TRIM21-mediated ubiquitination of SQSTM1/p62 abolishes its Ser403 phosphorylation and enhances palmitic acid cytotoxicity

ABSTRACT

Long-chain free fatty acids (FFAs) accumulation and oxidative toxicity is a major cause for several pathological conditions. The mechanisms underlying FFA cytotoxicity remain elusive. Here we show that palmitic acid (PA), the most abundant FFA in the circulation, induces S403 phosphorylation of SQSTM1/p62 (sequestosome 1) and its aggregation, which sequesters KEAP1 and activates the non-canonical SQSTM1-KEAP1-NFE2L2 antioxidant pathway. The PA-induced SQSTM1 S403 phosphorylation and aggregation are dependent on SQSTM1 K7-D69 hydrogen bond formation and dimerization in the Phox and Bem1 (PB1) domain, which facilitates the recruitment of TBK1 that phosphorylates SQSTM1 S403. The ubiquitin E3 ligase TRIM21 ubiquitinates SQSTM1 at the K7 residue and abolishes the PB1 dimerization, S403 phosphorylation, and SQSTM1 aggregation. TRIM21 is oxidized at C92, C111, and C114 to form disulfide bonds that lead to its oligomerization and decreased E3 activity. Mutagenizing the three C residues to S (3CS) abolishes TRIM21 oligomerization and increases its E3 activity. TRIM21 ablation leads to decreased SQSTM1 K7 ubiquitination, hence elevated SQSTM1 S403 phosphorylation and aggregation, which confers protection against PA-induced oxidative stress and cytotoxicity. Therefore, TRIM21 is a negative regulator of SQSTM1 phosphorylation, aggregation, and the antioxidant sequestration function. TRIM21 is oxidized to reduce its E3 activity that helps enhance the SQSTM1-KEAP1-NFE2L2 antioxidant pathway. Inhibition of TRIM21 May be a viable strategy to protect tissues from lipotoxicity resulting from long-chain FFAs.

Abbreviations: ER: endoplasmic reticulum; FFA: free fatty acid; HMOX1/HO-1: heme oxygenase 1; IB: immunoblotting; IF: immunofluorescence; IP: immunoprecipitation; KEAP1: kelch like ECH associated protein 1; MASH: metabolic dysfunction-associated steatohepatitis; MEF: mouse embryonic fibroblast; NFE2L2/Nrf2: NFE2 like BZIP transcription factor 2; PA: palmitic acid; PB1: Phox and Bem 1; ROS: reactive oxygen species; SLD: steatotic liver disease; SQSTM1: sequestosome 1; TBK1: TANK-binding kinase 1; TRIM21: tripartite motif containing 21.

Introduction

Intracellular accumulation of long chain free fatty acids (FFAs) is associated with cellular dysfunction and contributes to the pathogenesis of several diseases including insulin resistance, diabetes, cardiovascular diseases, inflammation, and liver diseases. Serum levels of FFAs are elevated in patients with metabolic dysfunction-associated steatohepatitis (MASH) and steatotic liver disease (SLD) [1,2]. Palmitic acid (PA) is the most abundant FFA in the circulation. It induces cell death in many cell types including pancreatic $\mathit{\beta }$-cells, cardiac myocytes, microvascular endothelial cells, and hepatocytes [3–6]. PA-induced hepatic lipotoxicity is considered one of the factors that play a crucial role in the pathogenesis of SLD [6,7].

Several mechanisms have been reported to mediate PA-induced lipotoxicity, including MAPK/JNK-dependent BAX activation [8], PPP2/PP2A-mediated FOXO3/FoxO3A dephosphorylation that leads to increased transcription of BCL2L11/Bim [9], lysosomal permeabilization and CTSB (cathepsin B) activation [10], endoplasmic reticulum (ER) stress [11], and generation of reactive oxygen species (ROS). Many of these molecular events are inter-connected, and ROS are believed to play a major role in FFA lipotoxicity resulting from disrupted intracellular membrane integrity, mitochondrial dysfunction, calcium imbalance, and excessive ER stress response [12,13].

The ubiquitin-binding protein SQSTM1/p62 (sequestosome 1), among its numerous functions, critically regulates both proteostasis and redox balance, by sequestering certain proteins in aggregates and delivering them to phagophores for degradation. One of the client proteins sequestered by SQSTM1 is KEAP1 (kelch like ECH associated protein 1), a negative regulator of the antioxidant response that suppresses the antioxidant transcription factor NFE2L2/Nrf2 (NFE2 like bZIP transcription factor) [14,15]. This sequestration function of SQSTM1 relies on its dimerization via the hydrogen bond between lysine (K)7 and aspartate (D)69 residues in the N-terminal Phox and Bem1p (PB1) domain [16]. We recently reported that TRIM21 (tripartite motif containing 21), a RING finger domain-containing ubiquitin E3 ligase, directly interacts with and ubiquitinates SQSTM1 at K7 via K63-linkage, which abolishes the K7-D69 hydrogen bond hence SQSTM1 oligomerization, aggregation, and sequestration functions [17]. TRIM21-mediated SQSTM1 K7 ubiquitination leads to the failure of KEAP1 sequestration and suppressed antioxidant response. Conversely, TRIM21-deficient cells display increased SQSTM1 oligomerization, protein aggregation, KEAP1 sequestration, NFE2L2 activation, and antioxidant response [17].

In addition to the K7 ubiquitination that negatively regulates SQSTM1 dimerization and sequestration, SQSTM1 is regulated at the post-translational level via several other mechanisms. Its sequestration and protein clearance activity is positively regulated via phosphorylation at multiple sites: at human SQSTM1 S403 (S405 in mouse) by casein kinase 2 (CK2), TBK1, MAP3K7/TAK1, or ULK1 [18–22], at S349 (S352 in mouse) in an MTORC1-dependent manner [23], at S407 (S409 in mouse) by ULK1 [22]. Since some of these phosphorylation events are associated with SQSTM1 sequestration and antioxidant function, the molecular relationship between SQSTM1 K7 ubiquitination and the phosphorylation events remains elusive. In this study, we set out to study this question regarding the regulation of SQSTM1 aggregation, sequestration function, and its antioxidant capacity in response to PA exposure.

Results

PA-induced SQSTM1 Ser403 phosphorylation is dependent on its K7-D69 dimerization

To study the regulation of SQSTM1 phosphorylation, we treated several liver cancer cell lines (SK-HEP-1, HepG2, Huh7, and SNU449) with cell stress inducers including the lysosome inhibitor chloroquine, proteasome inhibitor MG132, ER stress inducer thapsigargin, and palmitic acid (PA). Phosphorylation of S349 and S403 was induced to various levels by these agents, with the most robust induction by PA in all cell lines tested (Figure 1A). This PA-induced SQSTM1 phosphorylation correlated with increased antioxidant response indicated by elevated expression of HMOX1 (heme oxygenase 1) upon PA treatment (Figure 1A).

Figure 1.

