SCIENCE CHINA Chemistry •ARTICLES• . . . . . . . . . . . . . . . .. . .. .. .. . .. .. ... .. .. . .. .. ... .. .. . .. .. .. . .. .. ... .. .. . .. .. .. . .. .. ... .. .. . .. .. .. . .. .. ... .. .. . .. .. ... .. .. . .. .. .. . https://doi.org/10.1007/s11426-018-9283-3 ............................

A new class of HIV-1 inhibitors and the target identification via proteomic profiling Ying-Zi Ge1†, Bin Zhou1†, Ruo-Xuan Xiao1, Xiao-Jing Yuan1, Hu Zhou1, Ye-Chun Xu1, Mark A. Wainberg2, Ying-Shan Han2* & Jian-Min Yue1* 1 State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai 201203, China; McGill University AIDS Centre, Lady Davis Institute for Medical Research, Jewish General Hospital, Montreal, Quebec H3T 1E2, Canada

2

Received April 4, 2018; accepted May 15, 2018; published online June 21, 2018

Anti-HIV screening with the MT-4/MTT assay on a focused library of structurally diverse natural products has led to the discovery of a group of steroids with potent activities, which include four new ergostane-type steroids, named amotsterols A−D (1−4), together with two known analogs. Among them, the most potent amotsterol D (4) exhibited anti-HIV activity against wildtype and some clinically relevant multidrug resistant HIV-1 strains. Subsequent studies on its target identification through a proteomic approach found that compound 4 might target PKM2, a rate limiting enzyme of glycolysis, in host cells to restrict HIV replication. The docking model of compound 4 to PKM2 showed that the two hydroxyl groups of 4 form hydrogen bonds with the two parallel Y390 in each subunit of PKM2 separately, and the ring C of 4 is sandwiched between the two parallel aromatic rings of F26. The identified hit compound may have the potential to be further developed as a novel anti-HIV agent. These results demonstrated that an integrated approach, which combines new chemical structures and phenotypic screening with a proteomic approach, could not only identify novel HIV-1 inhibitors, but also elucidate the unknown targets of compound interactions in antiviral drug discovery. ergostane-type steroids, anti-HIV, target identification, PKM2 Citation:

Ge YZ, Zhou B, Xiao RX, Yuan XJ, Zhou H, Xu YC, Wainberg MA, Han YS, Yue JM. A new class of HIV-1 inhibitors and the target identification via proteomic profiling. Sci China Chem, 2018, 61, https://doi.org/10.1007/s11426-018-9283-3

1 Introduction Standard antiretroviral therapy (ART) usually uses at least three different HIV drugs to treat HIV-infected individuals. Since ART was reported in 1996, 29 antiviral drugs have been approved for use in ART in clinics by the Food and Drug Administration of the United States. Based on their modes of action that target either different stages of the HIV life cycle or host cellular factor, these drugs are classified into six different categories, including nucleoside reverse transcriptase inhibitors (NRTIs), non-nucleoside reverse †The authors contributed equally to this work. *Corresponding authors (email: [email protected], on behalf of Dr. Mark A. Wainberg; [email protected])

transcriptase inhibitors (NNRTIs), entry inhibitors, fusion inhibitors, integrase inhibitors (II), and protease inhibitors (PIs). So far, ART has greatly reduced HIV-caused death and has improved quality of life for people infected with HIV. However, HIV can develop resistance to almost all classes of currently used antiviral drugs. Besides, long-term drug toxicities also remain major problems [1,2]. Thus, there is still a need for continuing discovery and development of new anti-HIV agents that could have improved drug resistance profile and/or reduced toxicity. Natural products had been the most successful source of potent drugs [3], such as artemisin that revolutionized the treatment of malaria [4]. However, natural products for drug discovery has been greatly neglected since the 1990s, be-

© Science China Press and Springer-Verlag GmbH Germany, part of Springer Nature 2018 . . . . . . . . . . . . . . . . . . . . .

chem.scichina.com link.springer.com

2.

...........................

Ge et al.

Sci China Chem

cause the pharmaceutical industry moved to high-throughput screening (HTS), which usually screen large synthetic chemical libraries against potential therapeutic targets. In spite of great efforts in HTS-based drug discovery in the past two decades, the development rates were decreased dramatically. Recently there have been renewed interests in natural products as the sources of new and novel drugs [5,6]. Due to the structural novelty and ethnomedical application, natural products derived from plants have great potential as a source of anti-HIV drugs that might possess antiviral activity via unique modes of action [7,8]. The discovery of new anti-HIV compounds largely depends on valid biochemical and cell-based assays and a source of unique chemicals. HTS has been successfully used for the identification of a few HIV inhibitors for known molecular targets in the HIV life cycle [9]. But HTS-based approach involves huge upfront cost due to use of large synthetic chemical libraries. Now drug screens are moving away from screening large libraries of compounds to smaller focused libraries (100 to 3000 compounds) using biochemical and/or cell-based assays [4]. Recently, there is a resurgence in phenotypic drug screens, as this approach may be more likely to yield first-in-class drug leads with novel modes of action [10−12]. A cell-based MT4/3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, that measures HIV-induced cytopathic effect (CPE), has been used successfully to identify new HIV inhibitors [13]. This cell-based phenotypic screening approach has potential to identify first-in-class lead compounds that target any stage of the HIV life cycle or cellular factors in host cells, which could restrict HIV replication. Subsequent identification of targets for the hit compounds via this approach remains an important and challenging task. Drug affinity responsive target stability (DARTS) is a relatively quick approach to identify potential protein targets for small molecules [14−16]. DARTS has a great advantage that it does not require any chemical modification of an active compound for target identification. Aiming to identify antiHIV natural products with novel modes of action, we hypothesized that one ideal solution would be an integrated approach to combine phenotypic screening of a unique natural product library, followed by a chemical proteomic approach that uses DARTS for target identification. In this article, we applied the MT-4/MTT assay to identify antiviral compounds from a focused library of 220 natural products and employed DARTS to identify the potential therapeutic targets in the host cells. Our efforts have led to the discovery of a group of structurally related steroids with remarkable anti-HIV activity, which include four new ergostane-type steroids, named amotsterols A−D (1−4), together with two known analogs 5 and 6 (Figure 1). All the compounds were obtained from the Chinese medicinal plant of Amoora tsangii. The most promising compound 4 has

Figure 1

..............................................

