RESULTS
Gossypol Acetate (GA) was identified as a VCP inhibitor which binds to interface between N and D1 domain
To identify novel modulator of VCP, we established a high-throughput screening (HTS) biochemical assay for VCP ATPase activity measurement. Upon screening our in-house natural small molecular library of 700 compounds (at 10 µM each) (Supplementary Fig. 1A), 9 hits were identified with each inhibiting VCP ATPase activity by ≥50% at 10 µM. The most potent hit was gossypol acetate (Fig. 1A), an acetic acid mixture with gossypol, which almost completely inhibited the ATPase activity of VCP at 10 µM (Supplementary Fig. 1A). Gossypol is an active ingredient from the seeds or root bark of malvaceae or other plants. Gossypol acetate has better absorption, longer serum half-life and less side effects than gossypol itself and has been widely used in the clinic (27 ). We thus focused on gossypol acetate for further study. Gossypol acetate inhibited VCP enzymatic activity with an IC50 of 6.53 ± 0.6 µM (Fig. 1B). To confirm the binding of gossypol acetate to VCP, we performed partial trypsin digestion assay using purified 6xHis-VCP protein in the presence of gossypol, gossypol acetate or the allosteric VCP inhibitor NMS873 (28 ) (Fig. 1C). We observed significant difference between each inhibitor-treated group and the DMSO control group in the cleavage profile. Gossypol acetate as well as gossypol reduced VCP’s sensitivity to trypsin digestion compared to DMSO (Fig. 1C, red arrow), confirming their direct binding to VCP. Notably, the digestion patterns of gossypol and gossypol acetate group that were quite identical as expected, were quite different from that of NMS873, suggesting that the influence of NMS873 on VCP conformation is different from gossypol or gossypol acetate. Gossypol acetate inhibited ATPase enzymatic activity of both VCP-FL(full length) and VCP-N+D1 domain, but not the one of VCP-D1+D2 domain (Fig. 1D), suggesting that inhibition of VCP enzymatic activity by gossypol acetate required both N and D1 domain of VCP.
To verify the binding of gossypol acetate to VCP, we carried out thermal shift assay based on ligand-induced thermal stabilization of target proteins. Thermostability of 6XHis-VCP protein was gradually increased with increased concentration of gossypol acetate when temperature was fixed at 70℃(Supplementary Fig. 1B). At 5 μM concentration of gossypol acetate, VCP was significantly more stabilized compared to the DMSO control group. Next, VCP protein melting temperature (Tm) was determined upon incubating with 5 μM gossypol acetate or DMSO control. Gossypol acetate significantly increased VCP’s Tm from 64.2 ± 0.1℃ to 71.1 ± 0.7℃ (Fig. 1E, 1F), confirming the direct binding of gossypol acetate to VCP. Consistent with the result that VCP enzymatic inhibition by gossypol acetate required the presence of both N and D1 domains, thermal shift assay also confirmed that gossypol acetate bound to N+D1 domain. Incubation with 5 μM gossypol acetate significantly increased thermal stability of VCP-N+D1 domain (Tm increased from 56.9±0.1℃ to 62.3±1.6℃) (Fig. 1G, 1H). We then used Isothermal Titration Calorimetry (ITC) to quantify the interaction between VCP and gossypol acetate. The calorimetry experiments showed that KD value is about 6.9 μM for the binding between gossypol acetate and VCP (Supplementary Fig. 1C).
Next, we attempted to gain additional insight into the interaction by docking gossypol to the crystal structure of VCP (PDB code 5DYG) in a simulation experiment. The result of molecular docking, which suggested that gossypol bound at the interface between VCP N domain and D1 domain (Supplementary Fig. 1D), was consistent with the experimental results that inhibition of VCP ATPase activity by gossypol acetate requires both N and D1 domain. Given that the function of VCP relies on the formation of hexamers, we next used native gel electrophoresis assay to determine whether gossypol acetate affects VCP hexamer formation. Native gel electrophoresis of the product of VCP incubated with gossypol acetate showed that the oligomeric state of VCP complex was unchanged (Supplementary Fig. 1E). Together, these results demonstrated that gossypol acetate is bound to the interface between N and D1 domains of VCP, thereby inhibiting the VCP ATPase activity in vitro without affecting its oligomeric state.
Gossypol acetate reduced the mHTT protein level and neuronal toxicity
As a crucial regulator of proteostasis, VCP regulate degradation of damaged protein in proteasome and autophagy pathway. It has been implicated in the pathology of HD as well as other neurodegenerative disease, thus molecular modulator targeting VCP might be of therapeutic benefits for treatment of neurodegenerative diseases including Huntington’s disease. We tested whether gossypol acetate could facilitate degradation of the toxic, aggregation-prone mHTT and improve neuron survival in HD cellular models.
