ligands has met several disappointments. HSP70 functional promiscuity and the nature of the nucleotide binding site, highly conserved among HSP70 members and similar to that of other ATP-regulated proteins, pose selectivity 73



Perspective

Figure 10. Chemical structures of HSP70 activators myricetin and 115-7c and of the inhibitor 116-9e developed by the Gestwicki group, together with their binding sites on HSP70.

issues.150 Some compounds were in fact developed to target the ATP binding site of the protein but showed low potencies and pleiotropic effects on cells,143,155 which translated into toxic effects. To proceed toward the identification of novel modulators of HSP70, Gestwicki and collaborators utilized a “gray box” approach: their method consists of screening plant-derived extracts against an in vitro reconstituted mixture of bacterial HSP70 (namely, DnaK) and its partner HSP40 (namely, DnaJ). In contrast to cell-based phenotypic screenings, which may reveal effects due to both on-target and off-target effects, perturbing the components of a multiprotein complex in vitro in a purified and reconstituted form156,157 enables identification of the actual target in a framework that approximates native, physiological conditions. In general, one component may have measurable enzymatic activity (e.g., ATPase), while the other proteins in the complex may play ancillary roles that impact reaction kinetics, turnover rates, or substrate affinities. In their “gray box” experiment, Chang and collaborators screened ligands in the presence of high concentrations of ATP ([ATP] ≫ KM) and against the combined DnaK-DnaJ complex.156,157 The first choice was made to favor the discovery of allosteric rather than competitive ligands, while the second, linking measured effects to the presence of a stimulatory cochaperone, was selected to directly include the impact of DnaJ on observed activities. As a consequence,

compounds selected as active in this screen should expectedly be the ones influencing the activity of the full complex. This approach permitted the identification of myricetin, one of the first allosteric inhibitors of HSP70 able to inhibit its activity by about 75% (Figure 10). The binding site of myricetin was determined by NMR; the ligand was found to engage a previously unknown pocket between the IB and IIB domains of HSP70. As an allosteric modulator, myricetin prevented DnaJ-mediated stimulation of ATPase activity, with minimal impact on HSP70 intrinsic turnover rate and on GrpE (a DnaK cochaperone) binding to DnaK, which occurs at a different location. Chemical screens against multiprotein complexes were further used to uncover probes that may be able to control HSP70 within the intact chaperone machinery.158−161 These studies indicated that some compounds may impact ATPase activity of HSP70 only in the presence of specific cochaperones. One such compound, dihydropyrimidine 1157c (Figure 10), was shown to bind to a site close to the region engaged by J-domains of DNAJ for stimulation of HSP70 activities. In this context, 115-7c activities mirrored those of the cochaperone HSP40. TROSY, docking studies, and sitedirected mutagenesis further corroborated these observations and led to the indication that the small molecule modulator and DnaJ may act in concert.159 The structural model was used to introduce steric hindrance on the initial lead to outcompete DnaJ interaction. Interestingly, the authors developed an 74



Perspective

Figure 11. Chemical structures of MKT-077, JG-98, and YM-08 developed by the Gestwicki group together with the binding site on HSP70.

analog labeled 116-9e (Figure 10), which in the ATPase assay had little effect on the DnaK-GrpE combination but was found to inhibit the DnaK-DnaJ complex formation by 80%. These data indicated selective interference with J-domain in the stimulation of HSP70 ATPase activity. The possibility of activating or inhibiting HSP70 functions in a controlled fashion, by using small molecules as regulators, opens up fascinating therapeutic perspectives. This is mainly the case in Alzheimer and tau-related diseases, where neurodegeneration and toxicity can be traced back to common protein misfolding events. In this context, deciphering the unclear pathogenic role of the chaperone machinery can generate precious information eventually leading to novel pharmacological treatments. Jinwal et al. used selective HSP70 inhibitors (e.g., myricetin) and activators (e.g., 115-7c) to gain insight into these mechanisms.145 In a cell-based model of Alzheimer disease, inhibitors led to rapid tau degradation by the ubiquitin proteasome system (UPS). In contrast, activators preserved tau levels. These results were apparently paradoxical, as activation of HSP70 was expected to increase tau clearance. This led the authors to the interesting hypothesis that activation may mimic the effects of stress-induced increase in HSP70 levels, a mechanism needed to ensure that more HSP70 is available to bind larger amounts of unfolded substrates. Because of the very slow kinetics of processes

related to tau accumulation, an acute increase in HSP70 levels would shift the equilibrium toward larger amount of HSP70/ tau complexes, suggesting that apparent reduction in free tau levels is only due to stoichiometric considerations and not to end-point pharmacological response of the activator. In a model where HSP70 is genetically overexpressed, the authors also indicated that pharmacological activation of ATPase activity favors formation of a profolding complex as prelude to tau accumulation and hyperphosphorylation. In contrast, in the same conditions, myricetin and other HSP70 ATPase inhibitors determine a selective increase in the amount of low-affinity HSP70/tau complexes that have to be cleared by the UPS. Hence, whereas the overexpression of HSP70 or overactivation of its ATPase activity does not consequently alter the balance of folding/degradation, its inhibition is directly connected to increased tau binding for exclusive degradation. These considerations, suggested by the use of selective allosteric modulators of HSP70 and of its functional complexes, may also have therapeutic implications: higher levels of HSP70 can in fact be induced by the use of HSP90 inhibitors (the two chaperone systems are connected through HSF1)68,162 such as celastrol or geldanamycin. This would provide more HSP70/tau complexes to target for efficacious 75



Perspective

Figure 12. Chemical structures of allosteric HSP70 ligands developed by the Chiosis group.

and maintained a brain/plasma (B/P) ratio greater than 0.25 for 18 h, with no renal toxicity. Interestingly, MKT-077 and YM-1 substitute for the cochaperone Hip in promoting HSP70-dependent nNOS maturation.167 Thus, MKT derivatives showed interesting effects against tau aggregates in several models. Another series of derivatives was developed for cancer treatment, designed to improve the metabolic stability of MKT077. The most potent molecules, such as JG-98, turned out to bind at the same site and to be 3-fold more active than MKT-077 against breast cancer cell lines with EC50 values in the mid-nanomolar range. In spite of a modest destabilization of chaperone clients, like the isoform 1 of AKT and Raf-1, the ligand was able to induce apoptosis in cancer cells.168 From the mechanistic point of view, JG-98 effects were due to allosteric perturbation of HSP70 binding with Bag3, a cochaperone that is necessary to promote cell survival pathways (Figure 11).141 Chiosis and co-workers combined structural homology modeling, binding site prediction, and druggability score evaluations to design allosteric ligands of HSP70. Their model represented the more open ADP-bound state of the chaperone, revealing pockets, which were not visible in the crystal structures of ATP-bound homologs. One of the pockets, located in the NBD, had a high druggability score and was designated as the chosen pocket to target allosteric

inhibitors of HSP70 ATPase, augmenting tau clearance by ubiquitination. This strategy would also represent a possible alternative use of HSP90 inhibitors that showed toxicity when administered for long times: one would use HSP90 inhibitors for a limited amount of time to upregulate HSP70, which would be consequently inhibited to degrade toxic aggregates. Another class of HSP70 allosteric ligands is represented by the derivatives of MKT077, a molecule belonging to the class of rhodacyanines. MKT077 was shown to have interesting anticancer effects against a number of human cancer cell lines, with IC50 between 0.35 and 1.2 μM.163,164 NMR, biochemical, and computational studies were used to show that the compound has affinity for the ADP-state but not for the ATP-state of the protein and that it engages a pocket, which is close but not superimposable to the ATP-binding site in the NBD165 (Figure 11). Due to the presence of a charged pyridinium, MKT077 has poor pharmacological properties, which may explain the failure in phase I cancer clinical trials and the very limited performances in passing the blood−brain barrier (BBB) in tests against neurodegeneration models. MKT077 was thus evolved into neutral analogues (YM series).166 The resulting compounds retained affinity for HSP70 in vitro and reduced pathogenic tau in brain slices. In vivo, one of the YM compounds, YM-08, crossed the BBB 76



Perspective

Figure 13. Chemical structures of small molecules targeting the SBD of HSP70, together with crystal structures of the respective complexes. PDB codes are 4WV7 (novolactone) and 4R5G (PES-16).

modulator of HSP70. Indeed, biochemical, structural, and functional analysis revealed that this allosteric inhibitor could covalently bind at the interface between the SBD and the NBD, disrupting interdomains communication and locking HSP70 in a conformation not compatible with ATP-induced substrate release. This, in turn, resulted in the inhibition of HSP70 refolding activity. Finally, it is worth mentioning that a small molecule inhibitor targeting the SBD has shown interesting activities against human HSP70 in cultured tumor cells and the bacterial HSP70.174,175 2-Phenylethynesulfonamide (PES) was selective for HSP70 vs HSP90 and induced disruption of the association of HSP70 with its cochaperones as well as with clients (Figure 13). PES determined tumor cell death through protein aggregation, impaired autophagy, and effects on lysosomes. The molecule was also tested in an in vivo tumor model, proving able to enhance survival in mice with Myc-induced lymphomagenesis. Overall, while the SBD has been a rather unexplored site so far, it may offer novel possibilities for allosteric interventions on the HSP70 system.175 Overall, allosteric modulation may hold lots of promise for targeting HSP70 pathways, defying chemical intervention based on ATP competitive molecules. Furthermore, it is interesting to observe how HSP70 modulation via stimulators or inhibitors generated novel hypotheses, in particular on the mechanisms of tau-aggregate degradation and processing. In perspective, a combination (possibly with coordinated administration times) of HSP90 and HSP70 modulators/ inhibitors could prove useful in the treatment of difficult molecular diseases.

HSP70 mechanisms. Interestingly, this site exposed a Cys residue, which could be targeted by a Michael acceptor establishing a covalent link between a hypothetical ligand and the protein. On this basis, they designed and synthesized 2,5′thiodipyrimidines (Figure 12), carrying an acrylamide moiety to covalently bind the allosteric cystein residue (Cys267).169 Cell experiments, pulldowns with biotinylated ligands, and trypsin digestion followed by MS fragment identification verified that YK5 bound to the allosteric pocket and could covalently inhibit HSP70. This was further tested in cancer cells, where the cytotoxic activity was linked to the disruption of the formation of active oncogenic HSP70/HSP90/client protein complexes.169 YK5 was subsequently optimized with extensive SAR studies,170 reaching potencies (Figure 12) between high nanomolar to low micromolar range and leading to a reduction of oncoproteins typically protected by HSP70. The inhibitors containing acrylamide, which ensures covalent binding to HSP70, were further modified to identify derivative 27c (Figure 12), which had biological effects comparable to those observed for the irreversible inhibitors at comparable concentrations.171 Interestingly, the affinity and selectivity of these compounds for HSP70 were further exploited for the development of chemical tools that enable the analysis of the HSP70-regulated proteome in tumor vs normal cells as well as comparative analysis of complexes in different tumor types. In this context, YK5 was further modified with suitable spacers and terminal biotin. The new derivatives allowed the isolation of large molecular assemblies with HSP70 in complex with oncoclient proteins, effectively facilitating their characterization via biochemical techniques.172 Using a chemogenomics screening on yeast cells, Hassan et al.173 identified novolactone (Figure 13) as a potential 77