SQSTM1 S403 phosphorylation is dependent on its K7-D69 dimerization. (A) SK-HEP-1, HepG2, Huh7, and SNU449 cells were treated with chloroquine (10 μM), MG132 (10 μM), thapsigargin (1.5 μM) or PA (500 μM) for 12 h. Cells were lysed in RIPA (1% SDS) and subjected to IB with respective antibodies. Relative densitometry ratios of p-SQSTM1^S403 to total SQSTM1, and that of HMOX1 to actin, were calculated by normalizing the ratio to that of the vehicle treated cells. Shown is the average of three immunoblots. (B) sqstm1^−/− MEFs reconstituted with vector, Flag-SQSTM1 WT, Flag-SQSTM1^S403A, Flag-SQSTM1^S403E, Flag-SQSTM1^S349A or flag- SQSTM1^S349E mutants were treated with BSA or PA (500 μM) for 12 h. Cells were lysed in RIPA lysis buffer with 1% triton X-100 and subjected to soluble/insoluble fractionation. The insoluble fractions were dissolved in RIPA buffer containing 1% SDS. Both the triton X-100 soluble fractions and corresponding volume of insoluble fractions were probed for indicated proteins. Relative densitometry ratio of insoluble to soluble SQSTM1 is shown. Note that S403A profoundly decreased SQSTM1 insoluble fraction (highlighted in red). (C) SNU449 cells stably expressing flag-SQSTM1 WT or Flag-SQSTM1^K7R were treated with BSA or PA (500 μM) for 12 h. Cells were subjected to immunoprecipitation (IP) and immunoblot (IB) with flag and p-SQSTM1^S403 antibodies. (D and E) sqstm1^−/− MEFs reconstituted with vector, flag-SQSTM1 WT, K7R, D69A, S403A or S403E were treated with BSA or PA (500 μM) for 12 h. Cells were probed for respective proteins (D). Both the triton X-100 soluble fractions and corresponding volume of insoluble fractions were subjected to indicated antibodies (E). Relative densitometry ratios were calculated by normalizing to the ratios of WT cells. Note that SQSTM1 S403 phosphorylation is profoundly induced in WT SQSTM1, and that SQSTM1 and KEAP1 aggregation is high in SQSTM1 WT and S403E cells (highlighted in red). Shown are the mean of three independent immunoblots. (F) sqstm1^−/− MEFs reconstituted with vector, flag-SQSTM1 WT, K7R, or D69A were treated with BSA or PA (500 μM) for 12 h, then subjected to immunofluorescence (IF).

Since dietary fatty acids play an important role in the pathogenesis of MASH and SLD, and PA is strongly lipotoxic to the liver and to cultured cells [24,25], we went on to study the regulation of SQSTM1 phosphorylation and its function in PA lipotoxicity. To minimize confusion, hereafter we do not distinguish the phosphorylation sites between human and mouse SQSTM1, and refer to them with the human sites S349, S403, and S407, respectively. Both S349 and S403 have been described to involve in SQSTM1 aggregate formation and selective autophagy [18,23]. We performed a test by reconstituting sqstm1 knockout MEFs with WT SQSTM1 and the phosphorylation inactive alanine mutants S349A and S403A, as well as the phosphorylation mimicry glutamate mutants S349E and S403E (Figure 1B). Using the detergent solubility assay (lysing cells in 1% Triton X-100), we observed that while PA treatment led to increased SQSTM1 aggregates (insoluble fraction), the aggregation was more profoundly diminished in cells reconstituted with the S403A mutant (Figure 1B), indicating that SQSTM1 S403 phosphorylation is required for its aggregation. Therefore, in the current study, we focused on the regulation of S403 phosphorylation.

Since the sequestration function of SQSTM1 also relies on its dimerization via the hydrogen bond between lysine (K)7 and aspartate (D)69 residues in the N-terminal Phox and Bem1p (PB1) domain [16], we determined whether the PB1 dimerization is required for SQSTM1 S403 phosphorylation. Flag-tagged WT SQSTM1 or the N-terminal dimerization-defective mutants K7R and D69A were ectopically expressed in SNU449 cells. While PA induced S403 phosphorylation of endogenous SQSTM1, only Flag-tagged human WT SQSTM1 but not the K7R and D69A mutants were phosphorylated (Figure 1C). To rule out the possible effect of the ectopically expressed SQSTM1 on endogenous SQSTM1, we also reconstituted Flag-tagged human SQSTM1 WT, K7R, and D69 in sqstm1^−/− MEFs. Like in the liver cancer cell lines, PA induced S403 phosphorylation of Flag-tagged WT SQSTM1 but not the K7R and D69A mutants in the reconstituted MEFs (Figure 1D). The specificity of S403 phosphorylation antibody was verified by the phospho-inactive S403A and constitutive active mutant S403E (Figure 1D). These results demonstrate that PA-induced SQSTM1 S403 phosphorylation is dependent on K7-D69 mediated SQSTM1 dimerization in the PB1 domain.

Using the 1% Triton X-100 solubility test, we observed that while PA treatment led to increased SQSTM1 aggregates (insoluble fraction), the aggregation was abolished in cells reconstituted with the K7R and D69A mutants, indicating the K7-D69 dimerization is required for PA-induced SQSTM1 aggregation (Figure 1E). Consistent with the earlier result (Figure 1B), while S403A abolished SQSTM1 aggregation, S403E was induced to aggregate by PA (Figure 1E), indicating that SQSTM1 S403 phosphorylation is required for its aggregation. KEAP1, which retains the antioxidant transcription factor NFE2L2 in the cytosol, was induced to be sequestered in the insoluble fraction by WT SQSTM1 and the S403E mutant, but not the K7R, D69A, and S403A mutants (Figure 1E). Interestingly, immunofluorescence (IF) showed that upon PA treatment, SQSTM1 formed large size aggregates that were positive for S403 phosphorylation, whereas both the aggregation and S403 phosphorylation were abolished in the K7R and D69 mutants (Figure 1F). These results indicate that PA-induced SQSTM1 aggregation and sequestration are dependent on the K7-D69 mediated PB1 dimerization and S403 phosphorylation.

TRIM21 ubiquitinates SQSTM1 at K7 to abolish its S403 phosphorylation and aggregation

Our above results suggest that SQSTM1 S403 phosphorylation relies on its PB1 domain dimerization, which facilitates SQSTM1 oligomerization and sequestration function. To further illustrate that PA could induce SQSTM1 dimerization and oligomerization, we co-expressed SQSTM1 tagged with either HA or Flag in MEFs. PA treatment led to SQSTM1 dimerization indicated by increased co-IP of the differentially tagged SQSTM1 (Figure 2A). The SQSTM1 dimerization at both basal and PA-induced levels was decreased with the dimerization-defective K7R mutant (Figure 2A). The PA-induced SQSTM1 oligomerization could also be visualized by SDS-PAGE in non-reducing condition, which was diminished in the K7R mutant (Figure 2B). On the other hand, although SQSTM1^S403A mutant was defective in aggregation (Figure 1B, E), it still retained the ability to dimerize (Figure 2C). These results indicate that the K7-D69 mediated PB1 domain dimerization is induced by PA, which occurs prior to SQSTM1 S403 phosphorylation.

