The dance of proteostasis and metabolism: Unveiling the caloristatic controlling switch

Abstract

The heat shock response (HSR) is an ancient and evolutionarily conserved mechanism designed to restore cellular homeostasis following proteotoxic challenges. However, it has become increasingly evident that disruptions in energy metabolism also trigger the HSR. This interplay between proteostasis and energy regulation is rooted in the fundamental need for ATP to fuel protein synthesis and repair, making the HSR an essential component of cellular energy management. Recent findings suggest that the origins of proteostasis-defending systems can be traced back over 3.6 billion years, aligning with the emergence of sugar kinases that optimized glycolysis around 3.594 billion years ago. This evolutionary connection is underscored by the spatial similarities between the nucleotide-binding domain of HSP70, the key player in protein chaperone machinery, and hexokinases. The HSR serves as a hub that integrates energy metabolism and resolution of inflammation, further highlighting its role in maintaining cellular homeostasis. Notably, 5′-adenosine monophosphate-activated protein kinase emerges as a central regulator, promoting the HSR during predominantly proteotoxic stress while suppressing it in response to predominantly metabolic stress. The complex relationship between 5′-adenosine monophosphate-activated protein kinase and the HSR is finely tuned, with paradoxical effects observed under different stress conditions. This delicate equilibrium, known as caloristasis, ensures that cellular homeostasis is maintained despite shifting environmental and intracellular conditions. Understanding the caloristatic controlling switch at the heart of this interplay is crucial. It offers insights into a wide range of conditions, including glycemic control, obesity, type 2 diabetes, cardiovascular and neurodegenerative diseases, reproductive abnormalities, and the optimization of exercise routines. These findings highlight the profound interconnectedness of proteostasis and energy metabolism in cellular function and adaptation.

Graphical abstract

Introduction

The heat shock response (HSR) is a highly evolutionarily conserved cellular manifestation primarily devoted to re-establishing cellular homeostasis after stressful situations. Although the HSR acts as a frontline defense against factors that can denature proteins, that is, proteotoxic challenges (e.g., heat, heavy metals), it was evident since the pioneering studies of Professor Ferruccio Ritossa1, 2, 3 that this response is also triggered by disruptions in energy metabolism, such as oxygen deprivation and uncoupling of ATP synthesis. This insightful belief of Ritossa becomes a cogent line of reasoning that gains clarity when considering the following factors: protein synthesis demands significant energy expenditure; rectifying misfolded proteins via the HSR requires ATP energy; expediting mitochondrial ATP production amplifies the generation of proteotoxic reactive oxygen species (ROS). Hence, it is now more than expected that the HSR had evolved alongside the regulatory principles of cellular energy metabolism.

Evidence suggests that the emergence of proteostasis-defending systems dates back to last universal common ancestor (LUCA), more than 3.6 billion years ago,⁴ exactly at the same time that sugar kinases started evolution to improve glycolytic efficiency, approximately 3.594 billion years ago.⁵ Therefore, it is not surprising that the evolution of the nucleotide-binding domain (NBD) of the 70 kDa heat shock protein (HSP70, the working force of protein chaperone machinery) had introduced subdomains (Ia and IIa, discussed below) that are spatially identical to the NBD of hexokinases⁶ and other sugar kinases.⁷

It has been known for a long time that the HSR acts as a hub to integrate energy metabolism and the resolution of inflammation because the HSR pathway is anti-inflammatory per se.⁸ For instance, 5′-adenosine monophosphate-activated protein kinase (AMPK), which is the master cellular energy sensor activated during metabolic stress (energy paucity), simultaneously assumes the role of a central regulator of anti-inflammatory responses through the inhibition by phosphorylation of glycogen synthase kinase-3β (GSK-3β). This phosphorylation event liberates the heat shock transcription factor 1 (HSF1), initiating the HSR because GSK-3β constitutively inhibits HSF1.8, 9, 10 Conversely, the NAD⁺-dependent protein deacetylase of class III family sirtuin-1 (SIRT1), which is activated by calorie restriction, enhances the HSR by increasing HSF1 protein expression and DNA-binding activity onto heat shock genes.¹¹ Actually, heat shock and calorie restriction act synergistically to arm an HSR.¹² In addition to enhancing the expression of protein chaperones, HSF1 directly increases the expression of an ample array of genes implicated in energy metabolism, such as the SIRT1 itself, AMPK, and peroxisome proliferator-activated receptor-γ coactivator-1α, just to mention a few.13, 14 Apart from the expression of glucose-regulated chaperones (e.g., GRP78 and GRP75), the HSR is also connected to energy metabolism through the hexosamine biochemical pathway (HBP, discussed below), a metabolic shunt from glycolysis that leads to the blockade of GSK-3β activity and enhanced DNA-binding activity of HSF1.¹⁵

While the role of AMPK in activating the HSR during predominantly proteotoxic stress is well-established, a growing body of evidence is challenging this notion by highlighting AMPK’s potential to exert opposing effects. A case in point is the involvement of the mRNA-binding protein HuR, also known as embryonic lethal abnormal vision Drosophila homolog-like protein-1. HuR plays a pivotal role in stabilizing SIRT1 mRNA, consequently promoting the expression and activity of HSF1, a key player in the HSR. To effectively carry out its functions on target transcripts, HuR must translocate from the nucleus to the cytoplasm. Herein lies an intriguing paradox: contrary to its expected role, AMPK, which is typically known to promote SIRT1 activation during an HSR-dependent anti-inflammatory response,⁸ actually hinders the nuclear export of HuR,16, 17, 18 thereby attenuating the vigor of the HSR. Reinforcing this line of thought, Dai and colleagues19, 20 have provided compelling evidence indicating that AMPK, when confronted with predominantly metabolic stress, actively suppresses the proteostasis-preserving facet of the HSR. In contrast, during a bona fide HSR event (predominantly proteotoxic), the orchestrated activation of protein phosphatase 2A by HSF1 leads to the inhibition of AMPK,²¹ subsequently fostering an elevation in HSP70 expression.²²

These findings underscore a nuanced equilibrium governing the orchestration of the proteotoxic stress response and the metabolic stress response, operating at the interface of AMPK and HSF1 to regulate the HSR under diverse stress conditions. This intricate modulation of cellular stress responses was meticulously explored by Swan and Sistonen²³ based on the findings of Dai’s group¹⁹ and now is referred to as caloristasis.²⁴ In fact, predominantly proteotoxic stress and predominantly metabolic stress work in opposition to maintaining a very delicate thermodynamical poise that warrants the steady state between energy homeostasis and proteostasis so that overall cellular homeostasis may take place whatever the environmental or intracellular conditions. The central caloristatic controlling switch that integrates proteostatic and metabolic stressful conditions is summarized in Figure 1.

Fig. 1.

The heat shock response and the central caloristatic controlling switch. This switch orchestrates a delicate dance between metabolic stress, characterized by the activation of AMP-activated protein kinase (AMPK), and proteotoxic stress, marked by the accumulation of misfolded proteins and oxidative pressure. During periods of predominantly metabolic stress, AMPK takes center stage. It exerts its influence by phosphorylating heat shock transcription factor 1 (HSF1) at Ser121, effectively hindering the transcription of HSF1-dependent genes. On the other hand, predominantly proteotoxic stress ushers in a different set of players. The accumulation of misfolded proteins and heightened oxidative stress directly activate HSF1. Consequently, under proteotoxic stress conditions, a substantial portion of HSF1 is liberated from chaperone machines, effectively blocking AMPK and allowing for the expression of over 5200 genes directly regulated by HSF1. AMPK, 5′-adenosine monophosphate-activated protein kinase; GSK-3β, glycogen synthase kinase-3β; HSP70, the 70 kDa family of heat shock proteins; PGC-1α, peroxisome proliferator-activated receptor-γ coactivator-1α; SNS, sympathetic nervous system.

Taking into account the above findings, the purpose of the present work is to illustrate the currently recognized interconnections between the HSR as a safeguarding mechanism for proteostasis and the pathways that govern energy metabolism. As we shall discuss herein, understanding the caloristatic controlling switch is a fundamental prerequisite that may enable us to delve into the intricacies underlying a wide array of conditions, ranging from glycemic control, obesity, type 2 diabetes mellitus, cardiovascular and neurodegenerative diseases, to human reproduction abnormalities and well-structured exercise routines.

Unraveling the mosaic: convergence of insights leading to contemporary understanding of the HSR

In order to understand the logic of the HSR, it is interesting to note the history of heat shock proteins (HSPs) that can be traced back to the observations of Ferruccio Ritossa, a young geneticist at the time, which laid the foundation for our comprehension of these important proteins. He noticed that polytene chromosomes of salivary glands of Drosophila busckii fruit-fly larvae exhibited a novel pattern of puffing when heat-shocked from normal (25 °C) to 30 °C (or higher) in an incubator.¹ It is worth noting that polytene chromosomes are formed through the successive duplication of each chromosomal element (chromatid) without their separation from the chromosomes of diploid nuclei.²⁵ Consequently, the newly formed chromatids remain connected lengthwise and collectively give rise to cable-like structures known as polytene chromosomes. Importantly, back in 1962, the study of gene expression was limited to systems such as Drosophila spp. polytene chromosomes, where gene activity was visible as chromosomal puffing under the light microscope.²⁶ Messenger RNA had only just been discovered by Professor Matthew Meselson’s laboratory in 1961.²⁷

Ritossa supposed that the rapid appearance of specific puffs and disappearance of others in 3–4 min after the inadvertent but serendipitous heat shock of Drosophila cells and tissues would be related to the synthesis of mRNA since tritiated cytidine incorporations into the puffs were abolished by RNAse treatment.¹ He also realized that the effect of heat shock, leading to the appearance of some puffs and reversal of others that were present under normal conditions, was mimicked by the treatment of cells, organs, or whole organisms of Drosophila spp. with uncouplers of ATP synthesis, such as 2,4-dinitrophenol and sodium salicylate.1, 2, 3 Moreover, although some effects could be observed when larvae grown at 19 °C were exposed to 25 °C, they were typically of low intensity and efficiency. This suggested that variations in puffing patterns might not be solely dependent on a temperature increase of 5–6 °C, but rather on the rapid attainment of a specific temperature threshold. Therefore, Ritossa conjectured that these responses should be of general importance, especially because he had the understanding of Drosophila spp. as “somehow between bacteria and man” and also because similar results were obtained with anaerobiosis, thus linking the observed heat shock response to energy production.²⁶

