AutophagySMDB: a curated database of small molecules that modulate protein targets regulating autophagy

ABSTRACT

Macroautophagy/autophagy is a complex self-degradative mechanism responsible for clearance of non functional organelles and proteins. A range of factors influences the autophagic process, and disruptions in autophagy-related mechanisms lead to disease states, and further exacerbation of disease. Despite in-depth research into autophagy and its role in pathophysiological processes, the resources available to use it for therapeutic purposes are currently lacking. Herein we report the Autophagy Small Molecule Database (AutophagySMDB; http://www.autophagysmdb.org/) of small molecules and their cognate protein targets that modulate autophagy. Presently, AutophagySMDB enlists ~10,000 small molecules which regulate 71 target proteins. All entries are comprised of information such as EC50 (half maximal effective concentration), IC50 (half maximal inhibitory concentration), Kd (dissociation constant) and Ki (inhibition constant), IUPAC name, canonical SMILE, structure, molecular weight, QSAR (quantitative structure activity relationship) properties such as hydrogen donor and acceptor count, aromatic rings and XlogP. AutophagySMDB is an exhaustive, cross-platform, manually curated database, where either the cognate targets for small molecule or small molecules for a target can be searched. This database is provided with different search options including text search, advanced search and structure search. Various computational tools such as tree tool, cataloging tools, and clustering tools have also been implemented for advanced analysis. Data and the tools provided in this database helps to identify common or unique scaffolds for designing novel drugs or to improve the existing ones for autophagy small molecule therapeutics. The approach to multitarget drug discovery by identifying common scaffolds has been illustrated with experimental validation.

Abbreviations: AMPK: AMP-activated protein kinase; ATG: autophagy related; AutophagySMDB: autophagy small molecule database; BCL2: BCL2, apoptosis regulator; BECN1: beclin 1; CAPN: calpain; MTOR: mechanistic target of rapamycin kinase; PPARG: peroxisome proliferator activated receptor gamma; SMILES: simplified molecular input line entry system; SQSTM1: sequestosome 1; STAT3: signal transducer and activator of transcription

Introduction

Autophagy is a catabolic trafficking pathway deployed by cells for the destruction of misfolded protein aggregates and dysfunctional organelles, via regulated lysosomal degradation. In eukaryotes, autophagy occurs constitutively to maintain cellular homeostasis; however, it can be induced under several stimuli such as starvation, oxidative stress, and hormonal imbalance. Both its inadequate and overdriven functions thwart cell survival [1,2]. Due to the identification of a full set of essential autophagy genes and proteins in yeast, and their human orthologues, a greater molecular understanding of this fundamental process has been developed [3]. Research into autophagy has revealed its cardinal role in various pathophysiological processes such as cancer, cardiovascular diseases (including ischemic injury and myocardial infarction), and a number of neurodegenerative disorders (including Alzheimer disease [AD], Parkinson disease [PD], and Huntington disease [HD]) and also in infection [4–13]. This dynamic cellular homeostasis process is known to be controlled via a range of signaling proteins and transcription factors [14,15]. For example, MTOR (mechanistic target of rapamycin kinase), phosphoinositide 3-kinase (PI3K), AMP-activated protein kinase (AMPK, also named as protein kinase AMP-activated, or PRKA) and MAPKs (mitogen-activated protein kinases) are signaling proteins that regulate this process; whereas TFEB (transcription factor EB), E2F (E2F transcription factor), HIF1A/HIF-1 alpha (hypoxia inducible factor 1 subunit alpha), and NFKB/NF-ᴋB (nuclear factor kappa B) are examples of transcription factors. Moreover, druggable nuclear receptors such as VDR (vitamin D receptor), RAR (retinoic acid receptor), PPARG/PPAR-γ (peroxisome proliferator activated receptor gamma) and PPARA/PPAR-α (peroxisome proliferator activated receptor alpha), NR3C1/glucocorticoid receptor (nuclear receptor subfamily 3 group C member 1), AR (androgen receptor), and NR1D1/Rev-erb-α (nuclear receptor subfamily 1 group D member 1) regulate autophagy along with the ATG (autophagy related) proteins [16].

A growing body of evidence depicting its pathophysiological role, demands a meticulous understanding of this network of proteins and associated small molecules modulating autophagy [17–19]. Unfortunately, there is not any platform available on which all the information regarding these is accessible. Presently there are 3 autophagy databases: the Autophagy database, the Human Autophagy Database, and the Autophagy Regulatory Network. The Autophagy database comprehensively provides information on the ATG proteins across species [20]. The Human Autophagy Database is a database for integration and annotation of human genes described to regulate autophagy [21]. The Autophagy Regulatory Network provides information on both curated and predicted interactions of autophagy components, as well as lists several transcription factors and miRNAs that modulate autophagy [22]. However, none of these databases provide information on the small molecules that may regulate autophagy.