Figure 2.

TRIM21 inhibits SQSTM1 S403 phosphorylation and aggregation. (A) flag-SQSTM1, HA-SQSTM1 WT and K7R mutants were co-expressed in indicated combinations in HEK293T cells, and then treated with BSA or PA (500 μM) for 12 h. Cell lysates were subjected to flag IP, and IB with flag and HA antibodies. Densitometry ratios were calculated using image studio, and normalized to that of the WT SQSTM1 cells without PA treatment. Shown is the mean of three independent experiments. (B) sqstm1^−/− MEFs reconstituted with vector, flag-SQSTM1 WT, or Flag-SQSTM1^K7R were treated with BSA or PA (500 μM) for 12 h. Cell lysates were mixed with loading buffer containing or not containing DTT were subjected to SDS-PAGE and IB. (C) HA- or flag-tagged SQSTM1 WT or S403A mutants were expressed in HEK293T cells. Cells were subjected to flag IP, then probed for flag and HA. Densitometry ratio was calculated using image studio, and normalized to that of the WT SQSTM1 cells without PA treatment. Shown is the mean of three independent experiments. (D) sqstm1^−/− MEFs reconstituted with vector, flag-SQSTM1 WT or Flag-SQSTM1^K7R were treated with BSA or PA (500 μM) for 12 h. Cell lysates were subjected to flag IP, and IB with flag and ubiquitin antibodies. (E) Trim21^+/+ and trim21^−/− MEFs were treated with BSA or PA for 12 h. Cells were lysed in RIPA (1% SDS) and subjected to IB. Relative densitometry ratio from three independent experiments is shown. (F) HepG2 cells stably expressing vector, flag-SQSTM1, or Flag-SQSTM1^K7R were treated with BSA or PA. Cells were lysed in RIPA buffer with 1% triton X-100 and subjected to soluble/insoluble fractionation. The insoluble fractions were dissolved in RIPA buffer containing 1% SDS. Both soluble and insoluble fractions were subjected to IB. Relative densitometry ratio from three independent experiments is shown. (G) HepG2 cells infected with control shRNA or two independent shRNA for TRIM21 were treated with BSA or PA and probed with indicated proteins. Relative densitometry ratio from three independent experiments is shown. (H) HepG2 cells stably expressing control or shTRIM21 were treated with BSA or PA. The cell lysates mixed with loading buffer containing or not containing DTT were probed with indicated proteins. (I) AML12 stably expressing vector, flag-SQSTM1 WT, or Flag-SQSTM1^K7R were treated with BSA or PA. Both 1% triton X-100 soluble and insoluble fractions were subjected to IB. Relative densitometry ratio from three independent experiments is shown. (J) AML12 stably expressing vector or HA-TRIM21 were treated with BSA or PA. Total cell lysates (in 1% SDS) and triton X-100 soluble and insoluble fractions were subjected to IB with indicated proteins. Relative densitometry ratio from three independent experiments is shown.

Our previous study demonstrated that SQSTM1 is ubiquitinated at K7 by the ubiquitin E3 ligase TRIM21, which abrogates SQSTM1 oligomerization [17]. Indeed, PA induced SQSTM1 ubiquitination, which was abrogated in the K7R mutant (Figure 2D). To test whether TRIM21 affects SQSTM1 S403 phosphorylation by disrupting its oligomerization, we treated Trim21^+/+ and trim21^−/− MEFs with BSA or PA. As expected, PA induced SQSTM1 S403 phosphorylation, which was further enhanced in trim21^−/− MEFs (Figure 2E), indicating that abolishment of K7 ubiquitination enhances its oligomerization and subsequent S403 phosphorylation.

The effect of TRIM21 on SQSTM1 oligomerization and phosphorylation was also examined in the human liver cancer cell line HepG2. Like in MEFs, PA induced SQSTM1 aggregation in the detergent-insoluble fraction that was positive for S403 phosphorylation (Figure 2F). The S403 phosphorylation was abolished and aggregation reduced in the SQSTM1^K7R mutant (Figure 2F). Silencing TRIM21 led to increased S403 phosphorylation (Figure 2G) and oligomerization in non-reducing SDS-PAGE (Figure 2H). Likewise, PA treatment of the mouse liver cancer cell line AML12 led to increased insoluble WT SQSTM1 but to a significantly less extent of K7R (Figure 2I). TRIM21 overexpression inhibited SQSTM1 S403 phosphorylation and SQSTM1 aggregation (Figure 2J). Together, these results indicate that TRIM21-mediated K7 ubiquitination abolishes its S403 phosphorylation and aggregation function.

PA-induced SQSTM1 S403 phosphorylation is mediated by TBK1

Several kinases have been reported to regulate SQSTM1 S403 phosphorylation including CK2, TBK1, ULK1, and MAP3K7 [18–22]. Interestingly, in MEFs, PA-induced SQSTM1 S403 phosphorylation was robustly inhibited by the pharmacological TBK1 inhibitor B×795but not by the other kinase inhibitors (Figure 3A). This is consistent with a previous report that TBK1 phosphorylates S403 in a MASH model [26].

Figure 3.

TBK1 mediates pa-induced SQSTM1 phosphorylation. (A) MEFs were treated with PA (500 μM) alone, or in combination with CK2 inhibitor (CX-4945, 10 μM), ULK1 inhibitor (SBI-0206965,10 μM), MAP3K7 inhibitor (5Z-7-oxozeaenol, 10 μM), or TBK1 inhibitor (BX-795, 5 μM) for 12 h. Whole-cell lysates were subjected to IB with indicated antibodies. Western bands were quantified by image studio, and relative ratio of phospho-SQSTM1 S403 to total SQSTM1 was calculated and normalized to that of the untreated control cells. Shown is the mean of three independent immunoblots. (B) sqstm1^−/− MEFs reconstituted with vector, flag-SQSTM1 WT, Flag-SQSTM1^S407A or Flag-SQSTM1^S403E,S407A were treated with PA or in combination with TBK1 inhibitor (BX795) or ULK1 inhibitor (SBI-0206965). Cell lysates were subjected to IB and probed for indicated proteins. The average densitometry ratio of three independent experiments is shown. (C) MEFs were treated with PA alone or with TBK1 inhibitor (BX795) at indicated concentrations. Cells were probed with indicated proteins. Densitometry ratio was calculated, and shown is the mean of three independent experiments. (D) sqstm1^−/− MEFs reconstituted with vector, flag-SQSTM1 WT, or Flag-SQSTM1^K7R were treated with PA and BX795, separately or together. Cell lysates were probed for indicated proteins. The average densitometry ratio of three independent experiments is shown. (E) HEK293T cells were co-transfected with TBK1 and flag-SQSTM1 WT or Flag-SQSTM1^K7R mutant. Cell lysates were probed for indicated proteins. The average densitometry ratio of three independent experiments is shown. (F) Trim21^+/+ and trim21^-/- MEFs were treated with PA and B×795at indicated concentrations. Cell lysates were probed with indicated proteins. The average densitometry ratio of three independent experiments is shown. (G) HA-SQSTM1 WT or HA-SQSTM1^K7R mutant were con-transfected with flag-TBK1. After 24 h, cells were treated with BSA or PA for 16 h. Cell lysates were subjected to flag-IP and IB. The average densitometry ratio of three independent experiments is shown. (H) SNU449 stably expressing flag-SQSTM1 WT or Flag-SQSTM1^K7R were treated with BSA or PA. Cells were subjected to if with flag and TBK1 antibodies. (I) Co-ip assay to detect the interaction between SQSTM1 and TBK1 under the condition of TRIM21 overexpression. The average densitometry ratio of three independent experiments is shown.

