The ubiquitin-specific protease USP8 directly deubiquitinates SQSTM1/p62 to suppress its autophagic activity

ABSTRACT

SQSTM1/p62 (sequestosome 1) is a critical macroautophagy/autophagy receptor that promotes the formation and degradation of ubiquitinated aggregates. SQSTM1 can be modified by ubiquitination, and this modification modulates its autophagic activity. However, the molecular mechanisms underpinning its reversible deubiquitination have never been described. Here we report that USP8 (ubiquitin specific peptidase 8) directly interacted with and deubiquitinated SQSTM1. USP8 preferentially removed the lysine 11 (K11)-linked ubiquitin chains from SQSTM1. Moreover, USP8 deubiquitinated SQSTM1 principally at K420 within its ubiquitin-association (UBA) domain. Finally, USP8 inhibited SQSTM1 degradation and autophagic influx in cells with wild-type SQSTM1, but not its mutant with substitution of K420 with an arginine. Taken together, USP8 acts as a negative regulator of autophagy by deubiquitinating SQSTM1 at K420.

Abbreviations: BafA₁: bafilomycin A₁; BAP1: BRCA1 associated protein 1; DUB: deubiquitinating enzyme; ESCRT: endosomal sorting complex required for transport; HTT: huntingtin; K: lysine; KEAP1: kelch like ECH associated protein 1; MAP1LC3/LC3: microtubule associated protein 1 light chain 3; MEF: mouse embryonic fibroblast; shRNA: short hairpin RNA; SQSTM1: sequestosome 1; Ub: ubiquitin; UBA: ubiquitin-association; UBE2D2: ubiquitin conjugating enzyme E2 D2; UBE2D3: ubiquitin conjugating enzyme E2 D3; USP: ubiquitin specific peptidase; WT: wild-type

Introduction

SQSTM1/p62 functions as a signaling hub and a selective autophagy receptor [1–3]. SQSTM1 possesses multiple domains that mediate the interaction with different signaling proteins to regulate multiple cellular functions, including cell survival, inflammation, amino acid sensing and the oxidative stress response [2–4]. In addition, SQSTM1 acts as a selective autophagy receptor that is involved in both inclusion body formation and autophagy [5,6]. SQSTM1 mediates these functions mainly through its Phox1 and Bem1p (PB1) domain, ubiquitin-association (UBA) domain and LC3-interacting region (LIR) [2,3]. SQSTM1 forms homo- and/or hetero-dimers with other members of autophagic receptors via its PB1 domain, facilitating its oligomerization and interaction with ubiquitinated proteins [7,8]. Deletion of the PB1 domain abrogates the self-association of SQSTM1, reducing its sequestrating activity and cellular functions [9,10]. SQSTM1 interacts with ubiquitinated cargoes via its UBA domain and recruits them via its LIR into growing autophagosomes for lysosomal degradation [5,6,10]. Therefore, SQSTM1 is selectively incorporated into the autophagosome and then degraded [6,10]. Accordingly, SQSTM1 is widely used as an indicator of autophagic flux [11,12]. SQSTM1 plays specific and indispensable roles in selective autophagy, especially in response to proteotoxic and oxidative stresses [13,14]. Importantly, dominantly inherited missense or deletion mutations within the UBA domain of SQSTM1 are associated with degenerative diseases, which are commonly accompanied with the accumulation of SQSTM1 aggregates and ubiquitinated inclusions [15,16]. Of note, autophagy is responsible for the degradation of SQSTM1; therefore, impairment of autophagy is usually accompanied by massive accumulation of SQSTM1 followed by formation of aggregate structures positive for SQSTM1 and ubiquitin, which contributes to pathogenesis of diseases [5].

A number of key components of the autophagy pathway undergo the modification of ubiquitination, and this modification has been shown to play critical roles in the regulation of autophagy. For example, the activity of BECN1/Beclin-1, an important regulator in autophagy initiation and progression, is tightly controlled by ubiquitination. Ubiquitination of SQSTM1 also occurs, and is significantly elevated upon ubiquitin stress such as treatment of the proteasome inhibitor MG132 and heat shock [17–19]. Ubiquitination has been shown to modulate the activity of SQSTM1. Recent studies reported that several ubiquitin ligases, depending upon their site of ubiquitination, inhibited or facilitated SQSTM1’s function. The E3 ubiquitin ligase TRIM21 directly interacts with and ubiquitinates SQSTM1 at lysine 7 (K7) within its PB1 domain [17]. This ubiquitination markedly impairs SQSTM1 oligomerization and subsequently inhibits its sequestration function. The E3 ligase RNF166 catalyzes K29- and K33-linked polyubiquitination of SQSTM1 at residues K91 and K189 [20]. RNF166-mediated ubiquitin ligase activity facilitates SQSTM1’s role in the xenophagic degradation of intracellular bacteria. The HECT E3 ubiquitin ligase NEDD4 interacts with and ubiquitinates SQSTM1 for inclusion body autophagy [21]. KEAP1-CUL3 (cullin 3) ubiquitinates SQSTM1 at K420 within its UBA domain to enhance SQSTM1’s sequestering activity and autophagic degradation [22]. The ubiquitination of K420 may disrupt dimerization, liberating SQSTM1’s ability to recognize polyubiquitinated cargoes for selective autophagy [18,22]. Ubiquitin stress induces the ubiquitination of K420 to enhance the flux of both bulk and selective autophagy [18].

