Aim for the Core: Suitability of the ubiquitin-independent 20S proteasome as a drug target in neurodegeneration

Abstract

Neurodegenerative diseases are a class of age-associated proteopathies characterized by the accumulation of misfolded and/or aggregation-prone proteins. This imbalance has been attributed, in part, to an age-dependent decay in the capacity of protein turnover. Most proteins are degraded by the ubiquitin-proteasome system (UPS), which is composed of ubiquitin ligases and regulatory particles, such as the 19S, that deliver cargo to the proteolytically active 20S proteasome core. However, a subset of clients, especially intrinsically disordered proteins (IDPs), are also removed by the action of the ubiquitin-independent proteasome system (UIPS). What are the specific contributions of the UPS and UIPS in the context of neurodegeneration? Here, we explore how age-associated changes in the relative contribution of the UPS and UIPS, combined with the IDP-like structure of many neurodegenerative disease-associated proteins, might contribute. Strikingly, the core 20S proteasome (20S) has been shown to predominate in older neurons and to preferentially act on relevant substrates, such as synuclein and tau. Moreover, pharmacological activation of the 20S has been shown to accelerate removal of aggregation-prone proteins in some models. Together, these recent studies are turning attention to the 20S proteasome and the UIPS as potential therapeutic targets in neurodegeneration.

INTRODUCTION TO THE PROTEASOME

The proteasome is a central protein degradation machine in eukaryotes¹. Through hydrolysis activities, it removes damaged proteins and ensures the delivery of amino acids to support ongoing biosynthesis. In addition, the proteasome has been co-opted for more specialized tasks in regulating the cell cycle, differentiation, the inflammatory response, antigen presentation and apoptosis^2,3. To enable these functions, the proteasome makes up a staggering 1 to 2% of the entire proteome in healthy cells. However, a decline in proteasome activity has been broadly implicated in ageing and age-associated diseases, including neurodegeneration. Presumably, this decline contributes to a catastrophic imbalance in proteostasis and accumulation of damaged and/or misfolded proteins. In this review, we explore the structure-function of the proteasome and its implications in the onset and progression of neurodegenerative disease. In addition, we focus on emerging therapeutic opportunities through pharmacological activation of this degradation machine.

The 20S proteasome (20S) is a barrel-shaped complex comprised of four heptameric rings: two stacked β-rings that are sandwiched by two α-rings (Fig 1A). Three of the seven subunits (β1, β2, & β5) that make up the β-ring are proteases that hydrolyze peptide bonds of substrates. These active sites are sequestered in the interior of the 20S chamber, such that substrates must first traverse through the exterior α-rings. In its closed state, the α-rings have a narrow pore that occludes the entry of most proteins⁴. Thus, one key to understanding proteasome regulation is to learn how substrates are granted access to the proteolytic chamber. Substrates are targeted to the proteasome through two major pathways, the ubiquitin-proteasome system (UPS) and the ubiquitin-independent proteasome system (UIPS). Proteasome activators (PA), which are predominantly multi-protein complexes, help facilitate degradation by the 20S. There are many types of PAs and the specific one that is bound determines whether that 20S proteasome is coupled to the UPS or UIPS (Fig 1A). However, most of the PAs share a conserved tripeptide sequence, the HbYX (Hydrophobic-TYRosine-unspecified residue ‘X’), at their C-termini that interacts with pockets in the α-rings of the 20S to allosterically open the pore⁵.

Figure 1. Structure, Function, and Substrate Profile of the Proteasome Degradation Pathway.

(A) Components of the ubiquitin proteasome system (UPS) and the ubiquitin (Ub)-independent proteasome system (UIPS). (B) Diagram of substrate turnover by the different proteasome pathways. The 26S proteasome degrades over 90% of the proteome, including intrinsically disordered proteins (IDPs) and folded proteins, by the Ub- and ATP-dependent UPS. The UIPS targets substrates independent of Ub-conjugation and can effectively degrade IDPs, which constitute up to 41% of the proteome, but not folded proteins due to their three-dimensional structure. Cellular stresses, including mutations, post-translational modifications (PTMs) and oxidative damage, can partially unfold structured proteins making them susceptibility to turnover by the UIPS. Partially unfolded proteins (and IDPs) can self-associate to form insoluble aggregates, which cannot be processed by the proteasome, and are predominantly cleared from the cell through the lysosome-autophagy pathway.