■

Perspective

STRUCTURAL AND COMPUTATIONAL APPROACHES FOR THE DISCOVERY AND DESIGN OF CHAPERONE ALLOSTERIC MODULATORS The results discussed above clearly show that allosteric modulators of HSP90 and HSP70 have reached a stage where they can be aptly exploited for the investigation of molecular mechanisms of client binding and processing, as well as for drug development projects. The discovery and optimization of new allosteric effectors would greatly benefit from the availability of detailed structural information on their complexes with the target chaperone. In this respect, recent progress in structural biology has allowed overcoming of limitations imposed by the intrinsic flexibility of HSP90 and HSP70 protein families in their complex endogenous environment. Because of the extraordinary power of cryoEM, it has been possible to glimpse into how HSP90:Cdc37:Cdk4176 and HSP90:HSP70:Hop:GR20 are organized and how the HSP70 and HSP90 ATPase cycles might be coupled. However, while an increasing number of crystal structures have appeared that show the full-length chaperones bound to mimics of ATP in the nucleotide active site, very little or no experimental information is available to provide insight at atomistic resolution on the complexes with allosteric modulators (inhibitors or accelerators). This is particularly true for HSP90, where available structural characterization of the modes of binding of allosteric modulators is mainly due to NMR177,178 and cross-linking experiments.179 In the case of HSP70, X-ray information has been obtained for covalent173 and noncovalent174,175 ligands, while NMR has been used to generate binding modes for rhodacianines as 115-7c and related compounds.145,159,165 The difficulties in obtaining experimental molecular details of allosteric ligand−chaperone binding stem from the expected highly dynamic nature of the resulting complexes, making the latter refractory to crystallization. Indeed, kinetic and binding experiments30,122,123,128,132,159 indicate that while allosteric perturbation on the one hand modifies the timing of the conformational cycle, on the other hand it reshapes protein− protein interaction surfaces guiding the protein in preferential partner selection. To gain atomistic insights into the determinants of chaperone dynamics and how they are related to and influenced by allosteric ligand binding, we can turn to computational/theoretical approaches.180−184 Whereas plastic and finely regulated systems like HSP90 and HSP70 are a challenging task, in particular for all-atom models, they also represent prime test cases to evaluate the reach and significance of computational approaches. Herein, we will focus on results reported for the full-length proteins and their complexes. The dynamics of HSP90 and their relationships to allosteric regulation have been studied with a multiplicity of approaches, ranging from plain all-atom MD simulations to biased methods, from network analysis to coevolution studies. Using statistical mechanics based methods, we carried out comparative analyses of MD trajectories of multiple structural representatives of HSP90 (yeast HSP90, mammalian Grp94, bacterial HtpG), in different nucleotide states (ATP-bound, ADP-bound, apo state).47,126,127 Through this approach, we were able to identify potential points of allosteric control distal from the nucleotide binding-site. Specifically, we made use of the previously described CP analysis,47 to identify (groups of)

mechanically connected amino acids, and of the analysis of the mechanical strain experienced by the protein in the presence of different nucleotides (ATP, ADP, apo state). This latter calculated how much the instantaneous distance of a residue from neighboring amino acids differs from the time-average. By evaluating the time evolution of the parameters in MD runs with different ligands, it was possible to identify distal regions that undergo consistent nucleotide-dependent local deformations in different HSP90 homologs. Statistical and comparative analysis of the trajectories also unveiled the residues that move in a cooperative manner and may give the highest contribution in modulating functional motions. Interestingly, analysis of mechanical coordination and structural deformation revealed that the ligand-dependent structural modulations mostly consist of relative rigid-like movements around two common hinges of a limited number of quasi-rigid domains shared by all three proteins. The first hinge, whose functional role has been demonstrated by several experimental approaches, is located at the boundary between the N-terminal and niddle-domains. The second hinge is located at the end of the H4−H6 threehelix bundle, located at the boundary between the M-large and M-small structural subdomains, where three Phe residues (Phe484, Phe488, and Phe432) form a well-packed hydrophobic core which unfolds/unpacks going from the ATP- to the ADP-state. Interestingly, the latter hinge was located in the vicinity of Cys597 (human HSP90α numeration), whose Snitrosylation was experimentally observed to alter both the Cterminal and N-terminal association properties of HSP90, affecting its conformational equilibrium, the ATPase cycle and recruitment of clients.185 The results of computational analyses were used to guide the design of point mutations of residues distal from the ATP site, whose dynamics turned out to be most affected by the nucleotide exchange. In particular, it was interesting to observe that Phe to Ala mutations at the end of the aforementioned three-helix bundle could stimulate the ATPase and closure kinetics of the chaperone. From the conformational dynamics point of view, such mutations increased the overall flexibility in the chaperone, speeding up the search for the closed, catalytically competent state. In contrast, additional mutations, located in the CTD were found to rigidify the protein leading to inhibition. Furthermore, mutations in allosteric points also showed an effect on cochaperone recognition and binding.128 Seifert and Graeter186 made use of MD simulations of HtpG followed by force-distribution analysis to identify internal allosteric communication pathways that connect the nucleotide binding site to a region in the middle domain, which is 2−3 nm distant from the catalytic site. In this model, the long helix in the LMD (helix H18 in the original paper) serves as dynamic hinge region, consistent with the previously defined model,47 and its terminal part nicely overlaps with the client protein binding site mapped out by Genest and co-workers.31 Importantly, the work of Seifert and Graeter186 stresses the importance of the detailed molecular characterization of the mechanism: force transmission is in fact triggered only by ATP, which induces the bending of the proximal H3 helix in the NTD and represents the major force channel into the MD. In the ADP and apo states, this allosteric mechanism is different. This strategy, which has been extended to other allosterically regulated systems, indicates how an external perturbation, represented, for example, by ligand binding, can be efficiently modeled as an external force acting on the macromolecule, making techniques able to take into account 78



Perspective

also reconnected with the role of specific sites of HSP90 posttranslational modifications in regulating chaperone function.195 Overall, computational analyses carried out by different groups using different approaches converge in indicating preferential coordination pathways between the N-terminal ATP binding site and the client binding site located between the M-small and C-terminal domains of HSP90. The dynamics of such coordination pathways are specifically dependent on the type of bound nucleotide, establishing a direct connection between the site of ATP hydrolysis and the region where client recognition and processing takes place. Thus, it is conceivable to try and exploit this knowledge for the discovery of small molecule modulators targeting HSP90 and its complexes. In terms of allosteric lead discovery, information on the points of regulation defined through coordination propensity (CP) and strain analysis was used to define potential complementary interactions with the allosteric site47,126,127,196 and then translated into pharmacophore models used for virtual screening of large libraries of drug-like molecules.126 This approach revealed a series of compounds based on the benzofuran moiety (see above, allosteric activators paragraph) that proved to be genuine allosteric activators of the chaperone with important biological effects when tested in cancer cell models.123,124,197 Furthermore, the structural information on the potential allosteric site could be combined with docking approaches to generate SARs for known allosteric compounds and to guide the design of novel derivatives.95,198,199 In particular, analysis of the fitness of known active ligands (inhibitors or activators) for the allosteric pocket was combined with the characterization of their effects on HSP90 structural dynamics. The investigation of the mechanisms of formation/disappearance of pockets around the allosteric site, driven by conformational HSP90 response to first-generation ligands, was then exploited to define the positions where the addition/variation of specific chemical functionalities on existing scaffolds allowed for optimization of binding interactions. This approach was used to improve on the stimulating properties of activators,123,124,131 as well as to ameliorate potencies of coumarin-based allosteric inhibitors.200 Simulative approaches201,202 based on the use of MM-GBSA (molecular mechanics generalized Born surface area) calculations combined with a decomposition of the effective energy (i.e., the sum of gas phase and solvation free energy) of dimerization on a per residue level were used to detect the amino acids that are important to stabilize the C-terminal dimerization interface. This information was next translated into peptidic constructs that were successfully used to block CTD dimerization.201,202 A similar approach was used by Wang and co-workers203 to target the HSP90-Cdc37 protein−protein interaction (PPI). Using interface mimicry and analysis of peptide binding patterns with MD simulations, the authors identified the short sequence HFGMLRR (Pep-5) as the one able to bind HSP90 with a Kd of 6.90 μM and to interfere with the recruitment of Cdc37. Molecular simulations are playing a more and more important role also in shedding light on the mechanisms of allosteric regulation of HSP70: analyses of allosteric (DnaK) and nonallosteric homologues (HSP110-Sse1) of the chaperone in different nucleotide states could identify relevant populations on the free energy landscape, underlining the high flexibility and plasticity of this chaperone.204−209 Communica-

this perturbation a very general approach to study the mechanisms involved in protein allostery.187 All-atom MD simulations were also used to study the role of the unusual asymmetry in the dimeric structure of TRAP1 (the mitochondrial HSP90 homolog) on the mechanisms of ATP processing and conformational cycle regulation.188,189 Using a combination of X-ray crystallography, FRET analyses, DEER, mutational and biochemical experiments, Agard and coworkers proved that ATP hydrolysis on the two protomers is sequential and deterministic.188 Microsecond all-atom MD simulations showed that the ATP g-phosphate in the buckled protomer has fewer water molecules in proximity than the one in the straight protomer. At higher temperature, the same trend is maintained. While the two nucleotide binding pockets are essentially identical and the average rmsd between the NTD of the two protomers is 1.74 A, water dynamics in the active site (not only occupancy) was different between the two protomers, in particular near the two ATP b- and gphosphates. Most of the waters turned out to spend only a few nanoseconds in the buckled protomer. In contrast, in the straight protomer, there are longer-lived waters positioned between E130 and the g-phosphate and also more waters positioned above E130. The asymmetry in water occupancy and dynamics between two protomers was found to correlate well with the observed preferential hydrolysis of the buckled protomer ATP. Expanding on the issue of how structural asymmetry may determine function in TRAP1, the internal dynamics of the mitochondrial chaperone was investigated applying the mechanical, coordination, and fluctuation measures introduced for HSP90.47,126,127 The substructures at the interface between the MD and the CTD of TRAP1 were shown to be specifically modulated by the type of the nucleotide and by the specific protomer to which it was bound. Direct coordination was defined between the nucleotide site and distal site in the MD of TRAP1: in particular, the most intense coordination was observed between the two ATP pockets and elements of large the LMD of the straight protomer B.189 This asymmetric communication between the nucleotide and a previously identified region of the chaperone which was shown to bind client proteins31 supports the observation that structural asymmetry plays a role in functional regulation and tuning of the chaperone properties. Verkhivker and co-workers used a combination of MD simulations and computer science techniques, ranging from network modeling to community analysis, to identify preferential control points and communication mechanisms in HSP90.190−194 In particular, structural stability analysis using force constant profiling of the inter-residue fluctuation distances permitted identification of networks of conserved structurally rigid residues that were defined as mediators of allosteric communication. Moreover, by connecting network analysis to conformational landscape concepts, the authors established a relationship between structural stability and the centrality in networks of allosterically relevant residues involved in chaperone regulation: in this picture, allosteric interactions appear to be mediated by modules of structurally stable residues. Importantly, stable and connected residues corresponded to regions in the M-domain, relevant for client binding.191 Interestingly, this kind of analysis was also used for studying large HSP90 complexes.193 In this work, MD simulations and network analysis revealed the central role of Cdc37 in mediating client recognition and defining allosteric regulation of the chaperone−kinase complex. The analysis was 79



Perspective

regulation and fine-tuning of their functions in both normal and disease conditions. The development of chemical tools and drug-like molecules that not only abolish the activities but rather modulate in vitro/in vivo conformational substates of the chaperone is playing an increasingly prominent role in elucidating the coupling between ATP processing, the conformational cycle, and client remodeling. At the cell level, chemical modulators (effectors) may even help direct cells toward different destinies. In this context, the main chemical challenges are represented by (i) the need to establish computational methods that include the explicit characterization of functional dynamics and its influence on possible interaction sites of the target protein; (ii) the development of novel classes of compounds that directly modulate the distribution of functional substates through rationally selected functional groups; (iii) the characterization of (potentially unexpected) effects of targeting unexplored allosteric sites. Achieving these goals will provide new chemical tools that can be used to take action at different stages, from ATP processing phase to multiprotein complex assembly. By exerting modular tuning of the chaperone machinery, it may conceivably be possible to elucidate the intricate mechanisms behind clients folding and cochaperones recognition. Such approaches will provide important complements to more classical molecular biology methods. The availability of these data represents a new exciting possibility to design molecules that perturb enzymatic ATPase activities or modulate protein− protein interfaces, generating a direct link between structural information and functional outcomes. In conclusion, we expect that the integration of structural dynamics studies with the development of new compounds will contribute to disentangle the intricate relationships between the selection of states through allosteric binding and the resulting functional implications. The synergy we propose will enable a dynamic chemical and structural picture for deciphering HPS70 and HSP90 functions as well as a rationalization of chaperone modulation operated by small molecules within a physiologically dynamic framework.

