The ubiquitin E2 enzyme UBE2QL1 mediates lysophagy

Abstract

Although damaged lysosomes with ruptured membranes can be repaired, these dangerous organelles are also selectively eliminated by autophagic degradation termed lysophagy. This process is initiated by ubiquitination of lysosomal proteins. In this issue of EMBO Reports, Koerver et al [1] identify the E2 enzyme UBE2QL1 that catalyzes ubiquitination of damaged lysosomes. Without this enzyme, the clearance of ruptured lysosomes is compromised not only upon lysosomal damage but also under normal conditions, revealing its adaptive and constitutive functions.

During macroautophagy (hereafter, autophagy), soluble proteins are sequestered by autophagosomes in a mostly random manner. However, certain proteins and large materials, such as organelles and intracellular bacteria, are selectively engulfed by autophagosomes. Selective autophagy of organelles aims to remove harmful and/or unnecessary organelles and therefore constitutes an important part of the intracellular quality control system 2. For example, autophagy eliminates dysfunctional mitochondria and damaged lysosomes that could release reactive oxygen species and hydrolytic enzymes, respectively. These target organelles are typically recognized by autophagosomal ATG8 homologs (ATG8s), that is, LC3 and GABARAP family proteins, by one of two ways. Some selective substrates including organelles have LC3‐interacting regions (LIRs; also called ATG8‐interacting motifs (AIMs) in yeast) and directly bind to autophagosomal ATG8s. One example is the yeast mitochondrial protein Atg32, which has this motif and serves as an autophagy receptor 2. Alternatively, selective substrates without LIRs can be recognized via soluble autophagy adaptors (also called autophagy receptors) that have both LIR and ubiquitin‐binding regions. These adaptors include optineurin, NDP52, TAX1BP1, SQSTM1 (also called p62; it is also an important autophagy substrate), and NBR1 2. In this case, ubiquitination of organellar surface proteins is a prerequisite to autophagy.

Lysosomal membrane rupture or permeabilization is an unfavorable event that would eventually cause cell death. It occurs under pathological (induced by drugs, silica or uric crystals, oxidative stress, amyloid fibrils, etc.) and perhaps even physiological conditions 3. Ruptured lysosomal membranes can be repaired by an ESCRT‐dependent mechanism 4 (Fig 1). However, following extensive damage (or simply used in parallel), damaged lysosomes are subjected to autophagy, which is termed lysophagy 5, 6 (Fig 1). After damaged lysosomes are fully contained within autophagosomes, remaining intact lysosomes will fuse with these autophagosomes. Because lysosomes are extensively ubiquitinated during lysosomal membrane damage and no intrinsic autophagic receptor has been identified so far for the lysosomal membrane, ubiquitination of lysosomal membrane proteins and subsequent recruitment of autophagy adaptors and VCP/p97, a ubiquitin‐binding protein, are considered to be crucial for lysophagy 3 (Fig 1). Yoshida et al 7 identified the FBXO27‐containing SCF (SKP1–CUL1–F‐box) ligase as a specific E3 enzyme that catalyzes ubiquitination of damaged lysosomes. FBXO27 acts as a lysosomal damage sensor because it can interact with lysosomal luminal N‐glycoproteins, such as the luminal part of LAMP2, only after membrane permeabilization. Besides FBXO27, the RING E3 ligase LRSAM1 is involved ubiquitination of lysosomes damaged by bacteria 8. The TRIM16 E3 ligase also recognizes lysosomal membrane damage through binding with galectin‐3 that can interact with luminal β‐galactosides exposed to the cytosol 9. However, it has been hypothesized that these are not the only mechanisms of lysosomal ubiquitination induced by membrane permeabilization 3.

Figure 1. Lysophagy induced by lysosomal membrane rupture.

When lysosomal membranes are ruptured by membrane stresses or even spontaneously, lysophagy factors such as the E2 enzyme UBE2QL1 and E3 enzymes SCF^{FBXO 27}, TRIM16, and LRSAM1 are recruited to ubiquitinate membrane proteins. In some cases, membrane rupture can be sensed by galectins, which bind glycosylated proteins. Ubiquitinated proteins recruit autophagy adaptors such as TAXBP1 and SQSTM1 as well as other factors such as VCP and PLAA. Gal8 recruits the autophagy adaptor NDP52 independently of ubiquitin. Autophagy adaptors interact with ATG8s and RB1CC1 in addition to TRIM16 interacting with ULK1 and ATG16L1, leading to induction of lysophagy. Once damaged lysosomes are fully engulfed by autophagosomes, they fuse with intact lysosomes to degrade the sequestered contents. Lysosomal membrane rupture can also be repaired by the ESCRT machinery, probably when damage is not severe.

To identify a novel mediator of lysosomal ubiquitination following lysosomal membrane damage, Koerver et al 1 carried out imaging‐based siRNA screening targeting 37 human E2 enzymes for their ability to ubiquitinate lysosomal membranes damaged by treatment with the lysosomotropic agent L‐leucyl‐L‐leucine methyl ester (LLOMe) in HeLa cells. They identified UBE2QL1, an E2 enzyme that is not well characterized, whose depletion successfully inhibited damage‐induced lysosomal ubiquitination. By contrast, depletion of autophagy factors such as ATG5 and ATG7 increased the number of ubiquitinated lysosomes, confirming the protective role of autophagy against lysosomal damage. UBE2QL1 is diffusely present in the cytosol in untreated cells and translocates to lysosomes damaged by not only LLOMe but also tau amyloid fibrils 1.