2

Structures of compounds (1–6) (color online).

potent anti-HIV activity against wild-type and clinically relevant multidrug resistant HIV strains. Further chemical proteomic studies using DARTS showed that 4 could putatively target biosynthetic pathway in the host cells. These results suggested that 4 may have the potential to be further developed as a novel anti-HIV agent. Herein we present the chemistry, HIV inhibition, a brief structure-activity relationship, target identification, and molecular docking of these natural products.

2 Experimental 2.1

General procedures

The optical rotations of isolated compounds were obtained on a Perkin-Elmer 341 polarimeter at room temperature (USA). UV spectra were measured on a Shimadzu UV-2550 spectrophotometer (Japan). IR spectra were recorded on a Perkin-Elmer 377 spectrometer with KBr disks. NMR spectra were carried out on a Bruker AM-400 spectrometer (Germany) with TMS as internal standard. Electrospray ionization mass spectrometry (ESI(±)MS) and HRESI(±)MS were determined on a Bruker Daltonics Esquire 3000 plus instrument and a Waters Q-TOF Ultima mass spectrometer (USA), respectively. Semipreparative high performance liquid chromatography (HPLC) was performed on a Waters 1525 pump equipped with a Waters 2489 detector and an YMC-Pack ODS-A column (10 mm×250 mm, S-5 μm, 12 nm). Silica gel (300–400 mesh), Silica gel H (Qingdao Haiyang Chemical Co. Ltd., China), C18 reversed-phase silica gel (250 mesh, Merck), Sephadex LH-20 gel (Amersham Biosciences) and MCI gel (CHP20P, 75–150 μM, Mitsubishi Chemical Industries, Ltd., Japan) were used for column chromatography, and pre-coated silica gel GF254 plates

3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Ge et al.

Sci China Chem

(Qingdao Haiyang Chemical Co. Ltd., China) were used for thin layer chromatography (TLC). All solvents used were of analytical grade (Shanghai Chemical Plant, China), and solvents used for HPLC were of HPLC grade (J&K Scientific Ltd. , China). Plant Material. The twigs of Amoora tsangii were collected from Jianfengling on Hainan Island, China, in September of 2010 and were identified by Prof. Shi-Man Huang of Department of Biology, Hainan University, China. A voucher specimen has been deposited in Shanghai Institute of Materia Medica, Chinese Academy of Sciences (accession number: Ats-20091Y). Extraction and isolation. The air-dried powder of twigs of A. tsangii (5 kg) was extracted three times with 95% EtOH at room temperature to provide a crude extract (300 g), which was subsequently suspended in H2O and partitioned with EtOAc. The EtOAc-soluble fraction (120 g) was subjected to MCI-gel CC eluted with a MeOH/H2O (30:70 to 100:0, v/v) to afford three main fractions A−C. The fraction A (40 g) was subjected to a silica gel column with petroleum ether (60−90 °C)/Me2CO (15:1 to 0:1, v/v) to give five major fractions (A1−A4). Fraction A3 (5.4 g) was subjected to a RP-18 silica gel column (MeOH/H2O, 30:70 to 100:0, v/v) to afforded six subfractionsA3a−A3f. Compound 3 (23 mg, tR: 12.42 min) and 6 (7.5 mg, tR: 15.77 min) were obtained from the fraction A3f (900 mg) via semi-preparative HPLC (3 mL/min, isocratic elution with MeCN/H2O, 70:30). Fraction A4 (5.0 g) was separated over a column of silica gel CC (CH2Cl2/MeOH, 200:1 to 10:1) to afford three major fractions A4a−A4c. The fraction A4c (1.2 g) was sequentially purified by Sephadex LH-20 gel (MeOH) and semipreparative HPLC (3 mL/min, isocratic elution with MeCN/ H2O, 65:35) to give compound 5 (6.3 mg, tR: 10.89 min). Fraction B (30 g) was subjected to a column of silica gel (CH2Cl2/CH3OH, 200:1 to 10:1) to afford six major fractions B1−B6. The subfraction B1 (5.3 g) was fractionated on a RP18 silica gel column (MeOH/H2O, 30:70 to 100:0, v/v) to yield five fractions. The B1c fraction (1.5 g) was first separated on a silica gel column eluded with petroleum ether (60−90 °C)/Me2CO (12:1 to 0:1, v/v) to gain a major component B1c2 (180 mg), which was then purified by semipreparative HPLC (3 mL/min, isocratic elution with MeCN/ H2O, 80/20) to yield compound 4 (16 mg, tR: 13.97 min). The fraction B1d (1.1 g) was purified similarly as that of B1c to afford a major component B1d2 (150 mg), and which was then purified semipreparative HPLC (3 mL/min, isocratic elution with MeCN/H2O, 80:20) to afforded compounds 1 (9.1 mg, tR: 13.44 min) and 2 (5.4 mg, tR: 14.02 min). Amotsterol A (1): White, amorphous powder; [α]22D −161 (c 0.5, CHCl3). UV (MeOH) λmax (log ε): 202 (4.08) nm. IR (KBr) νmax 3440, 2935, 1707, 1442, 1388, 1282, 1167, 1005, 887 cm−1. 1H and 13C NMR (CDCl3), see Table S1 (Supporting Information online); ESI(+)MS m/z 911.8 [2 M

..............................................