tion patterns, internal fluctuations analysis, and characterization of local deformations determined by ATP/ADP exchange were reconnected to the mechanisms of open-toclosed conformational transition.210,211 The results of these simulations were nicely consistent with independent results that identified important hinge residues presiding nucleotide processing and exchange,212,213 relevant dynamic domains,214 and inhibition mechanisms.215 In general, the results converge on highlighting the importance of the linker between the NBD and SDB as a dynamic allosteric switch,36,216 while the nucleotide-dependent domain motions of the IA and IIA lobes of the NBD behaved as the initiators of the functionally oriented conformational reorganizations. Recently, the interaction between the HSP70 and HSP40, fundamental for ATPase regulation, was studied by Malinverni et al.217 using an elegant combination of molecular simulations based on coarse-grained and atomistic models with coevolutionary sequence analysis. Focusing on a procariotic model of DnaK(HSP70)-DnaJ(HSP40) interaction, the authors were able to identify an interaction surface formed by helix II of the J-domain of DnaJ and a structurally contiguous region of DnaK conserved throughout evolution. The latter surface involved lobe IIA of the NBD, the interdomain linker, and the β-basket of the SBD. This integrated approach proved able to rationalize a number of experimental observations, at the same time providing a new high-resolution dynamic model to target for designing active inhibitors of DnaJ-DnaK interaction. Another example of integrated approach to the study of HSP70 functional dynamics comes from the Carlson lab.212 They combined MD simulations, mutagenesis, and enzymatic assays to identify the key residues responsible for regulating ATP hydrolysis. Focusing on the NBD of the HSP70 chaperone DnaK in the apo, ATP-bound, and ADP-bound states, the authors showed that each nucleotide state populates a distinct conformation. Interestingly, this was supported by biochemical mutation and binding data. MD simulation results also identified a shearing motion between subdomains IA and IIA, which governed the structural interconversion. By coupling MD simulation data with evolutionary analysis, G228 and G229 came out as hinges for the nucleotide dependent closing/opening motions: alanine mutations of these hinge residues caused reduction of DnaK chaperone activity in vitro and in vivo. It is tempting, at this point, to suggest that this wealth of information on the correlations among structure, communication mechanisms, dynamics, function, and ligand binding constitutes a solid basis to predict the allosteric response of HSP70: upon targeting regions endowed with functionally oriented dynamic properties, it may be possible to rationally optimize existing allosteric effectors and to discover novel chemical entities. The new molecules would have interesting applications in understanding how the chaperone functional cycle is linked to clients/cochaperones recruitment and, hence, to cell viability. Success in this kind of design would open up new possibilities for the development of useful chemical tools and, in the long term, for the discovery of new therapeutic drugs.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; [email protected]. ORCID

Giulia Morra: 0000-0002-9681-7845 Silvia Rinaldi: 0000-0002-1088-7253 Federica Compostella: 0000-0003-4721-0358 Giorgio Colombo: 0000-0002-1318-668X Notes

The authors declare no competing financial interest. Biographies Mariarosaria Ferraro got her Masters degree in Chemistry and Pharmaceutical Technologies at the University of Bologna in 2012, under the supervision of Prof. Andrea Cavalli. For her Ph.D. she focused on modeling lipid rafts and membrane proteins involved in drug addiction. In 2016, after a period as a visiting student in the laboratory of Prof. Ross Walker at the San Diego Supercomputer Center, she obtained her Ph.D. in Drug Discovery at the University of Genoa in collaboration with Istituto Italiano di Tecnologia (IIT) under the supervision of Dr. Giovanni Bottegoni. She is currently a postdoctoral researcher at ICRM CNR in the group of Prof. Giorgio Colombo, where she works on anticancer compounds targeting molecular chaperones.

■

CONCLUSIONS AND PERSPECTIVES HSP90 and HSP70 are essential molecular machines that control the folding process in order to activate a broad and disparate array of substrate “client” proteins. Despite their fundamental roles, many questions are still open regarding the 80



Perspective

Her research focuses on the synthesis of compounds with biological and pharmaceutical applications with particular attention to glycoconjugates and saccharide antigens.

llda D’Annessa obtained her M.Sc. in Bioinformatics in 2007 and her Ph.D. in Biochemistry and Molecular Biology in 2010 at the University of Rome Tor Vergata. She worked in the group of Prof. A. Desideri at the Department of Biology. Here, she used computational biophysics approaches to study the functional properties of protein−DNA interactions, focusing mainly on human topoisomerase IB as a target for anticancer therapy, through the characterization of its interaction with inhibitor compounds and functional consequences of mutants with altered sensitivity. In 2016 she joined ICRM CNR as a postdoc in the group of Prof. Giorgio Colombo, focusing on the study of HSP90/cochaperone−client interactions and in the development of allosteric modulators.

Giorgio Colombo is Professor of Organic Chemistry at the University of Pavia and Senior Researcher at ICRM CNR in Milano, where he set up and heads the Biocomputing laboratory. He received his M.Sc. and Ph.D. in Chemistry from the University of Milano, under the supervision of Profs. C. Scolastico and A. Bernardi. Before establishing his group, he worked with Prof. Ken Merz at Penn State University and Prof. A.E. Mark at the University of Groningen on computational approaches to enzyme reactivity and protein folding. His research focuses on the study of the dynamic and energetic determinants of protein structural organization and molecular recognition and on the development of methods to design molecules with useful biological properties. He applied these concepts to chaperones and structural vaccinology.

Elisabetta Moroni got her Ph.D. (2006) in complex systems in postgenomic biology at the University of Torino, focusing on developing effective energy functions to evaluate DNA−protein binding, under the supervision of Prof. M. Caselle. Next, she joined the group of Profs. L. Belvisi and A. Bernardi at the University of Milano as a postdoc working on the development of anticancer drugs. She then moved to the pharma company NMS (Nerviano, Italy) where she worked on the design of kinase inhibitors. She continued her research activity at ICRM CNR in the group of Prof. Colombo, focusing on theoretical methods to study functional mechanisms of proteins and the design of cancer therapeutics, before taking up her current position as researcher at IRCCS MultiMedica in Milano.

■

ACKNOWLEDGMENTS We acknowledge funding from AIRC (Associazione Italiana Ricerca sul Cancro) through Grant IG 20019 and from NTAP (Neurofibromatosis Therapeutic Acceleration Program) through Project “TRAPping the metabolic adaptations of plexiform neurofibroma”.

■

ABBREVIATIONS USED ER, endoplasmic reticulum; Grp94, glucose-regulated protein 94; TRAP1, tumor necrosis factor type 1 receptor-associated protein; SAXS, small-angle X-ray scattering; NTD, N-terminal domain; MD, middle domain; LMD, large middle domain; SMD, small middle domain; CTD, C-terminal domain; SDH, succinate dehydrogenase; 17-AAG, 17-(allylamino)-17demethoxygeldanamycin; 17-DMAG, 17-dimethylaminoethylamino-17-demethoxygeldanamycin; HER2, human epidermal growth factor receptor 2; GIST, gastrointestinal stromal tumor; HSF1, heat shock factor 1; PBMC, peripheral blood mononuclear cell; PTK, protein tyrosine kinase; PSTK, protein serine/threonine kinase; Raf-1, RAF proto-oncogene serine/ threonine-protein kinase; HSR, heat shock response; LNCaP, androgen receptor-dependent prostate cancer cell line; LAPC4, androgen receptor prostate cancer cell line; AKT, protein kinase B; HIF-1, hypoxia-inducible factor 1; SPR, surface plasmon resonance; ER, estrogen receptor; DRG, dorsal root ganglia; MCR, mitochondrial respiratory capacity; IGF-2, insulin-like growth factor 2; HOP, HSP organizing protein; IP6K2, inositol hexakisphosphate kinase 2; FKBP52, FK506binding protein 52; EGCG, epigallocatechin 3-gallate; PAINS, pan assay interference compounds; HtpG, high temperature protein G; GR, glucocorticoid receptor; CP, coordination propensity; HSC70, heat shock cognate 71; NBD, nucleotide binding domain; SBD, substrate binding domain; NEF, nucleotide exchange factor; TROSY, transverse relaxationoptimized spectroscopy; UPS, ubiquitin proteasome system; nNOS, neuronal nitric oxide synthase; PES, 2-phenylethynesulfonamide; cryoEM, cryogenic electron microscopy; DEER, double electron−electron resonance; MM-GBSA, molecular mechanics generalized Born surface area; PPI, protein−protein interaction

Giulia Morra obtained her M.Sc. (1998) in Physics in Milan and her Ph.D. in Chemistry (2005) at FU Berlin under the supervision of E. W. Knapp, focusing on electrostatics and MD simulations of proteins. In 2006 she joined the Biocomputing laboratory at ICRM CNR as a postdoc under the supervision of Giorgio Colombo, where she worked on developing methods for the analysis of allosteric phenomena in proteins, focusing on HSP chaperone proteins. In 2011 she was appointed as permanent research staff at the same institute. Antonella Paladino got her Ph.D. in Computational Biology in 2011 at the University of Naples “Federico II”. During her scientific training, she joined the Molecular Modeling and Bioinformatics group at the IRB of Barcelona and the Computational Biophysical Chemistry laboraotory at the University of the Basque Country. Since 2013 she has been a postdoctoral researcher in the group of Prof. Giorgio Colombo. Her research is aimed at investigating the relationship between the structure, the dynamics. and the function of biomolecules through the application of novel computational techniques. Silvia Rinaldi obtained her M.Sc. from the University of Perugia (2009) in the Center of Excellence for Innovative Nanostructured Materials. After receiving her Ph.D. in 2013 in the laboratory of Prof. Olivucci at the University of Siena working on computational photochemistry and photobiology, she moved to ICRM-CNR as a postdoctoral researcher in the group of Prof. Giorgio Colombo. Her research interests are focused on the computational modeling of protein dynamics, analysis of molecular recognition processes, and enzyme catalysis. Federica Compostella received her Laurea degree in Chemistry at the University of Milano. She worked in the laboratory of Profs. Scolastico and A. Bernardi at the University of Milano on the synthesis of compounds of pharmaceutical interest. After a period as a research fellow for the pharmaceutical companies Recordati SPA and Dompè Farmaceutici SPA, she became Assistant Professor at the University of Milano. She set up and is currently in charge of “Chemistry and Introductory Biochemistry” courses at the recently founded International Medical School of the University of Milano.

■

REFERENCES

(1) Whitesell, L.; Bagatell, R.; Falsey, R. The stress response: implications for the clinical development of HSP90 inhibitors. Curr. Cancer Drug Targets 2003, 3, 349−358.

