Inhibition of HSP90β Improves Lipid Disorders by Promoting Mature SREBPs Degradation via the Ubiquitin-proteasome System

Rationale: Heat shock protein 9 (HSP90) are a family of the most highly expressed cellular proteins and attractive drug targets against cancer, neurodegeneration diseases, etc. HSP90 proteins have also been suggested to be linked to lipid metabolism. However, the specific function of HSP90 paralogs, as well as the underlying molecular cascades remains largely unknown. This study aims to unravel the paralog-specific role of HSP90 in lipid metabolism and try to discover paralog-specific HSP90 inhibitors. Methods: In non-alcohol fatty liver disease (NAFLD) patients, as well as in diet induced obese (DIO) mice, expression of HSP90 paralogs were analyzed by immunohistochemistry and western blot. In hepatocytes and in DIO mice, HSP90 proteins were knockdown by siRNAs/shRNAs, metabolic parameters, as well as downstream signaling were then investigated. By virtue screening, corylin was found to bind specifically to HSP90β. Using photo-affinity labeling and mass spectrum, corylin binding proteins were identified. After oral administration of corylin, its lipid lowering effects in different metabolic disease mice models were evaluated. Results: We showed that hepatic HSP90β, rather than HSP90α, was overexpressed in NAFLD patients and obese mice. Hepatic HSP90β was also clinical relevant to serum lipid level. Depletion of HSP90β promoted mature sterol regulatory element-binding proteins (mSREBPs) degradation through Akt-GSK3β-FBW7 pathway, thereby dramatically decreased the content of neutral lipids and cholesterol. We discovered an HSP90β-selective inhibitor (corylin) that only bound to its middle domain. We found that corylin treatment partially suppressed Akt activity only at Thr308 site and specifically promoted mSREBPs ubiquitination and proteasomal degradation. Corylin treatment significantly reduced lipid content in both liver cell lines and human primary hepatocytes. In animal studies, we showed that corylin ameliorated obesity-induced fatty liver disease, type 2 diabetes and atherosclerosis. Principle conclusions: HSP90β plays a parolog-specific role in regulating lipid homeostasis. Compound that selectively inhibits HSP90β could be useful in the clinic for the treatment for metabolic diseases.


Viability assay
Viability assay was detected by MTT assay. Briefly, HL-7702 and HH cells were seeded at the density of 2×10 4 cells/well in 96-well plate and maintained under 5% CO 2 at 37 °C. After the treatment, cells were treated with corylin as indicated. After 24 h, 10 μl of MTT (5 mg/ml) was added and incubated for 4 h. The cytotoxicity of corylin was determined by microplate reader (Multiskan FC).

Measurement of de novo fatty acid and cholesterol synthesis
After the cells were treated, acetic acid sodium salt 1-14 C (20 μCi/100 mm 2 dish) was directly added and incubated for additional 2 h. The cells were washed and dissolved with 0.1 N NaOH, and cell suspensions were autoclaved for alkaline saponification.
Then cholesterol were extracted in petroleum ether and evaporated to dryness with N 2 .
After addition of 12 N HCl, fatty acids were extracted in petroleum ether and evaporated to dryness. The lipids were resolved by thin-layer chromatography (Silica gel 60, Merck). The radioactive products were identified by comparison with unlabeled standards and visualized with iodine vapor.

qRT-PCR
Total RNA was extracted from HL-7702, HH cells or mice livers using Trizol reagent (Invitrogen) according to the manufacturer's instructions. RNA concentrations were equalized and converted to cDNA using a kit (Hiscript II reverse transcriptase, Vazyme). Gene expression was measured by qRT-PCR (Roche) using SYBR-green (Hiscript II reverse transcriptase, Vazyme). Expression was normalized to GAPDH.
The sequences of primers used in the experiments were listed in Supplementary   Table 4.