At that time, Michael Ashburner, who had been studying the puffing patterns of Drosophila spp. since the late 1950s at the University of Cambridge (UK), was investigating the induction of specific puffs during the stages of development of the animal28, 29 as well as by temperature shock, hormonal treatment, and anaerobiosis.³⁰ He had some suspicion about the likely expression of particular proteins related to these puffs due to the following observations. Incubation of salivary glands in vitro with antibiotic inhibitors of protein synthesis (such as cycloheximide and puromycin) did not affect puffing in D. melanogaster, but injecting cycloheximide or puromycin into 90-h L3 larvae had significant inhibitory effects on puffing. Also, these treatments rapidly inhibited the incorporation of amino acids into proteins (in less than 1 min), although changes in puffing patterns took approximately 3 h to become apparent.³⁰ Then, Alfred Tissières, on a sabbatical visit to the laboratory of Ursula M. Tracy at the California Institute of Technology, confirmed the appearance of new bands separated by SDS-PAGE in ³⁵S-methionine-labled tissues of Drosophila melanogaster; the same was observed when the whole animal was injected with the radiotracer.³¹ By using ³H-uridine labeling, they were also able to associate the synthesis of new polypeptides with the puffing activity of polytene chromosomes. They were impressed by the remarkably high rate of protein synthesis observed in a single band, later identified as ~70 kDa.³² This one band alone accounted for approximately 15% of all the labeled bands induced by heat shock. In addition, as previously noticed by Ritossa, heat shock induced the appearance of new proteins and the disappearance of others.³¹

The Ashburner group working on D. melanogaster, D. hydei, and D. simulans confirmed the above observations and added new information to them, showing that actinomycin D (a transcription inhibitor that binds to DNA duplexes preventing RNA polymerase elongation) blocks the synthesis of heat-induced proteins only if added at the beginning of the shock.³² Furthermore, in this latter study, Ritossa’s former notion that energy-threatening situations and heat shock should mandatorily share some fundamental principle in common was also corroborated. Accordingly, Lewis and co-workers³² tested the effects of uncouplers of oxidative phosphorylation (2,4-dinitrophenol), inhibitors of electron transport (rotenone), or recovery from prolonged anaerobiosis (2 h under nitrogen atmosphere) and, in fact, observed exactly the same patterns of puffing paralleled by the synthesis of new proteins, in particular, one of 70–72 kDa, which was always present in such preparations.

Susan Lindquist (then Susan Lee Lindquist McKenzie) discovered at Meselson’s laboratory that Drosophila cells exhibited the same response to heat shock as reported by Tissières and colleagues in 1974,³¹ with translational and transcriptional mechanisms governing the response.33, 34, 35, 36 These findings were and still are considered the most robust and widespread change in eukaryotic gene expression so far. She also revealed the unexpected capacity of eukaryotic cells to discriminate between coexisting mRNAs and independently regulate their translation. In other words, during heat shock, cells efficiently translate heat shock mRNAs while blocking normal mRNAs from translation, yet holding them ready for reactivation after heat shock. Lindquist’s observations then explained the findings of Ritossa and Tissières’ group that heat shock induces an extraordinary appearance of some puffs and the vanishing of others in polytene chromosomes, now under the molecular biology point of view.

Notwithstanding, Susan Lindquist was, to the best of our knowledge, the first to employ the term “heat shock proteins” to this set of heat-induced polypeptides and to emphasize the importance of the 70,000-dalton HSPs.³³ Since then, many groups have exploited details on the expression of heat shock proteins or heat shock-induced peptides, including Tissières’ laboratory³⁷ and Ashburner’s group, the latter introducing the acronym HSP to differentiate such polypeptides: HSP70, HSP80, and so on.³⁸

At that time, there was a primary source of confusion surrounding HSPs with molecular weights around 70 kDa. Mirault and colleagues³⁷ conducted two-dimensional gel experiments that revealed multiple polypeptides in this range, spanning from 70 to 72 kDa. Genetic and biochemical evidence indicated that these were very similar proteins. Therefore, two possible explanations were postulated: post-translational modification (PTM) of a parent polypeptide or each protein being the product of a particular gene copy due to multiple copies of the coding sequences.³⁸ The latter has been proven to be the case, as approached below.

Lindquist’s research group made significant contributions to the understanding of the role of HSPs in preventing and repairing stress-induced protein damage and its consequent cell toxicity. Her findings demonstrated that HSPs play a critical role in turning off the mechanisms that are activated in response to stress at every level. As HSPs restore protein homeostasis (proteostasis) and reset the damaged regulatory systems that inhibit normal mRNA functions, the elegant logic of the regulatory circuitry is revealed. By restoring regulatory systems to their normal state, HSPs eliminate their own advantage and switch off the response.³⁹ Lindquist’s work, in conjunction with the seminal work of Spradling, Pelham, Lis, and Wu on transcription, has resulted in the HSR being regarded as one of the most beautiful and complete examples of eukaryotic gene regulation.³⁹

More recently, there have been notable observations regarding the HSR. The universality of the response and the conservation of inducible genes across different species throughout evolution have been recognized as significant findings. These observations indicate that HSPs play fundamental roles in biological processes,⁴⁰ as previously assumed by Ritossa in his initial works1, 2, 3 that were not widely accepted at that time, by the way.²⁶ Indeed, the HSR has proved to be a universal phenomenon observed in every organism where it has been sought, ranging incredibly from eubacteria to archaebacteria, and from mice to soybeans.⁴¹ Moreover, it has been known for a long time that the HSR is influenced by the metabolic state of the cells. For instance, in the yeast Saccharomyces cerevisiae, the HSR in fermenting cells grown at 25 °C is transient at 36 °C and sustained at 40 °C, while in respiring cells, the response is transient at 34 °C and sustained at 36 °C. These findings resemble in much the effects of hyperbaric oxygen therapy observed in mammalian cells,²⁴ as we shall approach later. Additionally, the synthesis of HSPs is determined by previous incubation temperatures: in Drosophila cells, when the temperature is abruptly increased to high levels, the maximum response occurs at 37 °C, and only minimal synthesis is observed at 39 °C. However, when the temperature is gradually increased (a regimen more likely to reflect natural environmental exposure), the response is extended over several degrees. Apart from a few exceptionally rare and evolutionarily justifiable cases (discussed next), these discoveries have been replicated in several other organisms.⁴¹

These proteins are widely regarded as among the most highly conserved proteins in existence.⁴² Under normal growth temperatures, the HSP70 gene is transcribed at low levels in human cells. However, a transient heat shock at 43 °C can induce the gene by over 20-fold.⁴³ Importantly, the unusually high degree of conservation observed in the amino acid sequence of human HSP70,⁴³ which was found to be 73% identical to that of the Drosophila HSP70 and 47% identical to the Escherichia coli dnaK (!), suggests that HSP70s were conserved throughout evolution for a specific and important reason over the last 3.6 billion years.

Apart from certain stenothermal organisms, which are only capable of surviving within a narrow range of ambient temperatures (such as those found in the deep sea and polar environments), all eurythermal organisms (including birds and mammals) that have been studied so far exhibit the inducible production of proteins encoded by the HSP70 and HSP90 gene families in response to elevated temperatures. Interestingly, the stenothermal Antarctic fish Emerald rockcod (Trematomus bernacchii) possesses an exceptionally unusual trait: the animal does not express inducible forms of HSPs when subjected to heat or heavy metal stress. Instead, when challenged with non-lethal thermal stress, this fish increases the synthesis of the constitutive HSP70. This unique trait is exclusively found in animals that inhabit subzero, thermally stable waters.⁴⁴ It is believed that these animals lost the HSR between 14 and 25 million years ago. This timeframe coincides with the opening of the Drake Passage, leading to a significant cooling of Antarctic waters.⁴⁴ As a result, this evolutionary adaptation occurred relatively recently in terms of both evolution and geological history. A similar phenomenon is observed in the freshwater coelenterate species Hydra attenuata when it faces thermal or heavy metal stress. Notably, only the constitutively expressed HSP60 (but not HSP70), which functions as the primary chaperone, experiences an enhanced synthesis in such conditions. This heightened synthesis of HSP60 coincides with an increased thermotolerance, allowing the organism to better withstand subsequent challenges. Quirkier, its congener species Hydra oligactis displays an extreme sensitivity to thermal stress and fails to express any HSP in response to heat at all.⁴⁵ Indeed, many animals, particularly the ectothermic ones, evolved resourceful strategies to survive adverse ambient conditions, and all of them employ changes in chaperone functions.⁴⁶

Shortly after cloning, it became evident that the human HSP70 gene belongs to a multigene family, similar to the 70 kDa multigene families found in Drosophila and yeast,⁴⁷ reaffirming one of the hypotheses raised by Ashburner’s laboratory.³⁸ Notably, HSPA1A, the gene responsible for coding HSP70, is an intronless gene. This lack of intervening sequences is not just a coincidence, but a significant characteristic for genes whose transcripts accumulate (or must accumulate) rapidly in the cytoplasm. In contrast, most eukaryotic genes containing intervening sequences require proper processing before their transcripts can be transported into the cytoplasm. This is not the case for hspa1a, and it explains the rapid transcription and accumulation rates observed in Drosophila HSP70 mRNAs.⁴⁷

Before delving further into this discussion, it is pertinent to highlight that the term “proteotoxicity” was initially coined by analogy to “genotoxicity.” It was first introduced by Professor Larry Hightower⁴⁸ to characterize the harm inflicted upon proteins by various chemical and physical agents.