In addition to S349 and S403, S407 has been shown to be phosphorylated [22]. To test the possibility that the phospho-S403 antibody we used also recognizes phospho-S407, we generated S407A and S403E S407A mutants and expressed them in the sqstm1 KO MEFs. PA treatment induced S403 phosphorylation in WT SQSTM1 and the S407A mutant, but not in the S403E S407A double mutant (Figure 3B). This PA-induced S403 phosphorylation in the WT and S407A mutant was inhibited by the TBK1 inhibitor BX795, but not by ULK1 inhibitor (Figure 3B). These results strongly indicate that PA specifically induces S403 phosphorylation, which is mediated by TBK1.

Moreover, the B×795inhibition of PA-induced S403 phosphorylation was dose-dependent (Figure 3C). Consistent with the results that SQSTM1 S403 phosphorylation is abolished in K7R (Figure 1C–1F, Figure 2F), PA treatment led to S403 phosphorylation of reconstituted WT SQSTM1, which was also inhibited by BX795, whereas S403 phosphorylation did not occur in reconstituted SQSTM1^K7R (Figure 3D).

To further confirm that TBK1 is the key kinase for SQSTM1 S403 phosphorylation, we co-expressed TBK1 with SQSTM1 WT or K7R in HEK293T cells. Ectopically expressed TBK1 led to S403 phosphorylation, which was markedly higher in WT SQSTM1 comparing with the K7R mutant (Figure 3E). The enhanced S403 phosphorylation in the trim21^−/− MEFs was abolished by B×75treatment (Figure 3F). These results demonstrate that TBK1 is responsible for PA-induced SQSTM1 S403 phosphorylation. To explore the underlying mechanism that SQSTM1 K7 ubiquitination inhibits S403 phosphorylation, we performed co-IP assay to test the binding ability between SQSTM1 and TBK1. PA enhanced the interaction of TBK1 and WT SQSTM1, whereas the interaction between TBK1 and SQSTM1^K7R was markedly lower and was not enhanced by PA (Figure 3G). Consistently, immunofluorescence (IF) shows that upon PA treatment, WT SQSTM1 and TBK1 were induced to co-localize in the SQSTM1 aggregates, whereas SQSTM1^K7R and TBK1 remained diffused in the cytosol (Figure 3H). These results indicate that K7-mediated SQSTM1 PB1 dimerization is required for TBK1 interaction and S403 phosphorylation. Consistently, ectopic expression of TRIM21 reduced SQSTM1-TBK1 interaction (Figure 3I). Taken together, these results indicate that TRIM21 negatively regulates SQSTM1 S403 phosphorylation by disrupting SQSTM1 dimerization and its interaction with TBK1.

TRIM21 is oxidized to oligomerize and lose its E3 function

We have previously shown that TRIM21 ubiquitinates SQSTM1 at the K7 residue to prevent SQSTM1 dimerization in the PB1 domain and its aggregation and sequestration function, thereby negatively regulates antioxidant response [17]. Lipotoxicity caused by FFA has been largely attributed to oxidative stress resulting from enhanced mitochondrial respiration and hydrogen peroxide production from peroxisomes [27,28]. Since TRIM21 acts as a regulator of antioxidant response, we tested whether TRIM21 function is regulated in response to oxidative stress.

During our study, we noticed that TRIM21 showed a slow-migrating form in SDS-PAGE under the non-reducing condition (Figure 4A). This slow-migrating species was reversed to the original form by the reducing agent DTT (Figure 4A). We also found that the retarded migration was abolished in the C-terminal (Δ8-251) half but retained in the N-terminal half of TRIM21, which was also sensitive to DTT but not SDS (Figure 4A, B). These results indicate that TRIM21 May be oxidized to form covalent bonds that are resistant to SDS but sensitive to reducing conditions. Consistent with this theory, co-IP of differentially tagged TRIM21 showed that TRIM21 could oligomerize, and the oligomerization was further enhanced by PA or H₂O₂ treatment (Figure 4C). To further study the biochemical nature of this TRIM21 oligomerization, we mainly used H₂O₂ treatment since it can effectively induce oxidative stress in both cultured cells and cell lysates. Indeed, H₂O₂ enhanced the oligomerization of WT TRIM21 and the N-terminal half in a dose-dependent manner (Fi. 4D).

Figure 4.

TRIM21 oligomerization inhibits its E3 ubiquitin ligase activity. (A) flag-TRIM21 WT, Flag-TRIM21^∆8-251 or Flag-TRIM21^∆252-476 were expressed in HEK293T cells. Cell lysates mixed with loading buffer containing or not containing DTT were subjected to IB. Flag-TRIM21^∆8-251 lose the oligomerization of TRIM21. (B) the lysates from 293T cell overexpressed with flag-TRIM21 WT, Δ8-251, and Δ252-476 were mixed with loading buffer without reducing agent, with 100 mM DTT, or with extra SDS (2%). The samples were subjected to western blotting. (C) flag-TRIM21 and HA-TRIM21 were transfected to HEK293T for 24 h. Cells were treated with PA (500 mM) or H₂O₂ and subjected to flag IP, then IB with flag or HA antibodies. (D) HEK293T cells were transfected with flag-TRIM21 WT, flag-TRIM21[∆8-251] or flag-TRIM21[∆252-476], then treated with H₂O₂ at indicated concentrations for 1 min. Cell lysates mixed with loading buffer not containing DTT were probed with flag antibody. (E) flag-TRIM21 was transfected into HEK293T cells and harvested after 48 h. Cell lysates were treated with H₂O₂ (2 mM) and 15 mg/ml NEM for 1 h, then subjected to flag-ip, and resolved in non-reducing SDS-PAGE and stained with Coomassie Brilliant Blue R-250. TRIM21 oxidized oligomer and monomer are indicated. (F) appropriate lanes of the gel were excised. In-gel trypsin digestion was performed and the oxidized cysteines were reduced with DTT followed by iodoacetamide (IAM) alkylation. Both the NEM- and iam-modified peptide were identified and quantified by LC-MS/MS based on Label Free Quantification/lfq. The percentage of the oxidation for each peptide was calculated from the intensity of oxidized iam-peptide relative to the sum of the intensity of the reduced (NEM-) and oxidized (IAM-) peptide. Data from both untreated and H₂O₂-treated TRIM21 are plotted (top plot). Cys oxidation fold change was calculated from the oxidation % (bottom). (G) dimeric structure of TRIM21 (PDB: 5OLM). The RING domain is colored green and orange, and B-Box2 domain is colored blue. The B-Box2 contains two zinc-binding sites, I and II. The zinc ion in site I is coordinated by C92, C111, C114, and H95. Zinc ions are shown as spheres in wheat, and the side chains of H98, C92, C111, and C114 are shown as spheres. (H) HEK293T cells transiently expressing flag-TRIM21 WT or Flag-TRIM21^∆8-251 were treated with H₂O₂ (1 mM) for 18 h. Cell lysates were resolved on non-reducing SDS-PAGE. (I) HA- or flag-tagged TRIM21 WT or 3CS mutants were expressed in indicated combinations HEK293T cells. Cells were subjected to flag IP, then probed for flag and HA. (J) flag-TRIM21 weas co-transfected with his-ub in HEK293T cells. Cells were subjected to his pull down and IB. (K) his-TRIM21 WT or 3CS was co-transfected with HA-Ub in HEK293T cells. After 18 h, cells were subjected to Ni-nta pull-down and IB. (L) flag-TRIM21 WT or 3CS mutant was co-transfected with HA-Ub and his-SQSTM1 in HEK293T. After 18 h, cells were subjected to Ni-nta pull-down and IB. (M) flag-TRIM21 WT or 3CS was transfected to HEK293T. After 24 h, cells were treated with PA (500 μM) and subjected to flag IP and IB. (N) AML12 stably expressing vector, HA-TRIM21 WT, or HA-TRIM21[3CS] were treated with BSA or PA (500 μM). Triton X-100 (1%) soluble and insoluble fractions were subjected to IB for indicated proteins.

