International Journal of

Molecular Sciences Article

Identification of Direct Activator of Adenosine Monophosphate-Activated Protein Kinase (AMPK) by Structure-Based Virtual Screening and Molecular Docking Approach Tonghui Huang *, Jie Sun, Shanshan Zhou, Jian Gao

and Yi Liu *

Jiangsu Key Laboratory of New Drug Research and Clinical Pharmacy, School of Pharmacy, Xuzhou Medical University, Xuzhou 221004, China; [email protected] (J.S.); [email protected] (S.Z.); [email protected] (J.G.) * Correspondence: [email protected] (T.H.); [email protected] (Y.L.); Tel.: +86-516-8326-2137 (T.H.); +86-516-8326-2136 (Y.L.); Fax: +86-516-8326-2251 (T.H.); +86-516-8326-2136 (Y.L.) Received: 11 May 2017; Accepted: 27 June 2017; Published: 30 June 2017

Abstract: Adenosine monophosphate-activated protein kinase (AMPK) plays a critical role in the regulation of energy metabolism and has been targeted for drug development of therapeutic intervention in Type II diabetes and related diseases. Recently, there has been renewed interest in the development of direct β1-selective AMPK activators to treat patients with diabetic nephropathy. To investigate the details of AMPK domain structure, sequence alignment and structural comparison were used to identify the key amino acids involved in the interaction with activators and the structure difference between β1 and β2 subunits. Additionally, a series of potential β1-selective AMPK activators were identified by virtual screening using molecular docking. The retrieved hits were filtered on the basis of Lipinski’s rule of five and drug-likeness. Finally, 12 novel compounds with diverse scaffolds were obtained as potential starting points for the design of direct β1-selective AMPK activators. Keywords: Adenosine 50 -monophosphate-activated protein kinase; virtual screening; molecular docking; selective activator

1. Introduction Kidney disease associated with diabetes is the leading cause of chronic kidney disease (CKD) and end-stage kidney disease worldwide and nearly one-third of patients with diabetes develop nephropathy [1]. As the incidence of both types 1 and 2 diabetes rises worldwide, diabetic nephropathy (DN) is likely become a significant health and economic burden for society [2]. Current therapy for diabetic nephropathy includes glycemic optimization using antidiabetics and blood pressure control with blockade of the renin-angiotensin system [3]. However, these strategies are slow but cannot reverse or at least stop the disease progression [4]. Although several clinical trials are currently in progress, there are still no drugs approved for the treatment of DN. Among these ongoing phase 3 clinical trials, atrasentan is still in progress, while bardoxolone methyl and paricalcitol failed to meet the primary endpoint or was terminated on safety concerns [4,5]. Recently, there has been renewed interest in the development of direct β1-selective Adenosine monophosphate-activated protein kinase (AMPK) activators that have the potential to treat diabetic nephropathy [6]. AMPK is master sensor of cellular energy and plays a critical role in the regulation of metabolic homeostasis [7]. AMPK is a heterotrimeric kinase comprised of a highly conserved catalytic α subunit and two regulatory subunits (β and γ) [8]. The α subunit possess a N-terminal serine/threonine catalytic kinase domain (KD) that is followed by an autoinhibitory domain (AID) and a C-terminal Int. J. Mol. Sci. 2017, 18, 1408; doi:10.3390/ijms18071408

www.mdpi.com/journal/ijms

Int. J. Mol. Sci. 2017, 18, 1408

2 of 11

subunit and two regulatory subunits (β and γ) [8]. The α subunit possess a N-terminal serine/threonine catalytic kinase domain (KD) that is followed by an autoinhibitory domain (AID) Int. J. Mol. Sci. 2017, 18, 1408 2 of 11 and a C-terminal β subunit-binding domain [9]. The β subunit serves as a scaffold to bridge α and γ subunits that contains a glycogen binding domain (GBD) and a C-terminal domain [10]. The γ β subunit-binding domain The β subunitregion servesand as atwo scaffold to bridge α and γ subunits that subunit is composed of a β[9]. subunit-binding Bateman domains [11]. These seven contains a glycogen binding domain (GBD) and a C-terminal domain [10]. The γ subunit is composed subunits (α1, α2, β1, β2, γ1, γ2, and γ3) are encoded by separate genes, resulting in 12 different αβγ of a β subunit-binding region two Bateman domains [11]. These subunits (α1,are α2,not β1,fully β2, AMPK heterotrimers [12]. Theand distinct physiological functions of eachseven AMPK isoforms γ1, γ2, and γ3) arederive encoded by separate genes, resulting in 12 different αβγdifferent AMPK heterotrimers understood, but from differential expression patterns among tissues [13]. [12]. For The distinct functions AMPKevenly isoforms are not fully understood, but derive from instance, thephysiological α1 subunit appears toofbeeach relatively expressed in kidney, rat heart, liver, brain, differential expression patterns among tissues [13]. For instance, the in α1skeletal subunitmuscle, appearsheart, to be lung and skeletal muscle tissues, whiledifferent the α2 subunit is mainly expressed relatively evenly expressed in kidney, rat heart, liver, brain, lung and skeletal muscle tissues, while and liver tissues [14]. Among the two known β subunits, β1 subunit is highly abundant in kidneythe as α2 subunit by is mainly suggested mRNAexpressed levels [6].in skeletal muscle, heart, and liver tissues [14]. Among the two known β subunits, β1 subunit highly abundant in kidney as suggested mRNAto levels [6]. α1β1γ1 and More recently, a is direct AMPK activator PF-06409577 was by reported activate More recently, a direct AMPK activator PF-06409577 was reported to activate α1β1γ1 and α2β1γ1 AMPK isoforms with EC50 of 7.0 nM and 6.8 nM but was much less active against α2β1γ1 AMPK isoformsAMPK with EC nM EC50 and 6.8 nM but less active against α1β2γ1/α2β2γ1/α2β2γ3 isoforms with greater 4000was nMmuch [6]. Besides, compound 50 of 7.0 α1β2γ1/α2β2γ1/α2β2γ3 AMPKinisoforms with EC50 greater 4000 nMnephropathy. [6]. Besides,Compounds compound PF-06409577 exhibited efficacy a preclinical model of diabetic PF-06409577 exhibited efficacysimilar in a preclinical diabetic nephropathy.containing Compounds A-769662 and 991 possessed potency model towardofAMPK heterotrimers a β1A-769662 subunit and 991 possessed potency toward AMPKsite heterotrimers containing a β1 subunit as as PF-96409577 [15]. similar On the other hand, an allosteric of AMPK has been named allosteric drug PF-96409577 [15]. On the other hand, an allosteric site of AMPK has been named allosteric drug and metabolite site (ADaM site) [16], which was constructed by the catalytic kinase domain (KD) of and metabolite site regulatory (ADaM site) [16], which was constructed the catalytic kinase[13,17]. domain (KD) of α subunit and the carbohydrate-binding module by (CBM) of β subunit The three α subunit andAMPK the regulatory carbohydrate-binding module (CBM) subunit [13,17]. three known direct activators (PF-06409577 [6], A-769662 [18], and of 991β [19], Figure 1) allThe bound to known direct AMPK activators (PF-06409577 [6], A-769662 [18], and 991 [19], Figure 1) all bound to the the allosteric site and showed better potency for isoforms that contain the β1 subunit. This implies allosteric site and showed potency for isoforms that contain the β1 This implies that the allosteric site canbetter be used to design the selective activators of subunit. AMPK containing thethat β1 the allosteric site can be used to design the selective activators of AMPK containing the β1 subunit. subunit.