Initially identified as stress-responsive proteins necessary to cope with thermal and other proteotoxic stresses, the human family of HSPs was soon found to contain members that are constitutively expressed, such as HSC70, the cognate form of HSP70, also known as HSP73 encoded by HSPA8 gene, besides its most prominent stress-inducible HSP72, encoded by HSPA1A and HSPA1B genes.⁴⁹ Another feature of HSPs is their location in different compartments of the eukaryotic cell. In addition, they may or may not be induced by heat and this depends on tissue and metabolic state. For instance, glucose-regulated protein 78 (GRP78), a member of the HSP70 family, is situated within the endoplasmic reticulum (ER). Its expression is induced in response to diminished glucose levels. In contrast, another glucose-responsive HSP70, Glucose-Regulated Protein 75 (GRP75), resides in the mitochondrion. GRP75 similarly reacts to glucose scarcity but remains unresponsive to heat shock.41, 42

Insights into the true function of HSPs as proteostasis-safeguarding in the cells, nevertheless, emerged with the discovery that abnormal, denatured proteins can act also as stress signals, activating heat shock genes.⁵⁰ Inside cells, HSPs play a vital role in preventing protein aggregation when newly synthesized proteins are leaving the ER en route to their final destination, or when intracellular proteins are under immediate threat of misfolding due to various agents such as heat, alcohol, amino acid analogs or heavy metals, and post-ischemic reperfusion, which induce oxidative stress and menace protein physiology as well. However, it was difficult to solve the puzzle that connected the spatial conformation of polypeptides with the activation of specific genes. Even more: how did evolution prepare sensors, if any, capable of transducing the misfolding of a protein into a signal for the transcription of heat shock genes?

For a long time, it was believed that the likelihood of a stepwise random folding-denaturation-refolding process during the biosynthesis of a polypeptide chain from the amino to carboxy-terminus would be minimal. Moreover, it was widely recognized that there exists a delicate balance between a stable, native protein structure, and a random, biologically functionless one. Therefore, it was expected that some biological system would operate to maintain this balance thermodynamically.⁵¹ The question is that, in Anfinsen’s in vitro experiments with chemical denaturants, the probability of a polypeptide folding correctly after the denaturant is removed increases when the protein concentration is low, limiting inter-polypeptide interactions, and at low temperatures, which attenuates hydrophobic interactions. However, this is not the case for real cells in vivo. The high protein concentrations and temperatures present within the cell can cause premature interactions between newly synthesized polypeptides, resulting in misfolding and aggregation.⁵² This explains why, during the course of metazoan evolution, molecular chaperones emerged as a vital component in maintaining proper protein folding and preventing aggregation.⁴

The discovery of HSP chaperone cascades (next sections), nevertheless, does not alter Anfinsen’s thermodynamic folding principle,⁵¹ which states that the primary amino acid sequence encodes all the information necessary to determine the shape of the native state, and chaperones are not responsible for changing it.⁵³ Chaperones function in a manner similar to enzymes in that they do not interfere with the quantum of free energy (ΔG) involved in folding. Instead, they lower the activation energy required for proper folding or transport toward other cellular compartments. This is made possible because, like conventional enzymes, chaperones “embrace” specific segments of the polypeptide chain, reducing their mobility (entropy). As expected, this energetic movement against spontaneity (ΔS < 0) requires energy input obtained from ATP hydrolysis, but this is the faster step of the folding reaction. Instead, upon being released at their final destinations, Brownian mobility restores entropy (and free energy) to the medium, and this is the slow (bottleneck) step of the entire process.

For example, HSP70-HSP90 cascades significantly increase folding yields, but they basically do not affect the overall kinetics of the folding process.⁵⁴ This is because HSP70s dose-dependently block protein folding by imprisoning client proteins at the chaperone core, a very fast process that is ATP-dependent. HSP90s, on the other hand, promptly resume the folding course by taking over client proteins from HSP70s in an ATP-dependent manner as well. HSP70-HSP90 activities are limited to the early folding phase, during which their ATPase activity is required for a few seconds (k_cat of up to 0.79 s⁻¹).⁵⁵ Once this phase is complete, the client protein follows an Anfinsen folding trajectory (rate limiting for the velocity of folding; several minutes) that does not require the assistance of either chaperone.⁵³

For the inter-organelle transport of polypeptides or dispatch of them from the ribosomal exit tunnel, a pulling force (ΔG) is required, which is generated by the low intrinsic entropy state created by the limited freedom of movement of polypeptides bound to HSP70 near the membrane and the pore (translocon) through which the polypeptides must pass. As the polypeptide moves inward through the translocon, it increases the freedom of movement (ΔS > 0) of the substrate-bound HSP70, leading to an increase in entropy (resulting in a negative ΔG). This creates a one-way pulling motion, as entropy can only spontaneously increase.⁵⁶

The available data suggest a sequential occurrence of ATP and cochaperone-induced structural rearrangements in bacterial HSP70 (DnaK), leading to the resolution of previously unforeseen cochaperone and client-induced changes.⁵⁷ Peptides induce significant conformational alterations in DnaK·ATP prior to ATP hydrolysis, whereas protein clients induce smaller changes but exhibit greater effectiveness in stimulating ATP hydrolysis.⁵⁷ The analysis of the enthalpies of activation for the ATP-induced opening of the DnaK lid in the presence of client peptides indicates that the lid does not exert an enthalpic pulling force on bound clients. This finding suggests that entropic pulling serves as a major mechanism for client unfolding,⁵⁷ as previously proposed.53, 54 These recent findings provide valuable insights into the mechanics, allostery, and dynamics of HSP70 chaperones, as demonstrated.⁵⁷ In essence, chaperoning is a process fundamentally governed by thermodynamics, much like the entire process of triggering the HSR.

While the HSR also extends to the organism level,⁵⁸ HSPs play two distinct roles in terms of cellular protein protection⁵⁶: housekeeping and stress-activated functions. The housekeeping functions are related to (1) cooperation with other protein folding and quality-control machineries; (2) de novo protein folding; (3) protein translocation across membranes; (4) assembly/disassembly of protein complexes; (5) regulation of protein activity; and (6) protection from proteolysis. On the other hand, stress functions evolved to limit the noxious effects of unfolded proteins, which evoke inflammatory responses if accumulated.¹¹ They comprise (1) prevention of protein aggregation; (2) protein disaggregation, in cooperation with small HSPs (sHSPs) and HSP100 families; (3) protein refolding; and (4) protein degradation (e.g., autophagy and ubiquitin-proteasome system) to clear aberrant proteins and protein aggregates. Structure-function relations and HSP70-client protein interactions have also been detailed.⁵⁹ Finally, HSR regulatory mechanisms poise proteostasis protection with all metabolic pathways related to energy preservation (caloristasis), including animal reproduction and anti-inflammatory responses. This is why modern human diseases such as obesity, in which a surplus of energy is detected within cells, so dramatically disarrange proteostasis and HSR-dependent anti-inflammatory pathways.⁵⁸ After all, the HSR evolved to juxtapose proteotoxic stress with metabolic stress, specifically in response to indications of insufficient cellular energy, not energy surplus.

Comparing the initial observations described in this section with what is known at present, it is amazing what a close look at the result of an error in the laboratory gave us. On the occasion of the 50th anniversary of the discovery of the HSR, Professor Maria Gabriella Santoro talked through the significance of the discovery. In her reflections, she highlighted the first interpretation of the HSR given by Ritossa himself, which is worth mentioning: “It did not matter if this interpretation was true or false; it was a working link between imagination and reality, like love.” Ferruccio is not only a scientist and an artist; he is also a poet, she said.⁶⁰

Heat shock protein families

Before exploring the roles of HSPs, it is worth noting that, although the term “molecular chaperones” be commonly used in the field of HSPs, it was originally introduced by Laskey and colleagues⁶¹ to describe the remarkable capacity of these proteins to prevent incorrect ionic interactions between histones and DNA.

To cope with both housekeeping and stress-induced functions, evolution ameliorated, from ancient prokaryotic ancestors, eight families of HSPs. They were initially characterized by their molecular weight (i.e., HSP72, HSP90, etc.) and are now grouped into gene families and superfamilies, more consistently with HUGO Gene Nomenclature Committee as follows⁴⁹: HSPA (e.g., HSP70), HSPB (small HSPs), HSPC (e.g., HSP90), HSPD and HSPE (HSP60 and HSP10 chaperonins), HSPH (e.g., HSP110), DNAJ (J-proteins, formerly HSP40 cochaperones) alongside CCT, and other chaperonin-related genes.

To date, 13 different genes have been identified in the human genome that code for members of the HSPA superfamily: 11 for HSPB, five for HSPC, one for HSPD, one for HSPE, nine for the CCT chaperonin genes, and three others for chaperonin-like (one for MKKS and two for BBS).⁴⁹ Although the number of HSPA homologs/paralogs in the human genome is notable, it is the number of identified DNAJ genes in humans⁶² that truly stands out, with 50 genes identified (4 DNAJA members, 13 DNAJB, and 33 DNAJC). This high number is an indication of the significant role that DNAJ genes play in the chaperone machinery.⁶² DNAJ expression is found in all cellular compartments.⁶³

J-proteins are distinguished by the presence of a conserved J-domain responsible for recruiting and activating HSPA ATPase activity.⁴⁹ This domain, which is named after its founding member, the E. coli DnaJ cochaperone, contains a highly conserved ~70 amino acid signature region. Of particular importance is a His-Pro-Asp tripeptide, which is located within a loop connecting the two main helices (helix II and helix III). The His-Pro-Asp tripeptide motif is critical for the J-domain’s function in stimulating HSP70’s ATPase activity.⁶⁴

The HSPA/HSP70 superfamily, in particular, exhibits a high degree of diversity, which can be attributed to several factors, for instance, the various possible cellular locations, such as cytoplasm and nucleus (e.g., HSPA1A/HSP72 and HSPA1B/HSP70-2), ER (e.g., HSPA5/GRP78), and mitochondrion (e.g., HSPA9/GRP75). Despite the significant homology between the genes in this superfamily, HSPA members demonstrate a high degree of specialization, nonetheless. For example, the response to stress can vary significantly: HSPA8/HSC73 (expressed in both cytoplasm and nucleus)⁶³ is primarily a non-inducible housekeeping gene, whereas HSPA6/HSP70B′ exhibits strictly inducible expression, with minimal or no basal expression in most cells. Actually, there is a vast repertoire of combinations between intracellular location and inducibility among HSP members.⁶⁵ Studies have shown that proteasome inhibition is a potent activator of HSP70B′. In comparison to HSP72, which is the primary responder to increasing levels of proteotoxic stress among the HSPA/HSP70 family, HSPA6/HSP70B′ serves as a secondary responder. Interestingly, in cell models, HSPA6/HSP70B′ is induced by ZnSO₄ but not HSP72, indicating that HSPA6/HSP70B′ may have a stressor-specific primary role.⁶⁶

The ER HSPA5 member, also referred to as binding immunoglobulin protein (BiP) or GRP78 (78 kDa glucose-regulated protein), is a master regulator of the unfolded protein response (UPR), reducing ER stress levels and apoptosis due to an enhancement of the cellular folding capacity, including against prion replication.⁶⁷ The mitochondrial HSPA9, also known as GRP75 (75 kDa glucose-regulated protein), is a heat shock cognate protein that plays a capital role in cell proliferation, stress response, and maintenance of the mitochondrion. It is essential in increasing ER-mitochondria contact during palmitate-induced apoptosis in pancreatic β-cells.⁶⁸ The expression of both GRP78 and GRP75 is also activated by lowering glucose levels within the cell. Actually, the level of HSP expression may determine if a cell can be rescued from a death stimulus or must undergo apoptosis.⁶⁹

Chaperones also differ much in their preference for client proteins. For instance, while cytosolic HSPA/HSP70s have an inclination for binding leucine-enriched peptide motifs, which are abundant in aliphatic residues, the ER homolog BiP/HSPA5 has a preference for motifs with aromatic residues.⁵⁶ Apropos of HSPA gene family, its members possess high but different degrees of homology in respect of HSPA1/HSP72 (the most highly expressed one), namely: HSPA1B/HSP70-2, 99%; HSPA5/GRP78, 64%; HSPA6/HSP70B′, 85%; HSPA8/HSP73/HSC70, 86%; HSPA9/GRP75, 52%.