Ubiquitin-proteasome system (UPS)

Proteasomal degradation by the UPS first requires the conjugation of multiple ubiquitin (Ub) proteins onto the substrate, generating the polyUb signal that designates it as a substrate of the proteasome. Recent work has shown that conjugation of two or more polyUb chains is needed on the tagged substrate to efficiently interact with the UPS machine⁶. Thus, regulation of this pathway by the activity of the E1, E2 and E3 Ub ligases is a critical component of its function⁷, but will not be described in detail here. The canonical regulatory particle of the UPS is PA700 (or 19S), which is a 700 kDa proteasome activator complex that associates with the 20S to create the 26S proteasome (26S)⁸. PA700 is comprised of a ‘base’ and a ‘lid’. The lid contains subunits that bind to polyUb chains, as well as deubiquitinating enzymes (DUBs) that regulate association with the particle. The base contains the HbYX motifs that interact with the α-rings, and ATPases that unfold the substrate so that it can access the proteolytic chamber⁹. Recent reviews provide additional information about the structure of the 26S and its biological fuction⁶.

Ub-independent-proteasome system (UIPS)

Ub-independent degradation is coordinated by the 20S and may be amplified with UIPS-specific PAs, including PA200 and the heptameric PA28¹⁰. PA200 is a monomeric protein that uses a C-terminal HbYX motif to bind to and activate the 20S. PA28 is composed of multiple, different subunits (alpha, beta, and gamma) and it relies on an alternative (e.g. non-HbYX) motif for association with the 20S^11,12. The UIPS-specific PAs typically lack the unfolding activity of PA700; rather, they open the α-ring gate through a binding-induced conformational change and increase the flux of suitable substrates into the proteolytic chamber¹³. As discussed below, this mechanism restricts UIPS substrates to unfolded proteins that can fit into the channel without an active unfoldase. However, PA700-bound 26S has an open α-ring gate too and is thus capable of facilitating ubiquitin-independent substrate turnover¹⁴. The relative contributions of the 26S in the UPS and UIPS pathways remain unclear (Fig 1B); and, for simplicity, we will only mention the contributions of the 26S to the UIPS in passing in this review. Finally, the free 20S (no PA) is likely to be a contributor to the UIPS. Although the 20S has relatively low enzymatic activity in the absence of PAs (see below), some small or unfolded substrates may be able to traverse the closed gates and be degraded by the minimal machine.

Substrate-targeting by the UPS and UIPS

Over 90% of the human proteome is regulated by the UPS¹⁵. These substrates include a vast array of structured (or folded) proteins, intrinsically disordered proteins (IDPs) and proteins containing intrinsically disordered regions (IDRs). Structured proteins must be unfolded prior to their degradation and can therefore only be cleared by the UPS^16,17. Essentially, folded proteins cannot fit through the narrow axial pore of the 20S, making them inaccessible to degradation by the UIPS particles¹⁸. However, IDPs and IDR-containing proteins, which lack this three-dimensional structure, are thought to readily traverse the α-ring gate¹⁹. Twenty percent of cellular proteins are classified as IDPs and as many as 41% of the eukaryotic proteome is predicted to contain IDRs^20,21, suggesting that the substrate pool of the UIPS may be considerably large. These substrates are particularly relevant for this discussion because they include the proteins that accumulate in neurodegenerative disorders, such as amyloid beta, tau, TDP-43 and α-synuclein (Table I)^22,23.

Table I.

IDPs and associated neurodegenerative diseases. Adapted from Uversky⁴⁰.