Figure 1. 1. Structures Structures of of reported reported direct direct AMPK AMPK activators. activators. Figure

The present study aims to investigate details of the domain structure and identify new potential The present study aims to investigate details of the domain structure and identify new potential β1-selective AMPK activators. Hence, sequence alignment and structural comparison were used to β1-selective AMPK activators. Hence, sequence alignment and structural comparison were used to identify the key amino acids that are involved in the interaction with activators and structure identify the key amino acids that are involved in the interaction with activators and structure difference difference between different subunits. Furthermore, molecular docking was performed for virtual between different subunits. Furthermore, molecular docking was performed for virtual screening screening to discover direct β1-selective AMPK activators. The screened retrieved hits were then to discover direct β1-selective AMPK activators. The screened retrieved hits were then subjected to subjected to several filters such as estimated activity and quantitative estimation of drug-likeness several filters such as estimated activity and quantitative estimation of drug-likeness (QED) [20,21]. (QED) [20,21]. Finally, 12 compounds with diverse scaffolds were selected as potential hit Finally, 12 compounds with diverse scaffolds were selected as potential hit compounds for the design compounds for the design of novel β1-selective AMPK activators. These findings provided a useful of novel β1-selective AMPK activators. These findings provided a useful molecular basis for the design molecular basis for the design and development of novel β1-selective AMPK activators. and development of novel β1-selective AMPK activators. 2. Results Results and and Discussion Discussion 2. 2.1. Sequence Sequence Alignment Alignment and and Structural Structural Comparison Comparison 2.1. To reveal reveal the selective potency of activators against the To the possible possible molecular molecularmechanism mechanismfor forthe the selective potency of activators against β1-isoform of of AMPK, sequence between the β1-isoform AMPK, sequenceand andsecondary secondary structure structure elements elements comparison comparison between carbohydrate-binding module were investigated. As As shown in Figure 2, the carbohydrate-binding module of of β1 β1and andβ2β2subunits subunits were investigated. shown in Figure 2, sequences that were boxed blue were located within the range of 5 Å of active site. Sequence the sequences that were boxed blue were located within the range of 5 Å of active site. Sequence alignment reveals reveals that that β1 β1 and and β2 β2 subunits alignment subunits shares shares 77.1% 77.1% sequence sequence identity. identity. As As shown shown in in Figure Figure 3, 3, superposition with the two subunits reveals a deflexion of sheet1 in β2 subunit as compared with β1 superposition with the two subunits reveals a deflexion of sheet1 in β2 subunit as compared with β1 subunit. The Phe-82 of β1 subunit corresponded to Ile-81 in β2 subunit, as well as the Thr-85 to Ser-84,

Int. J. Mol. Sci. 2017, 18, 1408 Int. 18,18, 1408 Int.J. J.Mol. Mol.Sci. Sci.2017, 2017, 1408

3 of 11 3 3ofof1111

subunit. The Phe-82 of β1 subunit corresponded to Ile-81 in β2 subunit, as well as the Thr-85 to subunit. The Phe-82 of β1 which subunitmay corresponded Ile-81 in β2 of subunit, as the Thr-85 to Ser-84, Gly-86 to Glu-85, account fortothe deflexion sheet1 as in well β2 subunit. The large Gly-86 to Glu-85, which may account for the deflexion of sheet1 in β2 subunit. The large aromatic Phe Ser-84, Gly-86 Glu-85, and which may Thr account deflexionaofbinding sheet1 in β2 subunit. largeof aromatic Phetoresidues small and for Glythepresented surface more The capable residues and small Thr and Gly presented a binding surface more capablesurface of accommodating ligand. aromatic Phe residues and small Thr and Gly presented a binding more capable of accommodating ligand. accommodating ligand.

