Methods

Non–small cell lung cancer cell lines A549, H1299, and H1975 were cultured in RPMI 1640 medium supplemented with 10% (vol/vol) fetal bovine serum at 37°C in a humidified 5% CO₂ atmosphere incubator (Thermo Fisher Scientific Inc., Waltham, MA, United States). Cell lines were placed in 96-well plates at 4 × 10³ per well in triplicates and SNX-2112 (100, 10, 1, 0.1, 0.01 μM) treated cell lines for 72 h (quantitative IC₅₀ value). Cell lines were placed in triplicate at 4 × 10³ per well in 96-well plates and then treated by SNX-2112 (4 × 0.50 μM for A549, 4 × 1.14 μM for H1299, and 4 × 2.36 μM for H1975) or DMSO for 24, 48, or 72 h (cell viability assay); 2.5 h after adding CCK-8, OD₄₅₀ was detected by CCK-8 assay (7Sea, Shanghai, China) using an ELISA plate reader (BioTek Instruments, United States; Xue et al., 2016; Wei et al., 2018).

Non–small cell lung cancer cell lines A549, H1299, and H1975 were seeded in a 6-well culture plate at density 3 × 10⁵ cells per well, which were treated with different concentrations of SNX-2112 or DMSO for 72 h (4 × 0.50 μM for A549, 4 × 1.14 μM for H1299, and 4 × 2.36 μM for H1975). Cells were then gathered and washed with cold phosphate-buffered saline (PBS). And then at 4°C, 1 mL PBS and 2 mL 100% ethanol were was added. After centrifugation at 1,000 g for 5 min, they were washed with 2 mL PBS and resuspended with 400 μL PBS. In the end, 50 μL propidium iodide (PI) and 50 μL RNase were added, and cells were darkly cultivated for 30 min at 37°C. The effects of SNX-2112 on cell cycle distribution were analyzed with flow cytometry (Becton Dickinson, Franklin Lakes, NJ, United States; Olszewska et al., 2020)

Based on instructions from the manufacturer, the apoptosis rate was detected by the annexin V–fluorescein isothiocyanate (FITC)/PI Apoptosis Detection kit (556547, Becton Dickinson, United States). Cells were treated identically with mentioned as Cell Cycle Analysis. And then, cells were gathered and washed with 2 mL PBS and resuspended in 100 μL binding buffer. Lastly, 5 μL annexin V–FITC and 10 μL PI were added to the buffer and darkly cultivated at 4°C for 30 min. Effects of SNX-2112 on cell apoptosis were determined by flow cytometry (Becton Dickinson, Franklin Lakes, NJ, United States; Ye et al., 2017).

Thermal shift assay was applied for evaluating molecular interaction between target Hsp90^N and its ligand SNX-2112 using real-time PCR (7500, ABI Corp., United States). A 20-μL reaction system consisted of Hsp90^N (1 mg/mL) 2 μL, ligand SNX-2112 (dissolved in DMSO, 100 mM) 0.5 μL, buffer (pH 7.5, 20 mM Tris–HCl, 150 mM NaCl, and 10% glycerol) 10 μL, Protein Thermal Shift Buffer (Applied Biosystems, United States) 5 μL, and TSA dye (Applied Biosystems, United States) 2.5 μL, eight replicates. The samples were centrifuged at 1,000 revolutions/min (rpm) for 1 min. The samples were run from 25 to 95°C with a ramp rate of 1°C/min. The protein unfolds when it is heated and exposes hydrophobic regions to bind the environmentally sensitive TSA dye and fluoresces. The melting temperature (Tm) and melting temperature differences (ΔTm) were derived from the melting curve, which was related to the binding affinity of Hsp90^N-SNX-2112 (Andreotti et al., 2015).