LLOMe treatment induces both K48‐ and K63‐linked ubiquitination, but signal reduction resulting from UBE2QL1 depletion was greater for K48‐linked ubiquitination (K48‐Ub) than K63‐linked ubiquitination (K63‐Ub) 1. Consistently, the recruitment of UBE2QL1 and the appearance of K48‐Ub signal were well synchronized (peaked at 2–3 h after LLOMe treatment) and slightly slower than that of K63‐Ub signal (peaked at ~1 h). Thus, UBE2QL1 is involved primarily in K48‐Ub and partially in K63‐Ub. The ubiquitin signals were often detected inside lysosomes by immuno‐electron microscopy, suggesting that ubiquitination occurs on the luminal regions of lysosomal proteins (Fig 1).

Next, Koerver et al 1 searched for neighboring proteins using the APEX2‐dependent proximity labeling method and identified lysosomal membrane proteins, including LIMP2, NPC1, LAMP1, and LAMP2, as potential targets of ubiquitination. In addition, factors that have been linked to lysophagy such as galectins (Gal3 and Gal8), the autophagy adaptors TAXBP1 and SQSTM1, VCP/p97, and PLAA (a cofactor of VCP), were also identified in the screen (Fig 1). VCP is a multifunctional ubiquitin‐binding protein and is also important for lysophagy 3. Koerver et al 1 confirmed that VCP is recruited to lysosomes together with SQSTM1 and LC3 upon damage. The signals of VCP, as well as SQSTM1 and LC3, were reduced by depletion of UBE2QL1, suggesting that UBE2QL1‐dependent ubiquitination recruits VCP and SQSTM1 and induces autophagosome formation.

Finally, Koerver et al 1 investigated the physiological relevance of UBE2QL1‐dependent lysophagy. LLOMe treatment induces lysosomal membrane rupture, which can be detected as Gal3‐positive puncta. After washing out LLOMe, most of these puncta disappeared within 10 h in control‐knockdown cells. However, puncta disappearance was significantly delayed in UBE2QL1‐knockdown cells, leading to a higher rate of cell death. Thus, UBE2QL1 is indeed important for the clearance of damaged lysosomes. Importantly, this protective effect of UBE2QL1 was also observed in cells that were not treated with LLOMe; Gal3‐positive lysosomes accumulated spontaneously under normal culture conditions. This result suggests that UBE2QL1 has a constitutive housekeeping role in lysosomal homeostasis. Accordingly, mTORC1 inactivation and subsequent TFEB nuclear translocation, both of which are caused by lysosomal dysfunction, were observed in UBE2QL1‐knockdown cells. However, this constitutive effect may not be through lysophagy because the accumulation of Gal3‐positive lysosomes was not observed in ATG5‐ or ATG7‐knockdown cells. Thus, the ubiquitination of damaged lysosomes has likely additional roles that go beyond lysophagy induction. The physiological importance of UBE2QL1 was confirmed in vivo using Caenorhabditis elegans. A C. elegans mutant deficient for UBC‐25, a UBE2QL1 homolog, accumulated Gal3‐positive damaged lysosomes even under normal conditions, which was aggravated in scav‐3‐deficient worms with weakened lysosomal membrane integrity 1.

This study successfully demonstrates that UBE2QL1‐mediated ubiquitination of lysosomal proteins is crucial for lysophagy following various types of lysosomal damage. Furthermore, it also shows that this E2 enzyme has a constitutive role in lysosomal homeostasis even under normal conditions. This finding challenges the widely accepted model that lysosomal membranes are protected from degradation by their own enzymes. Instead, this study suggests that lysosomes can be spontaneously damaged and thus require constitutive repair or replacement.

This study, of course, also raises a number of questions. How UBE2QL1 is recruited to damaged lysosomes is especially intriguing. Recruitment may be triggered by the exposure of glycosylated luminal proteins, as indicated by the identification of Gal3 and Gal8 in the UBE2QL1 proximity labeling assay. However, the depletion of Gal3 or Gal8 did not affect UBE2QL1 recruitment 1, suggesting the presence of redundant or alternative mechanisms. For instance, UBE2QL1 itself might recognize membrane pores of damaged lysosomes or it might be recruited by an unknown E3 enzyme, whose identity is a question in itself. It is unlikely that the previously identified E3 enzyme FBXO27 functions with UBE2QL1 because it is not expressed in HeLa cells, which Koerver et al used in this study 1, 7. Another critical question is what the targets of UBE2QL1‐mediated ubiquitination are. LAMP1 and LAMP2, which are major lysosomal membrane proteins, could be good candidates. The clearance of whole organelles during lysophagy may be analogous to Parkin‐mediated mitophagy, where multiple proteins are ubiquitinated and substrate specificity is not strict. Finally, what are the adaptor proteins communicating the ubiquitin signal to the autophagy machinery? This study revealed that SQSTM1 and TAX1BP1 are recruited to damaged lysosomes, suggesting that they may be key adaptors. The identification of adaptors would be important because autophagy adaptors mediate not only LIR‐dependent recognition by ATG8s. Recent studies suggested that autophagy adaptors can also interact with RB1CC1 (also called FIP200) in the autophagy initiation complex 10. If this is also the case in lysophagy, it could be an important mechanism to initiate lysophagy upon membrane rupture.

EMBO Reports (2019) 20: e49104