81



Perspective

coordinated action of the HSP90 and HSP70 chaperone cycles. Cell 2014, 157, 1685−1697. (21) Hopkins, A. L. Network pharmacology: the next paradigm in drug discovery. Nat. Chem. Biol. 2008, 4, 682−690. (22) Whitesell, L.; Santagata, S.; Mendillo, M. L.; Lin, N. U.; Proia, D. A.; Lindquist, S. HSP90 empowers evolution of resistance to hormonal therapy in human breast cancer models. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 18297−18302. (23) van der Greef, J.; McBurney, R. N. Innovation - Rescuing drug discovery: in vivo systems pathology and systems pharmacology. Nat. Rev. Drug Discovery 2005, 4, 961−967. (24) Steensma, D. P. The ordinary miracle of cancer clinical trials. J. Clin. Oncol. 2009, 27, 1737−1739. (25) Vogelstein, B.; Kinzler, K. W. Cancer genes and the pathways they control. Nat. Med. 2004, 10, 789−799. (26) Ferguson, F. M.; Gray, N. S. Kinase inhibitors: the road ahead. Nat. Rev. Drug Discovery 2018, 17, 353−377. (27) Vetro, M.; Costa, B.; Donvito, G.; Arrighetti, N.; Cipolla, L.; Perego, P.; Compostella, F.; Ronchetti, F.; Colombo, D. Anionic glycolipids related to glucuronosyldiacylglycerol inhibit protein kinase Akt. Org. Biomol. Chem. 2015, 13, 1091−1099. (28) Neckers, L.; Blagg, B.; Haystead, T.; Trepel, J. B.; Whitesell, L.; Picard, D. Methods to validate HSP90 inhibitor specificity, to identify off-target effects, and to rethink approaches for further clinical development. Cell Stress Chaperones 2018, 23, 467−482. (29) Lamb, J.; Crawford, E. D.; Peck, D.; Modell, J. W.; Blat, I. C.; Wrobel, M. J.; Lerner, J.; Brunet, J. P.; Subramanian, A.; Ross, K. N.; Reich, M.; Hyeronimus, H.; Wei, G.; Armstrong, S. A.; Haggarty, S. J.; Clemons, P. A.; Wei, R.; Carr, S. A.; Lander, E. S.; Golub, T. R. The Connectivity Map: using gene-expression signatures to connect small molecules, genes, and disease. Science 2006, 313, 1929−1935. (30) Cesa, L. C.; Patury, S.; Komiyama, T.; Ahmad, A.; Zuiderweg, E. R. P.; Gestwicki, J. E. Inhibitors of difficult protein-protein interactions identified by high-throughput screening of multiprotein complexes. ACS Chem. Biol. 2013, 8, 1988−1997. (31) Genest, O.; Reidy, M.; Street, T. O.; Hoskins, J. R.; Camberg, J. L.; Agard, D. A.; Masison, D. C.; Wickner, S. Uncovering a region of heat shock protein 90 important for client binding in E. coli and chaperone function in yeast. Mol. Cell 2013, 49, 464−473. (32) Lorenz, O. R.; Freiburger, L.; Rutz, D. A.; Krause, M.; Zierer, B. K.; Alvira, S.; Cuéllar, J.; Valpuesta, J. M.; Madl, T.; Sattler, M.; Buchner, J. Modulation of the HSP90 chaperone cycle by a stringent client protein. Mol. Cell 2014, 53, 941−953. (33) McLaughlin, S. H.; Smith, H. W.; Jackson, S. E. Stimulation of the weak ATPase activity of human HSP90 by a client protein11Edited by G. von Heijne. J. Mol. Biol. 2002, 315, 787−798. (34) Liu, Y.; Gierasch, L. M.; Bahar, I. Role of HSP70 ATPase domain intrinsic dynamics and sequence evolution in enabling its functional interactions with NEFs. PLoS Comput. Biol. 2010, 6, e1000931. (35) Swain, J. F.; Dinler, G.; Sivendran, R.; Montgomery, D. L.; Stotz, M.; Gierasch, L. M. HSP70 chaperone ligands control domain association via an allosteric mechanism mediated by the interdomain linker. Mol. Cell 2007, 26, 27−39. (36) Zhuravleva, A.; Clerico, E. M.; Gierasch, L. M. An interdomain energetic tug-of-war creates the allosterically active state in HSP70 molecular chaperones. Cell 2012, 151, 1296−1307. (37) Nussinov, R.; Tsai, C.-J. Allostery in disease and in drug discovery. Cell 2013, 153, 293−305. (38) Nussinov, R.; Tsai, C.-J.; Liu, J. Principles of allosteric interactions in cell signaling. J. Am. Chem. Soc. 2014, 136, 17692− 17701. (39) Ostrem, J. M.; Peters, U. P.; Sos, M. L.; Wells, J. A.; Shokat, K. M. K-Ras(G12C) inhibitors allosterically control GTP affinity and effector interactions. Nature 2013, 503, 548−551. (40) Rettenmaier, T. J.; Fan, H.; Karpiak, J.; Doak, A.; Sali, A.; Shoichet, B. K.; Wells, J. A. Small-Molecule allosteric modulators of the protein kinase pdk1 from structure-based docking. J. Med. Chem. 2015, 58, 8285−91.

(2) Hartl, F. U.; Bracher, A.; Hayer-Hartl, M. Molecular chaperones in protein folding and proteostasis. Nature 2011, 475, 324−332. (3) Young, J. C.; Agashe, V. R.; Siegers, K.; Hartl, F. U. Pathways of chaperone mediated protein folding in the cytosol. Nat. Rev. Mol. Cell Biol. 2004, 5, 781−791. (4) Do, K.; Speranza, G.; Chang, L. C.; Polley, E. C.; Bishop, R.; Zhu, W. M.; Trepel, J. B.; Lee, S.; Lee, M. J.; Kinders, R. J.; Phillips, L.; Collins, J.; Lyons, J.; Jeong, W.; Antony, R.; Chen, A. P.; Neckers, L.; Doroshow, J. H.; Kummar, S. Phase I study of the heat shock protein 90 (HSP90) inhibitor onalespib (AT13387) administered on a daily for 2 consecutive days per week dosing schedule in patients with advanced solid tumors. Invest. New Drugs 2015, 33, 921−930. (5) Brehme, M.; Voisine, C.; Rolland, T.; Wachi, S.; Soper, J. H.; Zhu, Y.; Orton, K.; Villella, A.; Garza, D.; Vidal, M.; Ge, H.; Morimoto, R. I. A chaperome subnetwork safeguards proteostasis in aging and neurodegenerative disease. Cell Rep. 2014, 9, 1135−1150. (6) Finka, A.; Goloubinoff, P. Proteomic data from human cell cultures refine mechanisms of chaperone-mediated protein homeostasis. Cell Stress Chaperones 2013, 18, 591−605. (7) Shrestha, L.; Patel, H. J.; Chiosis, G. Chemical tools to investigate mechanisms associated with HSP90 and HSP70 in disease. Cell Chem. Biol. 2016, 23, 158−172. (8) Johnson, J. L. Evolution and function of diverse HSP90 homologs and cochaperone proteins. Biochim. Biophys. Acta, Mol. Cell Res. 2012, 1823, 607−613. (9) Fernandez-Fernandez, M. R.; Gragera, M.; Ochoa-Ibarrola, L.; Quintana-Gallardo, L.; Valpuesta, J. M. HSP70 - a master regulator in protein degradation. FEBS Lett. 2017, 591, 2648−2660. (10) Travers, J.; Sharp, S.; Workman, P. HSP90 inhibition: twopronged exploitation of cancer dependencies. Drug Discovery Today 2012, 17, 242−252. (11) Neckers, L.; Trepel, J. B. Stressing the development of small molecules targeting HSP90. Clin. Cancer Res. 2014, 20, 275−277. (12) Nylandsted, J.; Brand, K.; Jaattela, M. Heat shock protein 70 is required for the survival of cancer cells. Ann. N. Y. Acad. Sci. 2000, 926, 122−125. (13) Kaul, S. C.; Yaguchi, T.; Taira, K.; Reddel, R. R.; Wadhwa, R. Overexpressed mortalin (mot-2)/mtHSP70/GRP75 and hTERT cooperate to extend the in vitro lifespan of human fibroblasts. Exp. Cell Res. 2003, 286, 96−101. (14) Murphy, M. E. The HSP70 family and cancer. Carcinogenesis 2013, 34, 1181−1186. (15) Garrido, C.; Brunet, M.; Didelot, C.; Zermati, Y.; Schmitt, E.; Kroemer, G. Heat shock proteins 27 and 70: anti-apoptotic proteins with tumorigenic properties. Cell Cycle 2006, 5, 2592−2601. (16) Shelton, L. B.; Koren, J.; Blair, L. J. Imbalances in the HSP90 chaperone machinery: implications for tauopathies. Front. Neurosci. 2017, 11, 724. (17) Wang, A. M.; Miyata, Y.; Klinedinst, S.; Peng, H.-M.; Chua, J. P.; Komiyama, T.; Li, X.; Morishima, Y.; Merry, D. E.; Pratt, W. B.; Osawa, Y.; Collins, C. A.; Gestwicki, J. E.; Lieberman, A. P. Activation of HSP70 reduces neurotoxicity by promoting polyglutamine protein degradation. Nat. Chem. Biol. 2013, 9, 112−118. (18) Pratt, W. B.; Toft, D. O. Regulation of signaling protein function and trafficking by the HSP90/HSP70-based chaperone machinery. Exp. Biol. Med. 2003, 228, 111−133. (19) Rodina, A.; Wang, T.; Yan, P. R.; Gomes, E. D.; Dunphy, M. P. S.; Pillarsetty, N.; Koren, J.; Gerecitano, J. F.; Aldone, T. T.; Zong, H. L.; Caldas-Lopes, E.; Alpaugh, M.; Corben, A.; Riolo, M.; Beattie, B.; Pressl, C.; Peter, R. I.; Xu, C.; Trondl, R.; Patel, H. J.; Shimizu, F.; Bolaender, A.; Yang, C. H.; Panchal, P.; Farooq, M. F.; Kishinevsky, S.; Modi, S.; Lin, O.; Chu, F. X.; Patil, S.; Erdjument-Bromage, H.; Zanzonico, P.; Hudis, C.; Studer, L.; Roboz, G. J.; Cesarman, E.; Cerchietti, L.; Levine, R.; Melnick, A.; Larson, S. M.; Lewis, J. S.; Guzman, M. L.; Chiosis, G. The epichaperome is an integrated chaperome network that facilitates tumour survival. Nature 2016, 538, 397−401. (20) Kirschke, E.; Goswami, D.; Southworth, D.; Griffin, P. R.; Agard, D. A. Glucocorticoid receptor function regulated by 82



Perspective

from death through inhibition of the permeability transition. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 726−731. (63) Workman, P.; Al-Lazikani, B. Drugging cancer genomes. Nat. Rev. Drug Discovery 2013, 12, 889−890. (64) Neckers, L.; Workman, P. HSP90 molecular chaperone inhibitors: Are we there yet? Clin. Cancer Res. 2012, 18 (1), 64−76. (65) Workman, P.; Burrows, F. J.; Neckers, L.; Rosen, N. Drugging the cancer chaperone HSP90: combinatorial therapeutic exploitation of oncogene addiction and tumor stress. Ann. N. Y. Acad. Sci. 2007, 1113, 202−216. (66) Workman, P. Overview: translating HSP90 biology into HSP90 drugs. Curr. Cancer Drug Targets 2003, 3, 297−300. (67) Garcia-Carbonero, R.; Carnero, A.; Paz-Ares, L. Inhibition of HSP90 molecular chaperones: moving into the clinic. Lancet Oncol. 2013, 14, E358−E369. (68) Zou, J.; Guo, Y.; Guettouche, T.; Smith, D. F.; Voellmy, R. Repression of heat shock transcription factor HSF1 activation by HSP90 (HSP90 complex) that forms a stress-sensitive complex with HSF1. Cell 1998, 94, 471−480. (69) Conde, R.; Belak, Z. R.; Nair, M.; O’Carroll, R. F.; Ovsenek, N. Modulation of HSF1 activity by novobiocin and geldanamycin. Biochem. Cell Biol. 2009, 87, 845−851. (70) Powers, M. V.; Workman, P. Inhibitors of the heat shock response: Biology and pharmacology. FEBS Lett. 2007, 581, 3758− 3769. (71) Woodford, M. R.; Truman, A. W.; Dunn, D. M.; Jensen, S. M.; Cotran, R.; Bullard, R.; Abouelleil, M.; Beebe, K.; Wolfgeher, D.; Wierzbicki, S.; Post, D. E.; Caza, T.; Tsutsumi, S.; Panaretou, B.; Kron, S. J.; Trepel, J. B.; Landas, S.; Prodromou, C.; Shapiro, O.; Stetler-Stevenson, W. G.; Bourboulia, D.; Neckers, L.; Bratslavsky, G.; Mollapour, M. Mps1 mediated phosphorylation of HSP90 confers renal cell carcinoma sensitivity and selectivity to HSP90 inhibitors. Cell Rep. 2016, 14, 872−884. (72) Chiosis, G.; Neckers, L. Tumor selectivity of HSP90 inhibitors: the explanation remains elusive. ACS Chem. Biol. 2006, 1, 279−284. (73) Trepel, J. B.; Mollapour, M.; Giaccone, G.; Neckers, L. Targeting the dynamic HSP90 complex in cancer. Nat. Rev. Cancer 2010, 10, 537−549. (74) Graner, M. W. HSP90 and Immune Modulation in Cancer. In Advances in Cancer Research; Isaacs, J., Whitesell, L., Eds.; Academic Press, 2016; Vol. 129, Chapter 8, pp 191−224, DOI: 10.1016/ bs.acr.2015.10.001. (75) Yoshida, S.; Tsutsumi, S.; Muhlebach, G.; Sourbier, C.; Lee, M.-J.; Lee, S.; Vartholomaiou, E.; Tatokoro, M.; Beebe, K.; Miyajima, N.; Mohney, R. P.; Chen, Y.; Hasumi, H.; Xu, W.; Fukushima, H.; Nakamura, K.; Koga, F.; Kihara, K.; Trepel, J.; Picard, D.; Neckers, L. Molecular chaperone TRAP1 regulates a metabolic switch between mitochondrial respiration and aerobic glycolysis. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, E1604−1612. (76) Rasola, A. HSP90 proteins in the scenario of tumor complexity. Oncotarget 2017, 8, 20521−20522. (77) Rasola, A.; Neckers, L.; Picard, D. Mitochondrial oxidative phosphorylation TRAP(1)ped in tumor cells. Trends Cell Biol. 2014, 24, 455−463. (78) Liu, D.; Hu, J.; Agorreta, J.; Cesario, A.; Zhang, Y.; Harris, A. L.; Gatter, K.; Pezzella, F. Tumor necrosis factor receptor-associated protein 1(TRAP1) regulates genes involved in cell cycle and metastases. Cancer Lett. 2010, 296, 194−205. (79) Patel, P. D.; Yan, P.; Seidler, P. M.; Patel, H. J.; Sun, W.; Yang, C.; Que, N. S.; Taldone, T.; Finotti, P.; Stephani, R. A.; Gewirth, D. T.; Chiosis, G. Paralog-selective HSP90 inhibitors define tumorspecific regulation of HER2. Nat. Chem. Biol. 2013, 9, 677−684. (80) Patel, H. J.; Patel, P. D.; Ochiana, S. O.; Yan, P.; Sun, W.; Patel, M. R.; Shah, S. K.; Tramentozzi, E.; Brooks, J.; Bolaender, A.; Shrestha, L.; Stephani, R.; Finotti, P.; Leifer, C.; Li, Z.; Gewirth, D. T.; Taldone, T.; Chiosis, G. Structure-activity relationship in a purinescaffold compound series with selectivity for the endoplasmic reticulum HSP90 paralog Grp94. J. Med. Chem. 2015, 58, 3922−3943.

