Identification of Novel Vacuolin-1 Analogues as Autophagy Inhibitors by Virtual Drug Screening and Chemical Synthesis

Autophagy is a fundamental cellular degradation process which is essential for cell homeostasis, and dysfunctional autophagy has been associated with a variety of human diseases, such as cancer. Several autophagy chemical modulators have been applied in a number of preclinical or clinical trials against these autophagy related diseases, especially cancer. Small molecule vacuolin-1 potently and reversibly inhibits both endosomal-lysosomal trafficking and autophagosome-lysosome fusion, yet the molecular mechanisms underlying vacuolin-1 mediated autophagy inhibition remain unknown. Here, we first performed the virtual drug screening and identified 14 vacuolin-1 analogues as autophagy inhibitors. Based on these virtual screening results, we further designed and synthesized 17 vacuolin-1 analogues, and found that 13 of them are autophagy inhibitors and a couple of them are as potent as vacuolin-1. In summary, these studies expanded the pool of useful autophagy inhibitors and reveal the structural-activity relationship of vacuolin-1 analogues, which is useful for future development of vacuolin-1 analogues with high potency and for identification of the molecular targets of vacuolin-1.


Introduction
Autophagy contains at least three types, including macroautophagy, microautophagy, and chaperone-mediated autophagy, and macroautophagy (hereafter referred as autophagy) is the most common and best studied form. Autophagy is initiated when misfolded proteins or damaged organelles are sequestered by double-membrane autophagosomes, and is completed when autophagosomes fuse with lysosomes to form autolysosomes, inside which the sequestered contents are degraded by lysosomal acidic hydrolases into amino acids or fatty acids and recycled back to cytosol [1,2]. Autophagy is essential to maintain the cellular homeostasis by preventing damaged and harmful substances from accumulating inside cells, and thus is involved in a wide variety of cellular processes, from cell proliferation, differentiation, apoptosis, development, ageing, to immunity [3][4][5]. The initiation, elongation, membrane closure, and maturation of autophagic process are tightly controlled by the complicated interplay among several core molecular machineries, including the

Identification of Novel Vacuolin-1 Analogues by Virtual Drug Screening
Since the molecular targets of vacuolin-1 remains unknown, we performed a 2D similarity search from ZINC database which contains over 13 million compounds and J & K Chemical database which contains more than 700,000 molecules by the Pipeline Pilot 7.5 software, and identified 14 vacuolin-1 analogues with the 1,3,5-triazine scaffold (Table 1). We subsequently analyzed the effects of these compounds on autophagy in tandem fluorescence-tagged LC3B (tfLC3B) expressing HeLa cells by both western blot and confocal image analyses. As shown in Figure 1 and Figure S1, except VS11, the rest of the analogues all induced the accumulation of both lapidated LC3B-II and SQSTM1, an autophagy substrate, in a dose dependent manner, albeit exhibiting different potency. In addition, co-treatment of cells with the vacuolin-1 analogues (10 µM) and bafilomycin (BAF) (100 nM), an inhibitor of the vacuolar proton pump that blocks the fusion of autophagosomes with lysosomes, failed to further induce the accumulation of LC3B-II and SQSTM1 as compared to either drug alone ( Figure 2A). Likewise, treatment of cells with these vacuolin-1 analogues strongly induced the yellow LC3B-II puncta (representing the autophagosomes), not red-only LC3B-II puncta (representing autolysosomes) ( Figure 2B and Figure S2). Moreover, the induced LC3B-II puncta did not colocalize with lysosome-associated membrane protein 1 (LAMP1) (Figure 3). Taken together, these data indicate that the identified vacuolin-1 analogues via drug virtual screening, like vacuolin-1, can potently inhibit the fusion between autophagosome and lysosomes, resulting in the accumulation of autophagosomes thus blocking autophagic flux.
Comparing with vacuolin-1, VS1-VS8 whose R 1 groups were aromatic, maintained comparable autophagy inhibitory activity, whereas VS9-VS14 which bearing two morpholino groups showed obviously decreased potency. The difference of VS6 and VS7 suggested that electron-withdrawing R 2 substituted at the meta-position is more important for the autophagy inhibitory effects than para-position. autophagosomes), not red-only LC3B-II puncta (representing autolysosomes) ( Figure 2B and Figure S2). Moreover, the induced LC3B-II puncta did not colocalize with lysosome-associated membrane protein 1 (LAMP1) (Figure 3). Taken together, these data indicate that the identified vacuolin-1 analogues via drug virtual screening, like vacuolin-1, can potently inhibit the fusion between autophagosome and lysosomes, resulting in the accumulation of autophagosomes thus blocking autophagic flux. Comparing with vacuolin-1, VS1-VS8 whose R 1 groups were aromatic, maintained comparable autophagy inhibitory activity, whereas VS9-VS14 which bearing two morpholino groups showed obviously decreased potency. The difference of VS6 and VS7 suggested that electron-withdrawing R 2 substituted at the meta-position is more important for the autophagy inhibitory effects than para-position.
Comparing with vacuolin-1, VS1-VS8 whose R 1 groups were aromatic, maintained comparable autophagy inhibitory activity, whereas VS9-VS14 which bearing two morpholino groups showed obviously decreased potency. The difference of VS6 and VS7 suggested that electron-withdrawing R 2 substituted at the meta-position is more important for the autophagy inhibitory effects than para-position.