Figure 2. Sequence alignment of carbohydrate-binding module from the β1 and β2 subunits. Figure 2. Sequence alignment of carbohydrate-binding module from the β1 and β2 subunits. Figure 2. Sequence alignmentthat of have carbohydrate-binding module from the Colon β1 and β2 subunits. Asterisks indicate positions a single, fully conserved residue. (green) indicates Asterisks indicate positions that have a single, fully conserved residue. Colon (green) indicates Asterisks indicate positions that have a single, fully conserved residue. Colon (green) indicates conservation between groups of strongly similar properties. Period (yellow) indicates conservation conservation between groups of strongly similar properties. Period (yellow) indicates conservation conservation between groupssimilar of strongly similarBlank properties. Period indicates conservation between groups of weakly properties. character (red)(yellow) indicates conservation between between groups of weakly similar properties. Blank character (red) indicates conservation between between groups of weakly similar properties. Blank character (red) indicates conservation between groups of strongly different properties. groups of strongly different properties. groups of strongly different properties.

Figure 3. Structural comparison of the scope within 5 Å of the active site from β1 and β2 subunits. Figure Structural comparison of theand scope within Å of theThe active β1 andwere β2 subunits. The α 3. subunit was shown in cartoon colored by5the cyan. β1 site and from β2 subunits shown in Figure 3. Structural comparison of the scope within 5 Å of the active site from β1 and β2 subunits. The α subunit shown cartoon colored by the cyan. and β2 subunits were acids shownwere in cartoon and was colored by in green andand blue, respectively. The The sitesβ1with different amino The α subunit was shown in cartoon and colored by the cyan. The β1 and β2 subunits were shown in cartoon colored by green blue, respectively. Theby sites shown and in line. The Ser-108 wasand shown in stick and colored red.with different amino acids were cartoon and colored by green and blue, respectively. The sites with different amino acids were shown shown in line. The Ser-108 was shown in stick and colored by red. in line. The Ser-108 was shown in stick and colored by red.

The sheet 2 torsion may attribute to the amino acid sequence differences of the sites of 106 and 2 torsion attribute to acid sequence differences of the sites of 106 and 107.The Thesheet most notablemay is supposed tothe theamino Ser108 (red and stick), an autophosphorylation site, The most sheet notable 2 torsion attribute to the amino acid sequence an differences of the sites of 106 107. The is may supposed to the Ser108 (red and autophosphorylation phosphorylated serine (pSer108) formed hydrogen bonds withstick), Thr-21, Lys-29, Lys-31, His-109′,site, and and 107. The most notable is supposed to the Ser108 (red andThr-21, stick),Lys-29, an autophosphorylation site, phosphorylated serine (pSer108) formed hydrogen bonds with Lys-31, His-109′, and Asn-111′ enhancing the ADaM site stabilization [22], and the phosphate group contributed to the phosphorylated serine (pSer108) formed hydrogen bonds Thr-21, Lys-29, His-1090 , the and Asn-111′ the[23]. ADaM site stabilization [22], and with the phosphate groupLys-31, contributed binding 0enhancing of activators The Gln-109′ and Asn-111′ were mutated to His-109′ and Asp-111′,towhich Asn-111of enhancing the ADaM site stabilization [22], andmutated the phosphate group to the binding [23]. The Gln-109′ andgenerated Asn-111′ were to His-109′ andcontributed Asp-111′, which abolished activators original hydrogen bonds and a large conformational change. We speculated 0 0 0 0 binding oforiginal activators [23]. The Gln-109 and Asn-111a large were mutated to His-109 and Asp-111 , which abolished hydrogen bonds and generated change. We speculated that the above differences between β1 and β2 may affect theconformational binding of activators to AMPk isoforms. abolished original hydrogen bonds and generated a large conformational change. We speculated that that the above differences between β1 and β2 may affect the binding of activators to AMPk isoforms. the above differences between β1 and β2 may affect the binding of activators to AMPk isoforms. 2.2. Parameter Setting for Molecular Docking 2.2. Parameter Setting for Molecular Docking Docking Setting parameters, which exert an important influence on molecular docking-based virtual 2.2. Parameter for Molecular Docking Dockingwere parameters, which exert anThe important influence of onPF-06409577 molecular docking-based screening, optimized in advance. crystal structure bound to the virtual α1β1γ1 Docking parameters, which exert an important influence on molecular docking-based virtual screening, were optimized in advance. The crystal structure of PF-06409577 bound to the α1β1γ1 AMPK isoform (PDB ID: 5KQ5) and A-769662 bound to the α2β1γ1 AMPK isoform (PDB ID: 4CFF) screening, were optimized in advance. The crystal structure of PF-06409577 bound to the α1β1γ1 AMPK isoformas(PDB ID: 5KQ5)the and A-769662 bound towere the α2β1γ1 AMPK isoform (PDB ID: were 4CFF)as were chosen the reference, docking parameters adjusted until the docked poses AMPK isoform (PDB ID: 5KQ5) and A-769662 bound to the α2β1γ1 AMPK isoform (PDB ID: 4CFF) were chosen as the reference, the docking parameters were adjusted until the docked poses were close as possible to the original crystallized structures. The ring flexibility was mainly consideredasin were chosen as the reference, the docking parameters were adjusted until the docked poses were as close possible to the original crystallized structures. ring flexibility wasoverlay mainly considered in finalasoptimized docking parameters according to theThe default settings. The of the original close as possible to the original crystallized structures. The ring flexibility was mainly considered in final optimized docking accordingand to the settings. The overlay of the original ligand from X-ray crystalparameters (stick and magenta) the default conformation from Surflex-Dock results (stick final optimized docking parameters according to the default settings. The overlay of the original ligand ligand from X-ray crystal (stick and magenta) and the conformation from Surflex-Dock results (stick and green) were shown in Figure 4, in which the indole moiety of PF-06409577 and terminal benzene from X-ray crystal (stick and magenta) and the conformation from Surflex-Dock results (stick and and were shown in Figure 4, indeflection which the and indole moiety andinteraction terminal benzene ringgreen) of A-769662 generated a little there wasofnoPF-06409577 effect on the between green) were shown in Figure 4, in which the indole moiety of PF-06409577 and terminal benzene ring ring of A-769662 little The deflection and bond there was no effectappeared on the interaction compounds andgenerated the activea site. hydrogen interactions consistent between with the of A-769662 generated a little deflection and there was no effect on the interaction between compounds compounds and and the the active The hydrogen bond(RMSD) interactions appeared consistent with the original ligands rootsite. mean square deviation between these two conformations are original ligands and the root mean square deviation (RMSD) between these two conformations are