Deubiquitinating enzymes (DUBs) are proteases that conduct the reversal process of the ubiquitination of protein. Several DUBs have been reported to modulate autophagy by deubiquitinating the components of autophagy pathway. For example, USP14 regulates autophagy by negatively controlling K63-linked polyubiquitination of BECN1 [23]. USP19 modulates autophagy and antiviral immune responses by deubiquitinating BECN1 [24]. Currently, the specific deubiquitination enzyme for SQSTM1 has not yet been identified. Here we report that USP8 (ubiquitin specific peptidase 8) directly interacts with and deubiquitinates SQSTM1. USP8 removes the ubiquitination of SQSTM1 mainly at K420. USP8 knockdown resulted in autophagic degradation of SQSTM1 and increased autophagic flux. Our study identifies USP8 as a crucial and bona fide deubiquitinating enzyme for SQSTM1. Our results also strengthen the notion that the ubiquitination of autophagic receptors is an important regulatory mechanism by which selective autophagy is controlled.

Results

USP8 interacts with autophagy receptor SQSTM1

SQSTM1 was initially identified as a scaffold protein for atypical protein kinase C isoforms and mainly located in lysosome-targeted endosomes [25]. We reasoned that endosome-associated DUBs may regulate the ubiquitination of SQSTM1. USP8/UBPY locates to the endosome compartment and regulates the endosomal sorting of many proteins including epidermal growth factor (EGF) receptor [26,27]. We hypothesized that USP8 may interact with SQSTM1 and regulate its ubiquitination. We first examined whether USP8 can interact with SQSTM1. To this end, we transfected USP8-HA or the control vector into HEK293T cells which express endogenous SQSTM1 and USP8. We found that endogenous SQSTM1 was co-immunoprecipitated with anti-HA antibody only in USP8-HA transfected cells (Figure 1A). Furthermore, endogenous USP8 co-immunoprecipitated with anti-SQSTM1 antibody, but not control IgG (Figure 1B). In addition, His-USP8 purified from E.coli was specifically bound by purified GST-SQSTM1 protein, but not GST alone, indicating that USP8 directly interacts with SQSTM1 in vitro (Figure 1C). To determine which domain of SQSTM1 is required for the interaction with USP8, we constructed a panel of Flag-tagged SQSTM1 truncation mutants. We co-introduced Flag-tagged SQSTM1 or its truncation mutants with USP8-HA into HEK293T cells, and performed immunoprecipitation assay (Figure 1D). We found that USP8-HA failed to be immunoprecipitated by Flag-tagged SQSTM1 truncation mutant (loss of residues 85–233) using anti-Flag antibody (Figure 1D, lane 4). Again, MYC-tagged SQSTM1 truncation mutant losing this fragment failed to be immunoprecipitated by USP8-Flag using anti-Flag antibody (Figure 1E, lane 3). This fragment contains zinc finger (ZZ) and TRAF6-binding (TB) domains. The ZZ domain binds RIPK1 (receptor interacting serine/threonine kinase 1). Interestingly, more detailed mapping demonstrated that ZZ domain is necessary for the interaction between SQSTM1 and USP8 (Fig. S1). Finally, immunofluorescence demonstrated that endogenous USP8 co-localized with SQSTM1 at perinuclear region in HeLa cells (Figure 1F). Collectively, these data suggest that USP8 directly interacts with SQSTM1.

Figure 1.

USP8 interacts with SQSTM1. (A) Endogenous SQSTM1 was precipitated using HA beads in HEK293T cells transfected with USP8-HA expression construct. (B) Endogenous USP8 was precipitated using anti-SQSTM1 antibodies in MEFs. (C) Reconstituted GST fusion SQSTM1 directly interacted with His-USP8 in GST pulldown assay. (D) Schematic representation of Flag-tagged SQSTM1 or its truncated mutants (left panel). HEK293T cells were co-transfected with USP8-HA and Flag-tagged SQSTM1 or its truncated mutants, immunoprecipitated with Flag beads and immunoblotted with antibodies against HA and Flag. (E) HEK293T cells were co-transfected with USP8-Flag and myc-tagged SQSTM1 or its indicated mutants, immunoprecipitated with Flag beads and immunoblotted with antibodies against myc and Flag. HC, heavy chain. (F) Endogenous SQSTM1 colocalized with USP8 in HeLa cells revealed by immunofluorescence. Scale bar, 10 µm.

USP8 deubiquitinates SQSTM1 both in vitro and in vivo

Considering USP8 as a deubiqiuitinating enzyme, we next sought to determine whether USP8 functions as a bona fide SQSTM1 deubiquitinase both in vitro and in vivo. To examine whether USP8 directly deubiquitinates SQSTM1 in vitro, we purified ubiquitinated SQSTM1 from HEK293T cells co-transfected with SQSTM1-Flag and His-Ub using affinity purification with anti-Flag agarose beads in RIPA buffer, followed by elusion with Flag peptide. HA-tagged USP8 or the catalytic inactive mutant (USP8^C786A) were purified from HEK293T cells transfected with HA-USP8 or USP8^C786A expression vectors using affinity purification with anti-HA agarose beads and HA-peptide elution. The ubiquitinated SQSTM1 was incubated with purified wild-type USP8 or USP8^C786A for indicated times, followed by immunoblotting with anti-His antibody. USP8 significantly reduced the levels of ubiquitinated species of SQSTM1 (Figure 2A). This effect required the DUB activity of USP8, as USP8^C786A failed to reduce the ubiquitination of SQSTM1 (Figure 2A). Considering the fact that UBA domain of SQSTM1 is prone to bind with ubiquitinated proteins, SQSTM1-immunoprecipitiated complex may contaminate a bulk of ubiquitinated proteins. USP8 may deubiquitinate these contaminated ubiquitinated proteins, resulting in decreased ubiquitinated species as observed in Figure 2A. To avoid this possibility, we generated ubiquitinated SQSTM1 in vitro as we described previously [18]. His-tagged SQSTM1 was isolated from E.coli and incubated with UBE2D2 and UBE2D3, two E2 Ub-conjugating enzymes. UBE2D2 and UBE2D3 can interact with and ubiquitinate SQSTM1 in the presence of Ub. We observed that USP8 can also significantly reduce the levels of ubiquitinated species of SQSTM1 generated in vitro (Figure 2B), further indicating that USP8 directly deubiquitinates SQSTM1 in vitro. As a negative control, BAP1 (BRCA1 associated protein 1) deubiquitinase failed to reduce the levels of ubiquitinated species of SQSTM1 generated in vitro (Figure 2B), suggesting the specificity of USP8 in controlling ubiquitination of SQSTM1.