3

+Na]+. HR(+)ESIMS m/z 445.3321 [M+H]+ (calcd. for C28H45O4, 445.3318). Amotsterol B (2): White, amorphous powder; [α]22D −89 (c 0.2, CHCl3). UV (MeOH) λmax (log ε): 203 (3.78) nm. IR (KBr) νmax 3465, 2954, 2925, 1714, 1466, 1385, 1228, 1114, 1026, 889 cm−1. 1H and 13C NMR (CDCl3), see Table S1; ESI (+)MS m/z 467.4 [M+Na]+. HR(+)ESIMS m/z 445.3323 [M +H]+ (calcd. for C28H45O4, 445.3318). Amotsterol C (3): White, amorphous powder; [α]22D −177 (c 1.3, CHCl3). UV (MeOH) λmax (log ε): 233 (4.37) nm; IR (KBr) νmax 3431, 2956, 1720, 1671, 1456, 1387, 1286, 1232, 1061, 968, 889 cm−1. 1H and 13C NMR (CDCl3), see Table S1; ESI(+)MS m/z 481.4 [M+Na]+; HR(+)ESIMS m/z 459.3112 [M+H]+ (calcd. for C28H43O5, 459.3110). Amotsterol D (4): White amorphous powder; [α]22D −216 (c 1.2, CHCl3). UV (MeOH) λmax (log ε): 237 (4.60) nm; IR (KBr) νmax 3431, 2954, 2871, 1720, 1668, 1464, 1387, 1288, 1232, 1180, 1057, 952 cm−1. 1H and 13C NMR (CDCl3), see Table S1; ESI(+)MS m/z 911.9 [2 M+Na]+. HR(+)ESIMS m/z [M+H]+ (calcd. for C28H45O4, 445.3318). Cells and reagents. Cell lines (MT-2, MT-4) and the infectious molecular clone pNL4-3 and nevirapine (NVP) were all obtained through the NIH AIDS Research and Reference Reagent Program. The 293T cell line was obtained from the American Type Culture Collection (CRL-11268). MT-2 and MT-4 cells were cultured in RPMI 1640 medium at 37 °C and 5% CO2. Complete RPMI 1640 media were supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 50 U/mL penicillin and 50 g/mL streptomycin. Cord blood mononuclear cells (CBMCs) were obtained through the Department of Obstetrics, Jewish General Hospital, Canada, and cultured as previously described [17]. Zidovudine (AZT) was purchased from Sigma-Aldrich (USA). Raltegravir (RAL) was obtained from Merck-Frosst Canada (Canada). Ritonavir (RTV) was obtained from Abbott Canada (Canada). A library of 220 natural products that were isolated from plants in China was provided by Prof. JM Yue (Shanghai Institute of Materia Medica, Chinese Academy of Sciences). Generation of replication-competent HIV-1 and virus production. Replication-competent HIV-1 plasmid wild-type and various mutant pNL4-3, i.e., pNL4.3INB(G140S), pNL4.3INB(E92Q), pNL4.3INB(Q148H), pNL4.3INB (N155H) and pNL4.3INB(G140S/Q148R) were generated as previously described [17]. Library screening for HIV-1 inhibitors and cytotoxicity assays. To identify potential HIV-1 inhibitors, a collection of 220 natural products was screened using the cell-based MT4/MTT assay as described previously [17]. Briefly, each compound was tested at final concentrations of 1 and 10 μM. The hit compounds were selected to be tested in a ten-point serial dilution at 2- or 3-fold in duplicate. Anti-HIV activity was also examined in CBMC cell cultures as described

4.

...........................

Ge et al.

Sci China Chem

previously [17]. The cytotoxicity of the hit compounds was determined using the MTT assay for measurement of cell viability after incubation in serial dilutions of the compounds for three days. Both EC50s and CC50s were calculated by nonlinear regression, using Prism 5.0 software. Assessment of drug pretreatment on viral infection. MT-4 cells were pretreated with 4, AZT or RAL at 5 to 50 times of their EC50s for 24, 4 and 0 h at 37 °C. Cells were washed 3 times with phosphate-buffered saline (PBS) and were added with medium or medium with drugs before infection of HIV-1. Untreated cells with medium alone or with drugs were served as a negative control or positive control, respectively [18]. After incubation at 37 °C for 3 d, cell viability was measured by the MTT assay as above. Time-of-addition assay. 50 μL of exponentially growing MT-4 cells (5×105/mL) were plated onto 96-well plates. Reference and test compounds were added at various time points at or after infection (0, 2, 4, 6, 8, 12 and 24 h). The cells were infected with HIV-1 NL 4-3 at a multiplicity of infection (MOI) of 0.05. Reference compounds included AZT, RAL and RTV. The concentration of the different compounds added was 5 to 50 times of their EC50s. Cells infected with virus in the absence of compound was set as a control (100%). Protection of CPE in MT-4 cells day 3 after infection were measured and compared with the control. Target identification using drug affinity responsive target stability. Target identification using DARTS was performed as described previously. MT4 cells grown in 10 cm dish at day 3 were collected by spinning down cells at 500 r/min. Cell pellet was applied 600 μL cold lysis buffer (50 mM Tris-HCl pH 7.5, 200 mM NaCl, 0.5% Triton X-100, 10% glycerol, 1 mM dithiothreitol (DTT)) supplemented with protease and phosphatase inhibitors. The cells lysate was incubated on ice for 30 min, gently mixing every 5 min and centrifuged at 18000×g for 10 min at 4 °C. The supernatant was collected, and protein concentration was determined by bicinchoninic acid (BCA). Aliquots of the supernatant were treated with 4 (0.1 mM, 1 mM) or dimethyl sulfoxide (DMSO) (as a control) at room temperature for 1 h, respectively. Each sample was proteolysed by Pronase solution in reaction buffer (50 mM TrisHCl (pH 8.0), 50 mM NaCl, 10 mM CaCl2) for 30 min. The digestion was stopped by directly adding equal volume of 2× SDS loading buffer into each sample and heating to 70 °C for 10 min immediately. Samples (15 μL) were subjected to 10% sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE). Gels were stained by silver staining. 2.2

Mass spectrometry analysis

Gel bands were cut out and prepared for MS analysis with trypsin digestion as described [19]. The resulting peptides

..............................................

4

were analyzed by LC-MS/MS. The reverse phase high-performance liquid chromatography (RP-HPLC) separation was achieved on the Easy nano-LC system (Thermo Fisher Scientific, USA) using a self-packed column (75 μm×150 mm; 3 μm ReproSil-Pur C18 beads, 120 Å, Dr. Maisch GmbH, Germany) at a flow rate of 300 nL/min. The mobile phase A of RP-HPLC was 0.1% formic acid in water, and B was 0.1% formic acid in acetonitrile. The peptides were eluted using a gradient (2%–90% mobile phase B) over a 60 min period into a nano-ESI orbitrap Elite mass spectrometer (Thermo Fisher Scientific). The mass spectrometer was operated in data-dependent mode with each full MS scan (m/z 350 −1800) followed by MS/MS for the 15 most intense ions with the parameters: ≥+2 precursor ion charge, 2 Da precursor ion isolation window and 35 normalized collision energy of higher-energy collisional dissociation (HCD). Dynamic Exclusion™ was set for 30 s. The full mass and the subsequent MS/MS analyses were scanned in the Orbitrap analyzer with R=60000 and R=15000, respectively. The MS data were analyzed using the software MaxQuant [20] (http://maxquant.org/, version 1.5.1.0). Proteins were identified by searching MS and MS/MS data of peptides against a IPI mouse proteome database (ipi. mouse. v3.87). Trypsin/P was selected as the digestive enzyme with two potential missed cleavages. The search included variable modifications of methionine oxidation and N-terminal acetylation, and fixed modification of cysteine carbamidomethylation. Peptides of minimum 6 amino acids and maximum of two missed cleavages were allowed for the analysis. For peptide and protein identification, false discovery rate was set to 0.01. Data analysis. Data were analyzed using Prism 5.0 and expressed as means±standard deviation (SD). For each assay, three or more independent experiments, each in duplicate or triplicate, were performed, and the results analyzed. 2.3

Computational methods

Activator-bound structure of PKM2 (PDB code: 3GR4) was used as the receptor structure and was prepared using Protein Preparation and Grid Preparation tools in the Schrödinger Maestro interface [21]. Since the structure is a dimer-ofdimers, and the activator binding pocket is located at the subunit interaction interface, only Chains A and B were reserved for docking. Water molecules and ions were deleted since they were not conserved among different activatorbound structures. The active site was defined according to the activator in the complex structure. The 3D structure of compound 4 were generated and optimized with GAUSSIAN 09 at B3LYP/6-31G* level [22,23]. The output structure was then prepared with the Schrödinger program LigPrep. Docking was performed with Glide [24]. We carried out Standard Precision (SP) calculations with default

5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Ge et al.