The triggering of HSP production during proteotoxic stress depends on the launching of the activation of heat shock transcription factors (HSFs), with HSF1 being the most studied due to its direct link with proteotoxic stress and the universality of expression across species. Under non-stressful situations, HSF1 resides in the cytoplasm in a monomeric conformation that has no DNA-binding or transcribing activity. To be fully activated, HSF1 first needs to trimerize and subsequently gain DNA-binding activity, nuclear accumulation, and extensive PTMs, including multiple serine-phosphorylations, acetylation/deacetylation, or even sumoylation and ubiquitinylation, whose prevalence depends on the physiological context.70, 71 Furthermore, as mentioned above, a sophisticated caloristatic control system ultimately governs HSF1’s involvement in the intricate process of switching between proteostasis-preserving and energy-conserving mechanisms. In any way, the main impact of HSF1 activation is the high throughput of HSPs, which represent 5–10% of the total protein in most cells and their intracellular content can increase about 20 times in response to stressors.⁷²

The heat shock factors (HSFs)

Unlike invertebrates, which possess a single HSF, plants and vertebrates express multiple HSFs. Humans have six members in the HSF family, namely HSF1, HSF2, HSF4, HSF5, HSFX, and HSFY.⁷¹ Each HSF has unique and overlapping functions, tissue-specific expression patterns, and undergoes various PTMs while interacting with multiple protein partners.⁷³ In eukaryotes, HSF1 is involved not only in HSP gene expression and stress resistance but also in the expression of genes with roles in cell maintenance and differentiation, as well as in developmental processes.13, 14, 74

Some genes whose expression is dependent on the binding of activated HSF1 include members of different HSP families, HSPA/HSP70 (HSPA1A, HSPA1B, HSPA1L, and HSPA8/HSC73), HSPC/HSP90 (HSPC1/HSP90AA1 and HSPC3/HSP90AB1), HSPH (HSPH1/HSP105 and HSPH3/HSPA4L), DNAJ/HSP40 (DNAJA1, DNAJA4, DNAJB1), chaperonins (HSPE1/HSP10 and HSPD1/HSP60), small HSPs (HSPB1 and HSPB8), BAG3, and many other stress-responsive but not necessarily HSP genes.⁷⁵

In fact, HSF1 also directly regulates the expression and activity of key factors involved in cell differentiation and longevity, autophagy, mitochondrial and ribosomal biogenesis, immune responses, multidrug resistance, cancer progression, aging, and neurodegenerative diseases.13, 70, 71, 73, 76 This feat is accomplished through the activation of HSF1, which orchestrates the transcription of a remarkable array of more than 5200 genes. This comprehensive list includes all core chaperones and their associated cochaperones, the mRNA-binding protein HuR, and the transcription factor peroxisome proliferator-activated receptor-γ coactivator-1α, both of which play pivotal roles in the HSR. Additionally, HSF1 regulates the expression of pro-inflammatory genes such as cyclo-oxygenase-2, IL-1β, TNFα, and the master regulator of inflammation, NF-κB. Notably, HSF1 also influences metabolic processes by modulating AMPK, sirtuin-1, and GSK-3β. Moreover, it exerts control over cell cycle and differentiation through the regulation of c-Jun, Fos, Wnt 2, CD95/Fas, and cyclin-dependent kinase inhibitor 2B, among numerous others, as depicted on the far-right side of Figure 1.

In entirety, HSF1 exerts control over nearly 25,000 genes, either directly or indirectly, through its collaboration with other transcription factors. This control mechanism entails both the activation of specific genes and the inhibition of others.13, 14

Initially, HSF1 was identified as the primary regulator of the HSR. However, it is now known that HSF2 modulates HSF1-mediated expression of HSP genes by forming heterocomplexes. Upon exposure to heat shock, HSF1 and HSF2 accumulate in nuclear stress bodies and bind to satellite III repeats.⁷³ Of note, HSF1 stimulates the transcription of HSF2,13, 14 so that HSF1 and HSF2 present a cross-regulation. In summary, HSF1 regulates protein quality-control machinery and gene expression to support cell survival, and this is modulated by the cellular metabolic status as well.23, 24 On the other hand, HSF2 is highly expressed during early development and in the testis, while it has multiple roles as an activator of protein chaperone genes, and as a tumor suppressor. HSF2 works as an activator of protein chaperone genes when the temperature is in the febrile range. HSF4 is required for eye lens development and is also expressed in the heart, brain, skeletal muscle, and pancreas. Recent research has revealed that HSF5 activity plays a crucial role in orchestrating spermatogenesis in mammals, as well as overseeing the events of programmed meiotic sex chromosome remodeling and silencing that occur during meiosis.⁷⁷ HSF3 is present in mice but not in humans. HSFX’s function is unknown, and HSFY is primarily expressed in the testis and contributes to male fertility.⁷¹ In this review, we shall primarily focus on the HSF1-dependent HSR unless specified otherwise. Nevertheless, we kindly encourage readers to consult the excellent reviews published previously70, 71, 73 for additional insights.

HSFs must bind to specific DNA sequences called heat shock elements (HSEs), which are located in the promoter regions of heat shock and other HSF1-driven genes, in order to trigger physiological responses.⁷⁴ Canonical HSEs comprise at least three continuous inverted repeats of the pentanucleotide sequence, 5′-nGAAn-3′, alternating between 5′-nGAAn-3′ and 5′-nTTCn-3′, or vice versa, where n is any nucleotide.⁷⁸ Moreover, the varying HSE architecture affects HSF-DNA-binding affinity and the corresponding magnitude of response.⁷⁸ The noncanonical (gapped) spacing of nGAAn units in HSE functions to limit the magnitude of transcriptional activation of heat shock genes in response to heat and oxidative stress.⁷⁹ This unique feature adds another layer of complexity and refinement to the regulation of HSF1-driven genes.

Mammalian HSFs, particularly the two most studied members of the family, that is, HSF1 and HSF2, exhibit unanticipated complexity in their structure, DNA-binding selectivity, PTMs, interacting partners, and regulation.⁷¹ The distribution of HSEs in the promoter regions of heat shock genes affects the intensity of gene expression as well as the tendency to be activated by different types of stresses. Furthermore, mutations to the HSE are involved in aggregative diseases, such as Huntington’s disease.⁷¹

HSE architecture in heat shock and other HSF1-commanded genes, including features such as proximity to TATA box, number of repeats, and spacing, do determine the speed of gene transcription initiation, sensitivity to heat and other stressors, and the amount of HSP produced in species living in thermally diverse environments.74, 80, 81 In addition to specific promoter features, gene properties such as chromatin activation are also required for the activation of heat shock genes.82, 83 Furthermore, HSF1 activity can be suppressed at both the intra- and inter-molecular levels. At the inter-molecular level, molecular chaperones such as HSP70/HSPA, HSP90/HSPC, and TRiC/CCT interact with HSF1 to inhibit its activation, thereby preventing its binding to DNA and regulation of gene expression.⁸⁴

Although the signals that trigger HSF1 activation may vary in their nature, they share a common feature in that such signals typically (but not exclusively) result in elevated levels of misfolded proteins within the cell. However, conditions that can lead proteins to be unfolded, for example, heat and oxidative stress, equally trigger HSF1 activation. In other words, HSF1 can also be activated in anticipation of protein denaturation, which is typically a physiological trait always observed during the course of evolution of life.

HSF1 activation can be achieved by at least four known mechanisms, which are not necessarily mutually exclusive.⁷⁰ Under normal cellular conditions, HSF1 exists in a complex with cytoplasmic chaperones, specifically DNAJ/HSP40, HSPA/HSP70, and HSPC/HSP90 (mainly).⁸⁵ In this state, HSF1 occurs as a monomer without DNA-binding activity. Being part of a multichaperone complex, HSC/HSP90 is involved in sequestering HSF1 monomers in the absence of stress and contributes to the deceleration of HSF1 activity after enough HSPs have been induced following stress.70, 86 However, upon exposure to various stresses such as heat shock, inflammation, or unfolded proteins, HSF1 is released from the chaperone complex and translocates into the nucleus. Once in the nucleus, HSF1 undergoes trimerization, hyperphosphorylation, and subsequently binds to HSEs present in HSP genes.⁸⁷ Actually, hyperphosphorylation of the regulatory domain of HSF1 is necessary but not sufficient for the full activation of HSF1.⁸⁸

Eventually, the binding of activated HSF1 to HSEs, thus, initiates the transcription of genes encoding HSPs, including HSPA/HSP70 and HSPC/HSP90.⁸⁷ This is known as the classical chaperone displacement mechanism. A second possibility is the RNA thermometer model, in which HSF1 remains associated with the heat-sensing RNA molecule HSR-1 under non-stressful situations. Heat stress uncouples HSF1 from the complex at the same time that liberates HSR-1 for binding to the elongation factor eEF1A, thus impeding protein synthesis during the heat shock. A third mechanism is the intrinsic response, in which HSF1 can undertake self-assembly (trimerization) in the presence of heat or other types of stress. Although there may be undiscovered mechanisms for HSF1 self-assembly, oxidation of cysteine residues in the DNA-binding domain of HSF1 is of particular importance because it permits quick formation of disulfide bonds between HSF1 monomers with the consequent trimerization of it,⁷⁰ particularly during the course of oxidative stress and metabolic defects in the generation of NADPH (see below).