(41) Hardy, J. A.; Wells, J. A. Searching for new allosteric sites in enzymes. Curr. Opin. Struct. Biol. 2004, 14, 706−715. (42) Hilser, V. J. An ensemble view of allostery. Science 2010, 327, 653−654. (43) Panjkovich, A.; Daura, X. Assessing the structural conservation of protein pockets to study functional and allosteric sites: implications for drug discovery. BMC Struct. Biol. 2010, 10, 9. (44) Wei, G.; Xi, W.; Nussinov, R.; Ma, B. Protein ensembles: how does nature harness thermodynamic fluctuations for life? The diverse functional roles of conformational ensembles in the cell. Chem. Rev. 2016, 116, 6516−6551. (45) Panjkovich, A.; Daura, X. Exploiting protein flexibility to predict the location of allosteric sites. BMC Bioinf. 2012, 13, 273. (46) Boehr, D. D.; Nussinov, R.; Wright, P. E. The role of dynamic conformational ensembles in biomolecular recongition. Nat. Chem. Biol. 2009, 5, 789−796. (47) Morra, G.; Potestio, R.; Micheletti, C.; Colombo, G. Corresponding functional dynamics across the HSP90 chaperone family: insights from a multiscale analysis of md simulations. PLoS Comput. Biol. 2012, 8, e1002433. (48) Monod, J.; Wyman, J.; Changeux, J. P. On the nature of allosteric transitions-a plausible model. J. Mol. Biol. 1965, 12, 88−118. (49) Yu, E. W.; Koshland, D. E., Jr. Propagating conformational changes over long (and short) distances in proteins. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 9517−9520. (50) Cui, Q.; Karplus, M. Allostery and cooperativity revisited. Protein Sci. 2008, 17, 1295−1307. (51) Hilser, V. J.; Wrabl, J. O.; Motlagh, H. N. Structural and energetic basis of allostery. Annu. Rev. Biophys. 2012, 41, 585−609. (52) McClellan, A. J.; Xia, Y.; Deutschbauer, A. M.; Davis, R. W.; Gerstein, M.; Frydman, J. Diverse cellular functions of the HSP90 molecular chaperone uncovered using systems approaches. Cell 2007, 131, 121−135. (53) Cortes, S.; Baker-Williams, A. J.; Mollapour, M.; Bourboulia, D. Detection and analysis of extracellular HSP90 (eHSP90). Methods Mol. Biol. 2018, 1709, 321−329. (54) Zhang, G.; Liu, Z.; Ding, H.; Zhou, Y.; Doan, H. A.; Sin, K. W. T.; Zhu, Z. J.; Flores, R.; Wen, Y.; Gong, X.; Liu, Q.; Li, Y.-P. Tumor induces muscle wasting in mice through releasing extracellular HSP70 and HSP90. Nat. Commun. 2017, 8, 589. (55) Roy, N.; Nageshan, R. K.; Ranade, S.; Tatu, U. Heat shock protein 90 from neglected protozoan parasites. Biochim. Biophys. Acta, Mol. Cell Res. 2012, 1823, 707−711. (56) Neckers, L.; Tatu, U. Molecular chaperones in pathogen virulence: emerging new targets for therapy. Cell Host Microbe 2008, 4, 519−527. (57) Ali, M. M. U.; Roe, S. M.; Vaughan, C. K.; Meyer, P.; Panaretou, B.; Piper, P. W.; Prodromou, C.; Pearl, L. H. Crystal structure of an HSP90-nucleotide-p23/Sba1 closed chaperone complex. Nature 2006, 440, 1013−1017. (58) Dollins, D. E.; Warren, J. J.; Immormino, R. M.; Gewirth, D. T. Structures of GRP94-Nucleotide complexes reveal mechanistic differences between the HSP90 chaperones. Mol. Cell 2007, 28, 41−56. (59) Lavery, L. A.; Partridge, J. R.; Ramelot, T. A.; Elnatan, D.; Kennedy, M. A.; Agard, D. A. Structural asimmetry in the closed state of mitochondrial HSP90 (TRAP1) supports a two-step ATP hydrolysis mechanism. Mol. Cell 2014, 53, 330−343. (60) Southworth, D. R.; Agard, D. A. Species-Dependent ensembles of conserved conformational states define the HSP90 chaperone ATPase cycle. Mol. Cell 2008, 32, 631−640. (61) Sciacovelli, M.; Guzzo, G.; Morello, V.; Frezza, C.; Zheng, L.; Nannini, N.; Calabrese, F.; Laudiero, G.; Esposito, F.; Landriscina, M.; Defilippi, P.; Bernardi, P.; Rasola, A. The mitochondrial chaperone TRAP1 promotes neoplastic growth by inhibiting succinate dehydrogenase. Cell Metab. 2013, 17, 988−999. (62) Rasola, A.; Sciacovelli, M.; Chiara, F.; Pantic, B.; Brusilow, W. S.; Bernardi, P. Activation of mitochondrial ERK protects cancer cells 83



Perspective

(81) Davenport, J.; Manjarrez, J. R.; Peterson, L.; Krumm, B.; Blagg, B. S. J.; Matts, R. L. Gambogic Acid, a natural product inhibitor of HSP90. J. Nat. Prod. 2011, 74, 1085−1092. (82) Marcu, M. G.; Schulte, T. W.; Neckers, L. Novobiocin and related coumarins and depletion of heat shock protein 90-dependent signaling proteins. JNCI 2000, 92, 242−248. (83) Marcu, M. G.; Chadli, A.; Bouhouche, I.; Catelli, M.; Neckers, L. M. The heat shock protein 90 antagonist Novobiocin interacts with a previously unrecognized ATP-binding domain in the carboxyl terminus of the chaperone. J. Biol. Chem. 2000, 275, 37181−37186. (84) Kijima, T.; Prince, T. L.; Tigue, M. L.; Yim, K. H.; Schwartz, H.; Beebe, K.; Lee, S.; Budzynski, M. A.; Williams, H.; Trepel, J. B.; Sistonen, L.; Calderwood, S.; Neckers, L. HSP90 inhibitors disrupt a transient HSP90-HSF1 interaction and identify a noncanonical model of HSP90-mediated HSF1 regulation. Sci. Rep. 2018, 8, 6976. (85) Hessling, M.; Richter, K.; Buchner, J. Dissection of the ATPinduced conformational cycle of the molecular chaperone HSP90. Nat. Struct. Mol. Biol. 2009, 16, 287−293. (86) Burlison, J. A.; Neckers, L.; Smith, A. B.; Maxwell, A.; Blagg, B. S. J. Novobiocin: redesigning a dna gyrase inhibitor for selective inhibition of HSP90. J. Am. Chem. Soc. 2006, 128, 15529−15536. (87) Yu, X. M.; Shen, G.; Neckers, L.; Blake, H.; Holzbeierlein, J.; Cronk, B.; Blagg, B. S. J. HSP90 inhibitors identified from a library of novobiocin analogues. J. Am. Chem. Soc. 2005, 127, 12778−12779. (88) Zhao, H.; Michaelis, M. L.; Blagg, B. S. J. HSP90 modulation for the treatment of Alzheimer’s disease. Adv. Pharmacol. 2012, 64, 1− 25. (89) Lu, Y.; Ansar, S.; Michaelis, M. L.; Blagg, B. S. J. Neuroprotective activity and evaluation of HSP90 inhibitors in an immortalized neuronal cell line. Bioorg. Med. Chem. 2009, 17, 1709− 1715. (90) Forsberg, L. K.; Anyika, M.; You, Z.; Emery, S.; McMullen, M.; Dobrowsky, R. T.; Blagg, B. S. J. Development of noviomimetics that modulate molecular chaperones and manifest neuroprotective effects. Eur. J. Med. Chem. 2018, 143, 1428−1435. (91) Donnelly, A.; Blagg, B. S. J. Novobiocin and additional inhibitors of the HSP90 C-Terminal Nucleotide-binding Pocket. Curr. Med. Chem. 2008, 15, 2702−2717. (92) Donnelly, A. C.; Mays, J. R.; Burlison, J. A.; Nelson, J. T.; Vielhauer, G.; Holzbeierlein, J.; Blagg, B. S. The design, synthesis, and evaluation of coumarin ring derivatives of the novobiocin scaffold that exhibit antiproliferative activity. J. Org. Chem. 2008, 73, 8901−8920. (93) Shelton, S. N.; Shawgo, M. E.; Matthews, S. B.; Lu, Y.; Donnelly, A. C.; Szabla, K.; Tanol, M.; Vielhauer, G. A.; Rajewski, R. A.; Matts, R. L.; Blagg, B. S. J.; Robertson, J. D. KU135, a novel novobiocin-derived C-terminal inhibitor of the 90-kDa Heat Shock Protein, exerts potent anitproliferative effects in human leukemic cells. Mol. Pharmacol. 2009, 76, 1314−1322. (94) Zhao, H. P.; Donnelly, A. C.; Kusuma, B. R.; Brandt, G. E. L.; Brown, D.; Rajewski, R. A.; Vielhauer, G.; Holzbeierlein, J.; Cohen, M. S.; Blagg, B. S. J. Engineering an antibiotic to fight cancer: optimization of the Novobiocin scaffold to produce anti-proliferative agents. J. Med. Chem. 2011, 54, 3839−3853. (95) Zhao, H.; Garg, G.; Zhao, J.; Moroni, E.; Girgis, A.; Franco, L. S.; Singh, S.; Colombo, G.; Blagg, B. S. Design, synthesis and biological evaluation of biphenylamide derivatives as HSP90 Cterminal inhibitors. Eur. J. Med. Chem. 2015, 89, 442−466. (96) Garg, G.; Forsberg, L. K.; Zhao, H.; Blagg, S. J. Development of phenyl cyclohexylcarboxamides as a novel class of HSP90 C-terminal inhibitors. Chem. - Eur. J. 2017, 23, 16574−16585. (97) Ansar, S.; Burlison, J. A.; Hadden, M. K.; Yu, X. M.; Desino, K. E.; Bean, J.; Neckers, L.; Audus, K. L.; Michaelis, M. L.; Blagg, B. S. J. A non-toxic HSP90 inhibitor protects neurons from Abeta-induced toxicity. Bioorg. Med. Chem. Lett. 2007, 17, 1984−1990. (98) Kusuma, B. R.; Zhang, L.; Sundstrom, T.; Peterson, L. B.; Dobrowsky, R. T.; Blagg, B. S. J. Synthesis and evaluation of novologues as C-terminal HSP90 inhibitors with cytoprotective activity against sensory neuron glucotoxicity. J. Med. Chem. 2012, 55, 5797−5812.

