1. OxLDL stimulation reduced TFEB SUMOylation
TFEB has been reported to undergo SUMOylation and this modification has a key regulatory role in its transcriptional activity [23]. We confirmed this finding by co-transfecting TFEB and SUMO1 plasmids in 293T cells and observed that TFEB is indeed a SUMOylated protein (Figure-1A). Simultaneously, we identified that the SUMOylation site of human TFEB was lysine 361, since TFEB SUMOylation band was completely disappeared when lysine 361 was mutated to arginine (K361R) (Figure-1A). Although SUMOylation has proved a very important regulatory effect on the transcriptional activity of TFEB, the specific molecular mechanism under different stimulation conditions is not completely clear, including during macrophage foam cells formation.
To detect the SUMOylation of endogenous TFEB, we performed immune co-precipitation assay in bone marrow-derived macrophages (BMDMs). A band at approximately 100 kDa reflecting endogenous TFEB SUMOylation was detected (Figure-1B). WT and TFEB-KR BMDMs were confirmed by F4/80 and CD11b co-staining [25] (Figure-S1). To clarify the role of TFEB SUMOylation in AS, BMDMs were treated with OxLDL (50 mg/ml) for 24 hrs. The SUMOylation band of TFEB was significantly weakened after OxLDL stimulation (Figure-1C). These results suggested that TFEB SUMOylation was involved in the process of AS. Through protein sequence alignment, we found that TFEB SUMOylation site is highly conserved across various species (Figure-S2A). To further study the role of TFEB SUMOylation in the development of AS in vivo, we mapped the TFEB SUMOylation site in mice (Figure-1D) and constructed the TFEB SUMOylation deficiency mice (K346R) by using CRISPR/Cas9 base editing technology. We named it TFEB-KR mice (Figure-1E).
2. TFEB-KR: Ldlr−/− mice exhibited milder atherosclerotic lesions with high cholesterol diet
Since TFEB-KR mice had whole body TFEB SUMOylation deficiency, we performed bone marrow transplantation experiments in these mice prior to feeding them with a high cholesterol diet for 12 weeks (Figure-S3). The results revealed that TFEB-KR: Ldlr−/− mice exhibited less atherosclerotic plaques in the aorta when compared with the control mice (Figure-2A). The ratio of atherosclerotic plaques area to total aortic area in WT: Ldlr−/− mice was 20.1%, 19.3%, 26.0%, and 19.2% and that in TFEB-KR: Ldlr−/− mice was 12.3%, 16.3%, 16.2%, and 14.1%. In addition, we observed that the average aorta lesion in TFEB-KR: Ldlr−/− mice was approximately 30.3% less (Figure-2B) than that in the control mice. Furthermore, TFEB-KR: Ldlr−/− mice had thinner aortic sinus plaques and smaller atherosclerotic plaques (Figure-2C). MAC-3 is a glycoprotein that is expressed on the surface of monocytes and macrophages[26, 27]. Lysosomal associated membrane protein 1 (LAMP1) is a type 1 transmembrane protein that exists on the lysosomal membrane and can indirectly reflect the activity of lysosomes [28, 29]. TFEB-KR: Ldlr−/− mice exhibited stronger LAMP1 (red) but not MAC-3 (green) expression when compared with the control mice (Figure-2D/E). These results demonstrated that the number of macrophages infiltrating into atherosclerotic plaques was comparable between the two groups. However, macrophages in atherosclerotic plaques from TFEB-KR: Ldlr−/− mice exhibited higher lysosomal activity. Further, we measured the levels of triglyceride (TG), total cholesterol (TC), high-density lipoprotein (HDL), and low-density lipoprotein (LDL) in the serum collected from the two groups. The content of TG, TC, and LDL but not of HDL was slightly higher in TFEB-KR: Ldlr−/− mice (Figure- 2H). This may be due to the higher lysosomal activity of macrophages in the plaques from TFEB-KR: Ldlr−/− mice, which can decompose and metabolize more lipids accumulated under the vascular endothelium. Further, we monitored the body weight between the two groups. No significant difference existed in body weight between the two groups even in the presence of western diet (Figure-2F). Moreover, the content of IL-1β, interleukin-4 (IL-4), and interleukin-6 (IL-6) was evaluated in the two groups from the AS model mice. TFEB-KR: Ldlr−/− mice exhibited lower IL-1β and higher IL-4 levels even though without significant difference. However, no difference was observed in terms of IL-6 secretion level (Figure-2G). These results indicated that TFEB-KR: Ldlr−/− mice was more resistant to inflammation. Collectively, we identified that TFEB SUMOylation plays an important role in the process of AS, and TFEB-KR: Ldlr−/− mice had stronger ability to resist AS.