Recent research has highlighted the amenability of these protein targets to be regulated by an enormous number of small molecule modulators which would be critical for pharmacological intervention [23]. Of late, a number of resources have become available that give information on drugs or small molecules and their targets, such as drug bank, KEGG DRUG Database, and Orphan nuclear receptor ligand database (ONRLDB) [24–26]. PubChem and ChEMBL databases are known to harbor information on biologically active small molecules for various cognate targets which are having drug like properties. These databases compile information from high throughput screening analysis and are not adequately curated however, AutophagySMDB (Autophagy Small Molecule Database) specifically compiles curated information on all the known autophagy targets and related small molecules [27,28]. This platform can essentially prove to be an asset for the chemists and researchers in the field of drug discovery for autophagy related diseases. As the therapeutic value of autophagy inducers or inhibitors to human diseases is getting increasingly recognized, there is an urgent need of a database for effectual exploitation of existing information [29].

AutophagySMDB offers a platform for gathering and assessing significant information on proteins reported to regulate autophagy and their associated small molecule modulators from direct and indirect evidence, regulating autophagy. All small molecule entries include 2-dimensional (2D) and 3-dimensional (3D) structure files along with a comprehensive array of additional information including the molecular formula, International Union of Pure and Applied Chemistry (IUPAC) name, canonical SMILES (Simplified Molecular Input Line Entry System), function, PubMedID (PMID), U.S. Food and Drug Administration (FDA) approval status, and various experimental values. This database is designed to permit the selection of small molecules based on various properties and to discover common or unique scaffolds for designing new drugs or to improve existing ones.

Results

Database description

AutophagySMDB is a comprehensive database of all the known autophagy targets and their corresponding small molecule modulators. Autophagy can be modulated via several signaling components acting together with transcription factors and autophagy core machinery proteins. The layout of the AutophagySMDB home page depicts functions such as search, tools, browse, upload, literature and help (Figure 1). The database covers experimentally validated small molecules which modulate these signaling components, transcription factors, or autophagy core machinery proteins. High-throughput screening data were excluded to avoid the issues related with experimental invalidity. This database comprises of 71 target proteins and their ~10,000 small molecule modulators. The targets have been divided into 4 broad categories: signaling components (24 targets, ~2800 small molecules), transcription factors (31 targets, ~7000 small molecules), autophagy core machinery proteins (16 targets, ~30 small molecules) and miscellaneous (~35 molecules). This database also provides various user-friendly search options and several online tools for the analysis of small molecules. Consequently, our database provides a platform for specific and polypharmacological drug discovery for multiple diseases that are modulated by autophagy. Instructions for using AutophagySMDB have been described in a video on the help page of the website.

Figure 1.

Screenshot of the AutophagySMDB homepage. AutophagySMDB homepage displaying various functionalities such as search, tools, browse, upload and help.

Data retrieval from AutophagySMDB

The AutophagySMDB interface is comprised of several search options such as text search for both target and small molecule, advanced search, and structure search (Figure 2(a–d)). An explicit target search will provide results with a PDB image, and an expandable list of all cognate small molecules with their corresponding physiochemical and experimental properties. A small molecule search will display structure, image, IUPAC name, canonical SMILES, references and manually curated experimental information such as EC50, IC50, Ki and Kd, as well as physiochemical properties such as hydrogen donor and acceptor count, number of rotatable bonds, aromatic rings, molecular weight (MW) and program for prediction of octanol/water partition coefficients of organic compounds (XlogP) (Figure 2(e)). A specific small molecule search will provide information on targets and off-targets; this function will be useful for finding crossover multi-target small molecules. Link has also been implemented in the small molecule page to allow the user to access associated target pages or vice versa. Furthermore, by using the advanced search, users can be more specific in choosing small molecules based on the previously mentioned physiochemical or experimental properties. Specific, or combinations of properties can be selected via filters using an operator (e.g., equal to, not equal to, greater than or less than) and a value. This search can be further specified by selecting a specific target and/or function such as an activator, or inhibitor.

Figure 2.

Screenshots of the text, advanced and structure search from AutophagySMDB. (a and b) The dialog box for target text search (a), small molecule text search (b). (c and d) The dialog box for advanced search (c) and structure search (d). (e) Results of text search for a small molecule pyrazolopyrimidine_derivative_4, displaying QSAR properties of the small molecule, 2D, 3D structure download options along with IUPAC name, canonical SMILE, and experimental values like EC50, IC50 etc.