The above results strongly suggest that TRIM21 can be oxidized to form disulfide bonds within its N-terminal 8–251 amino acid residues. To determine the specific Cys I residues that are involved in disulfide bond formation, we performed liquid-chromatography (LC)-tandem mass spectrometry (MS/MS)-based disulfide bond identification. Flag-TRIM21 was expressed in HEK293T cells. Cell lysates were treated with H₂O₂ then immunoprecipitated with Flag-beads (Figure 4E). The slow-migration form was visualized by Coomassie blue stain then excised for mass spectrometry to identify the oxidized C residues (Figure 4E). Stepwise N-ethylamleimide (NEM) alkylation (represents free thiols), non-reducing SDS-PAGE separation, disulfide bond reduction (DTT) and iodoacetamide (IAM) alkylation (represents oxidized thiols) followed by in-gel trypsin digestion, and LC-MS/MS identified C92 with the highest ratio of oxidation (i.e., 20 fold, Figure 4F) upon H₂O₂ treatment, while several other C residues including C111 and C114 in the N-terminus of TRIM21 were at high-level of oxidation even before H₂O₂ treatment (Figure 4F). Interestingly, the C92, C111, and C114 residues are within the B-Box2 domain that contains a unique RING-like Zn²⁺-binding entity where a Zn ion is coordinated by C92, C111, C114, and one histidine (H95), which leads to decreased TRIM21 E3 ligase activity, as previously described [29,30]. We then generated a triple mutant of C92,111,114S (3CS) to test whether the three Cys residues are important for TRIM21 oxidation, oligomerization, and E3 ligase activity. Indeed, the TRIM21 3CS mutant virtually abolished TRIM21 oligomerization judging by the slow-migrating form in non-reducing gel and by co-IP of differentially tagged TRIM21 (Figure 4H, I). To investigate the effect of TRIM21 oligomerization on its E3 ligase activity, which can be reflected by TRIM21 autoubiquitination and ubiquitination of its substrates including SQSTM1, we co-transfected HEK293T cells with His-Ub and Flag-TRIM21 (Figure 4J), HA-Ub and His-TRIM21 (Figure 4K), or Flag-TRIM21, SQSTM1-His, and HA-Ub (Figure 4L). His affinity isolation showed that TRIM21 3CS mutant enhanced autoubiquitination and SQSTM1 ubiquitination as compared to TRIM21 WT, suggesting that TRIM21 3CS mutant has a stronger E3 activity (Figures 4J–L). Consistently, PA-induced TRIM21 oligomerization was abolished in the 3CS mutant (Figure 4M), and PA-induced SQSTM1 aggregation and phosphorylation was suppressed by WT TRIM21, which was further suppressed by the 3CS mutant (Figure 4N). Taken together, these results demonstrate that TRIM21 is oxidized to oligomerize via C92,111,114-mediated disulfide bond formation, which diminishes its E3 activity. The 3CS mutant fails to respond to oxidative stress and remains active in its E3 ligase activity that leads to SQSTM1 K7 ubiquitination and reduced aggregation.

Loss of TRIM21 protects against pa-induced lipotoxicity

Lipotoxicity has been characterized by increased ER stress, oxidative stress, activation of JNK, and ultimately cell death [6]. Previous studies reported that TRIM21 ubiquitinates SQSTM1 and negatively regulates SQSTM1 sequestration function, leading to impaired NFE2L2 antioxidant pathway [17,31,32]. Since the KEAP1-NFE2L2 antioxidant pathway plays a protective role against lipotoxicity and NAFLD [33,34], we assessed how TRIM21-mediated SQSTM1 ubiquitination affects lipotoxicity. Indeed, while PA induced profound death of Trim21^+/+ MEFs, cell death was significantly alleviated in trim21^−/− MEFs (Figure 5A–C). Consistently, PA-induced ROS was attenuated in trim21^−/− MEFs indicated by decreased DHE staining (Figure 5D), increased HMOX1 expression (Figure 5E), and reduced ER stress response and MAPK/JNK activation (Figure 5F). trim21^−/− MEFs reconstituted with TRIM21 WT and 3CS mutant led to increased cell death and decreased antioxidant response (Figure 5G). On the other hand, sqstm1^−/− MEFs were sensitized to PA-induced lipotoxicity (Figure 5H, I), which was accompanied by decreased HMOX1 expression and increased JNK activation (Figure 5J). Importantly, sqstm1^−/− MEFs reconstituted with SQSTM1 WT or S403E were more refractory to PA-induced ROS and cell death, whereas the SQSTM1^K7R and S403A mutants failed to do so (Figure 5K). Collectively, these results indicate that TRIM21 deficiency mediated protection against lipotoxicity is dependent on SQSTM1 phosphorylation and aggregation, which is essential for the activation of the noncanonical NFE2L2 antioxidant pathway.

Figure 5.