Sci China Chem

..............................................

5

settings. The OPLS-2005 force field was used for minimization and grid generation, while OPLS-2001 was used for docking. The top ranked pose was reserved for analysis.

3 3.1

Results and discussion Chemistry of compounds 1−4

Compound 1 was obtained as white, amorphous powder, and its HRESI(+)MS displayed a quasi-molecular ion at m/z 445.3321 [M+H]+ (calcd. 445.3318), corresponding to the molecular formula of C28H44O4 with seven indices of hydrogen deficiency. The IR spectrum exhibited the strong absorption bands at 3440 and 1707 cm−1, revealing the presence of hydroxy and carbonyl functionalities. The characteristic signals of three tertiary methyls (δH 0.90, 1.07, 1.27, each 3H, s), two secondary methyls (δH 1.02, 6H, d, J= 6.8 Hz), one terminal double bond (δH 4.64, 4.73, each 1H, brs), one oxygenated methine, one oxygenated tertiary carbon and two keto groups (δC 210.9 and 221.6) were distinguished from the 1D NMR spectra (Table S1). The two keto carbonyl groups and one terminal double bond accounted for three out of the seven indices of hydrogen deficiency, and the remaining ones thus required compound 1 being tetracyclic. The above-mentioned data suggested that compound 1 was likely an ergostane-type steroid. The planar structure was further constructed by the detailed analysis of the 2D NMR spectra (Figure 2(a)). One hydroxy group was placed to C-3 (δC 66.0) as supported by the 1H detected heteronuclear multiple bond correlation (HMBC) correlations from H-3 (δH 4.08) to C-1 (δC 31.7) and C-5 (δC 41.3), while the correlations from H3-21 (δH 1.25) to C-17 (δC 70.4), C-20 (δC 73.9) and C-22 (δC 40.8) located the other hydroxy group at C-20. The HMBC interactions from H2-6 (δH 1.98 and 2.23, m) and H-8 (δH 2.49, dd, J=10.8, 10.8 Hz) to C-7 (δC 210.9) assigned a keto group at C-7. The HMBC correlations from H2-15 (δH 2.92 and 1.79, m) and H17 (δH 2.16, m) to C-16 (δC 221.6) fixed the other keto group at C-16 in the five-membered ring. The HMBC correlations from H2-28 (δH 4.71 and 4.63, brs) to C-23 (δC 27.8), C-24 (δC 156.4) and C-25 (δC 34.4) suggested the presence of a terminal ∆24(28) double bond. The relative configuration of compound 1 was established mainly by the interpretations of 1D NMR data and ROSEY and/or NOESY data (Figure 2(b)). The Me-18 and Me-19 showed ROESY correlations with H-8, indicating they were axially β-oriented. The NOESY correlations of H-5/H-9 and H-9/H-14 then indicated that the H-5, H-9 and H-14 were cofacial and α-oriented. Hence the A−D rings were assigned as trans-fused. The H-3 was defined equatorially β-oriented by the NOESY cross-peaks of H-3 with H2-2 (δH 1.60 and 1.42, m), indicating the hydroxy group at C-3 was a rare α-configuration. The ROESY correlations of Me-18 and Me-21

Figure 2 Selected HMBC (a), ROESY and NOESY (b) correlations of 1 (color online).

with H-12β [25], together with the chemical shift of Me-21 (δH 1.25) [26,27] indicated a 20S configuration for 1. Thus, compound 1, named amotsterol A, was elucidated as shown. Compound 2 shared the molecular formula of C28H44O4 with compound 1, as indicated by the protonated molecular ion at m/z 445.3323 [M+H]+ (calcd. 445.3318) in the HRESI (+)MS. Comparison of the NMR data of 2 (Table S1) with those of compound 1, revealed they were structural analogues possessing the same ergostane skeleton. The 2D NMR spectra (Figures S12 and S13, Supporting Information online) further assembled the planar structure of compound 2. The HMBC correlations from H2-2 (δH 2.34, m) and H2-4 (δH 2.21 and 2.03, m) to C-3 (δC 211.5) indicated the presence of C-3 carbonyl group. The 7-OH was located by the HMBC correlations from H-7 (δH 3.80, m) to C-5 (δC 45.3) and C-9 (δC 39.2). The relative configuration of compound 2 was verified by ROESY spectrum (Figure S14). The H-7 was equatorially β-configured as deduced by the ROESY correlations of H-8 (δH 1.64, m) and H2-6 (δH 1.59 and 1.42, m) with H-7 (δH 3.80). The other chiral centers of compound 2 were identical with those of compound 1, as supported by the NMR and ROESY data. The structure of amotsterol B (2) was thus identified as shown. Compound 3 displayed a protonated molecular ion peak at m/z 459.3112 [M+H]+ (calcd. 459.3110) in the HRESI(+)MS corresponding to a molecular formula of C28H42O5. Comparison of its NMR data (Table S1) with those of 1 indicated it was also an ergostane-type steroid, and the only differences are the presence an α,β-unsaturated carbonyl group (δC 201.8, 128.6, and 165.6), and an extra hydroxy group (δC 75.5) in 3. The aforementioned groups were then fixed by analysis of the 2D NMR spectra (Figures S20 and S21). Two adjacent hydroxy groups were attached to C-3 and C-4 by the key HMBC interactions from H-4 (δH 4.21) to C-2 (δC 31.2), C-3 (δC 71.5) and C-10 (δC 38.3). The HMBC networks of H8 (δH 2.47, dd, J=11.3, 11.3 Hz), H-9 (δH 1.63, m) and H-14 (δH 1.66, m) to C-7 (δC 201.8); H-6 (δH 5.80, s) to C-4 (δC 75.5), C-8 (δC 44.4) and C-10 (δC 38.3); and H3-19 (δH1.36, s) to C-5 (δC 165.6) suggested the presence of an α,β-unsaturated keto group at C-7. The relative configuration of 3 was assigned by the ROESY spectrum (Figure S22) and NMR data. The ROESY interactions of H-9/H-1α and H-1α/ H-3 implied that the H-3 was axially α-oriented. Meanwhile, the ROSEY correlation between H-4 and H-6, as well as the

6.