A structure search can be done in 3 ways: (i) by drawing the 2D structure, (ii) by entering SMILES or (iii) uploading the .sdf file or .mol file, for the small molecule. This search can be performed via similarity or substructure. By entering the SMILES format or uploading the SDF file users can retrieve similar compounds. To retrieve compounds by structure search, users can draw the complete or partial structure, using a 2D structure drawing interface.

Tools

AutophagySMDB is provided with various tools such as clustering, catalog, tree view and advanced catalog.

Clustering analysis

Clustering is a method that categorizes small molecules based on their structural backbone (scaffold) and it is a fundamental method for chemists and researchers in the process of drug discovery. We have implemented ChemMine tool (clone) for this purpose. Users can perform binning clustering, hierarchical clustering and multidimensional scaling analysis (MDS).

Catalog tool

Catalog tool implemented in this database can be used to filter small molecules based on target, function, MW and physiochemical properties. Structures of all the small molecules were embedded in Catalog to allow the visualization and to compare structures of small molecules.

Tree tool

By using tree tool users can retrieve small molecules modulators of one or more selected targets in a single window. The search result will also include information on multiple targets being regulated by small molecules. In addition to it, similarity and substructure features have also been incorporated. By using similarity analysis, users can compare and retrieve percentage similar small molecules between any 2 targets by using the different similarity index viz 100%, 80%, 60% and 40%. A similar analysis can be done with the substructure feature by drawing a substructure of a molecule on the provided Jmol java applet. Server script saves the query as a temporary file and compares it in selected targets to find structurally similar molecules. Additionally, the user can also perform the similarity analysis by uploading either SDF file or SMILES of the molecule. Search output will be in the form of names of the small molecules with their associated targets. Analysis can be retrieved as a PDF file by export result option. For illustrating how useful the tree tool would be for the users in drug discovery we have done a sample exercise described in Figure 3(a,b).

Figure 3.

Similarity search using tree tool. (a) Screenshot of tree page displaying 100% similar small molecules of MTOR and AMPK, and (b) Screenshot of tree page displaying 80% similar small molecules of MTOR and AMPK.

Advanced catalog

The advanced catalog has also been employed wherein a user can select 2 targets and apply options such as union, common, A-B, and B-A between those targets. Target selected first will be assigned as A and the target selected second will be assigned as B. Union option will give the complete list of all the small molecules in both the targets and common will give the result for existing common small molecules between both the targets. A-B and B-A options help to retrieve unique small molecules to each target. Similarity and substructure analysis can also be performed similarly to tree tool. After analysis output in advanced catalog depicts the structures of the small molecules unlike tree tool. Analysis can be retrieved as a PDF file by export result option. Figure 4(a,b) illustrate the functionality of advanced catalog.

Figure 4.

Similarity search using advanced Catalog. (a) Screenshot of advanced catalog page displaying 100% similar small molecules structures of MTOR and AMPK, and (b) Screenshot of advanced catalog page displaying 80% similar small molecules structures of MTOR and AMPK.

These tools together make database an interactive knowledgeable resource for multitarget drug discovery.

Convergence and dissonance of autophagy and disease

Most of the autophagy-modulating proteins listed in this database are also reported to influence various pathophysiological conditions such as cancer, infection, neurodegenerative and inflammatory diseases [30–32] (Table 1) [30,31,33–55].

Table 1.

List of target proteins known to modulate autophagy with their role in human diseases.

Target	Function	Disease Connection	References
MTOR	Phosphorylates ULK1 and inhibits autophagy	Cancer, diabetes, neurodegenerative disorders, obesity	[33]
phosphoinositide 3-kinase (PI3K)	Inhibit autophagy	Autoimmune disorders, cancer, inflammatory diseases	[34]
AMPK	Induce autophagy	Cardiovascular disease, neuronal injury, insulin resistance	[35,36]
MAPK1 (mitogen-activated protein kinase 1)	Regulate BECN1 expression and induce autophagy	Neurodegenerative diseases, cancer	[37]
BCL2	Inhibit autophagy	Neurodegenerative diseases, cancer	[38,39]
ROCK (Rho associated coiled-coil containing protein kinase)	Inhibit autophagy	Neurological disorders, cardiovascular disease, cancer	[40]
YWHAQ/14–3-3 protein (tyrosine 3- monooxygenase/tryptophan 5-monooxygenase activation protein theta)	Inhibit autophagy	Neurological disorders	[41]
ULK1 (unc-51 like autophagy activating kinase 1)	Induce autophagy	Crohn disease (CD)	[42]
PIK3C3/Vps34 (phosphatidylinositol 3-kinase catalytic subunit type 3)	Induce autophagy	Liver and heart disease	[43]
VDR (vitamin D receptor)	Induce expression of BECN1 and ATG5 and promote autophagy	Autoimmune disease, infection, cardiovascular disease	[30,44,45]
PPARG	Induce autophagy	Infection, cardiovascular disease	[31,46]
RAR (retinoic acid receptor)	Induce autophagy	Cancer, metabolic disease (diabetes and obesity)	[47]
NR1D1 (nuclear receptor subfamily 1 group D member 1)	Induce TFEB expression and autophagy	Infection, circadian disorders, obesity, diabetes	[48–50]
AR (androgen receptor)	Inhibit autophagy	Cancer, muscular atrophy	[51]
NR3C1 (nuclear receptor subfamily 3 group C member 1)	Induce autophagy	Asthma, COPD, ocular disorders	[52]
STAT3 (signal transducer and activator of transcription factor 3)	Inhibit autophagy	Autoimmune disease, leukemia, liver disease	[53–55]