Loss of TRIM21 alleviates lipotoxicity. (A) Trim21^+/+ and trim21^−/− MEFs were treated with BSA or PA (500 μM) for 24 h. Cells were observed under phase-contrast microscope. (B) Trim21^+/+ and trim21^−/− MEFs were treated with BSA or PA at indicated concentration for 24 h. Cell viability was determined by trypan blue staining. Shown are means plus standard deviation (std) of three independent experiments. (C) Trim21^+/+ and trim21^−/− MEFs were treated with BSA or PA (500 μM) for 24 and 48 h. Cell viability was determined by trypan blue staining. Shown are means plus std of three independent experiments. (D) Trim21^+/+ and trim21^−/− MEFs were treated with BSA or PA (500 μM) for 20 h and subjected to dihydroethidium (DHE) staining. Quantitative analysis of relative DHE fluorescence from three experiments is shown on the right. (E) relative mRNA level of HO-1 was measured by quantitative RT-PCR. Shown are the means plus std of three independent experiments. (F) Trim21^+/+ and trim21^−/− MEFs were treated with BSA or PA (500 μM) for 24 h and subjected to IB with indicated proteins. (G) trim21^−/− MEFs reconstituted with vector, HA-TRIM21 WT or 3CS were treated with BSA or PA (500 μM) for 36 h. Cell death was determined by trypan blue staining, and oxidative stress was determined by DHE fluorescence. Shown are means plus std of triplicate experiments. (H) Sqstm1^+/+ and sqstm1^−/− MEFs were treated with BSA or PA (500 μM) for 24 h. Cells were observed under phase-contrast microscope. (I) Sqstm1^+/+ and sqstm1^−/− MEFs were treated with BSA or PA at indicated concentrations for 24 h. Cell death was determined by trypan blue staining. Shown are the means plus std of triplicate experiments. (J) Sqstm1^+/+ and sqstm1^−/− MEFs were treated with BSA or PA (500 μM) for 24 h and subjected to IB. (K) sqstm1^−/− MEFs reconstituted with vector, SQSTM1 WT or indicated SQSTM1 mutants were treated with BSA or PA. Quantitative analysis of relative DHE fluorescence. Shown are means plus std of triplicate experiments. (L) cell death was determined by trypan blue staining. Shown are means plus std of triplicate experiments. Statistical analysis was performed by Two-tailed student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; n.S.: non-significant.

Discussion

In this study, we report that PA induces SQSTM1 S403 phosphorylation, aggregation, sequestration of KEAP1, and activation of the NFE2L2 antioxidant pathway in cultured liver cells and MEFs. The S403 phosphorylation is mediated by TBK1, whose interaction with SQSTM1 is dependent on SQSTM1 dimerization via the K7-D69 hydrogen bond formation in the PB1 domain. The ubiquitin E3 ligase TRIM21 negatively regulates the antioxidant response by ubiquitinating K7 and abolishes the K7-D69 hydrogen bond formation. In response to PA-induced oxidative stress, TRIM21 is oxidized and oligomerized via disulfide bond formation in the N-terminal cysteine cluster (C92, 111, and 114). TRIM21 oxidation and oligomerization abolish its E3 activity hence relieve its inhibition of the SQSTM1 dimerization and aggregation, thereby promote cellular antioxidant capacity (Figure 6).

Figure 6.

Schematic illustration of working model. SQSTM1 dimerizes via the K7-D69 hydrogen bond in the PB1 domain, which facilitates its recruitment of TBK1 that phosphorylates SQSTM1 S403, subsequent SQSTM1 aggregation, and sequestration of KEAP1, the negative regulator of the antioxidant transcription factor NFE2L2. This pathway is negatively regulated by the ubiquitin E3 ligase TRIM21 that ubiquitinates SQSTM1 K7 and abolishes its dimerization and aggregation. In response to oxidative stress caused by FFA, TRIM21 is oxidized to oligomerize, which abolishes its E3 ligase activity. Inhibition of TRIM21 may help increase cellular antioxidant capacity in the context of ffa-induced lipotoxicity and metabolic disorders.

SQSTM1 plays a critical role in many biological processes including selective autophagy, mTORC1 activation, phase separation, and antioxidant response. The antioxidant function of SQSTM1 is mainly attributed to its sequestration of KEAP1 that frees up NFE2L2 and allows it to translocate into the nucleus to drive antioxidant gene expression [14,15]. The SQSTM1 sequestration function is critically regulated at multiple levels. SQSTM1 K7 has been reported to form a hydrogen bond with D69 [16], which is essential for SQSTM1 dimerization and sequestration function [35]. SQSTM1 phosphorylation at S349, S403, and S407 sites have also been implicated in its sequestration activity [18–23]. We show here that the K7-D69 hydrogen bond formation and PB1 domain dimerization is required for the recruitment of TBK1 and phosphorylation of S403 (Figure 1C, D), and the S403 phosphorylation is required for PA-induced SQSTM1 aggregation and KEAP1 sequestration (Figure 1D–F). We were unable to assess S407 phosphorylation in our study due to the unavailability of the p-S407 specific antibody. While SQSTM1 S349 is also phosphorylated upon PA treatment, whether it is required for PA-induced SQSTM1 aggregation and sequestration remains to be determined.

We have previously shown that TRIM21 functions as a negative regulator of antioxidant response to maintain a homeostatic level of ROS in response to various oncogenic and pharmacological stress [17,31,32]. Our currently study shows that TRIM21 acts as a sensor for oxidative stress. Using LC-MS/MS based disulfide bond identification, we found that C92 in the TRIM21 B-Box2 domain is highly sensitive to oxidation induced by H₂O₂, while C111 and C114 are already in a highly oxidized state at the basal condition that can be further enhanced upon H₂O₂ treatment (Figure 4E–G). The oxidized cysteines may become solvent exposed and can therefore cross-link with cysteine residues from other TRIM21 molecules to form oligomers, resulting in reduced E3 ligase activity. It has been previously shown that the B-Box2 domain, upon chelating with Zinc ion via the C92, H95, C111, and C114 residues, represses the E3 ligase activity of TRIM21 by blocking the E2 binding site [29,30]. Our hypothesis that oxidation of the three C residues and the formation of DTT-sensitive TRIM21 oligomers lead to decreased E3 activity provides a regulatory mechanism for TRIM21 E3 activity in response to redox change. Our finding that TRIM21 3CS mutant has higher E3 ligase activity is consistent with the higher activity predicted upon abolishing the B-Box2 domain [30].

In the liver, various dietary, chemical, and genetic mouse models have been described for studying MASH and SLD that are associated with abnormal lipid metabolism [36–38]. While elevated circulating FFAs are believed to play a major role in hepatic lipotoxicity and concomitant hepatocytes injury, the precise cause for the pathological conditions is complex and may be dependent on the composition, amount, and exposure duration of various lipid metabolites [39]. It has been reported that SQSTM1 aggregates play a detrimental role in a choline-deficient high fat diet model [26], yet mice with adenoviral hepatic SQSTM1 overexpression are protected from a fasting/high carbohydrate-refeed model [34,40]. Our findings suggest a protective role of the SQSTM1/KEAP1/NFE2L2 non-canonical antioxidant pathway in diseases that are associated with FFA lipotoxicity. It will be interesting to determine the effect of TRIM21, SQSTM1 and their various mutants in vivo, with careful consideration of the models. Nonetheless, the protective effect of TRIM21 ablation in PA-induced lipotoxicity, together with previous findings that trim21^−/− mice are protected from various oxidative tissue damages [17,31,32], suggest that TRIM21 May be a viable therapeutic target for FA-associated metabolic diseases.