Int. J. Mol. Sci. 2017, 18, 1408

4 of 11

and the active site. The hydrogen bond interactions appeared consistent with the original ligands 4 of 11 and the root mean square deviation (RMSD) between these two conformations are 0.53 and 0.56 Å, respectively. The molecular docking results indicated that the Surflex-Dock was reliable and could be 0.53 and 0.56 Å, respectively. The molecular docking results indicated that the Surflex-Dock was used for the further virtual screening. reliable and could be used for the further virtual screening. Int. J. Mol. Sci. 2017, 18, 1408

Figure 4. Conformation comparison of the original ligand from X-ray crystal (magenta and stick) and Figure 4. Conformation comparison of the original ligand from X-ray crystal (magenta and stick) the conformation from Surflex-Dock result (green and stick). (A): PF-06409577; (B): A-769662. The and the conformation from Surflex-Dock result (green and stick). (A): PF-06409577; (B): A-769662. indole moiety of PF-06409577 and terminal benzene ring of A-769662 generated a little deflection in The indole moiety of PF-06409577 and terminal benzene ring of A-769662 generated a little deflection compared with the original conformation. The hydrogen bond was labeled by red dashed lines. in compared with the original conformation. The hydrogen bond was labeled by red dashed lines.

2.3. High-Throughput Virtual Screening Procedure 2.3. High-Throughput Virtual Screening Procedure To identify new potent activators of AMPK, virtual screening was performed on the active sites To identify new potent activators of AMPK, virtual screening was performed on the active as mentioned previously. A chemical library containing with 1,500,000 commercially available sites as mentioned previously. A chemical library containing with 1,500,000 commercially available compounds (ChemDiv database) was docked to the molecular models of α1β1γ1 and α2β1γ1 compounds (ChemDiv database) was docked to the molecular models of α1β1γ1 and α2β1γ1 AMPK AMPK isoforms in silico, respectively. Prior to docking, the ChemDiv database was split into eight isoforms in silico, respectively. Prior to docking, the ChemDiv database was split into eight subsets subsets for molecular docking. About 600 top ranked compounds with high total-scores were for molecular docking. About 600 top ranked compounds with high total-scores were screened and screened and subsequently checked for their binding modes and interactions with the active site, subsequently checked for their binding modes and interactions with the active site, especially the especially the hydrogen bonds formed with the residuals of Asp-88, Lys-29, Lys-31, and Gly-19. hydrogen bonds formed with the residuals of Asp-88, Lys-29, Lys-31, and Gly-19. Then the potential hit Then the potential hit compounds were evaluated for their drug-likeness model scores using compounds were evaluated for their drug-likeness model scores using Lipinski’s rule of five (Table 1). Lipinski’s rule of five (Table 1). Finally, six potential hits with new scaffolds could serve as activators Finally, six potential hits with new scaffolds could serve as activators for α1β1γ1 AMPK isoform and for α1β1γ1 AMPK isoform and six for α2β1γ1 AMPK isoform were visually chosen from the top six for α2β1γ1 AMPK isoform were visually chosen from the top potential hits. potential hits. Table 1. The docking scores and drug-likeness model scores of selected activators for AMPK (α1β1γ1 Table 1. The docking scores and drug-likeness model scores of selected activators for AMPK (α1β1γ1 and α2β1γ1). and α2β1γ1). Isoforms Isoforms

AMPK AMPK (α1β1γ1) (α1β1γ1) activators activators

AMPK (α2β1γ1) AMPK activators (α2β1γ1) activators

Compound Compound No. Total-Score Total-Score CrashCrash Polar No. F064-1335 F064-1335 10.5010.50 −2.35−2.353.54 M5653-1884 M5653-1884 10.2410.24 −1.43−1.432.90 D454-0135 10.20 −1.58 D454-0135 10.20 −1.58 3.41 M8006-4303 10.07 −1.83 M8006-4303 10.079.93 −1.83−1.394.29 F264-3019 F264-3019 9.9310.03 −1.39−1.433.19 F377-1213 PF-06409577 F377-1213 10.037.29 −1.43−0.071.32 PF-06409577 7.2910.96 −0.07−2.463.08 L267-1138 F684-0053 L267-1138 10.9610.60 −2.46−2.772.78 C804-0412 F684-0053 10.6010.15 −2.77−3.274.31 M5976-1661 9.46 −0.92 C804-0412 10.15 −3.27 3.39 M039-0295 9.35 −1.61 M5976-1661 9.46 −0.92 M5050-0116 9.27 −1.351.32 M039-0295 9.35 7.44 −1.61−1.461.48 A-769662 991 M5050-0116 9.27 8.38 −1.35−0.962.82 A-769662 7.44 −1.46 1.26 991 8.38 −0.96 4.08

Number MolLog Drug-Likeness Drug-Likeness Number of of Polar Similarity Similarity HBA/HBDMolLogP P Model Score Score HBA/HBD Model 3.54 0.44 0.44 3.79 −0.16 6/1 6/1 3.79 −0.16 2.90 0.50 0.50 6.07 0.22 5/1 5/1 6.07 0.22 3.41 0.44 6/2 4.03 −0.31 0.44 6/2 4.03 −0.31 4.29 0.58 6/1 1.01 1.02 6/1 6/1 1.01 1.02 3.19 0.58 0.46 5.44 1.00 6/1 6/1 5.44 1.00 1.32 0.46 0.44 4.12 0.16 3.08 0.44 0.93 3.80 0.71 6/1 3/3 4.12 0.16 3/3 4/1 3.80 0.71 2.78 0.93 0.52 6.05 − 0.08 4.31 0.52 0.54 2.04 0.54 4/1 7/3 6.05 −0.08 3.39 0.54 0.54 2.53 1.00 7/3 5/2 2.04 0.54 1.32 0.46 6/0 4.84 0.62 0.54 5/2 2.53 1.00 1.48 0.50 6/1 2.66 −0.20 6/0 7/2 4.84 0.62 2.82 0.46 0.49 5.13 0.38 6/1 5/3 2.66 −0.20 1.26 0.50 0.93 3.46 0.30 4.08 0.49 0.73 5.38 0.41 7/2 4/2 5.13 0.38 0.93 5/3 3.46 0.30 0.73 4/2 5.38 0.41