3. TFEB SUMOylation inhibited lysosomal biogenesis and activity
Lyso-Tracker is a fluorescent probe labeled with weak alkalinity; therefore, it can selectively stain acidic lysosomes[30, 31]. Upon OxLDL stimulation, TFEB-KR BMDMs exhibited stronger Lyso-Tracker and LAMP1 staining signals and formed larger vacuoles when compared with WT BMDMs (Figure-3A/B). Further, quantitative analysis revealed higher MFI of Lyso-Tracker in TFEB-KR BMDMs after OxLDL treatment (Figure-3C/D). Real-time quantitative PCR detection revealed that genes associated with lysosomal biogenesis (Clcn7, Lamp1, Ctsb, and Ctsd) and autophagy (Atg4b, Atg4d, Atg5 and Atg9a) were significantly upregulated in TFEB-KR BMDMs after OxLDL stimulation [32–36] (Figure-3E). The activity of cathepsin B (CTSB) can reflect the function of lysosome[37]. We observed that CTSB activity was approximately 26.1% higher in TFEB-KR BMDMs after OxLDL treatment (Figure-3G). LC3 is a class of microtubule associated protein, which plays a very important role in autophagy[38, 39]. We observed that the expression of LAMP1 and LC3 II was drastically increased in TFEB-KR BMDMs after OxLDL treatment (Figure-3F, Figure-S2B). These results indicated that TFEB-KR macrophages had higher lysosomal activity than wild type macrophages after OxLDL treatment.
4. TFEB-KR macrophages exhibited less lipid deposition upon OxLDL stimulation
BODIPY is used as a dye for natural oils and fats and a tracer for oils and other non-polar oils[40]. Oil red O and BODIPY staining revealed that lipid deposition was more pronounced in the wild type BMDMs (Figure-4A/B). We quantitatively analyzed the results of BODIPY staining using flow cytometry. The results revealed that the lipid content was approximately 31.6% less in TFEB-KR BMDMs (Figure-4C/D) than that in the WT group. Further, we measured the intracellular cholesterol content and observed that TFEB-KR BMDMs had less cholesterol (Figure-4E) than the WT group. These results suggested that TFEB-KR BMDMs had less lipid deposition after OxLDL stimulation.
In fact, the formation of macrophage foam cells is a way of metabolizing cholesterol by macrophages. The lipid metabolism process is very complex. It can be roughly divided into three steps: uptake, catabolism, and efflux of cholesterol[41–43]. Cholesterol uptake and efflux are mediated by a series of receptors. The major receptors related to cholesterol uptake include cluster of differentiation 36 (CD36), scavenger receptor class B member 1 (Scarb1), and macrophages scavenger receptor 1 (Msr1) [44–46]. The main receptor of cholesterol efflux is ATP binding cassette subfamily G member 1 (Abcg1) [47]. Upon OxLDL stimulation at 16 hrs, the mRNA levels of the above receptors were elevated in both WT and TFEB-KR groups and there was no difference between the two groups (Figure-4F). Therefore, TFEB SUMOylation did not affect the expression of receptors related to cholesterol uptake and efflux in macrophages. Further, the cholesterol uptake capacity was assessed between the WT and TFEB-KR BMDMs, which was observed to be comparable in the two groups (Figure-4G). Finally, we observed that the cholesterol efflux of TFEB-KR BMDMs was significantly increased upon OxLDL stimulation at 6 hrs (Figure. 4H). These results suggested that because of stronger lysosomal function, less lipid deposition was present in TFEB-KR macrophages.