Materials and methods

Reagents and antibodies

The following reagents were used: MG132 (MCE, HY-13259), chloroquine (MCE, HY-17589A), thapsigargin (MCE, HY-13433), palmitic acid (MCE, HY-N0830), CX-4945 (MCE, HY-50855), SBI-0206965 (MCE, HY-16966), 5Z–7-oxozeaenol (MCE, HY-12686), B×795(MCE, HY-10514), dihydroethidium (MCE, HY-D0079), BSA (Beyotime, ST025), Ni-NTA agarose (Qiagen 30,210), Trypan Blue staining solution (Meilunbio, MA0130), DAPI (Meilunbio, MA0128), H₂O₂ (Sigma-Aldrich 88,597).

The following antibodies were used: anti-SQSTM1/p62 (Novus, H00008878-M01; 1:4,000 for WB, 1:200 for IHC), anti-p-SQSTM1/p62 403 (Cell Signaling Technology [CST], 39786; 1:200 for IHC, 1:1,000 for WB), Anti-p-SQSTM1/p62 Ser349 (MBL, PM074; 1/1,000 for WB); anti-Flag (Sigma-Aldrich, F7425; 1:2,000 for WB), anti-HMOX1 (ABclonal, A1346, 1:1,000 for WB), anti-NQO1 (ABclonal, A19586; 1:200 for IHC, 1:1,000 for WB), anti-HA (CST, 3724; 1:1,000 for WB), anti-p-MAPK/JNK (CST, 9252; 1:1,000 for WB), anti-TRIM21 (Abcam, Ab207728; 1:1,000 for WB), anti-ubiquitin (Ptmbio, PTM-1107; 1:1,000 for WB). All the antibodies above are commercial antibodies.

Palmitic acid (PA) treatment

PA was dissolved in absolute ethanol at a stock concentration of 200 mM and then PA was diluted to 5 mM with 10% BSA (fatty acid-free, low-endotoxin; Sigma-Aldrich, A8806) in DPBS (Gibco 14,190–136). For the treatment of PA, the cells were incubated with PA (500 μM) in DMEM. BSA in DMEM was used as a control.

Cell culture and in vitro treatment

MEFs were generated as previously described [17], SNU449 (ATCC, CRL-2234), HEK293 (ATCC, CRL-1573), Huh7 (Glow Biologics, GBTC-099 H), and SK-HEP-1 (ATCC, HTB-52) cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco 11,965–092) supplemented with10% FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin. Mycoplasma testing has been periodically performed to avoid mycoplasma contamination. Cells were incubated at 37°C in a humidified atmosphere containing 5% CO₂. MEFs, SK-HEP-1, and SNU449 were stimulated with palmitic acid (dissolved in 10% fatty acid free BSA) at the indicated concentration. For the control group, cells were stimulated with fatty acid-free BSA (Beyotime, ST025). MG132 (MCE, HY-13259), thapsigargin (MCE, HY-13433), and chloroquine (MCE, HY-17589A) were applied at indicated concentration.

Transfection and infection for stable overexpression and knockdown

To establish overexpression cell lines, HEK293T seeded in 6-cm plates were transfected with VSVG, helper virus, and LPC retroviral expression constructs carrying TRIM21 or SQSTM1 mutants by Lipofectamine 2000 (ThermoFisher 11,668,019). After 24 h, the virus-containing cell culture supernatant was collected and replaced with complete medium, then filtered through 0.45-μm pore-size nylon filter (VWR 76,479–032). The target cells were cultured with viral medium containing 0.1% polybrene (Sigma-Aldrich, TR-1003-G) and subsequently selected by puromycin for 3 days.

Plasmids

The Flag-tagged TBK1 was generated by replacing the HA tag in pCDNA3.1-HA-TBK1 (Shhebio, P7179) with a Flag tag. Human TRIM21 expression constructs were previously described [17]. They were generated by using cDNA amplified by reverse transcriptase PCR and inserted into LPC vector. The primers used for TRIM21: 5’-CCCAAGCTTACCATGTACCCATACGATGTTCCAGATTACGCG ATGGCTTCAG CAGCACGC-3’ and 5’-CGGAATTCTCAATAGTCAGTGGATCCTTG-3’, for deletion of amino acid 8–251, 5’-CAGCACGCTTGGAAAGGAGT GAGTCCTGGAACC-3’ and 5’-CACTCCTTTCCAAGCGTGCTGCTGAAGCCAT-3’; for deletion of amino acid 252–476, 5’-GATAATTGTCCTGGGATCACAAGGATCCACTGAC-3’ and 5’-CCTTGTGATCCCAGGACAATTATCACCTCCTGCA-3’. Human SQSTM1 constructs were previously described [17]. SQSTM1 mutants were generated by standard site-directed mutagenesis using the following oligonucleotides: for SQSTM1^S403A, 5’-AGATGCTGGCCATGGGCTTCTCTGATGAAG-3’ and 5’-GAAGCCCATGGCCAGCATCTGGGAGAGGGACT CA-3’; for SQSTM1^S403E, 5’-CAGATGCTGGAGATGGGCTT CTCTGAT-3’ and 5’-AGAAGCCCATCTCCAGCATCTGGGAGAGGGAC TC-3’; For SQSTM1^K7R, 5’-CGCTCACCGTGAGAGCCTACC TTCTGGGCAAGGAG-3’ 5’-CAGAAGGTAGGCTCTCACGGTGAGCGACGC CAT-3’. For SQSTM1^S407A, 5’-CATGGGCTTCGCAGATGAAGG CGGCTGGCTCA-3’ 5’-TCATCTGCGAAGCCCATGGACAGCATCTGGGAGAG-3’. For SQSTM1^S403E,S407A, 5’-ATGCTGGAGATGGGCTTCG CAGATGAAGGCGGCTG-3’ 5’-TCTGCGAAGCCCATCTCCAGCATCTGGGAGA GGGACTCA-3’.

Cell viability assay

Trypan Blue staining was performed to measure the cell death. 0.1 ml of cells were mixed with 0.1 ml of Trypan Blue and incubated at room temperature for 5 min. Cells were counted under a phase-contrast light microscope.

Triton X-100 soluble/insoluble fractionation and immunoblotting

For immunoblotting, cells were lysed in RIPA buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.5 mM EGTA, 1% Triton X-100 [Sigma-Aldrich, T-9284], 140 mM NaCl, 1 mM PMSF. [Sigma-Aldrich, P7626], 1 mM Na₃VO₄ [Sigma-Aldrich 450,243], 0.1% sodium deoxycholate [Sigma-Aldrich 106,504]) supplemented with 1% SDS. After sonication on ice, samples were collected and quantified by a BCA kit (Meilunbio, MA0082). For the soluble and insoluble fractions, cells were lysed in RIPA buffer supplemented with 1% Triton X-100 and spun down at 12,000 g for 10 min. The supernatant was collected as the soluble fraction. The insoluble fractions were dissolved in RIPA buffer containing 1% SDS. A total of 10–30 ug protein were separated by SDS-PAGE and transferred to nitrocellulose membrane. The membranes were blocked with 5% skim milk in TBST (10 mM Tris, pH 7.4, 150 mM NaCl, 0.05% Tween-20 [Sigma-Aldrich, P1379]), incubated with indicated antibodies overnight at 4°C and incubated with secondary antibodies.