Western blot analysis
For whole cell lysate, cells were harvested and suspended in 150 μl of RIPA buffer (Beyotime, China) containing protease inhibitors (10 g/ml leupeptin, 5 g/ml pepstatin A, 25 g/ml ALLN, 1 mM PMSF). Protein concentration was determined according to BCA (Beyotime, China), then the extracts were mixed with 5×SDS loading buffer (Beyotime, China). After the mixtures were boiled at 95 °C for 10 min, they were subjected to SDS-PAGE, transferred to nitrocellulose membranes, and subjected to immunoblot analysis.
Nuclear extract was extracted by NE-PER nuclear and cytoplasmic extraction kit (Thermo). The pellets from 100-mm dishes of HL-7702 cells were harvested by trypsinization and collected by centrifugation at 500 g for 5 min and washed with PBS.
Cells were transferred to 1.5 ml microcentrifuge tubes followed by centrifugation at 500 g for 3 min. After that, the supernatant layer was carefully removed by using pipette and discarded. The remaining cell pellet at the bottom was allowed to dry.
Then cells were suspended in 500 μl of buffer A (10 mM HEPES, KOH (pH 7.6), 1.5 mM MgCl 2 , 10 mM KCl, 5 mM EDTA, 5 mM EGTA, 250 mM Sucrose) containing protease inhibitors as described above, and then the tubes were vortex at the highest speed for 15 s to suspend the cell pellet and were put on ice and incubated for 10 min.
Followed by the addition of ice-cold CER-II to the tubes and vortex again for 5 s and incubated again on ice for 1 min. Then, the tubes were vortex again and centrifuged at 13,000 g for 5 min. After centrifugation, the upper supernatant layer containing cytoplasmic extract was transferred to pre-chilled tubes. Cold NER was added to the remaining insoluble pellet, which contained nuclei. After NER was added, the tubes were vortex at the highest speed for 5 s and kept on ice for 10 min, this was repeated for three times. When the incubation finished, the tubes were centrifuged at 13,000 g for 10 min. The supernatant layer containing the nuclear extract was transferred to pre-chilled tubes. The next operation is the same as above.
Protein A/G plus agarose beads (Santa Cruz) were added at 4 °C for another 2 h. The immunoprecipitation beads were washed with cold PBS for five times, followed by western blotting analysis.

Knockout of HSP90β by CRISPR-Cas9
Targeting sequences were designed at CRISPR direct (

Cholesterol and TG measurement
For measurement of intracellular TC and TG, the cells were cultured in six-well plates and collected in 1 ml PBS. 100 μl of the total cell suspension were transferred to a new tube and centrifuged at 1,000 g for 5 min at 4 °C. Then this portion of the cells were lysed in lysis buffer (RIPA lysis buffer) and used for protein quantification. The remaining cell suspension was used for lipid extraction. After centrifugation at 1000 g for 5 min at 4 °C, the collected cells were mixed with 1ml of chloroform/methanol (2:1, v/v) adequately on a shaker for 3 h at 24 °C. Then 500 μl NaCl (0.1 M) was added into each reaction tube and mixed thoroughly, followed by centrifugation at 3700 rpm for 10 min. the lower organic phase was transferred and evaporated to dryness. The residual liquid was re-suspended in 50 μl of 1% Triton-X 100 in absolute ethanol, and the concentrations of TC or TG were measured using the TC or TG determination kit according to the manufacturer's instructions, respectively (Shanghai, China). For measurement of liver TC and TG, 40-50 mg of liver tissue was homogenized in 0.5 ml PBS. About 5 μl of the total homogenates were used for protein quantification. About 0.4 ml homogenates were mixed with 1.6 ml of chloroform/methanol (2:1, v/v) adequately for lipid extraction. The following experimental procedures were identical with measurement of hepatic TC and TG.

Virtual screening
The crystal structures of HSP90β (3PRY) and HSP90α

Akt kinase activity measurement
HL-7702 cells were treated with corylin or 17-AAG. After that, the cells were lysis and obtained the protein buffer. The Akt kinase activity was detected by AKT ELISA Kit (ZciBio). The experimental procedure is carried out according to the Kit instructions.