...........................

Ge et al.

Sci China Chem

small coupling constant of H-4 suggested that the H-4 was equatorially α-oriented. The stereochemistry of the other stereogenic centers of compound 3 were assigned to be the same as those of compound 1 based on the similar NMR data. Thus, compound 3, namely amotsterol C, was elucidated as shown. Compound 4 was found to possess a molecular formula of C28H44O4, as determined by the HRESI(+)MS ion peak at m/z 445.3312 [M+H]+ (calcd 445.3318). The NMR data (Table S1) of compound 4 were highly similar to those of compound 3, and the main differences were resulted from the presence of an additional methylene and a secondary methyl (δH 0.79, d, J=6.7 Hz) in 4 replacing an oxygenated methine and the terminal double bond in compound 3, respectively. The HMBC correlations (Figure S29) from H3-28 to C-24 (δC 39.3) and C-25 (δC 32.3), and from H2-4 (δH 2.55, m) to C-2 (δC 31.2) and C-10 (δC 38.5) confirmed the aforementioned deduction for the structure of 4. The relative configurations of the stereo centers except for the C-24 in compound 4 were assigned to be identical to those of compound 3 by the ROESY data (Figure S30) and the NMR data comparison. The stereochemistry of C-24 was then defined as 24R-configured according to an empirical role developed for this steroid class [28,29], in which, for a 24R-configuration, the ∆δC (δC-26−δC-27) is about 2; while the ∆δC is approximately 3 for a 24S-configuration. The ∆δC value measured for compound 4 was 2.2. Thus, the structure of amotsterol D (4) was elucidated as shown. Two known compounds were identified as (20S)-5,24(28)ergostadiene-3β,7α,16β,20-tetrol (5) [30], and stigmast-5ene-3β,7α,-20-triol (6) [31] by spectroscopic data and comparing with those of literature reported. 3.2 The anti-HIV activity and structure-activity relationship of compounds 1−6 As the starting point, we used an in-house collection of 220 structurally diverse natural products isolated from Chinese medicinal plants, and some of which with unique structures have been isolated and characterized in the previous studies

..............................................

6

[32−34]. A few of these compounds have also shown moderate anti-HIV activities in our initial screening using a MT4/ MTT assay [35−38]. In particular, a group of steroids 1−6 isolated from the plant of A. tsangii showed remarkable activity (Table 1). Of which, compound 4 showed the most potent anti-HIV activity with an EC50 of 0.09±0.01 μM, and exhibited a dose-dependent protection of MT4 cell CPE induced by a laboratory wild-type HIV-1 strain NL4-3 (Figure 3(a)). The cytotoxicity of compound 4 appeared at a CC50 of 4.95±0.51 μM, yielding a selectivity index of 55.0 (Table 1). The fact that CC50 of 4 much exceeded the EC50 in MT4 cells suggested that inhibition of HIV-1 infection was not caused by cytotoxicity induced by compound 4. The antiviral activity of 4 was also confirmed in primary CBMC cell culture (Figure 3(b)). Observation the structures and the anti-HIV activities of compounds 1−6 allowed to outline a gross structure-activity relationship for this class of anti-HIV agents, which are: (1) the changes of the side chain at C-17 and the oxidation patterns of the five-membered ring will not obviously affect the activities; (2) the presence of a 3-OH seems important for the activity as compared the anti-HIV potency of compound 2 with those of the other five ones; (3) the presence of an α,βunsaturated keto group at the B ring will dramatically increase the anti-HIV activity, e.g., in the cases of compounds 3 and 4, which also suggests the attendance of an 4-OH in the ring A of compound 3 as compared with 4 attenuated the Anti-HIV activities of 1−6 on HIV-1 NL 4-3 infected MT4

Table 1 cells

CC50 (μM)

SIa)

3.02±0.61

>25

>8.28

6.25±0.76

11.90±1.23

1.90

compounds

EC50 (μM)

1 2 3

0.54±0.07

18.55±1.71

34.35

4

0.09 ± 0.01

4.95 ± 0.51

55.00

5

1.75±0.21

11.35±1.31

6.49

6

1.22±0.11

25.8±2.7

21.15

NVPb)

0.12±0.02

>25

>208.33

a) SI (selective index)=CC50/EC50, b) NVP as a positive control.

Figure 3 (a) Dose-response CPE-protection curve of 4 in MT-4 cells infected with wild-type HIV-1 strain NL4-3, as measured by MTT assay. (b) Doseresponse CPE-protection curve of 4 in CBMC cells infected with wild-type HIV-1 strain NL4-3, as measured by MTT assay.

7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Ge et al.

Sci China Chem

..............................................

7

activity. 3.3 The activity of compound 4 against multidrug resistant HIV-1 strains To test whether 4 has activity against clinically relevant multidrug resistant HIV-1 strains, we further evaluated the anti-HIV activity of 4 in MT4 cells against some RI- and IIresistant HIV-1 strains. The results showed that 4 protected CPE at similar EC50 values as that against the wild-type strain (Table 2). These results also demonstrate that the inhibitory activity of 4 against HIV-1 seems to be not acting as a RI or II. 3.4

Target identification for compound 4

Compound 4 appeared to target cellular factors in the host cells to inhibit HIV-1 replication. As compound 4 has been shown to not act as RI or II, it might act via a novel mechanism of action. To further elucidate the underlying mechanism of compound 4, drug pretreatment studies were performed to determine whether 4 targeted a step of HIV life cycle or cellular factors in the host cells. Treatments of cells for 24 or 4 h, prior to virus infection, with 4, but not AZT (a RI) and RAL (an II), resulted in almost complete inhibition of viral induced CPE (Figure 4), demonstrating that 4 has induced an irreversible effect in the cells, that restrict virus replication. These results indicate that 4 might directly target cellular factors in the host cells. 3.5

Time-of-addition of compound 4

To determine if 4 inhibits possible step(s) of the HIV-1 replication, a time-of-addition assay was performed [39]. In this assay, a few representative HIV-1 inhibitors in different classes were used as reference compounds, including AZT (a Table 2