Isothermal titration calorimetry was used for further detecting Hsp90^N-SNX-2112 molecular interactions using ITC (ITC-200, Malvern Instrument Ltd., United Kingdom). SNX-2112 in DMSO was diluted with buffer for protein purification (pH 7.5, 20 mM Tris–HCl and 150 mM NaCl) to 500 μM. Fresh-purified Hsp90 was extensively dialyzed against the same buffer and concentrated to 50 μM. After being centrifuged and degassed, 2 μL aliquots of SNX-2112 were injected into Hsp90^N solution in the cell with an interval of 200 s and 750-rpm stirring speed. With Microcal Origin software, the experimental data were fitted with a bimolecular binding model with stoichiometry (n), enthalpy (ΔH°), and association constant (K_a) as adjustable parameters. The thermodynamic parameters ΔG° (free energy) and ΔS° (entropy) were derived from the equation -RT ln K_a = ΔG°= ΔH° - TΔS (He et al., 2014b).

Homo sapiens Hsp90^N gene containing residues 9 to 236 was artificially synthesized and cloned into plasmid pET-28a, which was transformed into Escherichia coli BL21 (DE3; TIANGEN Biotech Corp., Beijing, China) to express and purify Hsp90^N as reported by Cao (Cao et al., 2017).

With 5:1 molar ratio, SNX-2112 was added into Hsp90^N to incubate for 30 min at 4°C. Then, the mixture was centrifuged for 10 min at 3,000 g, and the supernatant was taken to mix with the same amount of a crystal precipitant [pH 8.5, 100 mM Tris–HCl, 200 mM MgCl₂, 25–30% PEG4000 (Stebbins et al., 1997)]. With the hanging drop vapor diffusion method, cocrystallization was carried out at 4°C for 3–7 days in an incubator controlled by a bath circulator (PolyScience 9712, PolyScience, United States). Complex crystal image of Hsp90^N-SNX-2112 was captured by a stereomicroscope (M165, Leica Microsystems, Germany).

Complex crystals were mounted with cryo-loop (Hampton Research Corp., Aliso Viejo, CA, United States), and then quickly soaked in the cryoprotectant solution containing 20–25% glycerol and crystal reagent mentioned previously (Prodromou et al., 2000). Then, crystals were flash-frozen in liquid nitrogen for XRD. With an ADSC Quantum 315r CCD detector (Wang et al., 2015), all data sets were collected at 100 K on Macromolecular Crystallography Beamline17U1 (BL17U1) at Shanghai Synchrotron Radiation Facility (SSRF, Shanghai, China; Wang et al., 2018).

Aquarium pipeline was applied for automatic processing of diffraction data (Yu et al., 2019b). Using the crystal structure of Hsp90^N-FS23 (PDB ID 5CF0) as the research model (Li et al., 2015), the complex crystal structure was determined and refined by molecular replacement with PHENIX software (Adams et al., 2010; Yu et al., 2019a). The initial model was rebuilt by Coot software (Emsley et al., 2010), and CCP4MG software was applied for graphing and molecular interaction analysis (McNicholas et al., 2011; Winn et al., 2011).

A series of new SNX-2112 derivatives were designed based on the complex crystal structure and molecular interaction analysis, which were evaluated by molecular docking with software SYBYL-X 2.0 (Tripos Associates Inc., St. Louis, MO, United States). The crystal structure of Hsp90^N-FS23 (PDB ID 5CF0) was as the docking model. First, a compound library of newly designed SNX-2112 derivatives was built, and new derivatives were conducted with geometric and force field optimization for 10,000–100,000 times. Second, the target Hsp90^N was optimized by adding hydrogen and charges. Last, molecular docking was carried out using the Surflex-Dock module of SYBYL software. Total score and Cscore were chosen as parameters to evaluate the binding affinity between the target Hsp90 and new ligands. CCP4MG software was applied for simulated complex three-dimensional (3D) structure reconstruction and molecular interaction analysis.

Statistical analysis was conducted with SPSS 13.0 software (International Business Machines Corporation, United States) and GraphPad Prism 5.01 software (GraphPad Software, San Diego, United States). Data were presented as mean ± standard deviation (SD). Differences between two groups were performed according to unpaired Student t test. P < 0.05 was deemed as statistically significant.