In the case of cancer, autophagy plays a bimodal role. Tumorigenesis signaling pathways are well related with the regulation of autophagy [56]. Several tumor suppressor genes inhibit MTOR signaling, and thereby stimulate autophagy: TP53/p53 (tumor protein p53) positively modulates autophagy. The proto-oncogene BCL2/Bcl-2 (BCL2, apoptosis regulator) inhibits autophagy by binding to BECN1 (beclin 1), and thereby promotes tumor growth [56]. Genetic deletion of the autophagic gene BECN1 enhances susceptibility to breast, prostate, and ovarian cancers in humans [57]. At the same time basal autophagy has been shown to be upregulated in hypoxic tumors and also in RAS-transformed cancer cells [58]. Besides its role in the clearance of misfolded proteins, autophagy plays a critical role in the clearance of aggregated proteins which are generally associated with several neurodegenerative diseases [59]. As far as host immune system is concerned, autophagy modulates several processes like antigen uptake, killing of pathogens, T cell homeostasis and also inflammation [12]. Polymorphisms in autophagy-related genes contribute to tissue specific inflammatory diseases like inflammatory bowel disease and systemic lupus erythematosus [60]. Autophagy is also a well-documented defense mechanism against several disease causing pathogens including Mycobacterium tuberculosis, Salmonella and Listeria monocytogenes etc [61–63]. In all, this database may prove to be an ideal resource which can be utilized for targeting these diseases via small molecule modulators through regulating autophagy.

Possible usage of autophagysmdb for specific or polypharmacological drug discovery

Polypharmacological drug discovery opens a novel platform for rational drug designing. Polypharmacological phenomena includes: (a) single drug acting on multiple targets of a unique disease pathway, or (b) single drug acting on multiple targets pertaining to multiple disease pathways (c) In addition, polypharmacology for complex diseases is likely to employ multiple drugs acting on distinct targets that are part of networks regulating various physiological responses [64]. The use of a single drug that targets multiple factors involved in a distinct pathological condition may increase the efficacy of the treatment and limits negative aspects of a conventional single-target drug or a combination of multiple drugs. Autophagy is a dynamic cellular homeostasis event controlled by a range of signaling components and transcription factors and its deregulation is implicated in various pathologic processes such as neurodegenerative diseases, infectious diseases, cardiovascular diseases, cancer, and aging. Thus, multitarget modulation of autophagy is of great interest. AutophagySMDB consists of comprehensive information on cellular proteins and their small molecule modulators regulating the autophagic process in mammalian cells. By using this resource and associated tools one can generate unique and common scaffolds information of small molecules regulating key autophagy target proteins.

We have illustrated a polypharmacological (multitarget) approach that helps in designing small molecules that modulates autophagy by using information in AutophagySMDB. In our study we exemplified the polypharmacological phenomena of designing single drug acting on multiple targets in a distinct disease state. Common scaffolds among 2 targets have been identified by the ChemMine tool analysis (Figure 5(a,b)).

Figure 5.

Common scaffolds among autophagy targets. (a) Structures of common scaffolds in MTOR inhibitors and AMPK activators, (b) Structures of common scaffolds in calcium channel blockers and CAPN inhibitors.