Figure 2.

USP8 deubiquitinates SQSTM1 both in vitro and in vivo. (A) Ubiquitinated SQSTM1-Flag was purified from HEK293T cells transfected with HA-Ub and SQSTM1-Flag using anti-Flag affinity purification. HA-tagged USP8 or USP8^C786A was purified from HEK293T cells using anti-HA affinity purification. Ubiquitinated SQSTM1-Flag was incubated with HA-tagged USP8 or USP8^C786A for indicated times, followed by immunoblotting using antibodies against Ub and USP8. (B) Recombinant SQSTM1-Flag were incubated with in vitro ubiquitination assay system containing E1, E2 (UBE2D2 or UBE2D3) and Ub, followed by purification using anti-Flag affinity purification. Ubiquitinated SQSTM1-Flag were incubated with recombinant USP8 or BAP1 protein for indicated times, followed by immunoblotting using anti-Ub antibodies. (C) HEK293T cells were co-transfected with His-SQSTM1 and Flag-tagged USP8 or USP8^C786A mutant. After 48 h, the cells were subjected to pull down using the Ni²⁺-NTA beads under denaturation conditions, followed by immunoblotting. (D) HEK293T cells were co-transfected with His-SQSTM1, HA-Ub and USP8-Flag for 36 h, followed by treatment with heat shock (HS) for 30 min or bortezomib (BTZ, 50nM) for 12 h. The cell lysates were subjected to pull down using the Ni²⁺-NTA beads under denaturation conditions, followed by immunoblotting. (E) Immunoblotting of ubiquitinated species of endogenous SQSTM1 in scr (scramble) or shUSP8-infected MEFs treated with BafA₁ (200 nM) for 6 h.

We next sought to determine whether USP8 affects the ubiquitination of SQSTM1 in vivo. To this end, we transfected His-SQSTM1 and USP8-Flag into HEK293T cells. His-SQSTM1 was purified using Ni²⁺-NTA at the stringent condition to ensure no other proteins retained, and the attached ubiquitin chains were probed using anti-Ub antibodies. We found that wild-type USP8, but not the USP8^C786A mutant, decreased the ubiquitin species of SQSTM1 (Figure 2C). Our previous work showed that SQSTM1 underwent auto-ubiquitination by UBE2D2 and UBE2D3 under stress conditions, such as proteasome inhibition and heat shock [18]. We also observed that USP8 was able to reduce the ubiqutination of SQSTM1 induced by these stress conditions (Figure 2D). Ubiquitinated SQSTM1, together with ubiquitinated cargo, is continually recruited into growing autophagosomes for lysosomal degradation. Under the basal condition, the ubiquitinated SQSTM1 is barely detected in cells. Treatment with bafilomycin A₁ (BafA₁), an inhibitor of autophagy by blocking the autophagosome/lysosome fusion, has been reported to lead to the accumulation of ubiquitinated SQSTM1 protein in cells [22]. We found that knockdown of endogenous Usp8 in BafA₁-treated MEFs significantly enhanced the ubiquitin species of endogenous SQSTM1 (Figure 2E). Collectively, these results reveal that USP8 is a bona fide deubiquitinase of SQSTM1.

USP8 preferentially removes k11-linked ubiquitin chains from SQSTM1

Ubiquitin contains 7 lysines through which polyubiquitin chains can be assembled, with specific linkages determining the fate of the substrates [28]. SQSTM1 has been reported to be modified by K29-, K33-, K48- and K63-linked polyubiquitin chains [17,19–22]. USP8 has been shown to remove K48-, K63-, K11- and K6-linked ubiquitin chains [29–31]. We next set out to examine which types of SQSTM1 ubiquitination could be affected by USP8. To this end, we transfected His-SQSTM1, USP8-Flag and various HA-tagged Ub mutants into HEK293T cells. His-SQSTM1 was purified using Ni²⁺-NTA under the stringent condition, and the attached Ub chains were probed using anti-HA antibodies. Overexpression of USP8 markedly inhibited wild-type and K11-linked ubiquitination of SQSTM1, and to a less extent, K48- and K63-linked ubiquitination (Figure 3A). In contrast, overexpression of USP8 had no appreciable effect on the ubiquitination of SQSTM1 with K6, K27, K29 and K33 linkages (Figure 3A). To exclude the possibility of decreased K11-, K48- or K63-linked ubiquitination was not due to USP8 overexpression in transient transfection, we compared the levels of the K11-, K48- and K63-linked ubiquitin species of SQSTM1 between scramble and USP8-knockdown cells. Surprisingly, we observed that only K11-linked ubiquitin chains were markedly elevated in USP8-knockdown cells (Figure 3B). Finally, using specific antibodies against K63- and K48-linked ubiquitin chains we found that USP8 knockdown did not increase K63- and K48-linked polyubiquitination of endogenous SQSTM1 in cells (Figure 3C). These results collectively indicate that endogenous USP8 is dispensable to remove K48- and K63-linked ubiquitin chains from SQSTM1. We conclude that USP8 preferentially removes K11-linked ubiquitin conjugates from SQSTM1.

Figure 3.