Protein	Disorder by prediction (%)a	Function	Disease(s)
Amyloid-β	16.9	Peptidic fragment of APP, which regulates synapse formation, and neuronal plasticity	Alzheimer’s disease Amyloidosis
Tau	77.6	Promote the assembly of and stabilizes neuronal microtubules	Tauopathies Alzheimer’s disease Corticobasal degeneration Pick’s disease Progressive supranuclear palsy
TDP-43	57.3	Transcriptional repression, pre-mRNA splicing, and translational regulation	Amyotrophic lateral sclerosis and frontotemporal lobar degeneration
α-Synuclein	90.7	Regulate synaptic vesicles	Alzheimer’s disease (α) Multiple system atrophy(α) Parkinson’s disease (α, β, γ) Diffuse Lewy body disease (α, β γ)
β-Synuclein	87.3
γ-Synculein	100
FUS	90.7	Transcriptional regulation (initiation & repression), and RNA-binding	Amyotrophic lateral sclerosis

^aDisorder was predicted by PONDR^® VSL2.

In cells, IDPs typically have shorter half-lives relative to structured proteins²⁴. The UPS and UIPS have both been shown to facilitate the rapid proteasomal degradation of IDPs, such as p53 and p73²⁵. Recognition of the IDPs by these pathways is mediated, in part, by disordered regions that act as signals (or degrons)²⁴. Evidence for the UIPS in this process comes from experiments in which removal of the ubiquitinated lysine has been found to have little effect on turnover^26,27. Thus, it seems that both the UPS and UIPS can contribute to the turnover of IDPs.

ROLE OF THE PROTEASOME IN AGEING AND NEURODEGENERATIVE DISEASES

Neurodegenerative disorders, such as Alzheimer’s disease (AD), Parkinson’s disease (PD) and amyotrophic lateral sclerosis (ALS), are characterized by the progressive structural and functional impairment of neurons, resulting in neuronal death²⁸. Although these diseases have different clinical symptoms, they are all associated with the accumulation of aggregated proteins^29,30. This observation implicates the age-dependent decline in proteostasis and the contributing role of proteasomal dysfunction³¹. Consistent with this idea, species with increased longevity and long-lived individuals within a given species exhibit higher proteasome activity and are less susceptible to diseases, including neurodegeneration^32,33.

Many of the proteins implicated in neurodegeneration are IDPs³⁴. Due to their conformational flexibility, IDPs are prone to aggregation³⁵ and present a particular risk in aged neurons, where the protein quality control network has deteriorated³⁶. Against this backdrop, the proteasome likely plays a critical role because it rids neurons of damaged and misfolded proteins that may cause neuronal dysfunction. The role of the UPS in this process has received substantial attention due, in part, to the early observation that ubiquitinated proteins are found within disease-associated aggregates in postmortem brains³¹. Indeed, many recent reviews have focused on the connection between the UPS and neurodegeneration³⁷. Similarly, the immunoproteasome, composed of alternative subunits³⁸, likely plays a role, especially in the immune regulation of the diseases. The structure-function of the immunoproteasome has also been reviewed³⁹ and will not be discussed here. Instead, we focus on the relatively under-studied position of the UIPS.

Cells employ the UIPS to cope with proteotoxicity

During neurodegeneration, neurons are challenged with a higher substrate load, potentially causing an imbalance in proteostasis and creating a feed-forward loop that compromises proteasome function⁴¹. Specifically, increased levels of misfolded proteins are though to “clog” the proteasome and impair its function⁴². There are several proposed mechanisms for this inhibition, including stabilization of the proteasome in its closed-gate conformation^43,44. The feed-forward aspect of the relationship then comes about when the resulting proteotoxicity triggers mitochondrial dysfunction and increases reactive oxygen species (ROS) and DNA damage⁴⁵. These processes contribute to further protein misfolding and inactivation of the proteasome through direct oxidation of its subunits⁴⁶. Cells have adopted approaches to counteract this problem; for example, a rise in ROS induces expression of PA28α, which promotes UIPS activity⁴⁷. Additionally, oxidative stress induces expression of 20S subunits through the oxidative stress sensor, Nrf2 (SKN-1)⁴⁸. Interestingly, the levels of the UPS-associated PA700 remain constant under those conditions⁴⁹, suggesting that oxidatively damaged proteins may be preferentially cleared by the UIPS. Consistent with this idea, oxidative stress also mediates the Ecm29-dependent disassembly of 26S (Fig 2A)⁵⁰. Collectively, these findings suggest that challenged cells rely on the UIPS to degrade oxidatively damaged and unfolded proteins.