HSF1-dependent HSR in cells from Drosophila spp., vertebrates and in unicellular eukaryotes is cell-autonomous,⁸⁹ that is, operated within the cell in response to stressful signals addressed to the cell. Nevertheless, a last known mechanism by which HSR and ER-UPR may be activated transcellularly at the organismal level, that is, cell-nonautonomously (please see ref.⁵⁸), was initially perceived in Caenorhabditis elegans⁷⁰ but was now realized to be a mechanism conserved from worms to mammals.⁹⁰

The chaperone machinery involved in the HSR

In order to perform chaperoning of intracellular polypeptides, chaperones definitely do not operate alone. They work in complexes known as chaperone machines. The HSPA/HSP70 machine consists of the core HSPA protein along with an array of different cofactors and nucleotide exchange factors (NEFs) that transiently bind in cooperation with core HSPA, such as DNAJ/HSP40, HSPH/HSP110, BAG-1, Hip, HSPBP1, CHIP, and HOP.⁵² HSPA1A/HSP70, which is a major work-power for proper protein folding, requires client proteins (substrates) to be presented in order to complete the process, although HSPA1A/HSP70 itself contains the necessary information for folding. This involves necessarily the collaboration of HSPA1A/HSP70 with cochaperones (e.g., HSP40/DNAJ/J-proteins) and several regulatory cofactors, such as NEFs, or the transfer of client proteins to other chaperones (e.g., HSPC/HSP90) to terminate the process.64, 91, 92 NEFs are responsible for promoting the opening of the HSPA/HSP70 nucleotide-binding cleft, which allows for the release of ADP. This, in turn, enables the rebinding of ATP and facilitates the release of substrates from HSPA/HSP70, ultimately promoting efficient chaperone activity.⁵⁶ HSPA1A/HSP70 is a notable example where ATP hydrolysis prompts a conformational shift that significantly enhances client-protein affinity.⁹³

As noticed by Ritossa in his first studies,1, 2, 3 the HSR may be activated by energy imbalances within cells. However, the integrity of the proteostasis network faces a precarious situation when confronted with energy deficits. This predicament not only triggers the activation of AMPK, which hampers the proteostasis-saving HSR as discussed in the Introduction section16, 17, 18, 19, 20, 21 but also precipitates a decline in ATP levels, as expounded upon by Morimoto.⁸⁹ This is particularly noteworthy because HSPD/HSP60, HSPA/HSP70, HSPC/HSP90, and HSPH/HSP110 are all dependent on their ATPase activity to facilitate protein folding or refolding.⁴ Interestingly, recent evidence shows that HSPH members work as NEFs for HSPA family members.49, 56 Some HSPA/HSP70 homologs, such as the ribosome-associated HSPA14 or the cytosolic HSPH/HSP110 and the ER HSPH4/HSP110 (GRP170) members, which are both homologs of HSPA/HSP70s but act as NEFs for HSPA/HSP70s, show no conservation in the interacting residues.⁵⁶

Chaperones, in addition to facilitating protein folding (“foldase” activity), also play a role in the ubiquitin-mediated proteasomal degradation of client proteins when they are irremediably unfolded. The regulation of chaperone-mediated protein degradation is influenced by cochaperones, such as the C-terminal HSP70 binding protein (CHIP). By binding to HSPA/HSP70 and HSPC/HSP90 chaperones via its tetratricopeptide repeat domain, CHIP is able to function as an E3 ubiquitin ligase using a modified RING finger domain (U-box). This unique combination of domains enables CHIP to effectively connect chaperone complexes with the ubiquitin-proteasome system.⁹⁴

Polypeptides in their native form are also transferred from HSP70 to HSP90 chaperones as a means of regulating protein activity. For example, HSP70 and HSP90 jointly control the biological activity of various target proteins through temporary interactions, including but not limited to nuclear receptors (e.g., steroid hormone receptors), kinases (e.g., eIF2α-kinase, cyclin-dependent kinases), and transcription factors (e.g., p53 and HSF1), as well as a diverse array of other proteins.⁵⁶ By means of this type of protein-protein interaction, HSPA/HSP70 forms a complex with tumor-related antigens via its polypeptide-binding domain, to elicit greater antigen-specific immune responses.⁹⁵ Moreover, HSPs may be involved in the binding of protein fragments from dead malignant cells, to present them to antigen-presenting cells via MHC class I and class II molecules, leading to the activation of anti-tumor CD8⁺ and CD4⁺ T cells.

Although not a direct focus of the present discussion, small heat shock proteins (sHSPs/HSPBs) are highly conserved across species and also play a critical role in stress tolerance. Many HSPBs/sHSPs have chaperone-like activity, which helps to prevent the aggregation of target proteins, keeps them in a folding-competent state, and facilitates their refolding either independently or in conjunction with other ATP-dependent chaperones. Mutations in human HSPBs/sHSPs have been linked to myopathies, neuropathies, and cataracts. Moreover, the expression of HSPBs/sHSPs is impaired in various diseases, such as Alzheimer’s, Parkinson’s, and cancer. The ability of HSPBs/sHSPs to bind Cu²⁺ thereby suppressing the generation of ROS may also have important implications for Cu²⁺ homeostasis and neurodegenerative diseases.⁹⁶

As stated above, the HSR is also intimately connected to inflammation. Accordingly, during either sterile tissue injury or pathogen-elicited Toll-like receptor-2 and receptor-4 triggered inflammatory responses, nuclear factors of κB family (NF-κB) are activated and signal to pro-inflammatory gene expression.⁵⁸ On the other hand, a physiological negative feedback system that “resolves” inflammation is also enabled via the HSR pathway. HSF1 may be directly activated by PGE_2-induced rise in temperature (fever) during nuclear factor NF-κB-elicited cyclo-oxygenase-2 induction. PGA₂, the dehydration product of PGE₂, is also able to counteract NF-κB downstream effects by directly blocking IκB kinase-β. HSP70, per se, blocks NF-κB activation and transcribing activity.

Following oxidative stress and the formation of reactive oxygen and nitrogen species, ROS/RNS, conformational changes in unfolded proteins can be relayed to HSF1 either directly or via changes in glutathione (GSH)/protein sulfhydryl redox status or, finally, after the activation of the UPR protocol that activates HSF1 through the SIRT1 pathway. This happens through UPR-mediated activation of HuR, an RNA-binding protein that stabilizes SIRT1 mRNA and, consequently, its expression. SIRT1 downstream signals can also be transmitted to the HSR pathway via metabolic alterations, such as those that increase nicotinamide dinucleotide redox status (↑NAD⁺/NADH ratio) or adenosine monophosphate to triphosphate (↑AMP/ATP) ratio and the consequent activation of 5′-AMPK. However, under predominant metabolic stress, activated AMPK impedes the transport of HuR toward the cytosol and may decrease the level of HSF1 activation, thus lessening the HSR and favoring energy preservation. AMPK, the master fuel sensing kinase, may also be turned on by calorie restriction, physical exercise, or the antidiabetic drug metformin, which is also the gold standard for treating metabolic disorders associated with abnormal ovary function, such as polycystic ovary syndrome. Activated AMPK phosphorylates and blocks GSK-3β, which constitutively inhibits HSF1, so that activated AMPK can cause disinhibition of HSF1, leading to an anti-inflammatory HSR. Glutamine metabolism, which is enhanced by muscle contraction, by increasing metabolic flux through the HBP, also blocks GSK-3β thereby enhancing HSR. Estrogen (E2) can activate HSR machinery both directly acting over HSF1 activation and by membrane surface estrogen receptors that signal through Erk1/2, p38 MAPK, and PI3K/Akt pathways leading to enhanced endothelial NO synthase (eNOS) and neuronal NO synthase activity and NO production which, in turn, unveils a discrete redox imbalance that activates HSF1. Additionally, E2 blocks senescence-associated secretory phenotype (SASP) that emanates from continuous activation of NLRP3 (nucleotide-binding and oligomerization domain-like receptor family pyrin domain-containing-3) inflammasome which, in turn, impairs HuR-dependent activation of HSR through the SIRT1-HSF1 pathway. SASP results from long-term ER stress and its consequent unremitted activation of UPR (e.g., obesity, insufficient physical activity). Similarly, HS treatment, even in the fever-like range, blocks NLRP3 inflammasome-dependent SASP, re-establishing HuR-SIRT1-HSF1 downstream pathways, so that HS itself can, paradoxically, re-establish the HSR.⁵⁸ Under predominant proteotoxic stress, the excess of misfolded proteins leads to the sequestration of protein chaperones (mainly HSP90/HSC) from the associated HSF1 thereby setting this transcription factor free to physically block AMPK. Therefore, in this case, AMPK is shut off and the HSR may be resumed to warrant cellular proteostasis. Please see Figure 2 for an overview of the interplay between the HSR and the resolution of inflammation and confront AMPK in metabolic stress.

Fig. 2.

Overview of how the HSR integrates the interplay between energy sensing (AMPK pathways) and proteotoxic sensing (HSF1 pathways). Stress-activation of the biosynthetic pathway that leads to HSP70 expression from HSF1 couples reproduction (estrogen, E2), exercise, energy balance and proteostasis to anti-inflammation via HSP70. AMPK, 5′-adenosine monophosphate-activated protein kinase; HSP70, the 70 kDa family of heat shock proteins; HSF1, heat shock transcription factor 1; SIRT1, NAD+-dependent deacetylase sirtuin-1; NLRP3, NOD-like receptor pyrin domain-containing protein-3 inflammasome; eNOS, endothelial nitric oxide synthase; ROS, reactive oxygen species.

Reproduced with adaption from: Nitric oxide-heat shock protein axis in menopausal hot flushes: neglected metabolic issues of chronic inflammatory diseases associated with deranged heat shock response. Hum Reprod Update. 2017;23:600–628. https://doi.org/10.1093/humupd/dmx020. Hum Reprod Update|© The Author 2017. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. Reprinted by permission of OUP on behalf of the Human Reproduction Update.