Quantitative real-time PCR

Total RNA was extracted with a kit (PureLinkTM RNA Mini Kit, 12183018A) and RNA concentration was measured by a spectrophotometry. Reverse transcription was carried out with 1 μg of total RNA using the HiScript II Q RT SuperMix for qPCR Kit (Vazyme, R223–01) in accordance with the manufacturer’s instructions. Synthesized cDNA was used for quantitative real-time PCR with the Hieff qPCR SYBR Green Master Mix (YEASEN, 11203ES08) in accordance with the manufacturer’s instructions.

Coimmunoprecipitation

Cells were scrapped off the plates, spun down, washed with cold PBS and lysed in IP lysis buffer (20 mM Tris, pH 7.5, 100 mM NaCl, 1% Triton X-100, 1 mM EDTA, 200 mM PMSF) supplemented with protease inhibitor cocktail (G-Biosciences, 786–326) on ice for 30 min. Cell lysates were cleared by centrifugation at 4°C to collect the supernatant. The lysates were incubated with primary antibodies and protein A/G magnetic beads (Dynabeads Protein A/G; ThermoFisher, 10015D) overnight at 4°C with agitation. The complexes were precipitated with magnetic stand, and washed twice with IP lysis buffer supplemented with 200 mM NaCl in a rotator for 5 min, washed twice with IP lysis buffer supplemented with 100 mM NaCl for 5 min, then once with IP lysis buffer for 5 min, and resuspended the complexes with IP lysis buffer, boiled in SDS sample buffer at 95°C for 5 min.

His affinity isolation

HEK293T cells were transfected with His-tagged constructs with Lipofectamine 2000. Two days post transfection, cells were harvested and one tenth of cells were lysed in RIPA buffer as the input. The remaining cells were dissolved in Buffer A (6 M guanidine-HCl, 0.1 M Na₂HPO₄/NaH₂PO₄, 10 mM imidazole, pH 8.0) with sonication. Each sample was added with Ni-NTA agarose equilibrated with Buffer A and incubated overnight at 4°C with agitation. Then the agarose beads were precipitated by centrifugation and washed with Buffer A, Buffer B (10 mM Tris-Cl, pH 8.0, 8 M urea, 0.1 M NaH₂PO₄) and buffer C (1:4 diluted Buffer B supplied with 25 mM imidazole). The precipitates were resuspended in RIPA buffer, mixed with SDS sample buffer, boiled at 95°C for 5 min.

Immunofluorescence

Cells were fixed in 4% paraformaldehyde for 15 min at room temperature and then washed with PBS. The cells were permeabilized with 0.2% Triton X-100 for 10 min and washed with PBS. Cells were blocked in 5% nonfat milk in PBS for 1 h at room temperature. After a brief wash with PBS, cells were incubated with primary antibodies in PBS overnight at 4°C. Cells were then washed with PBS for three times with PBS, followed by incubation with fluorophore-conjugated secondary antibodies in PBS for 1 h at room temperature. Then the cells were washed with PBS and incubated with DAPI for 5 min. The cells were mounted with a mounting medium Fluoromount-G (Southern Biotechnology Associates, 0100–01). The slides were observed and imaged using a Keyence microscope.

Mass spectrometry

To map the C residues that are involved in thiol oxidation and oligomer formation, we used LC-MS/MS-based redox thiol identification. Flag-TRIM21 was expressed in HEK293T cells. Cell lysates were treated with either H₂O₂ or buffer, then immunoprecipitated with Flag-beads. First, N-ethylamleimide (NEM; Sigma-Aldrich, E3879) was used to alkylate free thiols in TRIM31. After non-reducing SDS-PAGE separation to remove an excess of NEM, the high molecular weight oligomer form of TRIM21 was visualized by Coomassie Brilliant Blue stain then excised for in-gel trypsin digestion [41]. In brief, 5 mM dithiothreitol (DTT) was used to reduce the oxidized thiols, followed by 25 mM iodoacetamide (IAM; Sigma-Aldrich, I1149) alkylation in the dark at room temperature. Then TRIM21 was in-gel trypsin digested at 37°C overnight. The resulting peptides were analyzed on a Orbitrap Lumos Tribrid Mass Spectrometer coupled with an Ultimate 3000 RSLCnano LC system (Thermo Scientific). The peptides were resuspended in Solvent A (2% acetonitrile (ACN) in 0.1% formic acid (FA)) and separated on an Acclaim PepMap C₁₈ column (75 µm × 50 cm, 2 µm, 100 Å), using a 3-h binary gradient from 2–95% of Solvent B (85% ACN in 0.1% FA), at a flow rate of 300 nL/min. The eluted peptides were directly introduced into the MS system for data-dependent MS/MS analysis in the positive ion mode.

The raw LC-MS/MS data were submitted for database search using the SEQUEST search engine on the Proteome Discoverer (V2.4) platform. The MS/MS spectra were searched against a UniProt human FASTA database. Trypsin was selected as enzyme with two miss cleavages. Methionine oxidation, cysteine carbamidomethyl modification and NEM modification were selected as variable modifications. The protein false discovery rate (FDR) was estimated using the decoy databases containing reversed sequences of the original proteins. Proteins were identified with both protein and peptide FDRs at or less than 1%.

Both the NEM- and IAM-modified peptide were identified and quantified by LC-MS/MS based on Label Free Quantification/LFQ. The percentage of the Oxidation for each peptide was calculated from the intensity of oxidized IAM-peptide relative to the sum of the intensity of the reduced (NEM-) and oxidized (IAM-) peptides.

TRIM21 B-Box2 zinc finger structural illustration

The dimeric structure of TRIM21 was downloaded from the protein data bank (PDB) with the code 5OLM. The structure figure was generated using Pymol.

Densitometry quantification of western bands

The WB bands were quantified by Image Studio by subtracting the intensity of the background. The relative densitometry ratio of p-SQSTM1:SQSTM1, or that of insoluble to soluble SQSTM1, respectively, was calculated and normalized to that of the control sample (ratio set to 1.00). All westerns were run at least three times and the densitometry ratios shown are the mean of these multiple repeats.

Statistical analysis

Data were graphed and analyzed using two-tailed student’s t-test comparisons between two groups or one-way analysis of variance (ANOVA) with Tukey’s test for comparing more than two groups using GraphPad Prism software to determine statistical significance. Results were considered significant when p < 0.05.

Funding Statement

The study was supported by NIH [R01CA129536, R01AG072895, R37AA020518, and R01DK134737], Shanghai Pujiang Program [19PJ1401900], National Natural Science Foundation [31971161, 31970714, 82100293], Shenzhen Science and Technology Program [JCYJ20190807153203560], and Science and Technology Program of Jinhua Science and Technology Bureau [2023-3-037]. LC/MS/MS data were obtained with the instruments funded by NIH [NS046593 and 1S10OD025047].

Disclosure statement

No potential conflict of interest was reported by the author(s).