Figure 2. Proteasomal Regulation in Ageing and Neurodegenerative Disease Models.

(A) Partial list of stresses that effect UPS (orange) and UIPS (blues) function. (B) Summary of the cellular composition of proteasome complexes (20S, 26S, and UIPS-specific) in young vs. aged fibroblasts. See the text for citations. Briefly, proteasome complexes from young or old fibroblasts were placed into categories based on whether they were free 20S or whether they contain PA700 (26S/UPS) or PA28 (UIPS)^51,55. Note that this simplification does not account for the contribution of the UPS to ubiquitin-independent turnover. (C) Schematic representation of the general decline in proteasome activity during ageing. In theory, small molecules might be used to partially restore this activity to boost function. ^51,55

Aged cells contain a latent pool of free (e.g. unbound) 20S proteasome

What are the relative contributions of the free 20S, the UPS and the UIPS to proteotoxicity and neurodegeneration? The answers are not yet clear, but some clues come from reports that the relative levels of 20S and PA-bound 20S change with ageing and disease. Using label-free proteomics in nine different human cell lines, Fabre et al estimates that 21 – 35% of the total proteasome pool is PA700-bound (26S) and less than 10% is bound to UIPS-specific PAs, while the remaining ~66% is unbound 20S⁵¹ (Fig 2B). This is an interesting result because it suggests that cells may contain a latent pool of 20S proteasome that is not bound to any PA. Because PAs enhance the rate of proteolysis by as much as 20-fold⁵², these findings suggest that some cells have a pool of 20S that is poised to be activated. How does this ratio change during ageing and disease? It is known that this downregulation is mediated, in part, by diminished expression of proteasome genes themselves⁵⁴. To date, the distribution of PA-bound and unbound 20S has not been surveyed in a neurodegeneration model, but the ratio of unbound (e.g. free 20S) proteasome to UPS-specific 26S has been reported to increase with age. For example, in cultured human fibroblasts from old individuals, the amount of 20S proteasome is decreased by 2.5-fold whereas the level of PA700 is lowered by 6.5-fold, compared to fibroblasts from young individuals⁵⁵. Conversely, the levels of PA28 remained unchanged,⁵⁵ suggesting a relative switch from UPS to UIPS pathways during ageing (Fig 2B). Although it is difficult to conclude a direct cause-and-effect, this increased availability of 20S roughly coincides with a decrease in total proteasome activity in many cells, tissues, and organisms during ageing⁵³.

Proteasomal upregulation accelerates the clearance of pathogenic proteins

Together, these findings suggest that decreased proteasome activity, driven by a combination of reduced expression of 20S and an increase in free (i.e. non-PA bound) proteasome, may contribute to neurodegeneration. Accordingly, this model suggests that boosting proteasome activity may counteract the process. This goal could theoretically be achieved in a number of ways. For example, delivery of purified 20S to cells through direct injection has been shown to accelerate clearance of tau⁵⁶. Similarly, cells expressing a 20S mutant, in which the gating N-termini of the α-subunits are deleted, are also partially protected from tau aggregation⁵⁷. Pharmacologically, Finley et al identified small molecules that inactivate the deubiquitinating activity of USP14, the PA700-associated DUB, allosterically activate proteasomal degradation of polyUb-conjugated proteins and, consistent with the model, accelerate turnover of tau in vitro^58,59. These findings have motivated others to search for molecules that directly bind the 20S to promote its activity (see below). Such a strategy might take advantage of the fact that only <10% of total proteasomes in a neurodegenerative disease model are intact 26S and a large pool is free 20S (Fig 2B). Consistent with this idea, over-expression of PA28 enhances protein clearance in ageing models⁶⁰. The pharmacological equivalent of this approach would be to activate the 20S, in the absence of a PA. In the next sections, we discuss the opportunities and challenges of this possibility.