It merits note that human immunodeficiency virus-1 (HIV-1) has the ability to effectively co-opt the host cellular machinery, with a particular reliance on the host chaperone machinery for the assembly of its proteins.⁹⁷ Among the host proteins involved in this process, HSPBP1, which serves exclusively as a NEF, plays a crucial role in inhibiting HIV-1 long-terminal repeat (LTR) promoter activity. This is because HSPBP1 binds to κB sites at the LTR promoter, which in turn prevents the binding of NF-κB heterodimers (p50/p65) to the same region. Consequently, this leads to the repression of NF-κB-mediated activation of LTR-driven gene expression and HIV-1 proliferation.⁹⁸ The same mechanism is thought to take place during the HSR-mediated resolution of inflammation,⁵⁸ through the HSPA/HSP70-elicited negative regulation of inflammatory responses through, but not limited to, its negative effect on NF-κB signaling pathway.8, 87, 99, 100 In the same line of reasoning is the antiviral effect of the antibiotic geldanamycin, which binds to the ATP pocket of HSPC/HSP90, disturbing the binding of HSPC/HSP90 to HSF1 thereby increasing HSPA/HSP70 gene expression. Therefore, HSP90 inhibitors, such as geldanamycin, can activate a strong HSF1-dependent HSR, resulting in elevated levels of HSP70 and HSP90 itself.⁹⁷ Importantly, inhibition of HSP90 blocks NLRP3 inflammasome activation,¹⁰¹ thus preventing inflammasome-mediated SASP and chronic inflammatory diseases, by stimulating HSP70 expression.⁵⁸ It has also been suggested drug repositioning with HSPC/HSP90 inhibitors (particularly geldanamycin) for treating COVID-19 patients.¹⁰² Actually, several inhibitors of HSPC/HSP90 have been tested in clinical trials.103, 104 So far, none have reached the clinic, reflecting that inhibition of chaperones is a drastic measure.⁵³

With the improvement of experimental tools to identify PTM, it has become clear that not only HSF1, but many chaperones undergo extensive PTMs that determine their chaperone activity as well as interactions of HSP70, HSP90, HSP110, and J-proteins. However, the enzymes responsible for these modifications and the functional consequences that PTMs might have on these proteins remain largely unknown. In addition to the aforementioned transformations and PTM, phosphorylation has notably emerged as an important contributor to the biological activity of HSP70, HSP110, and HSP40/DNAJ interactions. These findings underscore an added layer of complexity in regulating HSP70 function.¹⁰⁵ Lastly, a noteworthy PTM has recently come to light, namely, the covalent modification of HSPs by AMP, a process known as AMPylation. This discovery further reinforces the intricate connection between the HSR and cellular energy levels. When AMPylated, BiP, for instance, exhibits an enhanced capacity for binding and releasing substrates, a diminished basal ATPase activity and a dampened response to ATP hydrolysis stimulation by J-protein cochaperones.¹⁰⁶ These findings compellingly support the notion that AMPylation serves as an inhibitory modification in cellular conditions characterized by reduced demands for anabolic processes or scarcity of energy resources. Kindly refer to Figure 2 for a more comprehensive overview of how the HSR intricately integrates the interplay between energy sensing (AMPK pathways) and proteostasis sensing (HSF1 pathways) defining the regulation of the poise amidst proteotoxic and metabolic stress responses. In any case, it is now evident that both predominantly metabolic stress and predominantly proteotoxic stress responses are governed by the cellular equivalent of a continuously adjustable rheostat, rather than operating as simple binary on/off switches. These regulatory mechanisms are finely attuned to respond to the swiftly fluctuating states of caloristasis, as postulated by Professor Larry Hightower’s group.²⁴

Beyond individual chaperone functions discussed here, the concept of chaperome analysis is gaining prominence as an approach to understanding the role of the HSR in various conditions, including health and disease. The chaperome comprises a diverse family of proteins, encompassing chaperones, cochaperones, and numerous other factors. When examining chaperome subnetworks in specific contexts such as aging or neurodegenerative diseases, it becomes evident that these subsets serve as protective buffers, preserving proteostasis amid proteotoxic stress.¹⁰⁷ Hence, the current trend is to prioritize the examination of chaperome networks over the exclusive emphasis on chaperome inhibitors or stimulators, placing them above genetics or client proteins in research priorities.¹⁰⁸

Coevolution of chaperome and energy-controlling systems

The history of life on Earth, as our present records suggest, was shaped by constant fluctuations in the supply of heat. Since ancient times, Earth experienced numerous and dramatic shifts in its heat dynamics, influencing the adaptation of primordial molecules and the emergence of the first life forms. In the early stages of prebiotic chemistry, energy from the environment was harnessed to construct more complex compounds, enabling the emergence of primordial biochemical systems aimed to protect themselves from processes like hydrolysis and other forms of energy dissipation, approximately 4.0 billion years ago. Notably, during the early Archaean era, Earth’s surface temperatures could have reached scorching levels of 80–100 °C, gradually cooling to 30–50 °C around 3.5 billion years ago.¹⁰⁹ Therefore, it comes as no surprise that the biochemical machineries that developed during the Paleoarchean eon, alongside the emergence of the first life forms, evolved to incorporate both heat-preserving and heat-dissipating pathways. This scenario brings to the stage the coevolution of the chaperome and energy-controlling systems which stand as a remarkable testament to life’s adaptability and resilience in the face of Earth’s ever-shifting thermal challenges.

In contrast to simple prokaryotic organisms, which house thousands of proteins, plants and animals boast an extensive repertoire of hundreds of thousands of proteins. These proteins exhibit greater length, possess multiple functional domains, and encompass a wide array of intricate folds and combinations, interspersed with repeated segments and beta-rich architectures that render them susceptible to misfolding and aggregation.⁴ Interestingly, the relative representation of core chaperones responsible for upholding the fidelity of protein folding in increasingly intricate proteomes underwent minimal change from prokaryotic to mammalian genomes.⁴ This observation is not unexpected, given the pivotal role of protein chaperoning in sustaining cellular vitality. Thus, the optimal evolutionary solution that originated in the LUCA, predating the schism between archaea and bacteria, has persevered largely unaltered since that epoch. Only marginal adaptations have been introduced to accommodate the escalating complexity of proteins over time.

Evidence suggests that the genesis of core chaperones can be traced back to the early diverging prokaryotes, marking the advent of HSP60 as the first core chaperone. HSP60, an ATP-powered unfoldase possessing a cage-like structure, emerged more than 3.6 billion years ago during the Paleoarchean eon, within the era of LUCA. Concurrently, the appearance of HSP20/GroES anti-aggregation chaperones may have occurred in tandem with HSP60 during this epoch. In contrast, the introduction of HSP70, HSP90, and HSP100 into the bacterial domain likely transpired around 3.25 billion years ago. Subsequently, the emergence of chaperones such as HSP110, Bag Hop, and other cochaperones unfolded within the Proterozoic era, around 2 billion years ago, during the era of the last eukaryotic common ancestor, as highlighted by Rebeaud and colleagues⁴ in their recent but seminal paper.

Key milestones in microbial evolution were influenced by critical events, including the extensive glaciation of the late Neoproterozoic Snowball Earth, approximately 750–580 million years ago, and the subsequent oxygenation event of the oceans and atmosphere approximately 630–551 million years ago. These events set the stage for the explosion of complex life forms in the Cambrian, about 540–520 million years ago.¹¹⁰ Particularly noteworthy is the Neoproterozoic oxygenation event (around 600 million years ago), a turning point when metazoans, and subsequently chordates, emerged and faced the imperative necessity to produce protective chaperones against oxidative stress that became pronounced.

In a manner akin to the discussed chaperome, reconstructions suggest a systematic transfer of sequences encoding glycolytic enzymes among diverse organisms. Similar to chaperones, there appears to be limited exchange between bacterial and eukaryotic domains.¹¹¹ Driven by negative selection and subtle enzymatic alterations, significant evolutionary shifts in the function of ADP-dependent sugar kinases conferred vital adaptive advantages, such as enhancing glycolytic efficiency in archaea roughly 3.594 billion years ago.⁵ Apropos, ADP-dependent sugar kinases stand out for their distinctive feature of phosphorylating glucose and fructose-6-phosphate using ADP, instead of ATP as the phosphoryl donor. They appear to represent not the last ancestral forms but rather a transitional link between glucokinase and phosphofructokinase found in modern archaea and higher eukaryotes alike.¹¹²

Glycolytic enzymes, being ubiquitously present, have been instrumental in deciphering evolutionary relationships between organisms. These enzymes serve as suitable evolutionary chronometers, exhibiting a rate of change slow enough to discern broad evolutionary patterns yet swift enough for precise classification of closely related organisms.¹¹³ The phylogenetic divergence of sugar kinases and chaperones traces back to their respective common ancestors approximately 3.5 billion years ago.⁵ Beyond their roles in cellular energy production, notably exemplified by hexokinase as the primary enzyme initiating intracellular glucose metabolism, nature seems to have intricately interwoven sugar metabolism and proteostasis protection within a singular molecule: HSP70. Notably, during the elucidation of the three-dimensional structure of the ATPase fragment of HSPA8 (HSC70), Flaherty and colleagues⁶ serendipitously uncovered striking structural (spatial) similarities, particularly in subdomains Ia and IIa of HSC70 (Figure 3), between the chaperone and hexokinase, despite disparities in their amino acid sequences. Furthermore, significant structural resemblances were found with other sugar kinases, including fucokinase, gluconokinase, xylulokinase, ribulokinase, and glycerokinase. These kinases are predicted to share subdomains with a comparable tertiary structure to the ATPase subdomains Ia and IIa of hexokinase (Figure 3), actin, and HSC70, exhibiting a similar ATP-binding pocket and the ability to undergo interdomain hinge motion during functional state transitions.⁷

Fig. 3.

Representative structures of HSP70. (a) Nuclear magnetic resonance/residual dipolar couplings/X-ray resolution of the tridimensional structure of prokaryotic HSP70 (Escherichia coli’s DnaK) in complex with ADP and a peptide (PDB #2KHO, residues 1-605) showing hexokinase subdomains (Ia and IIa) in the 45 kDa nucleotide-binding domain (NBD). The 30 kDa substrate-binding domain (SBD) is also given. (b) Spatial disposition of Ia and IIa (hexokinase subdomains of ATPase domain) and Ib and IIb (non-hexokinase subdomains) along with their respective residues on NBD of HSP70. (c) X-ray diffraction-solved structure of human HSP70 (HSPA1A/1B) NBD (PDB # 3AY9) in the ADP-bound conformation. (d) X-ray diffraction-solved structure of human HSP70 (HSPA1A) with unhydrolyzed ATP bound to NBD (PDB #5BPM). The N-terminal sequence (Ile9 - Asp10 - Leu11 - Gly12 - Thr13 - Thr14), which is highly conserved among all HSP70s, is depicted in the dark-blue ribbon. (e) X-ray diffraction-solved structure of bovine HSC70 (HSPA8) in the ADP-bound conformation (PDB #1NGB). Structure files were obtained from RCSB Protein Data Bank (PDB) at https://www.rcsb.org/structure/.