(99) Zhang, L.; Zhao, H.; Blagg, B. S. J.; Dobrowsky, R. T. Cterminal heat shock protein 90 inhibitor decreases hyperglycemiainduced oxidative stress and improves mitochondrial bioenergetics in sensory neurons. J. Proteome Res. 2012, 11, 2581−2593. (100) Urban, M. J.; Li, C.; Yu, C.; Lu, Y.; Krise, J. M.; McIntosh, M. P.; Rajewski, R. A.; Blagg, B. S. J.; Dobrowsky, R. T. Inhibiting heatshock protein 90 reverses sensory hypoalgesia in diabetic mice. ASN Neuro 2010, 2, e00040. (101) Ma, J.; Pan, P.; Anyika, M.; Blagg, B. S. J.; Dobrowsky, R. T. Modulating molecular chaperones improves mitochondrial bioenergetics and decreases the inflammatory transcriptome in diabetic sensory neurons. ACS Chem. Neurosci. 2015, 6, 1637−1648. (102) Zhang, X.; Li, C.; Fowler, S. C.; Zhang, Z.; Blagg, B. S. J.; Dobrowsky, R. T. Targeting Heat Shock Protein 70 to ameliorate cJun expression and improve demyelinating neuropathy. ACS Chem. Neurosci. 2018, 9, 381−390. (103) Yin, Z.; Henry, E. C.; Gasiewicz, T. A. (−)-Epigallocatechin-3gallate is a novel HSP90 inhibitor. Biochemistry 2009, 48, 336−345. (104) Khandelwal, A.; Hall, J.; Blagg, B. S. J. Synthesis and Structure activity relationships of EGCG analogues, a recently identified HSP90 inhibitor. J. Org. Chem. 2013, 78, 7859−7884. (105) Burlison, J. A.; Blagg, B. S. J. Synthesis and evaluation of Coumermycin A1 analogues that inhibit the HSP90 protein folding machinery. Org. Lett. 2006, 8, 4855−4858. (106) Lee, S.-C.; MIn, H.-Y.; Choi, H.; Kim, H. S.; Kim, K.-C.; Park, S.-J.; Seong, M. A.; Seo, J. H.; Park, H. J.; Suh, Y. G.; Kim, K. W.; Hong, H. S.; Kim, H.; Lee, M.-Y.; Lee, J.; Lee, H.-Y. Synthesis and evaluation of a novel Deguelin derivative, L80, which disrupts ATP binding to the C-terminal domain of Heat Shock Protein 90. Mol. Pharmacol. 2015, 88, 245−255. (107) Yi, F.; Regan, L. A novel class of small molecule inhibitors of HSP90. ACS Chem. Biol. 2008, 3, 645−654. (108) Yi, F.; Zhu, P.; Southall, N.; Inglese, J.; Austin, C. P.; Zheng, W.; Regan, L. An AlphaScreen-based high-throughput screen to identify inhibitors of HSP90-cochaperone interaction. J. Biomol. Screening 2009, 14, 273−281. (109) Kajander, T.; Sachs, J. N.; Goldman, A.; Regan, L. Electrostatic interactions of Hsp-organizing protein tetratricopeptide domains with HSP70 and HSP90: computational analysis and protein engineering. J. Biol. Chem. 2009, 284, 25364−25374. (110) Pimienta, G.; Herbert, K. M.; Regan, L. A compound that inhibits the HOP-HSP90 complex formation and has unique killing effects in breast cancer cell lines. Mol. Pharmaceutics 2011, 8, 2252− 2261. (111) Speltz, E. B.; Nathan, A.; Regan, L. Design of protein-peptide interaction modules for assembling supramolecular structures in vivo and in vitro. ACS Chem. Biol. 2015, 10, 2108−2115. (112) Peppard, J.; Glickman, F.; He, Y.; Hu, S. I.; Doughty, J.; Goldberg, R. Development of a high-throughput screening assay for inhibitors of aggrecan cleavage using luminescent oxygen channeling (AlphaScreen). J. Biomol. Screening 2003, 8, 149−156. (113) Vasko, R. C.; Rodriguez, R. A.; Cunningham, C. N.; Ardi, V. C.; Agard, D. A.; McAlpine, S. R. Mechanistic studies of Sansalvamide A-Amide: an allosteric modulator of HSP90. ACS Med. Chem. Lett. 2010, 1, 4−8. (114) Kunicki, J. B.; Petersen, M. N.; Alexander, L. D.; Ardi, V. C.; McConnell, J. R.; McAlpine, S. R. Synthesis and evaluation of biotinylated Sansalvamide A analogs and their modulation of HSP90. Bioorg. Med. Chem. Lett. 2011, 21, 4716−4719. (115) Alexander, L. D.; Partridge, J. R.; Agard, D. A.; McAlpine, S. R. A small molecule that preferentially binds the closed conformation of HSP90. Bioorg. Med. Chem. Lett. 2011, 21, 7068−7071. (116) Wang, Y.; McAlpine, S. R. C-terminal Heat shock protein 90 modulators produce desirable oncogenic properties. Org. Biomol. Chem. 2015, 13, 4627−4631. (117) Koay, Y. C.; Wahyudi, H.; McAlpine, S. R. Reinventing HSP90 inhibitors: blocking C-Terminal binding events to HSP90 by using dimerized inhibitors. Chem. - Eur. J. 2016, 22, 18572−18582. 84



Perspective

(118) Koay, Y. C.; McConnell, J. R.; Wang, Y.; Kim, S. J.; Buckton, L. K.; Mansour, F.; McAlpine, S. R. Chemically accessible HSP90 inhibitor that does not induce a heat shock response. ACS Med. Chem. Lett. 2014, 5 (7), 771−776. (119) Aldrich, C.; Bertozzi, C.; Georg, G. I.; Kiessling, L.; Lindsley, C.; Liotta, D.; Merz, K. M.; Schepartz, A.; Wang, S. The ecstasy and agony of assay interference compounds. ACS Infect. Dis. 2017, 3, 259−262. (120) Zorn, J. A.; Wells, J. A. Turning enzymes ON with small molecules. Nat. Chem. Biol. 2010, 6, 179−188. (121) Yokoyama, Y.; Ohtaki, A.; Jantan, I.; Yohda, M.; Nakamoto, H. Goniothalamin enhances the ATPase activity of the molecular chaperone HSP90 but inhibits its chaperone activity. J. Biochem. 2015, 157, 161−168. (122) Zierer, B. K.; Weiwad, M.; Rubbelke, M.; Freiburger, L.; Fischer, G.; Lorenz, O. R.; Sattler, M.; Richter, K.; Buchner, J. Artificial accelerators of the molecular chaperone HSP90 facilitate rate-limiting conformational transitions. Angew. Chem., Int. Ed. 2014, 53, 12257−12262. (123) Sattin, S.; Tao, J. H.; Vettoretti, G.; Moroni, E.; Pennati, M.; Lopergolo, A.; Morelli, L.; Bugatti, A.; Zuehlke, A.; Moses, M.; Prince, T.; Kijima, T.; Beebe, K.; Rusnati, M.; Neckers, L.; Zaffaroni, N.; Agard, D. A.; Bernardi, A.; Colombo, G. Activation of HSP90 enzymatic activity and conformational dynamics through rationally designed allosteric ligands. Chem. - Eur. J. 2015, 21, 13598−13608. (124) D’Annessa, I.; Sattin, S.; Tao, J.; Pennati, M.; Sànchez-Martìn, C.; Moroni, E.; Rasola, A.; Zaffaroni, N.; Agard, D. A.; Bernardi, A.; Colombo, G. Design of allosteric stimulators of the HSP90 ATPase as novel anticancer leads. Chem. - Eur. J. 2017, 23 (22), 5188−5192. (125) Weikl, T.; Muschler, P.; Richter, K.; Veit, T.; Reinstein, J.; Buchner, J. C-terminal regions of HSP90 are important for trapping the nucleotide during the ATPase cycle. J. Mol. Biol. 2000, 303, 583− 592. (126) Morra, G.; Neves, M. A. C.; Plescia, C. J.; Tsustsumi, S.; Neckers, L.; Verkhivker, G.; Altieri, D. C.; Colombo, G. DynamicsBased discovery of allosteric inhibitors: selection of new ligands for the C-terminal domain of HSP90. J. Chem. Theory Comput. 2010, 6, 2978−2989. (127) Morra, G.; Verkhivker, G. M.; Colombo, G. Modeling signal propagation mechanisms and ligand-based conformational dynamics of the HSP90 molecular chaperone full length dimer. PLoS Comput. Biol. 2009, 5, e1000323. (128) Rehn, A.; Moroni, E.; Zierer, B. K.; Tippel, F.; Morra, G.; John, C.; Richter, K.; Colombo, G.; Buchner, J. Allosteric regulation points control the conformational dynamics of the molecular chaperone HSP90. J. Mol. Biol. 2016, 428, 4559−4571. (129) Street, T. O.; Lavery, L. A.; Agard, D. A. Substrate binding drives large-scale conformational changes in the HSP90 molecular chaperone. Mol. Cell 2011, 42, 96−105. (130) Vettoretti, G.; Moroni, E.; Sattin, S.; Tao, J. H.; Agard, D. A.; Bernardi, A.; Colombo, G. Molecular dynamics simulations reveal the mechanisms of allosteric activation of HSP90 by designed ligands. Sci. Rep. 2016, 6, 23830. (131) Bassanini, I.; D’Annessa, I.; Costa, M.; Monti, D.; Colombo, G.; Riva, S. Chemo-enzymatic synthesis of (E)-2,3-diaryl-5-styryltrans-2,3-dihydrobenzofuran-based scaffolds and their in vitro and in silico evaluation as a novel sub-family of potential allosteric modulators of the 90 kDa heat shock protein (HSP90). Org. Biomol. Chem. 2018, 16, 3741−3753. (132) Zierer, B. K.; Rubbelke, M.; Tippel, F.; Madl, T.; Schopf, F. H.; Rutz, D. A.; Richter, K.; Sattler, M.; Buchner, J. Importance of cycle timing for the function of the molecular chaperone HSP90. Nat. Struct. Mol. Biol. 2016, 23, 1020−1028. (133) Roe, M. S.; Wahab, B.; Török, Z.; Horváth, I.; Vigh, L.; Prodromou, C. Dihydropyridines allosterically modulate HSP90 providing a novel mechanism for heat shock protein co-induction and neuroprotection. Front. Mol. Biosci. 2018, 5, 51.