Figure 4 Pretreatments of cells with drugs on HIV-1 infection. MT-4 cells were pretreated with test compounds for 24, 4, 0 h. Cells were washed 3 times with PBS, followed by addition of medium or medium with drug and infected with HIV-1 NL 4-3. 4 (0.5 μM), AZT (0.5 μM), RAL (0.5 μM). The cell viability, relative to the control, was determined by MTT assay. Data represent mean±SD calculated from three independent experiments with duplicate samples (color online).

nucleoside reverse transcriptase inhibitor), RAL (an integrase inhibitor), RTV (a protease inhibitor). Distinct from all three classes of reference inhibitors used, the inhibition of HIV-1 replication by 4 was not significantly decreased when it was added 12 hpi (hour post-infection) (Figure 5). This result suggested that 4 could target the cellular factors that interferes a late event of the HIV-1 replication or apoptosis of the host cells. 3.6 Compound 4 may target biosynthetic pathways in the host cells To investigate the possible protein targets of 4 in the host cells, we employed a drug affinity responsive target stability (DARTS) assay. DARTS is based on the principle that a ligand binding to its target protein leads to a thermos dynamically more stable complex which is resistance to protease degradation [14,40–42]. In order to test the interaction of compound 4 with the possible protein targets, lysates from MT-4 cells were incubated with 4 or DMSO (vehicle con-

Antiretroviral activity of 4 against some RI- or II-resistant HIV-1 strains in MT4 cells Compound

Resistance

4

AZT

NVP

RAL

87.60

1.91

116

1.57

RTB-K103N

86.63 (1.0)

30.0 (15.8)

3950 (34.1)

–

RTB-Y181C

118.53 (1.4)

18.2 (9.6)

2769 (23.9)

–

WT (HIV-1NL 4-3) RI res

INI res

EC50 (nM)a) (fold change)b)

Virus strain

RTB-K65R

139.30 (1.6)

79.9 (41.8)

183.8 (1.6)

–

HIV-1B-E92Q

102.9 (1.2)

–

–

24.1 (15.4)

HIV-1B-G140S

140.2 (1.6)

–

–

7.6 (4.8)

HIV-1B-N155H

126.9 (1.4)

–

–

51.5 (32.8)

HIV-1B-G140S/Q148R

105.3 (1.2)

–

–

18.6 (11.8)

a) EC50: 50% effective concentration on CPE protection, determined in MT-4 cells against different HIV-1 strains by MTT. b) Fold change in EC50 compared to that of the wild type HIV-1 NL4-3 strain. Data represents the means of results from at least three independent experiments in duplicate.

8.

...........................

Ge et al.

Sci China Chem

Figure 5 Time-of-addition analysis. MT-4 cells were infected with HIV-1 NL 4-3. Test compounds were added at the time of infection or at various time points post-infection (0, 4, 8, 12, 24 h). Final concentrations of each compound were 5 to 100-fold higher than their EC50s; i.e., 4 (0.5 μM), AZT (0.5 μM), RAL (0.5 μM), RTV (1 μM). The percent CPE protection relative to the control was determined by the MT-4/MTT assay. Data represent mean±SD calculated from three independent experiments with duplicate samples (color online).

..............................................

Figure 6 Compound 4 protected four proteins from proteolysis. Lysates from MT-4 cells were treated with 4 (0.1, 1 mM) or DMSO and digested with Pronase. Samples were run on a SDS-PAGE gel. Bands appeared in the sample with 4 at 1 mM were cut and proteins were identified by MS (color online).

Identification of proteins protected from proteolysis by 4

Table 3

trol). Samples were digested with different concentrations of protease and subjected to SDS-PAGE. Bands of the target protein incubated with the drug should be more intense than those of the relevant protein incubated with DMSO control. As shown in Figure 6, incubation with 4 at 1 mM resulted in four clear bands distinct from that with DMSO. Subsequent MS analysis identified a number of proteins, including PKM2, JUP, GAPDH, DSP, etc. (Table 3). The identified proteins are supposed to be potential target proteins or offtargets by 4. The mechanism of action of compound 4 appeared to target proteins in the host cells. This was demonstrated in the pretreatments of cells for 24 h with 4 resulted in inhibition of viral induced CPE (Figure 4). Second, time-of-addition studies showed that the activity of 4 was not significantly reduced when it was added even at 12 h post-infection to susceptible cells (Figure 5), suggesting that 4 targeted a late event in the HIV life cycle. Finally, a DARTS assay demonstrated that treatment with 4 protects a few proteins from proteolysis (Figure 6). These proteins were identified as pyruvate kinase isoform M2 (PKM2), JUP, GAPDH, DSP, etc. (Table 3). Of the identified proteins, PKM2, a rate limiting enzyme of glycolysis, is particularly interesting. A previous study has shown that HIV-1 Tat induces perturbations in glycolysis, i.e., PKM2 was detected up-regulated in T cells expressing Tat [43]. It was also reported that PKM2 facilitates viral replication in HIV-1 infected macrophages [44]. Recently it was demonstrated that PKM2 expression was up regulated in HIV-1 JRFL infected PBMCs and during reactivation of HIV-1 in latently infected U1 cells [45]. In addition, it has been reported that host cell kinases are important for the replication of a number of viruses and targeting host kinases can provide a broad-spectrum of an-

8

Band

MW (kDa)

Protein

Band 1

DSG1: 113.8

PKM2, Jup, DSG1

Band 2

58K (monomer)

PKM2, ALMS1, GPRIN1,

Band 3

36.1

GAPDH

Band 4

66.4 JUP

DSP, JUP

tivirals [46,47]. The time-of-addition study has shown that effect of 4 on inhibition of HIV-1 induced CPE coincide with a step after transcription in HIV-1 life circle (Figure 5). Based on these observations, it can be proposed that 4 might target PKM2 to restrict replication of HIV-1. However, DARTS assay detected a few proteins not previously known to be associated with HIV-1 replication (Figure 6, Table 3). Now it has been realized that drugs can have more than one effective targets [42,48], thus we could not exclude that 4 might target additional proteins other than PKM2. Since host-targeting drugs are unlikely to generate resistance, further studies including derivatives of 4 in combination with ART might help devise antiretroviral therapies with improved resistance properties. However, the proposed mechanism mediating its antiviral activities at biochemical and molecular level have yet to be unequivocally elucidated. 3.7

Molecular docking of compound 4 with PKM2

To further rationalize how compound 4 could interact with PKM2, molecular docking studies were conducted. As was known, small-molecule activators could lock PKM2 into a high activity tetrameric conformation which is a dimer-ofdimers. The activator binding pockets are located at the interface of subunit interaction and lined with equivalent sets of residues provided by each of the PKM2 monomers. As revealed by crystal structures, the bound activator adopting

9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Ge et al.