PHARMACOLOGICAL ACTIVATORS OF THE 20S PROTEASOME

The proteasome has long been the subject of studies to identify inhibitors. Early work identified natural products, such as lactacystin, that inactivate it by mimicking peptide substrates and covalently modifying the β-subunits⁶¹. These compounds became widely used chemical probes and, subsequently, a few were approved for the treatment of multiple myeloma patients⁵³, where they seem to exploit the need of cancer cells for high protein turnover. More recently, other strategies have been developed to create inhibitors with different mechanisms⁶². Activators of the proteasome, in contrast, have received less attention; as mentioned above, such molecules would ideally bind to the unbound 20S and open its pore, acting as “artificial activators” to compensate for the loss of natural PAs and proteasome activity during ageing (Fig 3). However, the size and complexity of the 20S coupled with an incomplete understanding of the mechanisms of activation, has made the discovery of activators a difficult task. Furthermore, the field of 20S activators lacks the serendipitous discovery of natural product leads (such as lactacystin) as starting points. The next sections introduce the early chemical efforts to create proteasome activators and point out the substantial challenges that remain in the discovery, optimization and deployment of these molecules. For clarity, the sections are divided based on the chemical composition of the compounds.

Figure 3. Schematic representation of the general decline in proteasome activity during ageing.

In theory, small molecules might be used to partially restore this activity to boost function.

Denaturants

The first reported proteasome activator was the detergent, sodium dodecyl sulfate (SDS), which was identified in vitro by measuring the chymotryptic activity of the 20S against the fluorogenic peptide substrate, succinyl-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (LLVY-amc). Addition of SDS increases the turnover of LLVY-amc by 20-fold, but at levels above its critical micellar concentration. This result suggests that SDS acts as a detergent, likely activating by partially denaturing the α-ring gate and allowing access of substrate into the pore. Another sign of this mechanism is that stimulation occurs within a narrow concentration range: lower concentrations activate turnover, while higher levels inhibit activity, likely by more extensive denaturation of the 20S^63,64. Similar trends are seen for polycations (e.g. polylysine), polyanionic lipids (e.g. cardolipin & heparin), fatty acids (e.g. oleic, linoleic, & linolenic acids) and the natural product oleuropein^65–69 (Table IIA and B). Molecules with this type of concentration-activity signature are likely to be rather non-specific activators, with mechanisms that make them difficult to optimize.

Table II.

Activators of the 20S proteasome.

Small molecules

Betulinic acid is a triterpene natural product that stimulates the 20S at low micromolar concentrations⁷⁰. Unlike the detergents, betulinic acid preferentially stimulates the chymotryptic activity of the 20S (and not the other two activities). However, attempts to improve the potency of betulinic acid through medicinal chemistry have only yielded analogs that inhibit (rather than stimulate) proteasome activity, suggesting complex structure-activity relationships (SAR)⁷¹. More recent reports have turned to high-throughput chemical screens to identify alternative molecules. These assays often rely on measurement of proteasome activity through the hydrolysis of LLVY-amc, while secondary assays are used to measure degradation of biologically relevant substrates^72,73. For example, Trader et al screened a subset of the NIH Clinical Collection (NCC) library to discover two compounds, AM-404 and MK-866, that stimulate the 20S in vitro and enhance turnover of α-synuclein by 3 to 4-fold in cells⁷². In a parallel screen of the Natural Product Library (NPL) of the NCC, Coleman & Trader identified three additional molecules (denoted NPL-1, 2, & 3) that activate the 20S at low micromolar concentrations⁷⁴ (Table IIC). Beyond the specific molecules identified, one of the highlights of that work is the introduction of a rigorous suite of assays to triage non-specific mechanisms. However, additional medicinal chemistry and structural biology efforts will be needed to advance chemical probes with the desired selectivity and potency.