Importantly, the divergence among sugar kinases, actin, and HSC70 predates the prokaryote-eukaryote divergence, as HSC70 family members are found in both eukaryotic and prokaryotic organisms.⁷ This suspicion is further corroborated by conformational changes occurring within the ATPase domain that facilitate nucleotide-controlled communication with the substrate-binding domain of HSP70. Evidence indicates that the nucleotide-binding cleft undergoes opening and closure dynamics during the nucleotide-binding/release cycle of both HSP70 and hexokinase.114, 115 A pivotal mechanism underlying HSP70 allostery involves the transmission of information from the nucleotide-binding site to the interdomain linker. In the presence of ATP, the linker interacts with the periphery of the IIa β-sheet, establishing a structural connection between the linker and the nucleotide-binding site. Consequently, an allosteric communication pathway is established, facilitating the transmission of signals from the NBD to the substrate-binding domain via the interdomain linker.116, 117

It is intriguing that a 16 kDa nucleoside diphosphate (NDP) kinase, p16, from the Nm23/NDP kinase family, serves as an accessory protein to HSPA8 (HSC70). This protein can influence HSPA8’s oligomeric state and dissociate unfolded proteins from HSPA8 even in the absence of exogenous ATP.¹¹⁸ NDP kinases facilitate the exchange of terminal phosphates between different NDPs and triphosphates in a reversible manner, resulting in the production of nucleotide triphosphates. Interestingly, p16 is also stress-responsive and exhibits higher sensitivity to proteotoxic stress compared to HSC70 or HSP70 themselves. By converting HSPA8 oligomers into active monomers, p16 appears to regulate HSPA8 activity. Given that ATP/ADP exchange occurs slowly, and no homologs of the Gro-P like protein E NEF have been discovered in the cytoplasm, some substrates may be released very slowly without assistance from accessory proteins. In the presence of p16, however, substrates can be released from HSC70-ADP, and subsequent ATP/ADP exchange transforms HSC70 into a form that is most accessible for substrate binding, as described by Leung and Hightower in 1997.¹¹⁸ Moreover and importantly, phylogenetic analyses of NDP kinases have revealed that these kinases also share a common ancestor in LUCA, as demonstrated.¹¹⁹ This parallels what has been observed with chaperones and sugar kinases. Therefore, it is not unreasonable to consider that an ancestral form of p16 NDP kinase might have played a role in the evolutionary development of early ATP-fueled chaperones, providing an initial link between HSPA8/HSPA1A and energy metabolism. Notably, rat HSC70 proteins expressed in E. coli exhibit significant structural similarities with E. coli DnaK protein,¹²⁰ a much older protein in the Tree of Life. This may suggest that an ancestral ATPase activity may still be present in HSPA8, reinforcing the connection between chaperones and energy metabolism. Considering all the above observations, therefore, it is plausible that the 44 kDa ATPase fragment of present-day HSC70 had evolved from a common ancestor shared with hexokinases and actin. The question of whether HSC70 possesses some kinase activity toward glucose, however, remains a subject of debate, yet it warrants investigation.

In addition to the role of AMPK, which can either downregulate or upregulate the HSR through the GSK-3β/HSF1 duet, as previously discussed, another pathway involving glucose metabolism, known as the HBP, plays a significant role in HSR regulation. The HBP, which is a nutrient-sensing pathway with intricate ties to energy metabolism and the HSR, extends beyond the role of GSK-3β in glycogen synthesis.15, 121, 122 Elevated levels of circulating glutamine, as seen during exercise, lead to a notable diversion of fructose-6-phosphate from glycolysis into the HBP through its interaction with glutamine, catalyzed by GFAT (glutamine-fructose-6-phosphate transaminase). This results in the production of glucosamine-6-phosphate. The final metabolite of the HBP, UDP-N-acetylglucosamine (UDP-GlcNAc), exerts a dual effect on the HSR: firstly, by inhibiting GSK-3β and secondly, by covalently modifying HSF1. This modification involves O-linked N-acetylglucosaminylation, enhancing HSF1’s DNA-binding and transcriptional activities regarding heat shock genes. The heightened availability of glutamine stimulates increased flux through the HBP, resulting in elevated HSP70 expression.¹²³ A comparison between parts A and B of Figure 4 illustrates the flux balance analysis of the HBP¹⁵ under two distinct plasma and muscle glutamine concentrations in rats.¹²⁴ Additionally, the intensified flux through the HBP induces a minor redox imbalance, sufficient to upregulate the expression of genes responsible for redox protection, including those involved in the biosynthesis of GSH and glutamine itself.15, 124, 125 De novo GSH synthesis primarily occurs through transcriptional regulation, involving a series of signaling events culminating in the binding of nuclear factor-erythroid 2 p45-related factor 2 to promoter regions containing antioxidant response elements within the nucleus.126, 127, 128, 129, 130, 131 These findings highlight the intimate interplay between the HSR, energy metabolism, and redox protection (crucial for proteostasis). They also shed light on why glutamine emerges as a potent co-inducer of the HSR, that is, HFS1 must be previously recruited to have its function augmented by HBP. Please see Figure 5 for an integrative overview of caloristasis networks and energy regulation through glucose metabolism.

Fig. 4.

Heat shock response interplay with glutamine metabolism via hexosamine biosynthetic pathway (HBP). Depicted are the major routes of glucose utilization after its entry into cells. Soon after passing the hexokinase (HK) bottleneck, phosphorylated glucose may be diverted to glycolysis, glycogen synthesis, or the pentose-phosphate shunt (hexose-monophosphate shunt), in a proportion that depends on the cell type and physiological conditions. The present artwork is a graphic illustration of experimental values obtained from soleus and gastrocnemius muscles of 8-week trained (treadmill) rats treated or not with l-glutamine supplementations during the last 21 days.¹²⁴ Under l-glutamine supplementations, excess intramuscular l-glutamine supply enforces fructose-6-phosphate (F6P) to divert from glycolysis and enter the HBP (shaded box in the center) thus changing the heat shock response and antioxidant metabolism. The thicknesses of the arrows indicate the approximate proportion of each metabolite entering each given sub-pathway (parentheses). Data are given in μmol min⁻¹ mg protein⁻¹. ARE, antioxidant response elements; HSF1, heat shock transcription factor 1.

Reused from: Physiological regulation of the heat shock response by glutamine: implications for chronic low-grade inflammatory diseases in age-related conditions. Nutrire. 2016;41:17. https://doi.org/10.1186/s41110-016-0021-y. Under the open access Creative Common CC BY license 4.0.

Fig. 5.

Extended perspective on the heat shock response within the context of caloristasis. An illustration of potential alterations in glucose metabolism, diverging from the typical predominance of cytosolic glycolysis, is given. These deviations can lead glucose metabolism towards two critical pathways: the pentose-phosphate shunt, involved in lipid synthesis and antioxidant metabolism, or the hexosamine biochemical pathway, which modulates glycogen synthase-3β and, consequently, impacts HSF1 activity and the regulation of the heat shock response itself. The pivotal role of AMP-activated protein kinase (AMPK) in this process is highlighted. HSP70, the 70 kDa family of heat shock proteins; HSF1, heat shock transcription factor 1; HuR, human antigen R, a.k.a. ELAV-1, for embryonic lethal, abnormal vision, Drosophila, homolog-like protein-1; SIRT1, NAD⁺-dependent deacetylase sirtuin-1; UPR, unfolded protein response.

The HSR has not only evolved at the molecular level but also at the organismal level, forming a close linkage with the sympathetic nervous system (SNS), which serves as the principal physiological sentinel responsible for safeguarding homeostasis against all forms of external threats. As expected, the activation of the SNS in situations such as starvation, physical exercise (fight-or-flight stress), and even heat therapy itself11, 132 can elicit a potent and sustained HSR.⁵⁸ Indeed, the intricate connections among glucose levels, energy sensing, the HSR, and sympathetic activity converge within the ventromedial hypothalamus (VMH), a central hub orchestrating functions like feeding, fear, thermoregulation, and sexual activity. The VMH also serves as a site for a robust estrogen-induced HSP70 response.133, 134 Furthermore, norepinephrine can modulate the hypothalamic mechanisms responsible for fever induction, an HSR stimulator, particularly in the preoptic area of the hypothalamus.¹³⁵ It is noteworthy that acute hypoglycemia (metabolic stress) triggers the release of extracellular HSP72 and pro-inflammatory cytokines, such as IL-6.¹³⁶ Contrarily, glucose ingestion inhibits exercise-induced eHSP70 secretion¹³⁷ and abolishes the counterregulatory responses prompted by hypoglycemia via the SNS.¹³⁸ This modulation is reliant on the neuronal circuitry of VMH.¹³⁹

While HSPA/HSP70 chaperones are abundant and extensively studied as a major work-force of core chaperones, HSPC/HSP90 plays a crucial role in the cardiovascular system by binding to key proteins involved in vascular relaxation, such as endothelial nitric oxide synthase (eNOS/NOS3) and soluble guanylate cyclase.¹⁴⁰ Moreover, nitric oxide generated in blood vessels by eNOS serves as a potent inducer of HSP70 synthesis. This induction occurs because nitric oxide induces slight oxidative stress, activating HSF1 via disulfide bond formation.11, 70 Besides this, HSPC/HSP90 also chaperones essential proteins related to animal reproduction, including glucocorticoid receptor, progesterone receptor, estrogen receptor, androgen receptor, mineralocorticoid receptor, and HSF1 itself.85, 141

The interplay between reproduction, HSR, and energy control goes beyond and reveals how evolution brought together physiological functions that depend on energy perception and the HSR. Accordingly, a primary clue to this understanding comes from the interactions between steroid hormone receptors and HSPC/HSP90, which are instrumental in guiding steroid hormones to their specific target promoters. However, further away from their role as regulators of steroid hormone receptors, steroid hormones themselves play a pivotal role in governing a critical facet of the HSR that pertains to reproduction in mammals and birds. As reviewed by Miragem and Homem de Bittencourt,¹¹ E2 not only influences the function of steroid hormone receptors but also initiates HSR in hypothalamic neurons responsible for producing gonadotropin-releasing hormone (GnRH). In this context, GnRH neurons release GnRH into the median eminence, a process that triggers the secretion of luteinizing hormone by the pituitary gland and, consequently, induces ovulation/estrus. However, for the secretion of GnRH to occur, GnRH neurons depend on input from kisspeptin-neurokinin B-dynorphin (KNDy) neurons located in the infundibular nucleus (arcuate nucleus in other mammals). KNDy neurons, which dictate the rhythm and intensity of pulsatile GnRH secretion, are responsive to various factors, including temperature, prostaglandin E₂ (PGE₂) during immune responses, energy reserves signaled by adipokines, and the overall metabolic status as conveyed by the VMH through medial preoptic and median preoptic nuclei of the hypothalamus. Lastly, VMH serves as the trigger for activating SNS efferences in response to perceived threats to homeostasis. These inputs collectively regulate the pulsatile secretion of GnRH, ultimately leading to ovulation/estrus and heightened E2 production. Furthermore, E2, in turn, acts as a potent inducer of HSP70 expression in various regions of the hypothalamus that are involved in both reproduction and energy sensing. Please refer to Figure 2 for a schematic representation of the role of E2 within the HSR.