Two master key regulators, MTOR and AMPK, modulate autophagy in opposite directions i.e. MTOR inhibits autophagy and AMPK activates autophagy. Given the importance of MTOR and AMPK in modulating autophagy, these kinases act as an attractive target for polypharmacological drug designing wherein we can modulate activities of both the targets by using a single small molecule to induce autophagy. Our idea is to design one single potent drug that simultaneously inhibits MTOR and activates AMPK thereby promoting autophagy. For that we compared the inhibitors of MTOR and activators of AMPK and identified common scaffolds among them. To identify the common scaffolds among small molecule modulators of MTOR and AMPK we performed binning clustering (Tanimoto coefficient 0.4) analysis using Chemmine tool. Upon comparison we observed 28 similar structures among these targets. We found 2 most abundant scaffolds among the 28 similar structures; (E)-[(1E)-3-(morpholin-4-yl)prop 1-en-1yl](phenylmethylidene)amine and (2E)-1-(4-hydroxyphenyl)-3-phenylprop-2-en-1-one by using maximum common substructure (MCS) analysis provided in ChemMine tool (Figure 5(a)). One of the common scaffolds, i.e. (E)-[(1E)-3-(morpholin-4-yl)prop 1-en-1yl](phenylmethylidene)amine is found in the molecules; 3_6_dihydro_2H_pyran_1, N_7_methylimidazolopyrimidine_1,tetrahydroquinazoline_derivative_1, dihydrofuropyrimidine_derivative_1, 4_morpholino_6_aryl_1H_pyrazolo_3_4_d_pyrimidines_1 and ETP-45,658, which are known to be the direct binders of MTOR. So, we were intrigued to check the cross modulation of these molecules in directly activating the upstream target AMPK in the modulation of autophagy. ETP-45,658 was commercially available as an MTOR inhibitor and we observed that it significantly induced the phosphorylation of AMPK similarly as observed in glucose starvation (Figure 6(a)) [65]. ETP-45,658 being an MTOR inhibitor also reduced the phosphorylation levels of MTOR and its substrate RPS6KB1/p70S6K (ribosomal protein S6 kinase B1) (Figure 6(a)). LC3-I to LC3-II conversion and SQSTM1/p62 (sequestosome 1) protein levels were checked upon treatment with ETP-45,658 to monitor autophagy (Figure 6(b)). Further, to check whether the increase in autophagy was due to increase in autophagic flux, LC3-II accumulation and SQSTM1 levels were also monitored upon treatment with ETP-45,658 in presence of bafilomycin (Figure 6(c)). These experiments were done in accordance with the autophagy interpretation guidelines [66]. Western blot analysis corroborated with the common scaffold structural interpretation i.e. modulating the 2 different targets MTOR inhibition and AMPK activation simultaneously. Experimental validation of common scaffold analysis confirmed the use of this scaffold structural information for the polypharmacological approach of designing a more potent drug which may overall induce autophagy by simultaneously activating AMPK and inhibiting MTOR.

Figure 6.

ETP-45,658 induces AMPKphosphorylation in HEK-293 cells. (a) Western blot analysis of signaling components AMPK, p-AMPK, MTOR, p-MTOR, RPS6KB1 and p- RPS6KB1. Glucose-free media and rapamycin were used as positive control for AMPK phosphorylation and MTOR dephosphorylation respectively (b and c) Western blot analysis of LC3-II and SQSTM1 upon treatment with ETP-45,658 (10 µM) for 4 h in presence and absence of bafilomycin A₁ (400 nM). LC3-I to LC3-II conversion is illustrated by LC3-II:ACTB ratios in bar graphs as in insets. Data is representative and mean ± SD from 3 independent experiments. *p < 0.05 as compared to control.

Another common scaffold, (2E)-1-(4-hydroxyphenyl)-3-phenylprop-2-en-1-one is found in 2 small molecules, baicalein and rottlerin; baicalein is well known to activate AMPK and inhibit MTOR complex components and thereby activate autophagy [67]. However, rottlerin is well known to activate autophagy by inhibiting MTOR signaling [68]. Some reports have also shown the effect of rottlerin on AMPK activation as well [69]. These reports strengthen our analysis related to common scaffold identification. This analysis of finding scaffolds would be a stepping stone for multitarget drug discovery. These scaffolds are not complete molecules but they can be used as a starting point for fragment based drug designing.