Int. J. Mol. Sci. 2017, 18, 1408 Int. J. Mol. Sci. 2017, 18, 1408

5 of 11 5 of 11

2.4. Analysis of Binding Mode of Activators for α1β1γ1 AMPK Isoform 2.4. Analysis of Binding Mode of Activators for α1β1γ1 AMPK Isoform

Int. J. Mol. Sci. 2017, 18, 1408 of 11 The structures of retrieved hits as activators of α1β1γ1 AMPK isoform are shown in5 Figure 5. The structures of retrieved hits as activators of α1β1γ1 AMPK isoform are shown in Figure 5. Although these compounds possess different chemical scaffolds, they exhibit similar binding modes 2.4. Analysis of Binding Mode possess of Activators for α1β1γ1 AMPK Isoformthey exhibit similar binding modes Although these compounds different chemical scaffolds, at theatactive site. Among these compounds, compounds F064-1335 and M5653-1884 possess higher the active site. Among these compounds, compounds F064-1335 and M5653-1884 possess higher The structures of retrieved hits as and activators of α1β1γ1 AMPK isoform are shown in Figure 5. docking scores, compounds M8006-4303 F264-3019 have perfect drug-likeness model scores. docking scores, compounds M8006-4303 and F264-3019 have perfect drug-likeness model scores.

Although these compounds possess different chemical scaffolds, they exhibit similar binding modes at the active site. Among these compounds, compounds F064-1335 and M5653-1884 possess higher docking scores, compounds M8006-4303 and F264-3019 have perfect drug-likeness model scores.

Figure 5. Structures of retrieved hits targeting α1β1γ1 AMPK isoform from ChemDiv database. Figure 5. Structures of retrieved hits targeting α1β1γ1 AMPK isoform from ChemDiv database.

As shown in Figure 6A, the compound F064-1335 with the highest docking score (10.50) formed

Figure 5. Structures of retrieved hits targeting α1β1γ1 AMPK isoform from ChemDiv database. As shown in Figure 6A, theactive compound F064-1335 with the highest docking scoreestablished (10.50) formed several hydrogen bonds with site residues. The two oxygen atoms of sulfonamide several hydrogen bonds with with active site residues. twoLys-29, oxygen atoms of sulfonamide a a hydrogen bond network the side chain of The Lys-31, and Asn-111′. The carbonylestablished oxygen As shown in Figure 6A, the compound F064-1335 with the highest docking score (10.50) formed atom of the network ester group formed two chain hydrogen bonds with the and mainAsn-111 chain of0 . Lys-29, the anther hydrogen bond with the side of Lys-31, Lys-29, The carbonyl oxygen several hydrogen bonds with active site residues. The two oxygen atoms of sulfonamide established atom group of the ester group bound to the main chain Asn-48 by of a hydrogen bond, which atomoxygen the ester formed twowas hydrogen with the of main Lys-29, the anther oxygen aofhydrogen bond network with the side chainbonds of Lys-31, Lys-29, andchain Asn-111′. The carbonyl oxygen made the alkoxy group trend into a hydrophobic pocket formed by the Lys-51, Ile-52, Val-62, and atomatom of the group wasformed boundtwo to the main chain Asn-48 by a chain hydrogen bond, of ester the ester group hydrogen bondsof with the main of Lys-29, thewhich anthermade Leu-47. In addition, the carbonyl oxygen of benzoxazolone ring formed a hydrogen bond with the oxygengroup atom of the ester was bound pocket to the main chainby ofthe Asn-48 by a Ile-52, hydrogen bond,and which the alkoxy trend intogroup a hydrophobic formed Lys-51, Val-62, Leu-47. side chain of Arg-83′. made the alkoxy group trend into a hydrophobic pocket formed by the Lys-51, Ile-52, Val-62, and In addition, the carbonyl oxygen of benzoxazolone ring formed a hydrogen bond with the side chain 0 . In addition, the carbonyl oxygen of benzoxazolone ring formed a hydrogen bond with the Leu-47. of Arg-83

side chain of Arg-83′.

Figure 6. The binding modes of typical hit compounds for α1β1γ1 AMPK isoform. (A): F064-1335; (B): M5653-1884; (C): M8006-4303; (D): F264-0391. The α subunit was shown in cartoon and colored by cyan and the β subunit was shown in cartoon and colored by pink. The hydrogen bonds were Figure 6. The binding modes of typical hit compounds for α1β1γ1 AMPK isoform. (A): F064-1335; labeled with red dashed lines. Figure The binding modes of typical compounds for α1β1γ1 AMPKinisoform. (A):colored F064-1335; (B):6.M5653-1884; (C): M8006-4303; (D): hit F264-0391. The α subunit was shown cartoon and (B): M5653-1884; F264-0391. The subunitby was shown in cartoonbonds and colored by cyan and (C): the βM8006-4303; subunit was (D): shown in cartoon andα colored pink. The hydrogen were by cyan labeled and thewith β subunit waslines. shown in cartoon and colored by pink. The hydrogen bonds were labeled red dashed

with red dashed lines.