USP8 preferentially removes K11-linked ubiquitination of SQSTM1. (A) HEK293T cells were co-transfected with His-SQSTM1, USP8-Flag and HA-Ub or indicated mutants. Lysates, 48 h post-transfection, were subjected to pull down using the Ni²⁺-NTA beads under denaturation conditions, followed by immunoblotting. Representative blots are shown on the left. Quantification of relative intensity of ubiquitinated SQSTM1 is shown on the right. Data are depicted as mean ± s.e.m from 3 independent experiments. *P < 0.01 (two-tailed t-test). (B) Scramble (Scr) or shUSP8-infected HeLa cells were transfected with His-SQSTM1 and HA-K11-, K48- or K63-linked Ub for 48 h. Lysates were subject to pull down using the Ni²⁺-NTA beads, followed by immunoblotting using antibodies against His and HA. (C) Immunoblotting of K48- and K63- ubiquitination of endogenous SQSTM1 in scramble (scr) or shUSP8-infected MEFs treated with BafA₁ (200 nM) for 6 h. HC, heavy chain.

USP8 deubiquitinates SQSTM1 mainly at K420

Mass spectrometry analysis has identified multiple ubiquitination sites on SQSTM1 that include lysine residues within the PB1 and UBA domains [18,19]. We next determined which lysine residues of SQSTM1 could be affected by USP8. The residue K420 within UBA domain was recently reported to be the principal ubiquitinated site on SQSTM1 [22]. The ubiquitination of K420 enhances SQSTM1’s sequestering activity and selective autophagy [18,22]. Therefore, we focused on the effect of USP8 on the ubiquitination of K420. We mutated lysine residue at position 420 into arginine (R) to generate SQSTM1^K420R mutant. We transfected His-wild-type SQSTM1 or SQSTM1^K420R mutant, together with USP8-Flag and HA-Ub, into HEK293T cells. His-tagged SQSTM1 or SQSTM1^K420R mutant was purified using Ni²⁺-NTA under the stringent condition, and the attached ubiquitin chains were probed using anti-HA antibodies. Consistent with previous report, we found that the amount of HA-Ub attached to the SQSTM1^K420R mutant was dramatically reduced compare to that of wild-type SQSTM1 (Figure 4A, lane 1 and 3). In contrast to wild-type SQSTM1, USP8 overexpression had no obvious effect on the ubiquitination of the SQSTM1^K420R mutant (Figure 4A). Moreover, the SQSTM1^K420R mutant displayed reduced K11-linked ubiquitination (Figure 4B). However, wild-type SQSTM1 and the SQSTM1^K420R mutant harbored similar amounts of K48- and K63-linked ubiquitin chains (Fig. S2), suggesting that K420 does not undergo the modification of K48- or K63-linked ubiquitination. As expected, USP8 overexpression did not significantly affect K48- or K63-linked ubiquitination of the SQSTM1^K420R mutant (Fig. S2). KEAP1 (kelch-like ECH-associated protein 1) was recently shown to ubiquitinate SQSTM1 at K420 to enhance SQSTM1’s sequestering activity and autophagic degradation [22]. Thus, we sought to determine whether USP8 competes with KEAP1 for the ubiquitination of SQSTM1 at K420. As shown in Figure 4C, USP8 overexpression diminished the ubiquitination of wild-type SQSTM1 induced by KEAP1 overexpression. This effect was not observed in the SQSTM1^K420R mutant (Figure 4C). Therefore, USP8 antagonizes KEAP1 for the ubiquitination of SQSTM1 at K420. Collectively, these results indicate that USP8 removes the ubiquitination of SQSTM1 principally at K420.

Figure 4.

USP8 deubiquitinates SQSTM1 mainly at K420. (A) HEK293T cells were transfected with HA-Ub and His-tagged SQSTM1 or SQSTM1^K420R mutant with or without USP8-Flag for 48 h. The cells were subjected to pull down using the Ni²⁺-NTA beads, followed by immunoblotting. (B) HEK293T cells were transfected with HA-K11-linked-Ub and His-tagged SQSTM1 or SQSTM1^K420R mutant for 48 h. The cells were subjected to pull down using the Ni²⁺-NTA beads, followed by immunoblotting. (C) HEK293T cells were transfected with HA-Ub and His-tagged SQSTM1, together with indicated vectors for 48 h. The cells were subjected to pull down using the Ni²⁺-NTA beads, followed by immunoblotting.

USP8 knockdown promotes the autophagic degradation of SQSTM1

Ubiquitination of SQSTM1 at K420 has been reported to enhance the sequestering activity and autophagic degradation of SQSTM1 [18,22]. As we have reported above, USP8 deubiquitinated SQSTM1 at the K420 residue. We reasoned that USP8 may regulate the autophagic degradation of SQSTM1. We first examined the effect of USP8 overexpression on SQSTM1 protein abundance. As shown in Figure 5A, overexpression of USP8, but not the USP8^C786A mutant, increased the levels of endogenous SQSTM1 protein in HEK293T cells. Furthermore, knockdown of endogenous Usp8 in MEFs significantly reduced SQSTM1 protein expression (Figure 5B). The similar results were observed in 2 other mammalian cell lines (Fig. S3A). We also measured the effect of USP8 on the half-life of endogenous SQSTM1 protein. We treated scramble or Usp8-knockdown MEFs with a protein synthesis inhibitor cycloheximide (CHX). Knockdown of Usp8 remarkably shortened the half-time of SQSTM1 protein in MEF cells (Fig. S3B), demonstrating a critical role of USP8 in maintaining the stability of SQSTM1 protein. SQSTM1 degradation is mediated mainly through autophagy; therefore, disruption of autophagy pathway results in accumulation of SQSTM1 protein in cells. As expected, treatment with BafA₁ increased SQSTM1 protein level in MEFs (Figure 5C). Interestingly, as incubating time increased, BafA₁ treatment up-regulated SQSTM1 protein in Usp8-knockdown MEFs to the similar level as in control cells (Figure 5C). Finally, we found that Usp8 knockdown did not lead to the reduction of SQSTM1 protein levels in atg7^−/- MEFs, which are deficient in autophagy (Figure 5D). Taken together, these results demonstrate that USP8 protects SQSTM1 from autophagic degradation.