The influence of environmental signals on reproduction involving the HSR extends beyond humans and mammals. In the avian realm, for instance, gonadotropin (follicle-stimulating hormone and luteinizing hormone) synthesis and release are stimulated by GnRH, as observed in breeding birds.¹⁴² Seasonal breeding avian species further showcase the intricate interplay between environmental factors such as light, temperature, food availability, and mate presence, all meticulously integrated by HSR-eliciting KNDy neurons to regulate reproduction, by modulating the release of gonadotropins from the pituitary gland.¹¹ A practical example of this connection can be witnessed in poultry farming: anyone who raises chickens knows that, when the weather gets excessively hot (surplus energy) or cold (scarce energy availability), or when the chicken is too fat (excess energy storages) or too lean (low energy reserves), egg production takes a significant hit. Seasonal breeders like sheep exhibit reproductive quiescence during the spring and early summer months (hotter weather). The measurement of circannual time crucially relies on photoperiod changes, with the pineal gland assuming a pivotal role by secreting melatonin during the night. This translates environmental cues into physiological signals that find integration at the level of HSR-elaborating KNDy neurons.¹⁴³ In essence, the E2-sensitive KNDy neurons not only serve as conduits for integrating environmental energy signals into the territory of reproduction but also play a dual role by triggering the HSR, a cellular mechanism providing anti-inflammatory and proteostasis-saving functions. This allows for E2-mediated protection against neurodegenerative diseases during the reproductive phase of vertebrates. Therefore, animal reproduction and neuroprotection are dependent on energy sensing that triggers the cytoprotective HSR.

For quite some time, it has been established that heat plays a pivotal role in shaping various metabolic adaptations. For instance, in cold-acclimated trout exposed to temperatures as low as 4 °C, there is a redirection of glucose utilization. This shift leads to an enhancement in the rate of fat synthesis and promotes a more efficient energy production system, particularly through fat oxidation, as part of the organismal response to cold compensation.¹⁴⁴ Furthermore, evolution has masterminded adjustments within biochemical systems to ensure the preservation of fundamental structures and processes across diverse organisms, irrespective of their environmental milieu.¹⁴⁵

Much like the arguments in the Introduction section regarding the stenothermal Antarctic fish Emerald rockcod,⁴⁴ stenothermal Antarctic notothenioid fish populations exhibit deficiencies in their capacity for temperature-induced gene expression, notably the absence of a genuine HSR. These limitations may impair their ability to acclimate to elevated temperatures, with these animals remaining unable to acclimate to temperatures exceeding approximately 4 °C. Curiously, full HSR is observed in temperate New Zealand notothenioids, suggesting that the evolutionary adaptation to cold and stable thermal conditions has depleted the genetic resources of Antarctic fishes.¹⁴⁶ Conversely, the preservation of a robust HSR equips exotic invader species with the means to acclimate to new environments and potentially outperform their native counterparts in metabolic adaptations.¹⁴⁷

In light of the broader insights into the evolution of the chaperome and sugar kinases, a remarkable convergence emerges—both in the context of heat exchange between environmental factors and molecular entities, and subsequently, between the environment and cellular systems responsible for energy production and protection against proteotoxic stresses. This convergence becomes especially pronounced upon considering that AMPK, the central regulator of cellular energy sensing, features a glycogen-binding domain on its β regulatory subunit,¹⁴⁸ endowing it with the capacity to function as a glycogen sensor. Remarkably, it is of note that glucose can downregulate AMPK through protein phosphatase 2A, a phenomenon observed in various contexts, including rodent islets, β-cell strains,¹⁴⁹ and yeast.¹⁵⁰ This underscores the complexity of cellular responses to metabolic stress, which may hinge upon the intricate proteostasis of the cells. When cells confront another type of metabolic stress, suboptimal oxygen conditions, not only does total ATP turnover significantly decrease, but each ATP-consuming process experiences distinct impacts. In essence, only vital processes, such as the sodium/potassium pump necessary for maintaining membrane potential and cell volume, assume prominence as the primary energy sinks in the energy-restricted state.¹⁵¹ Similarly, in the context of hyperbaric oxygen therapy, as demonstrated in an elegant set of experiments by Tezgin and co-workers, application of high partial pressure of oxygen redirects mitochondrial metabolism towards cytosolic glycolysis, that is, the Warburg effect also known as aerobic glycolysis,¹⁵² serving as a means to maintain proteostasis without necessitating the activation of a conventional HSR.²⁴ Similarly and notably, nutrient deprivation induces the generation of proteotoxic ROS alongside AMPK-dependent phosphorylation of pyruvate kinase.¹⁵³ This phosphorylation inhibits pyruvate production, forcing cells to rely solely on cytosolic glycolysis, ultimately resulting in the Warburg effect, where lactate accumulates. Consequently, it is reasonable to anticipate that evolution has minimized the ATP-consuming chaperone machinery in such states, thereby justifying the discussed caloristatic control switch (Figure 1).

Conclusion

In the grand tapestry of cellular life, the evolution of the HSR and the emergence of sugar kinases appear as two threads, interwoven over billions of years, each influencing the other in a dance of molecular intricacy. Just as nature crafts a mosaic of elements into a harmonious whole, the HSR, originally designed to combat proteotoxic challenges, now stands as a sentinel guarding the sanctity of cellular energy metabolism. Actually, the HSR is a story of sentinels; sentinels of homeostasis. Professor Ferruccio Ritossa’s pioneering insights revealed that the HSR, like a sentinel, senses not only the denaturing fires of heat but also the flickering uncertainties of energy paucity, working as a sentinel ready to rise in defense against both. This revelation, echoing through the annals of scientific discovery, aligns the HSR as a guardian of cellular homeostasis, a shield against the chaos that threatens proteostasis and metabolic equilibrium.

As we trace the evolutionary footprints left by the LUCA, a disclosure emerges—the origins of the proteostasis-defending systems coincide with the nascent steps of sugar kinases, as if destiny bound them together in the crucible of time. The NBD of HSP70, the main work-power of the protein chaperone machinery, bears subdomains akin to those found in sugar kinases, a testament to the ancient partnership between protein folding and energy production.

Beyond the realm of molecular intricacies, the HSR extends its embrace to the wider landscape of cellular life, integrating energy metabolism and the resolution of inflammation. In a delicate choreography, AMPK, the sentinel of energy scarcity, assumes the mantle of a maestro, orchestrating a symphony of responses that can vary from energy-saving, proteostasis-protecting or resolution of inflammation. Meanwhile, SIRT1, awakened by the whispers of calorie restriction, joins the orchestra, amplifying the HSR’s melody. This interplay extends beyond chaperones, shaping the destiny of energy metabolism genes themselves. As glucose-regulated chaperones harmonize their tune with the HSR, the HBP adds its notes to the score, guiding GSK-3β to silence and HSF1 to rise. Yet, within this symphony, paradoxes arise, akin to a melody that shifts unexpectedly. AMPK, usually an ally in the HSR’s anti-inflammatory cause, becomes an enigmatic muse, impeding the nuclear journey of HuR, the stabilizer of SIRT1 mRNA. This paradoxical dance reveals the nuanced equilibrium of cellular stress responses, a dance governed by the dual maestros, AMPK and HSF1, creating a new composition—caloristasis, a delicate thermodynamic balance between energy homeostasis and proteostasis, woven with precision.

In this intricate mosaic, our goal was to shed light on the interplay between the HSR as a guardian of proteostasis and the pathways dictating energy metabolism. The caloristatic controlling switch, summarized in Figure 1, and in an extended view in Fig. 2, Fig. 5, may hold the key to unraveling the mysteries of obesity, diabetes, cardiovascular and neurodegenerative diseases, reproductive abnormalities, as well as pharmacological and non-pharmacological interventions including the artistry of well-structured exercise routines. Like a maestro conducting an orchestra, we stand on the precipice of discovery, ready to uncover the hidden harmonies that govern life’s symphony.

Funding and support

This work has been supported by The State of Rio Grande do Sul Foundation for Research Support (FAPERGS/Decit/SCTIE/MS/CNPq/SESRS n. 03/2017-PPSUS #17/2551-0001424-3 to MSK and #19/2551-0001713-8 to PIHBJ) and The Brazilian National Council for Scientific and Technological Development (CNPq, process #303853/2017-4 to PIHBJ, process #307926/2022-2 to TGH and process # 302959/2020-3 to MSK). Financial support from CAPES (Coordination of Superior Level Staff Improvement) is also acknowledged.

Author contribution

PIHBJ conceptualized the paper, authored the initial draft, and oversaw its finalization. All the authors were involved in co-writing this work. PIHBJ prepared the figures. All the authors have read and agreed to the submitted and published versions of the manuscript.

Declarations of interest

The authors declare no conflict of interest and no competing interests such as consultancies, financial involvement, patent ownership, etc. in relation to the work described. CNPq, FAPERGS, and CAPES (the funding organisms) had no involvement in the propositions presented in this manuscript.

Data availability statement

Data will be made available on request.