In a parallel effort, Jones et al screened the NCC and Prestwick libraries, revealing chlorpromazine as a putative agonist of the 20S. This compound stimulated the activity of the 20S by 20-fold in vitro. Unbiased docking of chlorpromazine to the 20S predicted possible interactions with the intersubunit pocket of the α-ring, reminiscent of the natural HbYX motif. Thus, this chemical series may take advantage of intrinsic allosteric mechanisms to promote turnover. Chlorpromazine is a well-known dopamine D2 agonist, but chemical modifications suggested that the SAR for proteasome activation was distinct. However, the potency of these compounds remains limited (Table IIC)⁷³.

Peptides

Another strategy for stimulating the proteasome is to directly create mimics of the HbYX motif. Indeed, peptides inspired by the HbYX-containing C-termini of PAs can stimulate the 20S in vitro at mid-micromolar concentrations⁵. Moreover, the potency of these HbYX peptides seems to depend on which subunits of PA700 they were taken from ^75,76. Some of these HbYX peptides can enhance proteasomal degradation of a model substrate and rescue the 20S from inhibition by toxic amyloid-β oligomers⁴³. While promising, peptide-based activators of the 20S have intrinsic challenges that will need to be overcome. For example, they have poor membrane permeability, uncertain selectivity and typically low metabolic stability. Despite these hurdles, the recently reported crystal structure of the HbYX peptide-bound eukaryotic 20S⁷⁷ might provide a structure-guided way to advance this approach. It is important to note that any compounds that bind this site would expectedly displace the UPS and UIPS-specific PAs. Thus, even though they might stimulate the pool of free 20S, their pharmacology is expected to be complex.

CONCLUSION and PROSPECTUS

There is mounting evidence that decreased activity of the proteasome contributes to neurodegeneration. Although decreased 26S is assuredly a key component of this decline, we have focused here on the contributions of the UIPS and the free 20S. This focus is based on evidence that both the PA28-bound and free 20S seem to become more prominent during ageing. Moreover, many of the neurodegenerative disease-associated proteins, such as tau, are IDPs and thus particularly good substrates for these complexes.

The increased level of free 20S during ageing has interesting implications for drug discovery. Specifically, molecules that boost the function of this “low activity” 20S pool might partially compensate for diminished proteasomal function. While a compelling idea, there are major challenges to pursuing this concept and the current molecules are not yet up to this task. One challenge is that many compounds can have detergent-like activity that is potent, but ultimately untenable for the creation of activators. Thus, any screening effort is highly likely to produce denaturation-like artifacts that must be removed through subsequent secondary assays and careful analysis of their dose dependence. In addition, one cannot use the typical approach of adding Tween or Triton-X in screening buffers to minimize the discovery of aggregators or pan-assay interference (PAINS) molecules⁷⁸. Thus, additional types of artifacts are likely to populate the list of apparent “hits”. These two challenges are expected to result in higher-than-normal failure rates during compound triage and special care will be needed to select high quality scaffolds.

Abbreviations

20S

core 20S proteasome

AD

Alzheimer’s disease

ALS

amyotrophic lateral sclerosis

ATP

adenosine triphosphate

BBB

blood-brain barrier

DUB

deubiquitinating enzyme

HbYX

Hydrophobic-Tyrosine-unspecified residue ‘X’

IDP

intrinsically disordered protein

IDR

intrinsically-disordered region

LLVY-amc

succinyl-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin

NCC

NIH (National Institute of Health) Clinical Collection

NPL

Natural Product Library

PA

proteasome activators

PAINS

pan-assay interference compounds

PD

Parkinson’s disease

ROS

reactive oxygen species

Rpt

regulatory particle of triple-ATPase

SAR

structure-activity relationship

tau

microtubule-associated protein tau (MAPT)

TDP-43

trans-activation response element (TAR) DNA-binding protein 43

Ub

ubiquitin

UIPS

Ubiquitin-independent proteasome system

UPS

ubiquitin-proteasome system

USP-14

Ubiquitin-specific-processing protease