Int. J. Mol. Sci. 2017, 18, 1408

6 of 11

The compound M5653-1884 with a considerable docking score (10.24) and the bind mode is shown in Figure 6B. Four hydrogen bonds were observed between the compound and the active site residues. One carbonyl oxygen atom of 1,3-indandione formed hydrogen bond with the side chain of Lys-31, Int. J. Mol. Sci. 2017, 18, 1408 6 of 11 another carbonyl oxygen atom formed a hydrogen bond with the main chain of Val-11. The carbonyl The with ahydrogen considerable docking score (10.24) andthe theside bindchain mode of is Lys-29 oxygen atom ofcompound the amideM5653-1884 group showed bond interactions with shown in Figure 6B. Four hydrogen bonds were observed between the compound and the active site and Asn-111. In addition, there was a hydrophobic effect with the side chain of Ile-46, Asn-48, Asp-88, residues. One carbonyl oxygen atom of 1,3-indandione formed hydrogen bond with the side chain of and Phe-88. Lys-31, another carbonyl oxygen atom formed a hydrogen bond with the main chain of Val-11. The As carbonyl shown oxygen in Figure 6C, the compound of M8006-4303 exhibited similar binding mode as atom of the amide group showed hydrogen bond interactions with the side chain of PF-06409577. The ethanol attached to the piperazineeffect group participated in of two hydrogen Lys-29 and Asn-111. Ingroup addition, there was a hydrophobic with the side chain Ile-46, bond interactions with side chain of Gly-19 and Lys-31. The carbonyl oxygen atoms of Asn-48, Asp-88, and the Phe-88. As shown in Figurea6C, the compound of M8006-4303 exhibited mode pyrrolidine-2,5-dione formed hydrogen bond interaction with the sidesimilar chainbinding of Lys-29. In as addition, PF-06409577. The ethanol group attached to the piperazine group participated in two hydrogen the oxygen atom of oxygen butyl associated with the benzene ring accepted a hydrogen bond from the bond interactions with the side chain of Gly-19 and Lys-31. The carbonyl oxygen atoms of main chain of Asn-48. Within the cavity of the active site, Ile-47, Asn-48, Lys-51, and Ile-52 probably pyrrolidine-2,5-dione formed a hydrogen bond interaction with the side chain of Lys-29. In addition, generated hydrophobic effect. butyl associated with the benzene ring accepted a hydrogen bond from the aoxygen atom of oxygen Thethe binding mode of compound prefect drug-likeness score (1.00) shown in main chain of Asn-48. WithinF264-3091 the cavity with of thea active site, Ile-47, Asn-48, Lys-51, andwas Ile-52 probably generated a hydrophobic effect. Figure 6D. The oxgen atom of an oxyethyl group on the benzene ring participated in a hydrogen bond Thechain binding of compound F264-3091 withatom a prefect drug-likeness scoreshowed (1.00) wastwo shown with the main of mode Val-11. The carbonyl oxygen of the amide group hydrogen in Figure 6D. The oxgen atom of an oxyethyl group on the benzene ring participated in a hydrogen bonds with the main chain of Gln-19 and side chain of Lys-31. In addition, one hydrogen bond was bond with the main chain of Val-11. The carbonyl oxygen atom of the amide group showed two formed hydrogen between bonds the side chain of Asn-48 and the oxyethyl group connected with flavone B-ring while with the main chain of Gln-19 and side chain of Lys-31. In addition, one hydrogen 0. the B-ring showed a stacked cation-π with thetheside chaingroup of Val-83 bond was formed between the sideinteraction chain of Asn-48 and oxyethyl connected with flavone B-ring while the B-ring showed a stacked cation-π interaction with the side chain of Val-83′.

2.5. Analysis of Binding Mode of Activators for α2β1γ1 AMPK Isoform 2.5. Analysis of Binding Mode of Activators for α2β1γ1 AMPK Isoform

The chemical structures of six compounds as activators of α2β1γ1 AMPK isoform are shown in Themolecular chemical structures six compounds as activators α2β1γ1 AMPK isoform are higher shown indocking Figure 7. The dockingofresults indicated that all of the compounds possess Figure 7. The molecular docking results indicated that all the compounds possess higher docking scores than A-769662 and 991. The binding modes of the representative compound M2958-7438 and scores than A-769662 and 991. The binding modes of the representative compound M2958-7438 and M5050-0116 in the in active site of AMPK areshown shown in Figure M5050-0116 the active siteα2β1γ1 of α2β1γ1 AMPKisoform isoform are in Figure 8. 8.

Figure 7. Structures of retrieved hits targeting α2β1γ1 AMPK isoform from ChemDiv database.

Figure 7. Structures of retrieved hits targeting α2β1γ1 AMPK isoform from ChemDiv database.

As shown in Figure 8A, six hydrogen bonds were formed between the compound M2958-7438 and active site residues, in which the barbituric acid ring formed three hydrogen bonds with the side chain of Asp-88, making prominent contributions to the high docking score (10.04). The oxygen atom of the anisole associated with the barbituric acid ring accepted a hydrogen bond from the side chain of Lys-29, and two oxygen atoms in the linker participated in two hydrogen bonds with the side chain of Lys-31. In addition, the barbituric acid ring generated a stacked cation-π interaction with the side chain of Arg-830 . The compound M5050-0116 with a docking score of 9.27 and formed four hydrogen bonds with Val-11, Leu-18, Lys-29, and Asn-1110 . As shown in Figure 8B, the Lys-29 of α subunit and

Int. J. Mol. Sci. 2017, 18, 1408

7 of 11

Asn-1110 of β subunit, simultaneously coordinated the oxygen atom of dibenzofuran with hydrogen bonds. The hydroxyl group attached on pyrimidine anchored in a suitable geometry and formed two hydrogen bonds with the main chain of Glu-19 and the side chain of Val-11. Additionally, the compound M5050-0116 exhibited hydrophobic interactions with several residues, which formed a hydrophobic pocket including Ile-46, Leu-47, Asn-48, Asp-88, and Phe-90. Int. J. Mol. Sci. 2017, 18, 1408

7 of 11

Figure 8. The binding modes of typical activators for α2β1γ1 AMPK isoform. (A): M2958-7438; (B): Figure 8. The binding modes of typical activators for α2β1γ1 AMPK isoform. (A): M2958-7438; M5050-0116. The α subunit was shown in the illustration and colored cyan and the β subunit was (B): M5050-0116. The α subunit was shown in the illustration and colored cyan and the β subunit was shown in the illustration and colored pink. The hydrogen bonds were labeled with red dashed lines. shown in the illustration and colored pink. The hydrogen bonds were labeled with red dashed lines.