Figure 5.

Knockdown of USP8 promotes autophagic degradation of SQSTM1. (A) Overexpression of wild-type USP8, but not its catalytic inactive mutant, enhanced the protein level of SQSTM1 in HEK293T cells. (B) Immunoblotting of SQSTM1 and USP8 protein in scramble (scr) or shUsp8-infected MEFs. (C) Immunoblotting of SQSTM1 protein in scr or shUsp8-infected MEFs treated with 200nM BafA₁ for indicated times. (D) Immunoblotting of SQSTM1 in scramble (scr) or shUsp8-infected wild-type and atg7^−/- MEFs.

USP8 acts as a negative regulator of basal autophagy

Ubiquitination of SQSTM1 at K420 within its UBA domain enhances SQSTM1-mediated recruitment of ubiquitinated cargoes into autophagosomes for degradation. SQSTM1 is a well-known selective autophagic substrate and SQSTM1 reduction is an indicator of increased autophagic flux [11,12]. USP8 deubiquitinates SQSTM1 at K420 and inhibits its autophagic degradation, suggesting that USP8 may play an important role in the formation of autophagosome and autophagic flux. BafA₁ prevents the autophagosome-lysosome fusion and subsequently results in the accumulation of LC3-II, which is tightly bound to the autophagosomal membranes and serves as a hallmark of autophagy [11,12]. As expected, BafA₁ treatment led to the accumulation of LC3-II (Figure 6A). Overexpression of USP8, but not the USP8^C786A mutant, decreased the levels of LC3-II (Figure 6A). Deregulated autophagy is associated with many neurodegenerative diseases including Huntington disease (HD). Mutant HTT (huntingtin) is an autophagy substrate and its degradation can be accelerated by rapamycin, an inducer of autophagy [10]. We next used EGFP-tagged mutant HTT (GFP-Q46) to check the autophagic process. As expected, rapamycin treatment promoted the degradation of GFP-Q46, and USP8 delayed this process in a dose-dependent manner (Fig. S4A).

Figure 6.

USP8 negatively regulates autophagy. (A) Overexpression of wild-type USP8, but not catalytic its inactive mutant, inhibited the accumulation of LC3-II in the presence of 200 nM BafA₁ for 8 h. (B) Knockdown of USP8 led to the accumulation of LC3-II in HeLa cells in the presence or absence of 200 nM BafA₁ for 8 h. (C) Knockdown of USP8 promoted the formation of SQSTM1 aggregates in HeLa cells treated with BTZ (50 nM) for 8 h, revealed by immunofluorescence. Scale bar, 10 µm. (D) Knockdown of USP8 promoted the formation of autolysosomes in HeLa cells at the basal condition. Scramble (scr) and USP8 knockdown HeLa cells were transfected with mTagRFP-mWasabi-LC3 expression vector. The representative images of mTagRFP (red) and mWasabi (green) spots are showed at the left. The red⁺green⁻ spots represented autolysosomes. Scale bar, 10 µm. The average number of red⁺green⁻ spots per cell was counted in 100 cells from each indicated sample and shown as mean+S.D. * p < 0.05; Student’s t test.

We next examined the effect of USP8 knockdown on autophagy. USP8 knockdown in HeLa cells had much higher levels of LC3-II compared to the wild-type control at the basal condition (Figure 6B). Moreover, BafA₁ treatment resulted in the accumulation of more LC3-II in USP8-knockdown HeLa cells compared to wild-type control (Figure 6B). Certain stressors, such as the proteasome inhibitor bortezomib, can induce SQSTM1 aggregates in cells [17,18]. The ubiquitination of K420 is critical for the formation of SQSTM1 aggregates [18,22]. USP8-knockdown HeLa cells developed larger aggregates of SQSTM1 than scramble cells after bortezomib treatment (Figure 6C). Finally, we investigated whether USP8 knockdown increased the formation of autolysosomes using an improved tandem fluorescent-tagged LC3 plasmid reporter (mTagRFP-mWasabi-LC3) [32]. Autolysosomes display red puncta (red⁺ green⁻) because mWasabi (green) is damaged due to their acidic environment. USP8-knockdown HeLa cells contained a much higher number of red puncta (red⁺ green⁻) than the wild-type control, indicating that USP8 knockdown remarkably increased the formation of autolysosomes (Figure 6D). Collectively, these results indicate that USP8 is a negative regulator of basal autophagy.

USP8 modulates SQSTM1 degradation and autophagy by directly deubiquitinating SQSTM1 at K420

Activation of autophagy can induce SQSTM1 degradation, and USP8 may regulate other substrates to affect SQSTM1 degradation and autophagy. Next, we sought to investigate whether USP8 affects SQSTM1 protein levels and autophagy by deubiquitinating SQSTM1 at K420. Knockdown of Usp8 had no significant effect on the levels of LC3-II in sqstm1^−/- MEFs (Figure 7A), suggesting a critical role of SQSTM1 in USP8-mediated autophagy. Furthermore, endogenous SQSTM1 was depleted from HEK293T cells using short hairpin RNA (shRNA), and the SQSTM1^K420R mutant was subsequently introduced into these cells to generate SQSTM1^K420R mutant-expressing HEK293T cells. We found that USP8 overexpression did not obviously change the levels of the SQSTM1^K420R mutant and LC3-II in these cells (Figure 7B). Finally, rapamycin treatment could not lead to the degradation of GFP-Q46 in either SQSTM1-knockdown or SQSTM1^K420R mutant-expressing HEK293T cells (Fig. S4B), suggesting an essential role of SQSTM1 in GFP-Q46 autophagic degradation. Again, overexpression of USP8 did not influence the levels of GFP-Q46 in either SQSTM1-knockdown or SQSTM1^K420R mutant-expressing HEK293T cells (Fig. S4B and S4C). Collectively, these results demonstrate that USP8 regulates SQSTM1 activity and autophagy by deubiquitnating SQSTM1 at K420.