In silico molecular docking research
To analyze the binding affinities of corylin to HSP90β and the possible binding sites, an in silico protein-ligand docking software AutoDock 4.2 program was applied. The docking steps were performed as the follows: Crystal structure file of HSP90β (3PRY) was downloaded from the RCSB protein data bank; (2)  (DE3) strain as C-terminal His-6-tagged fusion proteins by using the pET28a expression system (Novagen). The C-terminal tagged (His) 6 fusion proteins were purified by Ni 2+ -agarose affinity chromatography.

ATPase activity of HSP90β measurement
We used purified HSP90α and HSP90β (wild type or triple mutation) proteins to detect their chaperone activity in vitro by QuantiChrom™ ATPase Assay Kit (BioAssay Systems). The experimental procedure is carried out according to the Kit instructions.

Microscale thermophoresis analysis
Corylin, 17-AAG, AP-III-a4, compound A or neobavaisoflavone were titrated in Measurements were performed at room temperature and standard deviation was calculated from three independent experiments.

Photo-affinity labeling and pull down experiment
Photo-affinity-linker-coated (PALC) agarose beads and corylin-immobilized beads were prepared according to previous reported method. The protein extracts of HepG2 (2mg in 1ml NETN lysis buffer, 0.1% NP-40, 0.5 mM EDTA, 20 mM Tris-HCl, 150 mM NaCl 2 , protease inhibitor cocktail, pH=8.0) were first pre-incubated in the presence (3 replications) or absence (3 replications) of corylin (finally 0.5 μM) at 4 °C overnight, and then 12 μl of prewashed corylin-immobilized beads were added to each sample, and incubated at 4 °C for an additional 4 h. Subsequently, the beads were washed with lysis buffer for 3 times and eluted with 1×SDS-loading buffer (20 μl), then boiled for 5 min at 95 °C. The eluted proteins were analyzed according to digestion in gel method, and further analyzed by an EASY-nLC 1000 nano-flow LC instrument coupled to a Q Exactive quadrupole-orbitrap mass spectrometer (Thermo Fisher Scientific).
Subsequently, the beads were washed with lysis buffer for 3 times, and one group of corylin beads and blank beads were eluted with 100 μM corylin (in 40 μl PBS) and the other group of corylin beads were eluted with 40 μl PBS at 37 °C for 10 min. The eluted proteins were analyzed according to digestion in gel method, and further analyzed by an EASY-nLC 1000 nanoflow LC instrument coupled to a Q Exactive quadrupole-orbitrap mass spectrometer (Thermo Fisher Scientific). The screening principle is that statistical analyses were done with student's t-test vs blank (p*<0.05), the ratio of corylin-1/blank is greater than 1.2, and the ratio of coyrlin-1/corylin-2 is greater than 4.5. (blank: Intensity of the proteins, eluted by 100 μM corylin (PBS buffer) from blank beads; corylin_1: Intensity of the proteins, eluted by 100 μM corylin (PBS buffer) from corylin linked beads; corylin_2: Intensity of the proteins, eluted by PBS buffer from corylin linked beads).

Metabolic measurements
After receiving different treatments for 6 weeks, mice from each group were acclimated in a comprehensive lab animal monitoring system (Columbus Instruments, Columbus, OH) for 24 h according to the instructions of the manufacturer. After mice adapted to the metabolic chamber, volume of O 2 consumption and CO 2 production were continuously recorded over a 24 h period. RQ equals volumes of CO 2 released/volumes of O 2 consumed.

Adenovirus-mediated RNAi in mice liver
The adenoviral expression kit from Life Technologies was utilized to construct the adenovirus-producing plasmids containing a gene of shRNA HSP90β or LacZ. The adenovirus vectors were digested with Pac I, then transfected the 293A producer cell line in a 6-well-plate. The media was placed with DMEM containing 10% FBS and 1% penicillin/streptomycin the next day. The cells transferred to 10 cm tissue culture dishes after the transfection for 24 h. We replaced the culture media with fresh media every 2-3 days until cytopathic effect (CPE) was observed. We collected the cells when 80% CPE were observed and harvested adenovirus by repeatedly freezing at -80 °C and thawing at 37 °C for 4 times. We centrifuged cell lysates at 2,000g for 30 min at 25 °C and stored the supernatant containing adenovirus particles at -80 °C.
The adenoviruses were packaged in HEK293 cells and purified with CsCl ultracentrifugation. The viruses were tittered and administrated via caudal vein injection (5×10 9 pfu viruses per mouse).