As shown in Figure 8A, six hydrogen bonds were formed between the compound M2958-7438 2.6. andBiological active siteActivities residues, in which the barbituric acid ring formed three hydrogen bonds with the side chainThe of Asp-88, makingcompounds prominent contributions to the high docking scorewere (10.04). The oxygen six screened based on α1β1γ1 AMPK isoform evaluated for atom activities of the anisole associated with the barbituric acid ring accepted a hydrogen bond from the side chain against AMPK (α1β1γ1 isoform) at a dosages of 2 µM. A-769662, a known β1-selective AMPK activator, of Lys-29, and two oxygen atoms in the linker participated in two hydrogen bonds with the side was used as a control. The preliminary in vitro assay (Figure 9) indicated that most of the selected chain of Lys-31. In addition, the barbituric acid ring generated a stacked cation-π interaction with α1β1γ1 AMPK activators displayed promising activation potency against α1β1γ1 AMPK isoform. the side chain of Arg-83′. Compounds D454-0135 and F264-3019 displayed comparable activation activity against AMPK in the The compound M5050-0116 with a docking score of 9.27 and formed four hydrogen bonds with comparison with the known β1-selective AMPK activator A-769662. Morever, compounds M8006-4303 Val-11, Leu-18, Lys-29, and Asn-111′. As shown in Figure 8B, the Lys-29 of α subunit and Asn-111′ of and F264-3019 showed stronger activation against α1β1γ1 AMPK A-769662. Compound β subunit, simultaneously coordinated the activities oxygen atom of dibenzofuran withthan hydrogen bonds. The M563-1884 with a higher logP value and showed a relatively low activation activity among the assayed hydroxyl group attached on pyrimidine anchored in a suitable geometry and formed two hydrogen compounds. This implies that the lipophilicity may play an important role in the bioavailability bonds with the main chain of Glu-19 and the side chain of Val-11. Additionally, the compound for Int. J. Mol. Sci. 2017, 18, 1408 8 of 11 the compound. M5050-0116 exhibited hydrophobic interactions with several residues, which formed a hydrophobic pocket including Ile-46, Leu-47, Asn-48, Asp-88, and Phe-90. 2.6. Biological Activities The six screened compounds based on α1β1γ1 AMPK isoform were evaluated for activities against AMPK (α1β1γ1 isoform) at a dosages of 2 µM. A-769662, a known β1-selective AMPK activator, was used as a control. The preliminary in vitro assay (Figure 9) indicated that most of the selected α1β1γ1 AMPK activators displayed promising activation potency against α1β1γ1 AMPK isoform. Compounds D454-0135 and F264-3019 displayed comparable activation activity against AMPK in the comparison with the known β1-selective AMPK activator A-769662. Morever, compounds M8006-4303 and F264-3019 showed stronger activation activities against α1β1γ1 AMPK than A-769662. Compound M563-1884 with a higher logP value and showed a relatively low activation activity among the assayed compounds. This implies that the lipophilicity may play an important role in the bioavailability for the compound.

9. Activation of AMPK (α1β1γ1isoform) isoform) by screened compounds was measured using Figure 9. Figure Activation of AMPK (α1β1γ1 bythe the screened compounds was measured using Elisa Kit. Elisa Kit.

3. Materials and Methods 3.1. Sequence Alignment and Structural Comparison Sequence alignment is an essential method for similarity/dissimilarity analysis of protein, DNA, or RNA sequences [24]. The software used for sequence alignment tasks include HAlign, BioEdit, EMBL-EBI, T-Coffee, and CLUSTAL [25,26]. The crystal structure data of AMPK (α1β1γ1: 5KQ5, 4QFG; α1β2γ1: 4REW and α2β1γ1: 4CFF) were obtained from RCSB Protein Data Bank