Figure 7.

USP8 regulates autophagy by directly deubiquitinating SQSTM1 at K420. (A) Knockdown of Usp8 failed to increase the level of LC3-II in sqstm1^−/- MEFs. (B) Overexpression of USP8 did not affect the levels of SQSTM1^K420R mutant and LC3-II in SQSTM1-depleted HEK293T cells stably infected with the SQSTM1^K420R mutant.

Discussion

The ubiquitination modification of SQSTM1 has been demonstrated to modulate its function. Currently, several E3 ligases have been identified to ubiquitinate SQSTM1 at specific lysine residues to positively or negatively control the function of SQSTM1. However, deubiquitinases for SQSTM1 has never been reported. In this study, we report that USP8, a ubiquitin-specific protease, is a pivotal SQSTM1 deubiquitinase. We found that USP8 reduced the ubiquitination levels of SQSTM1 both in vitro and in vivo. USP8 deubiquitinated SQSTM1 principally at K420 within its UBA domain. Knockdown of Usp8 in cells led to autophagic degradation of SQSTM1 and autophagic influx at the basal condition. Taken together, this study identifies USP8 as the first deubiquitinase involved in regulating the ubiquitination and autophagic activity of SQSTM1.

The ubiquitination of autophagic receptors is emerging as a means of regulating their function. Ubiquitination of autophagy receptor OPTN (optineurin) by HACE1, a ubiquitin ligase, increases its association with SQSTM1, enhancing autophagic flux and suppressing tumor growth [33]. More recently, several ubiquitin ligases have been identified to promote the ubiquitination of SQSTM1 and regulate its sequestration activity. It is particularly interesting that depending upon their site of ubiquitination, these E3 ligases inhibit or facilitate SQSTM1’s function. TRIM21 ubiquitinates SQSTM1 on K7 within N-terminal PB1 domain, inhibiting SQSTM1 oligomerization and sequestration activity [17]. In contrast, KEAP1-CUL3 ubiquitinates SQSTM1 at K420 within its UBA domain to promote the sequestering function of SQSTM1 [22]. Our previous study also showed that ubiquitin stress induced the ubiquitination of SQSTM1 at K420 to activate its autophagy receptor function [18]. Importantly, K420 is the principal site for USP8-mediated deubiquitination. USP8 appears to function to antagonize KEAP1-CUL3-mediated ubiquitination at K420 to decrease the sequestering activity and autophagic degradation of SQSTM1 (Figure 4C). This may serve to enhance KEAP1 levels. Substrates of the KEAP1-CUL3 complex are GABPA and NFE2L2/NRF2, a crucial transcription activator of genes regulated by antioxidant response element [34]. Thus, USP8 may indirectly induce NFE2L2 degradation to inhibit the antioxidant response pathway. Previous studies have demonstrated that phosphorylation on Ser403 in the UBA domain was able to increase its binding affinity to the ubiquitylated substrates [35,36]. Our study further confirms that ubiquitination of K420 represents another pivotal regulatory mechanism by which the sequestration activity of SQSTM1 is regulated, although the underlying mechanism remains to be further defined. It is likely that ubiquitination and phosphorylation of SQSTM1 within the UBA domain may cooperate to control its sequestration activity.

USP8 has been extensively studied for its critical role in endosomal trafficking in the endosomal sorting complex required for transport (ESCRT) pathway and in regulating the endocytosis of the EGFR and other receptors [26,27]. More recently, we found that USP8 deubiquitinates GJA1/connexin 43 to repress its autophagy-mediated degradation [30]. USP8 was also reported to regulate mitophagy, the selective degradation of mitochondria by autophagy, by removing K6-linked ubiquitin conjugates from PRKN/parkin [29]. Our current study demonstrates that USP8 can inhibit selective autophagy by deubiquitinating SQSTM1. It is likely that USP8 may regulate multiple components of autophagy pathway to strictly modulate the autophagic processes. Further studies are needed to identify other autophagy-associated USP8 targets. Therefore, USP8 controls protein turnover by regulating the ESCRT and autophagy pathways. Of note, SQSTM1 is also implicated in the ESCRT pathway [37,38]. Endocytosed polyubiquitinated protein complexes can be recognized by SQSTM1 via its UBA domain and subsequently recruited into the autophagic degradation pathway. Thus, USP8 may indirectly affect the degradation of endocytosed proteins including GJA1 by modulating the activity of SQSTM1.

Aberrant activity and expression of SQSTM1 have been shown to be associated with many human diseases including cancers. It is becoming apparent that SQSTM1 is a crucial component of the cell transformation machinery [3]. USP8 gain-of-function mutations are prevailing in Cushing disease [39,40], which is also known as POMC/ACTH (proopiomelanocortin)-secreting pituitary adenomas that secret excessive cortisol. USP8 mutations may impair autophagic degradation of SQSTM1 and its accumulation subsequently contributes to the pathogenesis of Cushing disease. The important role of the USP8-SQSTM1 axis in the pathogenesis of cancer needs to be further explored.

Materials and methods

Cell culture and transfection

HEK293T (human embryonic kidney 293T), HeLa, MEF, atg7^−/- MEF and AtT20 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific, 11,965,118) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, 10,091,148) and penicillin- streptomycin [Thermo Fisher Scientific, 15,070,063]) and maintained at 37°C under 5% CO₂. Transient transfections of cells were performed with PEI (Poly Biosciences, 24,765–1) according to the manufacturer’s recommendations.