Fecal cholesterol and TG measurements
After receiving different treatments for 6 weeks, mice were kept into the metabolic chambers for 24 h to collect the feces which were lyophilized and ground up. About 250 mg mashed feces were extracted with 4 ml of methanol: chloroform (1:2, v/v) twice. The supernatants were pooled. Exactly 100 μl was removed and evaporated to dryness. TG was measured with determination kit. Then other 4 ml mixture of 5 N KOH: ethanol (10: 90, v/v) were added and heated at 70 °C for 1 h. After cooling to room temperature, 2 ml ultrapure water was added and saturated with sodium chloride, followed by solvent extraction with 3 ml hexane twice. The extracts were dried and re-dissolved with 50 μl hexane contained 50 μl 5-a-cholestane (1 mg/ml) was added.
Then GC-MS analyses were performed with an Agilent 7890B gas chromatograph interfaced to an Agilent 5977A mass-selective detector and equipped with HP-5ms Ultra Inert (30 m×250 mm×0.25 μm) column (Agilent Technologies, USA). The temperatures of the injector, interface, and ion source were 280 °C, 280 °C, and 230 °C, respectively. Helium was used as the carrier gas at a flow rate of 0.7 ml/min in constant flow mode.

The weight of fat analysis by NMR
To determine the fat content of animals, after six-week treatment, the mice were scanned with the minispec TD-NMR designed for experimental animals (Bruker, Germany), the fat content was calculated according to the measurement between the solid and liquid parts of the sample.

Serum and liver lipid determination
Serum TC and TG levels were measured according to the manufacturer's instructions (Kehua, China), and LDL-C and HDL-C levels were determined by the kits (Njjcbio, China). Liver tissues were homogenized and centrifuged. Supernatants were collected, and TC and TG were determined.

Glucose tolerance and insulin tolerance tests
Glucose tolerance tests (GTT) and insulin tolerance tests (ITT) were performed on mice fasted overnight with free access to water. Mice were injected with 0.75 U/kg insulin (Sigma) by i. p. or gavaged with 2 g/kg glucose (Sigma) by i. g. Glucose levels were measured from tail blood at 15, 30, 60, or 120 min after the injection. All animals were sacrificed 3 days after glucose tolerance or insulin tolerance tests, and blood and liver were harvested. Area under the curve (AUC) was calculated to quantify the GTT and ITT results.

Histological analysis of liver, adipose and aortas
Liver, WAT, BAT and aortas were fixed in 4% paraformaldehyde at 4 °C overnight and embedded in paraffin wax. Paraffin sections (5 μm) were cut and mounted on glass slides for H&E staining. Cryosections of livers were stained by oil red O and counterstained with hematoxylin to visualize the lipid droplets.

Immunohistochemistry
Immunohistochemistry was carried out using 3 µm thick sections fixed in 4% paraformaldehyde. After deparaffinization, rehydration and antigen retrieval, sections were incubated in blocking buffer containing 10% normal goat serum in PBS.
Sections were incubated with described antibodies followed by washing and incubated with HRP-tagged goat anti-rabbit secondary antibody. Samples were subsequently rinsed in wash buffer and incubated in diaminobenzidine (Sigma, St. Louis, MO) for 5 minutes and counterstained in hematoxylin. Tissue slides were scored in a blinded fashion. No staining was observed with negative control rabbit anti-IgG antibody. The images were measured blindly by one observer using Image-Pro Plus (Media Cybernetics, Silver Spring, USA).

Analysis of atherosclerotic lesions
To quantify atherosclerosis along the entire aorta, the aortic tree was dissected out and the lesions were stained with oil red O for 6 min, destained with 80% ethanol for 3 min. Sudanophilic lesions were assessed by computer-assisted image analysis.

Supplementary Table 3. The proteins pulled down by corylin in HepG2 cells
See the Excel Table (Supplementary Table 3) for details.