Autophagy acts as a protective mechanism during the initial stages of cancer and neurodegeneration. It has been reported that both calcium channels and CAPN (calpain) are hyperactivated and detrimental during these pathological conditions. We used small molecule information of AutophagySMDB to identify the common scaffolds in calcium channel blockers and CAPN inhibitors by using ChemMine tool. Upon comparison we observed 26 similar structures among these targets. We found 2 most abundant scaffolds among the 26 similar structures; 2-formamido-3-(4-methoxyphenyl)-N-(propan-2-yl)propanamide and 4-methyl-2-(methylamino)-N-(3-oxo-1-phenylbutan-2-yl)pentanamide (Figure 5(b)). Both of these common scaffolds i.e. 2-formamido-3-(4-methoxyphenyl)-N-(propan-2-yl)propanamide and 4-methyl-2-(methylamino)-N-(3-oxo-1-phenylbutan-2-yl)pentanamide are found in the molecules N_methyl_N_aralkyl_peptidylamines_derivatives_1, N_N_dialkyl_dipeptidylamines_derivatives_1, PD_167341, PD_173212 and CAT_811, which are well known to be the direct binders of calcium channels and CAPN, respectively. We investigated the cross modulation of calcium channel blockers in directly inhibiting the CAPN activity. PD_173212 was available commercially as a potent N-type calcium channel blocker. So, we checked for the direct inhibition of CAPN activity in the SH5YSY cell lysate and of the purified human CAPN upon treatment with PD_173212 in comparison to solvent control and a known CAPN inhibitor MDL-28,170. We observed that it was able to inhibit the CAPN activity significantly in comparison to solvent control (Figure 7(a,b)). Further, autophagy was monitored by conversion of LC3-I to LC3-II in SHSY5Y cells upon treatment with the PD_173212 and it was observed that PD_173212 was able to inhibit the CAPN activity significantly along with a slight increase in autophagy (Figure 7(c,d)). In order to improve its modulation of autophagy, we need to invoke fragment based drug discovery which employs scaffolds with special 3D binding modes; their fusion and designing to obtain a potent drug molecule. The common scaffold in PD_173212 can be fused to other suitable scaffolds to develop a novel potent drug or it can further be fragmentized to have a new scaffold altogether which can significantly induce autophagy by CAPN inhibition.

Figure 7.

Effect of PD-173,212 on CAPN activity. (a) SHSY5Y cell lysate was incubated with the indicated concentrations of PD-173,212 for 30 min and CAPN activity was assessed and represented as relative fluorescence units (360/450). A known inhibitor of CAPN, MDL-28,170, was used as a positive control. (b) Purified human CAPN (1 µg) was incubated at 37°C for 30 min with MDL-28,170 (30 µM), PD-173,212 (30 µM) and solvent vehicle control and CAPN activity was assessed using CAPN substrate (Suc-LLY-AMC) as described in the Materials and Methods section. (c and d) Western blot analysis of LC3-II upon treatment with PD-173,212 (30 µM) for 8 h in presence and absence of bafilomycin A₁ (400 nM). LC3-I to LC3-II conversion is illustrated by LC3-II:ACTB ratios in bar graphs as in insets. Data are representative and mean ± SD from 3 independent experiments. *p < 0.05 **p < 0.01 ***p < 0.001 as compared to control.

This exercise implies common scaffold structural information can be utilized in identifying the scaffolds which are common to distinct targets for the polypharmacological drug discovery. A similar analysis was done for identifying the unique scaffolds among these targets however, we were able to find 2 scaffolds; 3-[3,5-bis(methoxycarbonyl)-2,6-dimethyl-1,4-dihydropyridin-4-yl]-N-hydroxy-N-oxoanilinium and 1-(diphenylmethyl)-4-methylpiperazine in the calcium channel blockers only (Figure 8). Identification of the unique scaffold analysis would be beneficial in those cases where we need to design novel drugs which can target a specific autophagic modulator without affecting other modulators.

Figure 8.

Structures of unique scaffolds in calcium channel blockers.

Further, we have illustrated many small molecules having multiple targets which can be used as an initial point for drug discovery (Table 2) [68,70–89].

Table 2.

List of small molecules having multiple targets.

Small Molecules	Structure	Effect on Targets	Effect on Autophagy	PubMedIDs
Resveratrol		Induces AMPK Inhibits STAT3 Inhibits NLRP3	Induces autophagy	[70-72]
Raloxifene		Inhibits NR1H4/FXR Induces AMPK	Induces autophagy	[73,74]
PCI-24,781		Inhibits HDAC6 Induces AMPK	Induces autophagy	[75]
Nimodipine		Inhibits NR1H4 Calcium channel blocker	Induces autophagy	[73,76]
Nilotinib		Induces AMPK Inhibits ABL1/c-ABL	Induces autophagy	[77,78]
Niclosamide		Inhibits MTOR Inhibits STAT3	Induces autophagy	[68,79]
Belinostat		Inhibits HDAC6 Activates AMPK	Induces autophagy	[80,81]
Berbamine hydrochloride		Calcium channel blocker Inhibits CAPN	Induces autophagy	[82]
Curcumin		Inhibits EGFR Inhibits STAT3	Induces autophagy	[83,84]
Cryptotanshinone		Inhibits STAT3 Induces AMPK	Induces autophagy	[85,86]
Betulinic acid		Inhibits EGFR Inhibits CEBPB/C/EBP- beta	Dual modulator of autophagy	[83,87]
AZD8055		Inhibits MTOR Induces AMPK	Induces autophagy	[88,89]