Sci China Chem

two symmetric poses in the pocket [49−51]. Since the binding mode is the same in these two kinds of poses, we chose only one of them to analyze in the docking studies. The docking model of compound 4 to PKM2 is shown in Figure 7, where the ligand was accommodated through polar and hydrophobic interactions with pocket-lining residues. Particularly, the two hydroxyl groups made hydrogen bonds with the two parallel Y390 in each subunit separately. The ring C of compound 4 is sandwiched between the two parallel aromatic rings of F26.

4

Conclusions

We have used a small library of natural products to identify anti-HIVagents and their targets via an integrated phenotypic screening and proteomic approach. A group of steroids with potent anti-HIV activity were discovered, and the most promising compound 4 showed anti-HIV-1 activity with an

..............................................

9

EC50 at low micromolar range. Further studies on its mechanism of action showed that it might target PKM2 in the host cells. It would be interesting to further investigate the mechanism of the identified compound and optimization of its activity and cytotoxicity, specifically PKM2 involved in HIV-1 replication, and structure-activity relationship study of 4 derivatives. The molecular docking of compound 4 to PKM2 revealed that the two hydroxyl groups of 4 form hydrogen bonds with the two parallel Y390 in each subunit separately, and the ring C of 4 is sandwiched between the two parallel aromatic rings of F26. Compound 4 showed potent anti-HIV activities against HIV-1 wild-type and some clinically relevant multidrug resistant HIV-1 strains. We anticipate that the information obtained from this and future studies may help design novel antivirals that target cellular proteins to inhibit HIV replication with improved drug resistance profiles compared to currently used antivirals. Acknowledgements This work was supported by the National Natural Science Foundation of China (21532007, U1302222), the “Personalized Medicines—Molecular Signature-based Drug Discovery and Development”, Strategic Priority Research Program of the Chinese Academy of Sciences (XDA12020321) and the Canadian Institutes for Health Research (CIHR). We thank Prof. S.-M. Huang of Hainan University for the identification of the plant material. The content is solely the responsibility of the authors and does not necessarily represent the official views of the CIHR. Conflict of interest interest.

The authors declare that they have no conflict of

Supporting information The supporting information is available online at http://chem.scichina.com and http://link.springer.com/journal/11426. The supporting materials are published as submitted, without typesetting or editing. The responsibility for scientific accuracy and content remains entirely with the authors.

Figure 7 Structural analysis of the docking pose of compound 4 in PKM2 activator pocket. (a) Interaction between a dimer of the tetrameric PKM2 and compound 4. The two PKM2 monomers are represented as cyan and green cartoons. The bound compound 4 is shown in yellow surface. (b) The interactions between compound 4 and the surrounding residues. Compound 4 is represented as a ball and stick model with carbon, nitrogen and oxygen atoms colored yellow, blue and red, respectively. The residues from the two monomers that are involved in the interaction are labeled and colored green and cyan. Hydrogen bonds are indicated by yellow dashed lines. (c) Key interactions between compound 4 and PKM2 analyzed by Ligplot+ (color online).

1 Quashie PK, Mesplède T, Wainberg MA. Curr Infect Dis Rep, 2013, 15: 85–100 2 Wainberg MA, Mesplède T, Quashie PK. Curr Opin Virol, 2012, 2: 656–662 3 Newman DJ, Cragg GM. J Nat Prod, 2016, 79: 629–661 4 Tu Y. Nat Med, 2011, 17: 1217–1220 5 Harvey AL, Edrada-Ebel RA, Quinn RJ. Nat Rev Drug Discov, 2015, 14: 111–129 6 Shen B. Cell, 2015, 163: 1297–1300 7 López-Vallejo F, Giulianotti MA, Houghten RA, Medina-Franco JL. Drug Discovery Today, 2012, 17: 718–726 8 Andrae-Marobela K, Ghislain FW, Okatch H, Majinda RRT. Curr Drug Metab, 2013, 14: 392–413 9 Savarino A. Expert Opin Inv Drugs, 2006, 15: 1507–1522 10 Eggert US. The why and how of phenotypic small-molecule screens. Nat Chem Biol, 2013, 9, doi: 10.1038/nchembio.1206 11 Zheng W, Thorne N, McKew JC. Drug Discov Today, 2013, 18: 1067– 1073 12 Nijman SMB. Nat Chem Biol, 2015, 11: 942–948 13 Kim TH, Ko Y, Christophe T, Cechetto J, Kim J, Kim KA, Boese AS, Garcia JM, Fenistein D, Ju MK, Kim J, Han SJ, Kwon HJ, Brondani V, Sommer P. PLoS ONE, 2013, 8: e68767 14 Derry M, Somasagara R, Raina K, Kumar S, Gomez J, Patel M, Agarwal R, Agarwal C. Curr Cancer Drug Targets, 2014, 14: 323– 336

10

...........................

Ge et al.