5. TFEB SUMOylation inhibited the transcriptional activity of TFEB by preventing its binding activity
Chloroquine (CQ) can be used as a lysosomal inhibitor. Alkaline CQ interacts with acidic lysosomes and inhibits the activity of lysosomes[48]. It has been reported that TFEB translocated into the cell nuclei with its transcriptional activity enhanced upon CQ stimulation[15]. We overexpressed WT and TFEB-KR plasmids in 293T cells. The cells were further stimulated with CQ for 15 hrs, and as a result, both WT and TFEB-KR proteins were able to translocate into the nuclei (Figure-5A), and there was no difference in the ratio of TFEB translocation into the nuclei between the two groups (Figure-5B). By analyzing the structure of TFEB, we observed that TFEB SUMOylation site was close to the DNA binding domain [49] (Figure-5C). Therefore, we used chromatin immuneprecipitation to test whether the binding ability of TFEB to its target genes was different between the two groups. Mucolipin-1 (Mcoln-1), also known as transient receptor potential channel muolipin-1 (Trpml1), is a classical target for TFEB and plays an important role in maintaining the biological function of lysosome [50]. We evaluated Mcoln-1 level in both groups with or without OxLDL stimulation. The binding ability of TFEB to Mcoln-1 promoter was significantly increased in TFEB-KR macrophages upon stimulation (Figure-5D). These results indicated that TFEB SUMOylation inhibited its transcriptional activity by inhibiting its binding to its target genes.
6. TFEB SUMOylation inhibited the binding between TFEB and FACT complex to attenuate TFEB transcriptional activity
In order to further explore how TFEB SUMOylation affecting its transcriptional activity, we used mass spectrometry to find the key regulators affected by TFEB SUMOylation in this process. We overexpressed wild-type and mutant Flag-TFEB in 293T cells, and then enriched TFEB protein by immunoprecipitation (Figure-6A). Through mass spectrometry analysis, we found that TFEB SUMOylation affected its interaction with other proteins, among which the interaction with facilitates chromatin transcription (FACT) complex was the most affected (Figure-6B). FACT complex, a heterodimeric histone chaperone composed of structure specific recognition protein 1 (SSRP1) and SPT16 homolog, facilitates chromatin remodeling subunit (SUPT16H), mediates the disassembly and assembly of nucleosomes, thus improving the efficiency of transcription. A recent study showed that TFEB can regulate its transcriptional activity by binding to FACT, but the molecular mechanism is not fully understood[51]. To verify the influence of TFEB SUMOylation on the interaction between TFEB and FACT complex observed in MS analysis, we used sodium arsenite (NaAsO2) to stimulate 293T cells with overexpressed wild-type and mutant Flag-TFEB, respectively. NaAsO2 is a kind of oxidative stress stimulator known to induce efficient translocation of TFEB from the cytosol to the nucleus. Our results showed that TFEB SUMOylation did inhibit the interaction between TFEB and FACT complex, and this inhibition effect was more obvious after TFEB was activated (Figure-6C). Finally, we used luciferase reporter gene experiment to verify that TFEB SUMOylation affected its transcriptional activity through the FACT complex during the process of macrophage foam cells formation. As previously described, LAMP1 is a target gene for TFEB and then we constructed a related reporter gene plasmid (Figure-6D). We used OxLDL to stimulate macrophages to imitate the process of foam cells formation in RAW264.7 cells. Our results showed that TFEB SUMOylation inhibited the activity of LAMP1 promoter during the process of macrophage foam cells formation, but this inhibition completely disappeared when the FACT complex and TFEB were co-transfected (Figure-6E). The above results showed that TFEB SUMOylation attenuated its transcriptional activity by inhibiting its interaction with FACT complex during the process of macrophage foam cells formation.