Abbreviations: ABL1: ABL proto-oncogene 1, non-receptor tyrosine kinase; CEBPB: CCAAT enhancer binding protein beta; EGFR: epidermal growth factor receptor; HDAC6: histone deacetylase 6; NLRP3: NLR family pyrin domain containing 3; NR1H4: nuclear receptor subfamily 1 group H member 4.

Discussion

Given the importance of autophagy in many disease conditions and deeper investigation into its mechanism of action has revealed various small molecule modulators that could serve as potential tools for therapeutic purposes [12,56]. The detailed information regarding these small molecules is increasing day by day. Here we present a novel resource on autophagy regulating proteins and their small molecule modulators. AutophagySMDB is a comprehensive database featuring several target proteins from literature and small molecule modulating these proteins which are curated from all the direct and indirect evidence as complete as possible. Selective autophagy has been well reported and various proteins, e.g. SQSTM1, OPTN (optineurin), NBR1 (NBR1, autophagy cargo receptor), CALCOCO2 (calcium binding and coiled-coil domain 2), BNIP3L/NIX (BCL2 interacting protein 3 like), etc. regulating these pathways have been recently identified [90]. Information on small molecules targeting these proteins is mostly available from high throughput screening data therefore these proteins were not included in the database. However, literature related to small molecules targeting these selective autophagy modulators has been provided in the database in the search tab. We anticipate that this database opens new directions in autophagy research and will be valuable to the scientific community. This database allows users to access all the experimental information regarding small molecule interactions with their targets as well as their possible usage in drug development. To compile AutophagySMDB we reviewed over 2500 articles for all relevant information regarding the 71 targets and their small molecule modulators. Metric of relevant articles describing the role of autophagy targets and most cited small molecules has been provided in the database under the search tab. Due to the space constraint not all the reference literature has been included here, however, PubMedIDs of all the relevant articles surveyed for collecting and compiling the information have been provided in the database in the literature page.

To facilitate users in comparing, analyzing, and identifying scaffold similarity we have implemented structure, similarity, clustering, and cataloging tools into this database. Our database will allow users to design unique, common, or similar scaffolds. These functions will also assist in the rational designing of multitarget small molecules, which act on different target proteins that have similar biological functions. We have exemplified this approach to multitarget drug discovery by identifying common scaffold structural information using tools incorporated in the database and also validated it experimentally (Figures 5–7). The complexity of the biological systems limits the usage of small molecules as a therapeutic strategy, as their efficacy may vary among different systems. AutophagySMDB allows chemists and researchers in the field of autophagy to access information for polypharmacological drug designing and this information is not intended as a substitute for medical advice related to disease treatment. We have incorporated the most recent and highly curated information in the database. To cope up with the constantly increasing flow of information the data will be yearly evaluated, and updates will be done regularly for accuracy and to manifest advances in the field.

Materials and methods

Data collection

Data were collected from published literature and the information pertaining to small molecules was retrieved from the articles wherein small molecules were being tested in mammalian cells or in an in vivo setting. For each autophagy modulating target an exhaustive search for all experimental data and structural information was performed, excluding ambiguous data and bioassays. We have manually curated research articles, patents and reviews, as completely as possible. Additionally, we curated every reference cited from review articles. From each article, information regarding the small molecule of interest was gathered including, details of the targets, small molecule structure and function, EC50, IC50, Kd, and Ki, as well as the experimental procedure used to generate the information. The structures of the small molecules were either drawn using Marvin Sketch and saved as SDF files or files were downloaded from PubChem. FDA approval status of the small molecule has been collected from Drug Bank and mentioned in the comment section. Additionally, a list of FDA approved small molecules which modulate autophagy is mentioned under the ‘Search’ menu. Protein Data Bank (PDB) structures available for each target are mentioned under ‘Search’ menu and also included in the comment section of corresponding small molecules.

Database construction

AutophagySMDB web interface has been developed using the web server, Internet Information Server (IIS) 8.0 along with the database server, SQL server 2008 R2.

Clustering tool

The clustering tool ‘ChemMine’ is executed using the source code on the website (https://github.com/TylerBackman/chemminetools) [91]. Users can perform hierarchical clustering, MDS, and binning clustering of small molecules using the ChemMine tool (clone). SDF files, or the SMILES format, of the small molecules provided in the database can be used as the input for clustering analysis.

