HSP90

Heat Shock Protein 90: Structure + isoforms

HSP90 homologues are comprised of three conserved domains: N-terminal domain (NTD), C-terminal domain (CTD), and middle domain (MD) each with a specific function. There is a variable charger link domain present only in eukaryotes that links the NTD to the MD. HSPs are often classified based on their weight.

N-terminal Domain:

NTD contains an ATP binding motif so it is called the nucleotide-binding site. The ATP binding motif is highly conserved with the GHKL (Gyrase, HSP90, histidine kinase, and MutL) superfamily but differs from HSP70 or protein kianses. HSP90 ATPase activity is necessary for the chaperone cycle and for binding to client proteins. NTD has been utilized as a major target for HSP90 inhibitors such as Geldanamycin (GA) and Radicicol (RD) which compete with ATP for the binding pocket.

Charged Linker Region (CR):

Eukaryotic HSP90 contains a CR (absent in prokaryotes) which links the NTD to the MD. This domain has variable length and amino acid composition but is typically highly charged. This region is attributed to increased flexibility and dynamicity to help it manoeuvre in the crowded eukaryotic cell. Some studies suggest the charged region as a modulator of the chaperone activity, others neglect its potential influence on the HSP90 functions.

Middle Domain (MD):

MD modulates HSP90 function by binding the γ-phosphate of ATP specified for the NTD, to modulate its ATPase activity. Additionally, several studies demonstrated that this domain is implicated in binding co-chaperones like Aha1 and interacting with client proteins.

C-Terminal Domain (CTD):

CTD contains two key sites for calmodulin binding and another for HSP90 homodimersiation. CTD also contains a nucleotide-binding site which opens following the occupancy of the N-terminal site and serves as allosteric regulator of the N-terminal ATPase activity. The CTD domain differs from the NTD in terms of ligand specificity. The N-terminal binding site was found specific to adenosine nucleotides containing an intact adenine ring (such as nicotinamide adenine dinucleotides and adenosine polyphosphate alarmones), while the C-terminal binding site was much more unspecific by interacting with both purine and pyrimidine nucleotides (such as GTP and UTP nucleotides).

2′,3′-O-(2,4,6-trinitrophenyl)-nucleotides (TNP-ATP, TNP-GTP) and pyrophosphate were able to bind the C-terminal binding site with no need for occupied N-terminal site. Drugs like novobiocin (NB) and epigallocatechin-3-gallate (EGCG) could affect the function of HSP90 by targeting the CTD.

The CTD is included in the chaperone activity of HSP90 towards transcription factors and vesicular stomatitis virus G-protein. Targeting the CTD by the monoclonal anti-hsp90 antibody could hinder the interaction of HSP90 with many protein clients including steroid receptors and actin filaments.

Isoforms:

There are two main mammalian isoforms of HSP90, localized in the cytoplasm; the inducible form Hsp90α and the constitutive form HSP90β. Although these cytoplasmic HSP90 isoforms have recently shown distinctive functions, most reviews do not distinguish them due to their remarkable structural and functional similarity. Therefore, the name HSP90 has been used for both HSP90 α and β unless indicated. Studies dealing with the differentiation between cytoplasmic HSP90 isoforms, revealed that although both isoforms very often form dimers, HSP90α tends to dimerize frequently compared to HSP90β. In addition, the α and β isoforms vary in specific regions along the protein sequence, suggesting diverse functions of the two isoforms.

See Figure 1 for an image outlining the different roles of HSP90 isoforms.

Dimerization:

Dimerization is essential for proper functioning of HSP90. Mammalian HSP90 is a phosphorylated dimer containing 2–3 covalently-bound phosphate molecules per monomer. Structural studies revealed that dimeric forms of the protein have elongated structure. As a hydrophobic chaperone, heat shock increases its hydrophobicity. Another aspect is because HSP90 has an ATP binding site in its NTD, self-phosphorylation is likely to occur with profound conformational changes. Interestingly, although well known to have an ATP binding domain, HSP90 was found to have both ATPase and GTPase activity.

Stability:

Heat Shock Proteins are capable of withstanding higher temperatures than typical proteins and resist denaturation due to extreme temperatures and conditions. This resistance is attributed to their 3D structure that contains more abundant and stronger hydrophobic packing, hydrogen bonds, enhanced secondary proteins structure, and helix dipole stabilization.

The formation of fibrils often accompanies diseases, the most well-known of which include Alzheimer's disease, Type II diabetes, and spongiform encephalopathies (for example, Mad Cow Disease and Creutzfeldt-Jakob Disease/CJD).

Amyloid fibrils are formed when normally soluble proteins assemble and aggregate to form insoluble fibers which are resistant to degradation, via a process of polymerisation. These fibrils are deposited extracellularly in the tissues and result in pathogenic effects. Such agglomeration of fibrils is inherently stable, with structural studies revealing that these are mainly composed of β-sheet structures in a cross-β conformation. As such, the fibrils are self-assembled and proteins themselves can fold into alternative conformations which allows for this self-assembly.

​

There are 3 main proposed models for the self-assembly of amyloid fibrils from their monomers, these being:

1. The refolding model

2. The natively disordered model

3. The gain-of-interaction model

The refolding model is the most well-known model and it describes a situation wherein a protein refolds from its native state to an amyloid-forming state.

The natively disordered model suggests that the protein is natively disordered and folds directly into an amyloid-forming state.

The gain-of-interaction model is where part of one amyloid protein monomer becomes a domain in another amyloid protein monomer which eventually forms the amyloid fibril.

It should also be noted that while, on the whole, amyloid fibrils of the same protein form consistent structures, other fibril features such as morphology and properties can vary according to exposure to different environmental conditions during growth. Hence, minute changes to the growth conditions can have a significant effect on the potential cytotoxicity of the formed amyloid fibril.

​

The reasons why amyloid fibrils cause diseases are still unclear. In some cases, the deposits physically disrupt tissue architecture, suggesting disruption of function by some bulk process. An emerging consensus implicates prefibrillar intermediates, rather than mature amyloid fibers, in causing cell death, particularly in neurodegenerative diseases.

​

MECHANISMS OF TOXICITY:

Calcium dysregulation has been observed to occur early in cells exposed to protein oligomers. These small aggregates can form ion channels through lipid bilayer membranes and activate NMDA and AMPA receptors. Channel formation has been hypothesized to account for calcium dysregulation and mitochondrial dysfunction by allowing indiscriminate leakage of ions across cell membranes. Studies have also shown that amyloid deposition is associated with mitochondrial dysfunction and a resulting generation of reactive oxygen species (ROS), which can initiate a signaling pathway leading to apoptosis.

Beta‐amyloid (Aβ) accumulation is also commonly associated with Alzheimer's disease. Aβ toxicity results in an age‐related increase in cholinergic loss and microglial activation in the brain, along with neuronal loss. Studies have also shown that age plays a crucial role in susceptibility of the brain to certain Alzheimer's pathologies including accumulation of beta-amyloid.

All these mechanisms of toxicity are likely to play a role. In fact, the aggregation of a protein generates a variety of aggregates, all of which are likely to be toxic to some degree. The oligomers have also been reported to interact with a variety of molecular targets. The misfolded nature of protein aggregates causes a multitude of aberrant interactions with a multitude of cellular components, including membranes, protein receptors, soluble proteins, RNAs and small metabolites. Hence, it is unlikely that there is a unique mechanism of toxicity or a unique cascade of cellular events.