(134) Pullen, L.; Bolon, D. N. Enforced N-domain proximity stimulates HSP90 ATPase activity and is compatible with function in vivo. J. Biol. Chem. 2011, 286, 11091−11098. (135) Beebe, K.; Mollapour, M.; Scroggins, B.; Prodromou, C.; Xu, W.; Tokita, M.; Taldone, T.; Pullen, L.; Zierer, B. K.; Lee, M. J.; Trepel, J.; Buchner, J.; Bolon, D. N.; Chiosis, G.; Neckers, L. Posttranslational modification and conformational state of heat shock protein 90 differentially affect binding of chemically diverse small molecule inhibitors. Oncotarget 2013, 4, 1065−1074. (136) Daugaard, M.; Rohde, M.; Jaattela, M. The heat shock protein 70 family: highly homologous proteins with overlapping and distinct functions. FEBS Lett. 2007, 581, 3702−3710. (137) Kityk, R.; Kopp, J.; Sinning, I.; Mayer, M. P. Structure and dynamics of the ATP-bound open conformation of HSP70 chaperones. Mol. Cell 2012, 48, 863−874. (138) Mayer, M. P.; Bukau, B. HSP70 chaperones: cellular functions and molecular mechanism. Cell. Mol. Life Sci. 2005, 62, 670−684. (139) Kampinga, H. H.; Craig, E. A. The HSP70 chaperone machinery: J proteins as drivers of functional specificity. Nat. Rev. Mol. Cell Biol. 2010, 11, 579−592. (140) Bracher, A.; Verghese, J. GrpE, Hsp110/Grp170, Hspbp1/sil1 and bag domain proteins: nucleotide exchange factors for HSP70 molecular chaperones. Subcell. Biochem. 2015, 78, 1−33. (141) Li, X.; Colvin, T.; Rauch, J. N.; Acosta-Alvear, D.; Kampmann, M.; Dunyak, B.; Hann, B.; Aftab, B. T.; Murnane, M.; Cho, M.; Walter, P.; Weissman, J. S.; Sherman, M. Y.; Gestwicki, J. E. Validation of the HSP70−Bag3 protein−protein interaction as a potential therapeutic target in Cancer. Mol. Cancer Ther. 2015, 14, 642−648. (142) Soti, C.; Csermely, P. Chaperones and aging: role in neurodegeneration and in other civilizational diseases. Neurochem. Int. 2002, 41, 383−389. (143) Powers, M. V.; Jones, K.; Barillari, C.; Westwood, I.; Montfort, R. L. M. v.; Workman, P. Targeting HSP70: the second potentially druggable heat shock protein and molecular chaperone? Cell Cycle 2010, 9, 1542−1550. (144) Avila, J.; Lucas, J. J.; Perez, M.; Hernandez, F. Role of tau protein in both physiological and pathological conditions. Physiol. Rev. 2004, 84, 361−384. (145) Jinwal, U. K.; Miyata, Y.; Koren, J.; Jones, J. R.; Trotter, J. H.; Chang, L.; O'Leary, J.; Morgan, D.; Lee, D. C.; Shults, C. L.; Rousaki, A.; Weeber, E. J.; Zuiderweg, E. R. P.; Gestwicki, J. E.; Dickey, C. A. Chemical manipulation of HSP70 atpase activity regulates tau stability. J. Neurosci. 2009, 29, 12079−12088. (146) Shi, Y.; Mosser, D. D.; Morimoto, R. I. Molecular chaperones as HSF1-specific transcriptional repressors. Genes Dev. 1998, 12, 654− 666. (147) Karagoz, G. E.; Rudiger, S. G. HSP90 interaction with clients. Trends Biochem. Sci. 2015, 40, 117−125. (148) Mendillo, M. L.; Santagata, S.; Koeva, M.; Bell, G. W.; Hu, R.; Tamimi, R. M.; Fraenkel, E.; Ince, T. A.; Whitesell, L.; Lindquist, S. HSF1 drives a transcriptional program distinct from heat shock to support highly malignant human cancers. Cell 2012, 150, 549−562. (149) Le Breton, L.; Mayer, M. P. A model for handling cell stress. eLife 2016, 5, e22850. (150) Assimon, V. A.; Gillies, A. T.; Rauch, J. N.; Gestwicki, J. E. HSP70 protein complexes as drug targets. Curr. Pharm. Des. 2013, 19, 404−417. (151) Zheng, X.; Krakowiak, J.; Patel, N.; Beyzavi, A.; Ezike, J.; Khalil, A. S.; Pincus, D. Dynamic control of HSF1 during heat shock by a chaperone switch and phosphorylation. eLife 2016, 5, e18638. (152) Krakowiak, J.; Zheng, X.; Patel, N.; Feder, Z. A.; Anandhakumar, J.; Valerius, K.; Gross, D. S.; Khalil, A. S.; Pincus, D. HSF1 and HSP70 constitute a two-component feedback loop that regulates the yeast heat shock response. eLife 2018, 7, e31668. (153) Howe, M. K.; Bodoor, K.; Carlson, D. A.; Hughes, P. F.; Alwarawrah, Y.; Loiselle, D. R.; Jaeger, A. M.; Darr, D. B.; Jordan, J. L.; Hunter, L. M.; Molzberger, E. T.; Gobillot, T. A.; Thiele, D. J.; Brodsky, J. L.; Spector, N. L.; Haystead, T. A. J. Identification of an 85



Perspective

allosteric small-molecule inhibitor selective for the inducible form of Heat Shock Protein 70. Chem. Biol. 2014, 21, 1648−1659. (154) Mayer, M. P. Gymnastics of molecular chaperones. Mol. Cell 2010, 39, 321−331. (155) Rerole, A.-L.; Jego, G.; Garrido, C. HSP70: anti-apoptotic and tumorigenic protein. Methods Mol. Biol. 2011, 787, 205−230. (156) Chang, L.; Miyata, Y.; Ung, P. M. U.; Bertelsen, E. B.; McQuade, T. J.; Carlson, H. A.; Zuiderweg, E. R. P.; Gestwicki, J. E. Chemical screens against a reconstituted multiprotein complex: Myricetin blocks DnaJ regulation of DnaK through an allosteric mechanism. Chem. Biol. 2011, 18, 210−221. (157) Miyata, Y.; Chang, L.; Bainor, A.; McQuade, T. J.; Walczak, C. P.; Zhang, Y.; Larsen, M. J.; Kirchhoff, P.; Gestwicki, J. E. High throughput screen for Escherichia coli heat shock protein 70 (HSP70/ dnak): atpase assay in low volume by exploiting energy transfer. J. Biomol. Screening 2010, 15, 1211−1219. (158) Chang, L.; Bertelsen, E. B.; Wisén, S.; Larsen, E. M.; Zuiderweg, E. R.; Gestwicki, J. E. High-throughput screen for small molecules that modulate the ATPase activity of the molecular chaperone DnaK. Anal. Biochem. 2008, 372, 167−176. (159) Wisen, S.; Bertelsen, E. B.; Thompson, A. D.; Patury, S.; Ung, P.; Chang, L.; Evans, C. G.; Walter, G. M.; Wipf, P.; Carlson, H. A.; Brodsky, J. L.; Zuiderweg, E. R. P.; Gestwicki, J. E. Binding of a small molecule at a protein−protein interface regulates the chaperone activity of HSP70−Hsp40. ACS Chem. Biol. 2010, 5, 611−622. (160) Wisen, S.; Gestwicki, J. E. Identification of small molecules that modify the protein folding activity of heat shock protein 70. Anal. Biochem. 2008, 374 (2), 371−377. (161) Wright, C. M.; Chovatiya, R. J.; Jameson, N. E.; Turner, D. M.; Zhu, G.; Werner, S.; Huryn, D. M.; Pipas, J. M.; Day, B. W.; Wipf, P.; Brodsky, J. L. Pyrimidinone-peptoid hybrid molecules with distinct effects on molecular chaperone function and cell proliferation. Bioorg. Med. Chem. 2008, 16, 3291−3301. (162) Dickey, C. A.; Eriksen, J.; Kamal, A.; Burrows, F.; Kasibhatla, S.; Eckman, C. B.; Hutton, M.; Petrucelli, L. Development of a high throughput drug screening assay for the detection of changes in tau levels − proof of concept with HSP90 inhibitors. Curr. Alzheimer Res. 2005, 2, 231−238. (163) Wadhwa, R.; Sugihara, T.; Yoshida, A.; Nomura, H.; Reddel, R. R.; Simpson, R.; Maruta, H.; Kaul, S. C. Selective toxicity of MKT077 to cancer cells is mediated by its binding to the HSP70 family protein mot-2 and reactivation of p53 function. Cancer Res. 2000, 60, 6818−6821. (164) Propper, D.; Braybrooke, J.; Taylor, D.; Lodi, R.; Styles, P.; Cramer, J.; Collins, W.; Levitt, N.; Talbot, D.; Ganesan, T.; Harris, A. Phase I trial of the selective mitochondrial toxin MKT 077 in chemoresistant solid tumours. Annal of Oncology 1999, 10, 923−927. (165) Rousaki, A.; Miyata, Y.; Jinwal, U. K.; Dickey, C. A.; Gestwicki, J. E.; Zuiderweg, E. R. P. Allosteric drugs: the interaction of antitumor compound MKT-077 with human HSP70 chaperones. J. Mol. Biol. 2011, 411, 614−632. (166) Miyata, Y.; Li, X.; Lee, H. F.; Jinwal, U. K.; Srinivasan, S. R.; Seguin, S. P.; Young, Z. T.; Brodsky, J. L.; Dickey, C. A.; Sun, D.; Gestwicki, J. E. Synthesis and initial evaluation of YM-08, a bloodbrain barrier permeable derivative of the heat shock protein 70 (HSP70) inhibitor MKT-077, which reduces tau levels. ACS Chem. Neurosci. 2013, 4, 930−939. (167) Morishima, Y.; Lau, M.; Peng, H.-M.; Miyata, Y.; Gestwicki, J. E.; Pratt, W. B.; Osawa, Y. Heme-dependent activation of neuronal nitric-oxide synthase by cytosol is due to an HSP70-dependent, thioredoxin-mediated thiol-disulfide interchange in the heme/ substrate binding cleft. Biochemistry 2011, 50, 7146−7156. (168) Li, X.; Srinivasan, S. R.; Connarn, J.; Ahmad, A.; Young, Z. T.; Kabza, A. M.; Zuiderweg, E. R. P.; Sun, D.; Gestwicki, J. E. Analogues of the allosteric Heat Shock Protein 70 (HSP70) inhibitor, MKT-077, as anti-cancer agents. ACS Med. Chem. Lett. 2013, 4, 1042−1047. (169) Rodina, A.; Patel, P. D.; Kang, Y.; Patel, Y.; Baaklini, I.; Wong, M. J.; Taldone, T.; Yan, P.; Yang, C.; Maharaj, R.; Gozman, A.; Patel, M. R.; Patel, H. J.; Chirico, W.; Erdjument-Bromage, H.; Talele, T.