Tree

In order to generate the tree, the rotatable chart has been used from D3 chart library (https://d3js.org/). All targets are represented as nodes on the tree; by clicking on a specific node users can view all the small molecules associated with that particular target. By clicking on multiple nodes, either from the same category or different categories, users can view all small molecules between the selected targets.

Catalog and advanced catalog

The database provides catalog tools in order to filter and visualize the structures of compounds along with their physiochemical properties. The catalog is implemented in this database using Microsoft Silverlight PivotViewer which functions with the Java applet. Furthermore, our database features an advanced catalog where users can perform extensive cataloging by selecting 2 targets either within or between target categories, and can apply options such as union, common, A-B, and B-A.

Submitting a new target/molecule or program to the database

AutophagySMDB provides a provision for users to submit a new autophagy target or small molecule modulating autophagy under the upload section in the database. Instructions on how to upload a target and molecule have been provided in the help page of the database. Also, there is a provision for suggesting a new program or tool to be integrated in the upload section of the database for its upgradation. Users can suggest for new tools and programs along with their detailed configurations. Suggestions will be evaluated and if found compatible with the database configuration they will be added to the database to make it more valuable.

Cell lines

HEK-293 and SHSY5Y cell lines (ATCC) were cultured in Dulbecco’s modified Eagle media (DMEM) and DMEM/F12 media respectively supplemented with 10% fetal bovine serum and maintained under 5% CO₂ at 37°C in a humidified incubator.

Western blotting

Cell lysates were prepared by incubating the treated cells with cell lysis buffer on ice for 30 min and then centrifuged at 16,626 x g at 4°C. The supernatant was collected and the protein concentration was measured with the help of Bradford reagent (Sigma, B6916). For immunoblotting, the whole cell extract was resolved by SDS-PAGE and transferred to polyvinylidene difluoride membrane. Membranes were blocked with 5% bovine serum albumin (Merck Millipore, 12,659) in phosphate-buffered saline (pH 7.2) for 2 h and incubated overnight with primary antibodies. Primary antibodies were anti-LC3 (Sigma, L7543), anti-MTOR, p-MTOR, p-RPS6KB1, RPS6KB1 (Cell Signaling Technology, MTOR substrates antibody sampler kit- 9862), anti-SQSTM1 (BD Biosciences, 610,832) and anti-PRKAA1/2 (Santa Cruz Biotechnology, 25,792), p-PRKAA1/2 (Santa Cruz Biotechnology, 33,524), ACTB/β-actin (Santa Cruz Biotechnology, 47,778). Further, the membranes were incubated with HRP-conjugated secondary antibodies (Santa Cruz Biotechnology, 2313, 2314). The blots were visualized with Luminata Forte Western HRP substrate (Millipore, WBLUF0500).

CAPN activity assay

SHSY5Y cells were resuspended in lysis buffer and incubated on ice for 30 min and then centrifuged at 16,626 × g at 4°C. Protein content was measured using the Bradford reagent. CAPN activity was measured using Suc-LLY-AMC as the substrate provided in the CAPN activity assay kit (Merck Millipore, QIA120) as per the manufacturer’s instructions with minor modifications. CAPN activity was also measured with the purified human CAPN provided with the kit. After incubation, fluorescence was read in a Synergy H1 Hybrid Reader (BioTek Instruments, Inc., VT, USA) with excitation at 360 nm and emission at 450 nm and the enzyme activities were expressed as relative fluorescence units.

Disclaimer

The structural and experimental data presented here is collected from published literature. The appropriate references have been provided for all entries in AutophagySMDB. A lot of effort has gone in to make sure the data is scientifically correct. However, users are requested to refer to the original publications cited in the database for each entry. Small molecules are chemical compounds which can be promiscuous in nature. Given the complexity of the biological systems, the efficacy of these compounds can vary in different systems and they may have potential off-target effects. AutophagySMDB helps researchers to access the content for drug discovery purposes. AutophagySMDB is intended for educational and research purposes and not as a substitute for the medical advice or treatment.

Funding Statement

This work was supported by the Council of Scientific and Industrial Research (CSIR) 12th Plan Network project Genesis, Infectious Disease (BSC0121, BSC0210) to PG. This work is also supported by DBT-National Bioscience Award project (GAP-0162) to PG; Department of Biotechnology, Ministry of Science and Technology, National Bioscience Award project [GAP-0162].

Acknowledgments

We thank IMTECH, a constituent laboratory of the CSIR, for facilities and financial support.

Disclosure statement

No potential conflict of interest was reported by the authors.