How multi-component cascades operate in cells: lessons from the ubiquitin system-containing liquid-separated condensates

ABSTRACT

Membraneless condensates have recently caught the attention of biologists as hubs for cellular components required for catalysis of basic processes. Whether they are real has become the center of heated discussion where the main issues are their mechanism of assembly and function. A recent study describing these condensates as hubs for protein degradation by the ubiquitin system may shed a new light on this recent development in cell biology.

The ubiquitin-proteasome system (UPS) is one of the two main proteolytic machineries of the cell, alongside autophagy. The two systems share several components and characteristics, and communicate with one another. One of the shared components is the shuttle protein p62/SQSTM1. In the UPS, p62 was suggested to bind ubiquitinated substrates by its ubiquitin-associated domain (UBA), targeting them for degradation by the proteasome via interaction of its Phox and Bem1 (PB1) domain with the 19S RPN1 and RPN10 proteasomal subunits. In the case of autophagy, it was suggested that p62 acts to deliver ubiquitinated cargo to the growing autophagosome via its binding to the autophagic receptor microtubule-associated protein 1A/1B-light chain 3 (LC3), a process mediated by p62’s LC3-interacting region (LIR).¹ Interestingly, p62 also shuttles the proteasome – as a substrate – for degradation through autophagy,² similar to its role in the degradation of cellular organelles.¹ In our recent work,³ we show that p62 forms proteolytic active biomolecular condensates in the nucleus by liquid–liquid phase separation (LLPS). The formation of the droplets is probably due to its ability to oligomerize into flexible helical filaments through its PB1 domain and capture polyubiquitinated substrates by the UBA domain. Recruitment of the 26S proteasome to the p62 nuclear condensates along with all other essential components of the ubiquitination cascade allows effective removal of the ubiquitinated substrates.³

Although autophagy was shown to operate also in a semi-targeted manner,⁴ it is generally less selective, and serves for bulk removal of cytosolic proteins and organelles. The UPS, on the other hand, is highly specific and degrades proteins only following their specific marking by ubiquitin (Ub). Ub conjugation is mediated by a sequential activity of three enzymes: E1 – the Ub-activating enzyme (UBA1), a member of the E2 family of Ub-conjugating enzymes (UBCs), and a substrate-specific ubiquitin ligase, E3. Typically, the process repeats itself to generate a Ub chain which is recognized by the 26S proteasome.⁵ During the proteolytic process, the proximal Ub moiety(ies) is degraded with the substrate whereas the more distal ones are recycled by de-ubiquitinating enzymes (DUBs).⁶

While considering such a complex, processive, and timely regulated mechanism, one may hypothesize that such an orchestrated machinery will also be coordinated spatially and its different components should be highly concentrated. This is particularly required for the repeated cycle of ubiquitination and then for constantly monitoring the affinity of the growing Ub chain to the proteasome, a point where the ubiquitinated substrate is transferred from the conjugation machinery to the proteasomal “blender”. Further, unlike autophagy which is rather slow to respond to intra- and extra-cellular cues, the UPS is much faster, rapidly removing short-lived regulatory proteins (e.g. transcription factors) and constantly monitoring protein quality control (PQC). Therefore, such a mechanism is less likely to function in either a membrane-bound organelle or free in the cytoplasm or nucleoplasm. Membraneless condensates, on the other hand, allow for fast assembly and disintegration, as well as fast adaptation of contents to meet cellular needs, and importantly, efficient reactions.

In agreement with such hypothesis, we were able to capture discrete nuclear membraneless condensates, in which all the components of the ubiquitination cascade (for the particular substrate we study) – E1, E2 and E3, as well as the proteasome and DUBs – are present³ (Figure 1). We found that p62 is essential for the formation of such condensates and that the time constant of its recruitment into these foci is short – on the scale of seconds. Importantly, these condensates are proteolytically functional, facilitating the degradation of several substrates. Among them were NLS-GFP-CL1 – a GFP fused to a Nuclear Localization Signal (NLS) and to the CL1 degradation signal (originated from yeast), unincorporated proteasomal subunits, and c-Myc. Intriguingly, using different mutants of p62, we showed that the same motifs required for nuclear condensate formation – UBA and PB1 – are required for efficient degradation of c-Myc. Also, we identified in the droplets the chaperones – heat shock proteins (HSPs) 70 and 90 and the Ub ligase CHIP, suggesting that the droplets are part of the PQC machinery. While it is not surprising that such condensates are formed under stress (thermal and oxidative), it is important to emphasize that they exist under basal conditions, underscoring their role in maintaining protein homeostasis.

Figure 1.

Role of liquid-liquid phase separation in the formation of proteolytically active nuclear condensates by filamentous p62 A. The condensates contain essentially all the components of the Ubiquitin Proteasome System (UPS) including E1, E2, K48- and K63-based ubiquitin chains (presumably conjugated to target substrates), Deubiquitinating enzymes, and the proteasome (1), and all the components of a representative ubiquitin ligase – SKP1-CUL1-Fbxw7 (2). B. The condensates contain components of the Protein Quality Control (PQC) machinery, such as chaperones and the ubiquitin ligase Carboxy-terminus of Hsc70 Interacting Protein (CHIP). Their recruitment is stimulated by stress that induces generation of misfolded proteins. HeLa cells in which p62 was deleted by CRISPR were transfected with FLAG-p62-ΔNES (Nuclear Export Signal) (red) to ascertain its nuclear localization. All other components of the UPS and PQC system are in green

(The microscopic panels are taken from Ref. 3, and are published with permission from Proc. Natl. Acad. Sci. USA).

It was recently shown that in response to osmotic stress, the proteasome becomes concentrated within nuclear LLPS foci, containing also the shuttle protein RAD23B.⁷ We found that osmotic stress stimulates the formation of several types of nuclear condensates that either contain or lack proteasomes and p62. These droplets may represent intermediates of the mature, proteolytically active condensates.

It should be noted that the field of Liquid–Liquid Phase Separated droplets is still controversial.⁸ The controversy engulfs the mechanism of formation of these droplets, but more so their function. We were careful not to step into the minefield of the biophysical nature of what we have seen. Yet, intervention in protein–protein interaction by 1,6-hexanediol resulted in disintegration of the droplets.^3,7 Furthermore, besides our study,³ years ago, evidence for proteolytic functionality of condensates was reported by Yewdell who described proteasomal degradation of a misfolded viral antigen.⁹ Other functions such as protein synthesis have been described as well.¹⁰ All-in-all it looks that while the unknown still far exceeds the known on membraneless condensates, the evidence for their existence and important function accumulate.

Funding

Research in the laboratory of A.C. is supported by the Adelson Medical Research Foundation (AMRF), the Israel Science Foundation (ISF), the Israel Precision Medicine Partnership (IPMP) introduced by the ISF, and an Israel Cancer Research Fund (ICRF) Professorship.

Disclosure statement

No potential conflict of interest was reported by the authors.