T.; Young, J. C.; Chiosis, G. Identification of an allosteric pocket on human HSP70 reveals a mode of inhibition of this therapeutically important protein. Chem. Biol. 2013, 20, 1469−1480. (170) Kang, Y.; Taldone, T.; Patel, H. J.; Patel, P. D.; Rodina, A.; Gozman, A.; Maharaj, R.; Clement, C. C.; Patel, M. R.; Brodsky, J. L.; Young, J. C.; Chiosis, G. Heat Shock Protein 70 Inhibitors. 1. 2,5′Thiodipyrimidine and 5-(Phenylthio)pyrimidine Acrylamides as irreversible binders to an allosteric site on Heat Shock Protein 70. J. Med. Chem. 2014, 57, 1188−1207. (171) Taldone, T.; Kang, Y.; Patel, H. J.; Patel, M. R.; Patel, P. D.; Rodina, A.; Patel, Y.; Gozman, A.; Maharaj, R.; Clement, C. C.; Lu, A.; Young, J. C.; Chiosis, G. Heat Shock Protein 70 Inhibitors. 2. 2,5′Thiodipyrimidines, 5-(Phenylthio)pyrimidines, 2-(Pyridin-3-ylthio)pyrimidines, and 3-(Phenylthio)pyridines as reversible binders to an allosteric site on Heat Shock Protein 70. J. Med. Chem. 2014, 57, 1208−1224. (172) Rodina, A.; Taldone, T.; Kang, Y.; Patel, P. D.; Koren, J.; Yan, P.; DaGama Gomes, E. M.; Yang, C.; Patel, M. R.; Shrestha, L.; Ochiana, S. O.; Santarossa, C.; Maharaj, R.; Gozman, A.; Cox, M. B.; Erdjument-Bromage, H.; Hendrickson, R. C.; Cerchietti, L.; Melnick, A.; Guzman, M. L.; Chiosis, G. Affinity purification probes of potential use to investigate the endogenous HSP70 interactome in cancer. ACS Chem. Biol. 2014, 9, 1698−1705. (173) Hassan, A. Q.; Kirby, C. A.; Zhou, W.; Schuhmann, T.; Kityk, R.; Kipp, D. R.; Baird, J.; Chen, J.; Chen, Y.; Chung, F.; Hoepfner, D.; Movva, N. R.; Pagliarini, R.; Petersen, F.; Quinn, C.; Quinn, D.; Riedl, R.; Schmitt, E. K.; Schitter, A.; Stams, T.; Studer, C.; Fortin, P. D.; Mayer, M. P.; Sadlish, H. The Novolactone natural product disrupts the allosteric regulation of HSP70. Chem. Biol. 2015, 22, 87−97. (174) Leu, J. I. J.; Pimkina, J.; Frank, A.; Murphy, M. E.; George, D. L. A small molecule inhibitor of inducible Heat Shock Protein 70 (HSP70). Mol. Cell 2009, 36, 15−27. (175) Leu, J. I-J.; Zhang, P.; Murphy, M. E.; Marmorstein, R.; George, D. L. Structural basis for the inhibition of HSP70 and DnaK chaperones by small-molecule targeting of a C-terminal allosteric pocket. ACS Chem. Biol. 2014, 9, 2508−2516. (176) Verba, K. A.; Wang, R. Y.; Arakawa, A.; Liu, Y.; Shirouzu, M.; Yokoyama, S.; Agard, D. A. Atomic structure of HSP90-Cdc37-Cdk4 reveals that HSP90 traps and stabilizes an unfolded kinase. Science 2016, 352, 1542−1547. (177) Kumar Mv, V.; Ebna Noor, R.; Davis, R. E.; Zhang, Z.; Sipavicius, E.; Keramisanou, D.; Blagg, B. S. J.; Gelis, I. Molecular insights into the interaction of HSP90 with allosteric inhibitors targeting the C-terminal domain. MedChemComm 2018, DOI: 10.1039/C8MD00151K. (178) Sattin, S.; Panza, M.; Vasile, F.; Berni, F.; Goti, G.; Tao, J. H.; Moroni, E.; Agard, D.; Colombo, G.; Bernardi, A. Synthesis of functionalized 2-(4-hydroxyphenyl)-3-methylbenzofuran allosteric modulators of HSP90 activity. Eur. J. Org. Chem. 2016, 2016, 3349−3364. (179) Matts, R. L.; Dixit, A.; Peterson, L. B.; Sun, L.; Voruganti, S.; Kalyanaraman, P.; Hartson, S. D.; Verkhivker, G. M.; Blagg, B. S. J. Elucidation of the HSP90 C-terminal inhibitor binding site. ACS Chem. Biol. 2011, 6, 800−807. (180) van Gunsteren, W. F.; Bakowies, D.; Baron, R.; Chandrasekhar, I.; Christen, M.; Daura, X.; Gee, P.; Geerke, D. P.; Glättli, A.; Hünenberger, P. H.; Kastenholz, M. A.; Oostenbrink, C.; Schenk, M.; Trzesniak, D.; van der Vegt, N. F. A.; Yu, H. B. Biomolecular modeling: goals, problems, perspectives. Angew. Chem., Int. Ed. 2006, 45, 4064−4092. (181) van Gunsteren, W. F.; Dolenc, J.; Mark, A. Molecular simulation as an aid to experimentalists. Curr. Opin. Struct. Biol. 2008, 18, 149−153. (182) Pontiggia, F.; Colombo, G.; Micheletti, C.; Orland, H. Anharmonicity and self-similarity of the free energy landscape of protein G. Phys. Rev. Lett. 2007, 98, 048102. (183) Monticelli, L.; Tieleman, D. P.; Colombo, G. Mechanism of helix nucleation and propagation: microscopic view from microsecond 86



Perspective

time scale MD simulations. J. Phys. Chem. B 2005, 109, 20064− 20067. (184) Meli, M.; Morra, G.; Colombo, G. Investigating the mechanism of peptide aggregation: insights from mixed Monte Carlo-molecular dynamics simulations. Biophys. J. 2008, 94, 4414− 4426. (185) Retzlaff, M.; Stahl, M.; Eberl, H. C.; Lagleder, S.; Beck, J.; Kessler, H.; Buchner, J. HSP90 is regulated by a switch point in the Cterminal domain. EMBO Rep. 2009, 10, 1147−1153. (186) Seifert, C.; Gräter, F. Force distribution reveals signal transduction in E. coli HSP90. Biophys. J. 2012, 103, 2195−2202. (187) Seifert, C.; Grater, F. Protein mechanics: how force regulates molecular function. Biochim. Biophys. Acta, Gen. Subj. 2013, 1830, 4762−4768. (188) Elnatan, D.; Betegon, M.; Liu, Y.; Ramelot, T.; Kennedy, M. A.; Agard, D. A. Symmetry broken and rebroken during the ATP hydrolysis cycle of the mitochondrial HSP90 TRAP1. eLife 2017, 6, e25235. (189) Moroni, E.; Agard, D. A.; Colombo, G. The structural asymmetry of mitochondrial HSP90 (TRAP1) determines fine tuning of functional dynamics. J. Chem. Theory Comput. 2018, 14, 1033− 1044. (190) Blacklock, K.; Verkhivker, G. M. Differential modulation of functional dynamics and allosteric interactions in the HSP90cochaperone complexes with p23 and Aha1: a computational study. PLoS One 2013, 8, e71936. (191) Blacklock, K.; Verkhivker, G. M. computational modeling of allosteric regulation in the HSP90 chaperones: a statistical ensemble analysis of protein structure networks and allosteric communications. PLoS Comput. Biol. 2014, 10, e1003679. (192) Czemeres, J.; Buse, K.; Verkhivker, G. M. Atomistic simulations and network-based modeling of the HSP90-Cdc37 chaperone binding with Cdk4 client protein: a mechanism of chaperoning kinase clients by exploiting weak spots of intrinsically dynamic kinase domains. PLoS One 2017, 12, e0190267. (193) Stetz, G.; Verkhivker, G. M. Functional role and hierarchy of the intermolecular interactions in binding of protein kinase clients to the HSP90−Cdc37 chaperone: structure-based network modeling of allosteric regulation. J. Chem. Inf. Model. 2018, 58, 405−421. (194) Verkhivker, G. M. Dynamics-based community analysis and perturbation response scanning of allosteric interaction networks in the TRAP1 chaperone structures dissect molecular linkage between conformational asymmetry and sequential ATP hydrolysis. Biochim. Biophys. Acta, Proteins Proteomics 2018, 1866 (8), 899−912. (195) Stetz, G.; Tse, A.; Verkhivker, G. M. Dissecting structureencoded determinants of allosteric cross-talk between post-translational modification sites in the HSP90 chaperones. Sci. Rep. 2018, 8, 6899. (196) Colombo, G.; Morra, G.; Meli, M.; Verkhivker, G. M. Understanding ligand-based modulation of the HSP90 molecular chaperone dynamics at atomic resolution. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 7976−7981. (197) Vettoretti, G.; Moroni, E.; Sattin, S.; Tao, J.; Agard, D.; Bernardi, A.; Colombo, G. Molecular dynamics simulations reveal the mechanisms of allosteric activation of HSP90 by designed ligands. Sci. Rep. 2016, 6, 23830. (198) Zhao, H. P.; Moroni, E.; Colombo, G.; Blagg, B. S. J. Identification of a new scaffold for HSP90 c-terminal inhibition. ACS Med. Chem. Lett. 2014, 5, 84−88. (199) Zhao, H. P.; Moroni, E.; Yan, B.; Colombo, G.; Blagg, B. S. J. 3D-QSAR-assisted design, synthesis, and evaluation of Novobiocin analogues. ACS Med. Chem. Lett. 2013, 4, 57−62. (200) Moroni, E.; Zhao, H.; Blagg, B. S.; Colombo, G. Exploiting conformational dynamics in drug discovery: design of C-Terminal inhibitors of HSP90 with improved activities. J. Chem. Inf. Model. 2014, 54, 195−208. (201) Bopp, B.; Ciglia, E.; Ouald-Chaib, A.; Groth, G.; Gohlke, H.; Jose, J. Design and biological testing of peptidic dimerization

inhibitors of human HSP90 that target the C-terminal domain. Biochim. Biophys. Acta, Gen. Subj. 2016, 1860, 1043−1055. (202) Ciglia, E.; Vergin, J.; Reimann, S.; Smits, S. H. J.; Schmitt, L.; Groth, G.; Gohlke, H. Resolving hot spots in the C-Terminal dimerization domain that determine the stability of the molecular chaperone HSP90. PLoS One 2014, 9, e96031. (203) Wang, L.; Li, L.; Fu, W. T.; Jiang, Z. Y.; You, Q. D.; Xu, X. L. Optimization and bioevaluation of Cdc37-derived peptides: an insight into HSP90-Cdc37 protein-protein interaction modulators. Bioorg. Med. Chem. 2017, 25, 233−240. (204) Chiappori, F.; Merelli, I.; Milanesi, L.; Colombo, G.; Morra, G. An atomistic view of HSP70 allosteric crosstalk: from the nucleotide to the substrate binding domain and back. Sci. Rep. 2016, 6, 23474. (205) Maisuradze, G. G.; Senet, P.; Czaplewski, C.; Liwo, A.; Scheraga, H. A. Investigation of protein folding by coarse-grained molecular dynamics with the UNRES force field. J. Phys. Chem. A 2010, 114, 4471−4485. (206) Nicolaï, A.; Senet, P.; Delarue, P.; Ripoll, D. R. Human inducible HSP70: structures, dynamics, and interdomain communication from all-atom molecular dynamics simulations. J. Chem. Theory Comput. 2010, 6, 2501−2519. (207) Gołaś, E.; Maisuradze, G. G.; Senet, P.; Ołdziej, S.; Czaplewski, C.; Scheraga, H. A.; Liwo, A. Simulation of the opening and closing of HSP70 chaperones by coarse-grained molecular dynamics. J. Chem. Theory Comput. 2012, 8, 1750−1764. (208) Nicolai, A.; Delarue, P.; Senet, P. Conformational dynamics of full-length inducible human HSP70 derived from microsecond molecular dynamics simulations in explicit solvent. J. Biomol. Struct. Dyn. 2013, 31, 1111−1126. (209) Nicolai, A.; Barakat, F.; Delarue, P.; Senet, P. Fingerprints of conformational states of human HSP70 at Sub-THz frequencies. ACS Omega 2016, 1, 1067−1074. (210) Chiappori, F.; Merelli, I.; Colombo, G.; Milanesi, L.; Morra, G. Molecular mechanism of allosteric communication in HSP70 revealed by molecular dynamics simulations. PLoS Comput. Biol. 2012, 8, e1002844. (211) Stetz, G.; Verkhivker, G. M. Computational analysis of residue interaction networks and coevolutionary relationships in the HSP70 chaperones: a community-hopping model of allosteric regulation and communication. PLoS Comput. Biol. 2017, 13, e1005299. (212) Ung, P. M.-U.; Thompson, A. D.; Chang, L.; Gestwicki, J. E.; Carlson, H. A. Identification of key hinge residues important for nucleotide-dependent allostery in E. coli HSP70/DnaK. PLoS Comput. Biol. 2013, 9, e1003279. (213) Chang, L.; Thompson, A. D.; Ung, P.; Carlson, H. A.; Gestwicki, J. E. Mutagenesis reveals the complex relationships between atpase rate and the chaperone activities of Escherichia coli Heat Shock Protein 70 (HSP70/DnaK). J. Biol. Chem. 2010, 285, 21282−21291. (214) General, I. J.; Liu, Y.; Blackburn, M. E.; Mao, W.; Gierasch, L. M.; Bahar, I. ATPase subdomain IA is a mediator of interdomain allostery in HSP70 molecular chaperones. PLoS Comput. Biol. 2014, 10, e1003624. (215) Stetz, G.; Verkhivker, G. M. Probing allosteric inhibition mechanisms of the HSP70 chaperone proteins using molecular dynamics simulations and analysis of the residue interaction networks. J. Chem. Inf. Model. 2016, 56, 1490−1517. (216) English, C. A.; Sherman, W.; Meng, W.; Gierasch, L. M. The HSP70 interdomain linker is a dynamic switch that enables allosteric communication between two structured domains. J. Biol. Chem. 2017, 292, 14765−14744. (217) Malinverni, D.; Lopez, A. J.; De Los Rios, P.; Hummer, G.; Barducci, A. Modeling HSP70/Hsp40 interaction by multi-scale molecular simulations and coevolutionary sequence analysis. eLife 2017, 6, 20. (218) Kityk, R.; Kopp, J.; Mayer, M. P. Molecular mechanism of JDomain-triggered ATP hydrolysis by HSP70 chaperones. Mol. Cell 2018, 69, 227−237. 87


Allosteric Modulators of HSP90 and HSP70: Dynamics Meets Function

Recommend Documents