Antibodies and reagents

The primary antibodies used were as follows: anti-LC3B (Sigma-Aldrich, L7543), anti-USP8 (Sigma-Aldrich, SAB4200527), anti-ubiquitin (Ub; Sigma-Aldrich, SAB2702287), anti-His (Sigma-Aldrich, SAB2702218), anti-Flag (Sigma-Aldrich, F1084), anti-HA (Sigma-Aldrich, SAB1411738), anti-SQSTM1 (Abcam, ab91526), anti-K63-Ub (Abcam, ab179434), anti-K48-Ub (Abcam, ab140601), anti-TUBB/tubulin (Proteintech, 10,094–1-AP), anti-c-myc (Santa Cruz Biotechnology, sc-40), anti-ACTB/β-actin (Santa Cruz Biotechnology, sc-517,582) and anti-GAPDH (Santa Cruz Biotechnology, sc-137,179). The secondary antibodies used in western blotting were as follows: goat anti-rabbit IgG-HRP (Santa Cruz Biotechnology, sc-2004) and goat-anti-mouse IgG-HRP (Santa Cruz Biotechnology, sc-2005). Anti-FLAG® M2 Agarose Affinity Gel (A2220), 3× FLAG® Peptide (F4799), anti-HA (SAB1411738) and anti-HA-Agarose antibody (A2096) were from Sigma-Aldrich. Ni-NTA Agarose (R90115) and Pierce™ Glutathione Agarose (16,101) were purchased from Thermo Fisher Scientific. Bafilomycin A₁ (S1413) and bortezomib (PS-341) were from Selleckchem. Cycloheximide (C0934) was purchased from Sigma-Aldrich.

Construction of plasmids

The human USP8 cDNA was subcloned in pcDNA3.0 (Invitrogen Iife technologies, A-150,228) with a HA (pcDNA3.0/HA) or Flag (pcDNA3.0/Flag) epitope tag, or pET28a plasmids (Merck, 69,864-3CN). The human SQSTM1 cDNA were subcloned in pcDNA3.0 (Invitrogen Iife technologies, A-150,228) with a His tag (pcDNA3.0/His), pCMV3Tag4 (Agilent Technologies, 240,198), pET28a (Merck, 69,864-3CN) with Flag tag or pGEX4T-1 (GE Healthcare, 27–4580-01) plasmids. E1, UBE2D2 and UBE2D3 expression vectors were described previously [18]. Point mutations and truncations of SQSTM1 were created using the QuickChange site-directed mutagenesis method with KOD-Plus (Code No. KOD-201). The sequence of primers used in cloning is listed in Table S1. HA- wild-type Ub and its mutants (K6, K11, K27, K29, K33, K48 and K63), and KEAP1-myc expression vectors (21,555, deposited by Dr. Yue Xiong) were from Addgene. The SQSTM1^K420R mutant cDNA was sub-cloned into pLVX-Hygromycin (Clontech, 632,164). All constructs were confirmed by DNA sequencing.

Generation of knockdown cell lines

Lentivirus vectors encoding shRNA against USP8 were generated using the pLKO.1 vector with puromycin selection marker (Addgene, 8453; deposited by Dr. Bob Weinberg). The shRNA target sequences for mouse Usp8 are 5-TCAAGCAACAGCAGGATTATT-3 (shUsp8-1) and 5-CTCACATCTAATGCTTACAA-3 (shUsp8-2). The shRNA target sequence for human USP8 and SQSTM1 is 5-TCAAGCAACAGCAGGATTATT-3 (shUSP8-1) and 5-TCTGTCTCATAGTTGTGTTAA-3(shSQSTM1). The viral packaging was performed as described previously. The filtered virus supernatant was used to infect indicated cells. Three days post-infection, cells was selected in the presence of puromycin (1 µg/ml; Thermo Fisher Scientific, A1113802) for 5 days. Western blotting was performed to analyze the efficiency of USP8 knockdown.

Immunoprecipitation and western blotting

For immunoprecipitation (IP), whole-cell extracts were collected 36 h after transfection and lysed in IP buffer (1% [vol:vol] Triton X-100 [Sigma-Aldrich, 9002–93-1], 50 mM Tris-HCl, pH 7.4, 50 mM EDTA, 150 mM NaCl, 10 mM NaF, 10% glycerol [Sigma-Aldrich, 56–81-5] and fresh protease inhibitor cocktail [Roche, 04693132001]). After centrifugation for 10 min at 14,000 g, supernatants were collected and incubated with anti-FLAG® M2 Agarose Affinity Gel (Sigma-Aldrich, A2220) or anti-HA-Agarose antibody (Sigma-Aldrich, A2095). After 8 h of incubation, beads were washed 5 times with IP buffer. Immunoprecipitates were eluted by boiling with 1% (wt/vol) SDS sample buffer (60 mM Tris-HCl [Sigma-Aldrich, PHG0002], pH 6.8, 1% [wt:vol] SDS [Sigma-Aldrich, 74,255], 5% [vol:vol] glycerol, 0.005% [wt:vol] bromophenol blue [Sigma-Aldrich, B0126], 1% [vol:vol] 2-mercaptoethanol [Sigma-Aldrich, 07604]). Whole cell lysates were prepared in RIPA buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 10 mM NaF, 1 mM EDTA, 1% Triton X-100, 0.1% SDS, 2 mM sodium pyrophosphate, 1 mM DTT, 10% glycerol and fresh protease inhibitor cocktail).

For western blotting, immunoprecipitated proteins or whole cell lysates were resolved by SDS-PAGE. The proteins in the gel were transferred to PVDF membranes (Bio-Rad, 1,620,177). The immunoblot was performed as described previously, using Pierce™ ECL western blotting substrate (Thermo Fisher Scientific, 34,095). Each western blotting was repeated for several times and the representative results were shown.