Identification of Novel Vacuolin-1 Analogues by Chemical Synthesis
To further explore the SAR of vacuolin-1 on autophagy, we performed a systematic modification. The modifications could be divided into three parts based on the structure of vacuolin-1: the N,N-diphenylamino group modified analogues A1-A6, the morpholino group modified analogues B1-B5, and the aromatic moiety attached to hydrazine (henceforward hydrazine area) modified analogues C1-C6. The synthesis of vacuolin-1 analogues (Scheme 1) was carried out from trichlorotriazine to afford dichlorotriazine derivatives (1a-1g). Desired monoaddition products were obtained by controlling temperature and the excess amount of cyanuric chloride. Then 1a-1f were converted to monochlorotriazines (2a-2f and 2h-2k) after reacting with aliphatic amines with yields ranging from 67% to 86%. Compound 2g was obtained from the reaction of 1g with the corresponding aromatic amine under reflux due to low reactivity of the amine. Introduction of hydrazine on suitably derivatized monochlorotriazines provided 3a-3k. Notably, the use of excess hydrazine hydrate converted the alkynyl group in 2e to an alkyl group in 3e. Condensation of 3a-3k with 3-iodobenzaldehyde in the presence of acetic acid respectively provided vacuolin-1, A1-A6, B1, B2, B4, and B5. Deprotection of the Boc group in B2 by TFA successfully produced B3.

Identification of Novel Vacuolin-1 Analogues by Chemical Synthesis
To further explore the SAR of vacuolin-1 on autophagy, we performed a systematic modification. The modifications could be divided into three parts based on the structure of vacuolin-1: the N,N-diphenylamino group modified analogues A1-A6, the morpholino group modified analogues B1-B5, and the aromatic moiety attached to hydrazine (henceforward hydrazine area) modified analogues C1-C6. The synthesis of vacuolin-1 analogues (Scheme 1) was carried out from trichlorotriazine to afford dichlorotriazine derivatives (1a-1g). Desired monoaddition products were obtained by controlling temperature and the excess amount of cyanuric chloride. Then 1a-1f were converted to monochlorotriazines (2a-2f and 2h-2k) after reacting with aliphatic amines with yields ranging from 67% to 86%. Compound 2g was obtained from the reaction of 1g with the corresponding aromatic amine under reflux due to low reactivity of the amine. Introduction of hydrazine on suitably derivatized monochlorotriazines provided 3a-3k. Notably, the use of excess hydrazine hydrate converted the alkynyl group in 2e to an alkyl group in 3e. Subsequently, we again analyzed the effects of these 17 vacuolin-1 analogues on autophagy in tfLC3B-expressing HeLa cells by both western blot and confocal image analyses. As shown in Figure 4, S3, and Table 2, analogues B3, B4, B5, and C5 failed to induce the accumulation of both LC3B-II and SQSTM1 at all doses, analogues B1, C2, and C6 only weakly induced the accumulation of LC3B-II and SQSTM1, analogues A1, A2, A3, A6, C1, and C3 showed good potency, and analogues A4 and A5 were as potent as vacoulin-1. Similarly, the positive analogues, e.g., A5 and others, markedly induced the yellow LC3B-II puncta in HeLa cells ( Figure 5A and Figure S4), and co-treatment of cells with these analogues and bafilomycin (BAF) failed to further induce the accumulation of LC3B-II and SQSTM1 as compared to either drug alone ( Figure 5B). Therefore, these data indicate that the synthesized vacuolin-1 analogues are autophagy inhibitors with good potency comparable to vacuolin-1.
Consistent with virtual screening results ( Figure 1 and Table 1), removing one phenyl from diphenylamino group (A2-A6) maintained the ability to inhibit autophagy ( Figure 4 and Table 2). The results not only suggested that aromatic R 1 is important to keep autophagy inhibitory activity, but also implied that modification is tolerable on the phenyl group, particularly, analogues A5 exhibited slightly stronger inhibitory potency than vacuolin-1 ( Figure 4 and Table 2). Notably, small changes in morpholine caused a great loss of potency (B1-B5), indicating the crucial role of the morpholine ring in autophagy inhibition. As for the substituent in hydrazine area (C1-C6), moderate changes were acceptable, while bulky substituents killed the drug effects. The structure-activity relationship of vacuolin-1 analogues as autophagy inhibitors is summarized in Figure 6.