We first investigated whether gossypol acetate could promote degradation of autophagy substrate-mHTT proteins in HD patient iPS-derived Q47 striatal neurons generated as previous described (29). The well-established HTRF assay was used for measurement of mHTT level (30 ) and assessment of mHTT clearance by gossypol acetate treatment. Endogenous mHTT protein level was decreased dose-dependently by gossypol acetate treatment (Fig. 2A). The observed reduction is not due to cell loss, because the total protein concentration for mHTT measurement was controlled. Meanwhile, compound treatment slightly increased neuron cell number and improved cell morphology as shown in Fig. 2D, excluding compound toxicity at the dose used. Western-blot analysis was used to detect the level of mHTT and wtHTT protein. The 2B7 and MW1 antibodies were used, with MW1 specifically detecting the polyQ mHTT and 2B7 detecting both mHTT and wtHTT. As shown in Fig 2B, gossypol acetate induced obvious reduction of mHTT, while the level of total HTT were only slightly decreased (Fig. 2B). Cotreatment with the autophagy inhibitor CQ almost completely blocked mHTT degradation induced by gossypol acetate (Fig 2C), implying that the observed mHTT lowering mainly depends on autophagy pathway.
Since gossypol acetate induced mHTT degradation, we next examined whether it could improve neuronal survival. Withdraw of brain-derived neurophic factor (BDNF) induced neuronal apoptosis, shrinkage and loss, which are mHTT-dependent (Figure 2D). These phenotypes under stress condition were previously used for assaying huntington’s disease-associated cytotoxicity. Treatment of gossypol acetate significantly reduced caspase3 activation (Figure 2D-E) as well as rescued stress-induced neuronal shrinkage and loss (Figure 2D, 2F). The neuroprotective effect of gossypol acetate was further confirmed in the mouse striatal cell line STHdhQ7/Q111 (31 ). In this cell model, gossypol acetate remarkably induced mHTT degradation in a dose-dependent manner as shown in HTRF assay and Western-blot analysis (Supplementary Fig. 2A,B). Consistently, gossypol acetate reduced apoptosis (detected by caspase3 activation) induced by serum starvation (Supplementary Fig. 2C), indicating suppression of mHTT-induced cytotoxicity. In summary, gossypol acetate treatment induced degradation of mHTT proteins and significantly suppressed mHTT toxicity, which relies on autophagy.
Gossypol acetate increased the formation of mHTT-VCP-LC3 ternary complex
We next investigated the mechanism by which gossypol acetate induced mHTT degradation by harnessing the autophagic machinary. Previous studies showed that VCP colocalized with expanded polyQ in HD mouse brains and in postmortem patient brains (14, 32 ). We used pull-down experiment to confirm the VCP-mHTT interaction and further revealed that VCP’s N-terminal domain is the binding site for mHTT (Supplementary fig 3A). Meanwhile, it has been reported that VCP interacts with LC3 through LIR motif in VCP’s N terminal domain (26 ). However, it remains unclear whether VCP is in a complex with mHTT and LC3 and play a role in autophagy-mediated mHTT degradation.
We carried out in vitro pull-down experiments to determine whether VCP formed a complex with mHTT and LC3 and examined the effects of gossypol acetate on VCP interaction with LC3 or mHTT. First, gossypol acetate enhanced interaction between GST-LC3 and 6XHis-VCP in a dose-dependent manner (Fig. 3A). In contrast, the VCP enzymatic inhibitor MNS873 and DBEQ did not affect VCP’s interaction with LC3 (Supplementary Fig. 3B), in accordance with the observation that they have different binding sites on VCP as well as different trypsin digestion pattern Second, we examined whether gossypol acetate affected VCP’s interaction with mHTT or with wtHTT. The pull-down assay revealed that VCP’s interaction with mHTT is much stronger compared with its interaction with wtHTT. Gossypol acetate significantly increased VCP’s interaction with mHTT (Fig 3B), with little influence on VCP-wtHTT interaction. Next, we determined whether VCP was in a complex with mHTT and LC3 and the effects of gossypol acetate on complex formation. Pull-down experiment showed VCP did form a complex with mHTT and LC3 (Fig.3C). Consistent with the enhanced interaction of VCP-LC3 and VCP-mHTT respectively, gossypol acetate could also increase the complex formation of 6XHis-VCP, GST-LC3 and mHTT proteins (Fig.3C). However, in the absence of VCP, gossypol acetate did not obviously affect the binding between GST-LC3 and mHTT-Q72 (Supplementary Fig. 3C), suggesting that enhanced LC3-VCP- mHTT complex formation depends on gossypol acetate’s effect on VCP. VCP-FL, VCP-N domain and VCP-N+D1 domain interacted with LC3, respectively (Fig.3D), indicating that VCP interacts with LC3 through the N domain of VCP. Gossypol acetate increased LC3’s interaction with VCP-FL, VCP-N+D1 domain, but not VCP-N alone. This further suggested the enhanced VCP-LC3 interaction relies on gossypol acetate’s binding to VCP, which depends on the presence of both N and D1 domain (Fig.1). These data revealed that gossypol acetate enhanced formation of mHTT-VCP-LC3 complex, which may facilitate autophagy-mediated degradation of mHTT by increasing the efficiency of substrate recognition by autophagosomes.