GST affinity isolation

GST or GST-SQSTM1 fusion protein was purified from the strain of BL21. GST or GST-SQSTM1 was incubated with His-USP8 in GST pull-down buffer (10 mM Tris-HCl [pH 8.0], 1 mM EDTA, 100 mM NaCl, fresh protease inhibitor cocktail) overnight at 4°C, then prewashed Pierce™ Glutathione Agarose (250–500 μL beads per 100 mL of culture volume, 50% bead/buffer slurry; Thermo Fisher Scientific, 16,100) were added into each sample and tubes were rotated for 2 h at 4°C. After five times washing with GST pull-down buffer, samples were subjected to western blotting.

In vitro deubiquitination assay using ubiquitinated SQSTM1 purified from cells

Ubiquitinated SQSTM1 was isolated from HEK293T cells transfected with expression vectors for His-Ub and SQSTM1-Flag. Ubiquitinated SQSTM1 was purified from the cell extracts with ANTI-FLAG® M2 Agarose Affinity Gel in RIPA buffer. After extensive washing with RIPA buffer, the proteins were eluted with 3× FLAG® Peptide (Thermo Fisher Scientific, F4799). The recombinant USP8 or its mutant were purified from HEK293T cells overexpressing USP8-HA or its mutant using anti-HA-Agarose antibody (Sigma-Aldrich, A2095) in IP buffer (1%[vol/vol] Triton X-100, 50 mM Tris-HCl [pH 7.4], 50 mM EDTA, 150 mM NaCl, 10 mM NaF, 10% glycerol and fresh protease inhibitor cocktail). For in vitro deubiquitination assay, ubiquitinated SQSTM1 protein was incubated with USP8 or its mutant in the deubiquitination buffer (50 mM Tris-HCl [pH 8.0], 50 mM NaCl, 1 mM EDTA, 10 mM DTT, 5% glycerol) for the indicated time courses at 37°C. The ubiquitinated SQSTM1 were analyzed by immunoblotting.

In vitro deubiquitination assay using ubiquitinated SQSTM1 generated from in vitro reactions

E1, UBE2D2, UBE2D3, ubiquitin, SQSTM1-Flag and USP8-His were purified from E.coli. E1, UBE2D1/UBCH5A, UBE2D3/UBCH5C and ubiquitin were incubated with SQSTM1-Flag in ubiquitination buffer (25 mM Tris-HCl [pH 8.0], 5 mM MgCl2, 100 mM NaCl, 1 mM DTT, supplemented with 2mM ATP [Thermo Fisher Scientific, R0441]) to synthesize ubiquitinated SQSTM1 in vitro. The ubiquitinated SQSTM1 was enriched by anti-FLAG® M2 Agarose Affinity Gel in RIPA buffer. After extensive washing with RIPA buffer, the proteins were eluted with 3× FLAG® Peptide. The ubiquitinated SQSTM1 was then incubated with USP8 protein for indicated time courses, and analyzed by western blotting.

In vivo deubiquitination assay

HEK293T cells were transfected with indicated plasmids and harvested at 48 h after transfection, and 20% of the cells were used for direct immunoblotting and the rest of cells were harvested in denaturing buffer (6 M guanidine-HCl [Thermo Fisher Scientific, 50,933], 0.1 M Na₂HPO₄/NaH₂PO₄ buffer [Sigma-Aldrich, S8282], 0.01 M Tris-HCl [pH8.0], 5 mM imidazole, 10 mM 2-mercaptoethanol). The lysates were incubated with Ni²⁺-NTA-agarose beads (Qiagen, 30,210) for 3 h. Then, beads were washed with indicated buffer step by step as follows: buffer B (8 M urea [Sigma-Aldrich, U0931], 0.1 M Na₂HPO₄/NaH₂PO₄, 0.01 M Tris-HCl [pH8.0], 10 mM 2-mercaptoethanol); buffer C (8 M urea, 0.1 M Na₂HPO₄/NaH₂PO₄, 0.01 M Tris-HCl [pH 6.3], 10 mM 2-mercaptoethanol); buffer C plus plus 0.2% Triton X-100. Finally, ubiquitinated proteins were eluted by elution buffer (200 mM imidazole in 5% SDS, 0.15 M Tris-HCl [pH 6.7], 30% glycerol, 0.72 M 2-mercaptoethanol) and analyzed by immunoblotting with indicated antibodies.

Immunofluorescence

Cells grown on glass coverslips were fixed with 4% paraformaldehyde in PBS (Thermo Fisher Scientific, 10,010,049,) for 20 min, permeabilized with 0.1% Triton X-100 and blocked with 1% bovine serum albumin (Thermo Fisher Scientific, 37,520). After permeablization and blockage, cells were stained with the indicated primary antibodies overnight at 4°C, then followed by incubation with fluorescent-dye-conjugated secondary antibodies. Nuclei were counterstained with DAPI (Sigma-Aldrich, D9542). Imaging of the cells was carried out using a Leica laser-scanning confocal microscope. The images were acquired by confocal microscope (Leica TCS SP5). Antibodies are listed as follows: anti-USP8 (Sigma, SAB4200527; 1:50), anti-SQSTM1 (Abcam, ab91526; 1:200), donkey anti-rabbit IgG-PE (Santa Cruz Biotechnology, sc-3745; 1:1000), goat anti-mouse IgG-FITC (Santa Cruz Biotechnology, sc-2010; 1:1000).

Statistical analysis

Student’s t test was used for statistical analysis. P < 0.05 is considered to be significant.

Funding Statement

This work was supported by the National Key R&D Program of China (grant no. 2018YFC0115900) to C. Huang,National Natural Science Foundation of China (grant no. 31870872) to C.Huang; National Natural Science Foundation of China (grant no. 81702724) to H.Peng; National Outstanding Youth Foundation of China (grant no. 81725011) to Y. Zhao.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary Material

Supplemental data for this article can be accessed here.