2D Similarity Virtual Screening
The research was performed on ZINC database and two collections of the J & K Chemical database (K66-X4436 EXPRESS-Pick Collection and L62-X6857 2013feb_sc1). Calculated by Pipeline Pilot 7.5 software (Accelrys, San Diego, CA, USA), the similarity of molecules and vacuolin-1 will be scored by Tanimoto coefficient described by FCFP_4 fingerprint. Molecules with Tanimoto great than 0.6 were

2D Similarity Virtual Screening
The research was performed on ZINC database and two collections of the J & K Chemical database (K66-X4436 EXPRESS-Pick Collection and L62-X6857 2013feb_sc1). Calculated by Pipeline Pilot 7.5 software (Accelrys, San Diego, CA, USA), the similarity of molecules and vacuolin-1 will be scored by Tanimoto coefficient described by FCFP_4 fingerprint. Molecules with Tanimoto great than 0.6 were screened out.

Chemistry
All of chemicals and solvents used were obtained from J & K or Sigma. The solvents were dried by standard procedures. 1 H NMR spectra and 13 C NMR spectra were recorded at 400 MHz using Bruker Avance III spectrometer (Bruker BioSpin AG, Fällanden, Switzerland) in CDCl 3 or DMSO-d 6 solution with tetramethylsilane as the internal standard and chemical shift values were given in ppm. The NMR data was processed by software MestReNova (Ver.6.1.0, Mestrelab research S.L., Santiago de Compostela, Spain). The high resolution mass (HRMS) was measured on an FT-MS-Bruker APEX IV mass spectrometer.

General Procedure for the Synthesis of 1a-1g
An amine (1 eq.) in acetone solution was added dropwise to a stirred solution of cyanuric chloride (1.5 eq.) and K 2 CO 3 (1 eq.) in acetone at 0-5 • C. The resulting mixture was stirred at room temperature overnight. Ice water was added to the solution. The precipitate was collected by filtration and dried. The crude product was further purified by column chromatography.

General Procedure for the Synthesis of 2a-2k
General procedure for 2a-2f and 2h-2k: 1a-1g or 2h-2k (1 eq.), K 2 CO 3 (1 eq.) were suspended in DCM, cooled to 0-5 • C with an ice bath while stirred. Amine solution (1.2 eq., 0.3 mol/L solution of DCM) was added dropwise. The resulting mixture was stirred at room temperature for an additional 30 min, monitored by TLC until the starting material disappeared. The solvent was evaporated and the residue was purified by column chromatography.
Procedure for 2g: 1g (1 eq.) was dissolved in acetone, K 2 CO 3 (1 eq.) and amine (1.2 eq.) were added. The resulting mixture was stirred and refluxed overnight. The solvent was evaporated and partitioned between ethyl acetate and water. The organic phase was dried over with Na 2 SO 4 , filtered and concentrated in vacuo. The residue was purified by column chromatography.

General Procedure for the Synthesis of 3a-3k
2a-2k was dissolved in dioxane. Excess hydrazine hydrate (80%, m/m) was added to the solution while stirring. The reaction was monitored by TLC. After the staring material disappeared, water was added and then extracted using ethyl acetate (60 mL × 3). The combined organic phase was then washed with water, dried over with Na 2 SO 4 , filtered and concentrated in vacuo. The resulting solid (3a-3k) was used in the next condensing step without further purification.

General Procedure for the Synthesis of Vacuolin-1, A1-A6, B1-B5 and C1-C6
The reaction mixture under the conditions mentioned in Scheme 1 was stirred, and then evaporated to dryness under reduced pressure. The residue was purified by column chromatography.  13 13