Sci China Chem

15 Pai MY, Lomenick B, Hwang H, Schiestl R, McBride W, Loo JA, Huang J. Methods Mol Biol, 2015, 1263: 287–298 16 Lee S, Yoon KD, Lee M, Cho Y, Choi G, Jang H, Kim BS, Jung DH, Oh JG, Kim GW, Oh JW, Jeong YJ, Kwon HJ, Bae SK, Min DH, Windisch MP, Heo TH, Lee C. Br J Pharm, 2016, 173: 191–211 17 Han YS, Xiao WL, Xu H, Kramer VG, Quan Y, Mesplède T, Oliveira M, Colby-Germinario SP, Sun HD, Wainberg MA. Antivir Chem Chemother, 2015, 24: 28–38 18 Andjelic CD, Planelles V, Barrows LR. Mar Drugs, 2008, 6: 528– 549 19 Zhou H, Chen B, Li RX, Sheng QH, Li SJ, Zhang L, Li L, Xia QC, Wang HY, Zeng R. Rapid Commun Mass Spectrom, 2005, 19: 3569– 3578 20 Cox J, Mann M. Nat Biotechnol, 2008, 26: 1367–1372 21 Schrödinger Release 2010: Maestro, New York: Schrödinger, 2010 22 Becke AD. J Chem Phys, 1993, 98: 5648–5652 23 Frisch MJ, Trucks G, Schlegel HB, Scuseria G, Robb M, Cheeseman J, Scalmani G, Barone V, Mennucci B, Petersson G. Gaussian 09, revision C-01. Wallingford: Gaussian Inc, 2009 24 Friesner RA, Banks JL, Murphy RB, Halgren TA, Klicic JJ, Mainz DT, Repasky MP, Knoll EH, Shelley M, Perry JK, Shaw DE, Francis P, Shenkin PS. J Med Chem, 2004, 47: 1739–1749 25 Aiello A, Fattorusso E, Menna M. Steroids, 1991, 56: 513–517 26 Nes WR, Varkey TE, Crump DR, Gut M. J Org Chem, 1976, 41: 3429–3433 27 Nes WR, Varkey TE. J Org Chem, 1976, 41: 1652–1653 28 Sright JLC, McInnes AG, Shimizu S, Smith DG, Walter JA, Idler D, Khalil W. Can J Chem, 1978, 56: 1898–1903 29 Yue JM, Chen SN, Lin ZW, Sun HD. Phytochemistry, 2001, 56: 801– 806 30 Wu SB, Ji YP, Zhu JJ, Zhao Y, Xia G, Hu YH, Hu JF. Steroids, 2009, 74: 761–765 31 Zhao CC, Shao JH, Li X, Xu J, Zhang P. Arch Pharm Res, 2005, 28: 1147–1151 32 Zhan ZJ, Yang SP, Yue JM. J Org Chem, 2004, 69: 1726–1729 33 Wang XN, Fan CQ, Yin S, Gan LS, Yue JM. Phytochemistry, 2008, 69: 1319–1327 34 Zhang CR, Fan CQ, Dong SH, Liu HB, Zhou WB, Wu Y, Yue JM. Org Lett, 2011, 13: 2440–2443 35 Liu CP, Xu JB, Han YS, Wainberg MA, Yue JM. Org Lett, 2014, 16: 5478–5481 36 Xu JB, Zhang H, Gan LS, Han YS, Wainberg MA, Yue JM. J Am Chem Soc, 2014, 136: 7631–7633 37 Zhao JX, Liu CP, Qi WY, Han ML, Han YS, Wainberg MA, Yue JM. J Nat Prod, 2014, 77: 2224–2233 38 Yu JH, Wang GC, Han YS, Wu Y, Wainberg MA, Yue JM. J Nat Prod, 2015, 78: 1243–1252 39 Lalonde MS, Lobritz MA, Ratcliff A, Chamanian M, Athanassiou Z,

40

41

42

43 44 45 46

47 48

49

50

51

.............................................

10

Tyagi M, Wong J, Robinson JA, Karn J, Varani G, Arts EJ. PLoS Pathog, 2011, 7: e1002038 Lomenick B, Hao R, Jonai N, Chin RM, Aghajan M, Warburton S, Wang J, Wu RP, Gomez F, Loo JA, Wohlschlegel JA, Vondriska TM, Pelletier J, Herschman HR, Clardy J, Clarke CF, Huang J. Proc Natl Acad Sci USA, 2009, 106: 21984–21989 Chin RM, Fu X, Pai MY, Vergnes L, Hwang H, Deng G, Diep S, Lomenick B, Meli VS, Monsalve GC, Hu E, Whelan SA, Wang JX, Jung G, Solis GM, Fazlollahi F, Kaweeteerawat C, Quach A, Nili M, Krall AS, Godwin HA, Chang HR, Faull KF, Guo F, Jiang M, Trauger SA, Saghatelian A, Braas D, Christofk HR, Clarke CF, Teitell MA, Petrascheck M, Reue K, Jung ME, Frand AR, Huang J. Nature, 2014, 510: 397–401 Fleta-Soriano E, Martinez JP, Hinkelmann B, Gerth K, Washausen P, Diez J, Frank R, Sasse F, Meyerhans A. Microb Cell Fact, 2014, 13: 17 Jarboui MA, Bidoia C, Woods E, Roe B, Wynne K, Elia G, Hall WW, Gautier VW. PLoS ONE, 2012, 7: e48702 Barrero CA, Datta PK, Sen S, Deshmane S, Amini S, Khalili K, Merali S. PLoS ONE, 2013, 8: e68376 Sen S, Deshmane SL, Kaminski R, Amini S, Datta PK. J Cell Physiol, 2017, 232: 517–525 Mohr EL, McMullan LK, Lo MK, Spengler JR, Bergeron É, Albariño CG, Shrivastava-Ranjan P, Chiang CF, Nichol ST, Spiropoulou CF, Flint M. Antiviral Res, 2015, 120: 40–47 Perwitasari O, Yan X, O'Donnell J, Johnson S, Tripp RA. ASSAY Drug Dev Technol, 2015, 13: 638–649 Jurado KA, Wang H, Slaughter A, Feng L, Kessl JJ, Koh Y, Wang W, Ballandras-Colas A, Patel PA, Fuchs JR, Kvaratskhelia M, Engelman A. Proc Natl Acad Sci USA, 2013, 110: 8690–8695 Anastasiou D, Yu Y, Israelsen WJ, Jiang JK, Boxer MB, Hong BS, Tempel W, Dimov S, Shen M, Jha A, Yang H, Mattaini KR, Metallo CM, Fiske BP, Courtney KD, Malstrom S, Khan TM, Kung C, Skoumbourdis AP, Veith H, Southall N, Walsh MJ, Brimacombe KR, Leister W, Lunt SY, Johnson ZR, Yen KE, Kunii K, Davidson SM, Christofk HR, Austin CP, Inglese J, Harris MH, Asara JM, Stephanopoulos G, Salituro FG, Jin S, Dang L, Auld DS, Park HW, Cantley LC, Thomas CJ, Vander Heiden MG. Nat Chem Biol, 2012, 8: 839– 847 Kung C, Hixon J, Choe S, Marks K, Gross S, Murphy E, DeLaBarre B, Cianchetta G, Sethumadhavan S, Wang X, Yan S, Gao Y, Fang C, Wei W, Jiang F, Wang S, Qian K, Saunders J, Driggers E, Woo HK, Kunii K, Murray S, Yang H, Yen K, Liu W, Cantley LC, Vander Heiden MG, Su SM, Jin S, Salituro FG, Dang L. Chem Biol, 2012, 19: 1187–1198 Guo C, Linton A, Jalaie M, Kephart S, Ornelas M, Pairish M, Greasley S, Richardson P, Maegley K, Hickey M, Li J, Wu X, Ji X, Xie Z. Bioorg Med Chem Lett, 2013, 23: 3358–3363