Int. J. Mol. Sci. 2017, 18, 1408

8 of 11

3. Materials and Methods 3.1. Sequence Alignment and Structural Comparison Sequence alignment is an essential method for similarity/dissimilarity analysis of protein, DNA, or RNA sequences [24]. The software used for sequence alignment tasks include HAlign, BioEdit, EMBL-EBI, T-Coffee, and CLUSTAL [25,26]. The crystal structure data of AMPK (α1β1γ1: 5KQ5, 4QFG; α1β2γ1: 4REW and α2β1γ1: 4CFF) were obtained from RCSB Protein Data Bank [6,8,13,19], as well as the amino acid sequences of carbohydrate-binding module. The amino acid sequences of carbohydrate-binding module (CBM) on β subunits were used to study the differences. The sequence alignment between α1β1γ1 isoform (PDB: 4QFG) and α1β2γ1 (PDB:4REW) isoform was performed and edited using BioEdit software (version 7.1.8) [27], which is a user-friendly biological sequence alignment editor and analysis program. The crystallographic structures of AMPK for molecular docking studying were added to the hydrogen atoms and the charge was given to the Gasteiger-Huckel. The crystal structures comparison was conducted by Sybyl X 2.1 (Tripos Associates Inc., S.H. R.: St. Louis, MO, USA.) [28] and the binding modes were generated by PyMOL V0.99 (Schrödinger, New York, NY, USA.) [29]. The polar hydrogen atoms were added to the crystal structures of the AMPK via the biopolymer module and the Gasteiger-Huckel charges were loaded on the atoms of proteins. The protein peptide backbones were shown in cartoon and colored by different colors, the side chains of the nonconservative amino acids were shown in line and colored by chain. 3.2. Molecular Docking The virtual screening and molecular docking studies were performed using Surflex docking module in Sybyl X 2.1. There were still some deficiencies due to the fact that the receptor was regarded as a rigid structure. Therefore, it was essential to optimize the docking parameters, the co-crystallized ligand was extracted and re-docked into the active site of the AMPK with the varied parameters, and then the conformation of the original ligand and the re-docking ligand were compared. The binding site was defined as a sphere containing the residues that stay within 5 Å from the co-ligand. The maximum conformations per fragment and maximum number of rotatable bonds per molecule were 20 and 100, respectively. Furthermore, the options for pre-dock minimization and post-dock minimization of molecules were omitted, while other parameters were set as default options. The top 20 conformational poses were selected according to the docking score. Dock scores were evaluated by Consensus Score (CScore), which integrates the strengths of individual scoring functions combine to rank the affinity of ligands bound to the active site of a receptor. 3.3. High-Throughput Virtual Screening High-throughput virtual screening was regarded as an important tool to identify novel lead compounds suitable for specific protein targets [30], and the screened compounds can be easily obtained from commercial sources for biological evaluation as well [31]. The ChemDiv database was supplied by Topscience Co. (Shanghai, China), which includes 1,500,000 compounds was employed for virtual screening through Surflex docking module in Sybyl X 2.1. To accelerate virtual screening, the maximum quantity of conformations was reduced from 20 to 10, the maximum quantity of rotatable bonds was decreased from 100 to 50, and the top six conformations were collected. The same as the molecular docking studies, the default optimization of molecules before and after the docking was canceled. Other parameters were kept as default values. Compounds PF-06409577 and A-769662 were severed as reference molecules, respectively. The compounds with the docking score (≥8.0) were extracted for further analyzing the interactions between ligand and active site, to this end, 100 compounds were collected to calculate the drug-likeness model score. Drug-likeness model scores were computed for hit compounds using the MolSoft software (MolSoft, San Diego, CA, USA) [32].

Int. J. Mol. Sci. 2017, 18, 1408

9 of 11

3.4. The In Vitro Activation Assay The in vitro preliminary kinase assays human α1β1γ1 AMPK were carried out according to the previous experimental method [33]. The screened compounds and α1β1γ1 AMPK isoform were provided by Topscience Co. (Shanghai, China) and Huawei Pharmaceutical Co. Ltd. (Shanghai, China), respectively. Generally, each of the evaluated compounds was dissolved in 10% Dimethyl sulphoxide at 10 µM and diluted to a required concentration with buffer solution. Then, 5 µL of the dilution was added to a 30 µL kinase assay buffer and 5 µL AMPK isoform per well. The solution was mixed at 0 ◦ C for 30 min. Next, 5 µL of AMARA petide and 5 µL Adenosine triphosphate (ATP) were added to the well. The enzymatic reactions were conducted at 30 ◦ C for 30 min. The AMPK activity was determined by quantifying the amount of ATP remaining in assay solution with Kinase-Glo Plus luminescent kinase assay kit (Promega, Madison, WI, USA). The luminescent signal is correlated with the amount of ATP present, while inversely correlated with the kinase activity. The mean values from three independent experiments were used for the expression of relative activities. A-769662, a β1-selective AMPK activator reported by Abbott laboratories, was used as a control. 4. Conclusions In summary, the sequence alignment and structural comparison was performed to identify the AMPK domain structure detail, which provides a molecular basis of selective AMPK activators on β1-containing isoforms. The key amino acid residues (Phe/Ile82, Thr/Ser85, Gly/Glu86, Thr/Ile106, Arg/Lys107, Gln/His109, Asn/Asp111) may contribute to the selectivity and provide a foundation for structure-based design of new direct β1-selective AMPK activators. Furthermore, the structure-based virtual screening workflow for the identification of selective activators of AMPK (α1β1γ1 and α2β1γ1) was established and six potential hit compounds for α1β1γ1 isoform and α2β1γ1 isoform were obtained, respectively. The preliminary assay indicated that most of the selected α1β1γ1 AMPK activators displayed promising activation potency. Overall, these findings revealed extensive interactions of activators and AMPK for rational design of novel selective AMPK activators. Further in vitro testing of retrieved hits is still in progress in our laboratory. Acknowledgments: This work was supported by the Postdoctoral Science Foundation funded project (2017M611916), Natural Science Foundation of Jiangsu Province (Grants No. BK20140225), Science and Technology Plan Projects of Xuzhou (Grants No. KC16SG249), Scientific Research Foundation for Talented Scholars of Xuzhou Medical College (No. D2014008), Xuzhou Medical University School of Pharmacy Graduate Student Scientific Research Innovation Projects (No. 2015YKYCX014). Author Contributions: Tonghui Huang and Jie Sun performed the sequence alignment and structural comparison; Shanshan Zhou and Jian Gao performed virtual screening study; Yi Liu and Tonghui Huang analyzed the data; Tonghui Huang wrote the paper. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4. 5.

Gallagher, H.; Suckling, R.J. Diabetic nephropathy: Where are we on the journey from pathophysiology to treatment? Diabetes Obes. Metab. 2016, 18, 641–647. [CrossRef] [PubMed] International Diabetes Federation. IDF Diabetes Atlas—7th Edition; IDF: Brussels, Belgium, 2015. Available online: http://www.diabetesatlas.org (accessed on 9 June 2016). Quiroga, B.; Arroyo, D.; de Arriba, G. Present and future in the treatment of diabetic kidney disease. J. Diabetes Res. 2015, 2015, 1–13. [CrossRef] [PubMed] Chan, G.C.; Tang, S.C. Diabetic nephropathy: Landmark clinical trials and tribulations. Nephrol. Dial. Transpl. 2016, 31, 359–368. [CrossRef] [PubMed] Perezgomez, M.V.; Sanchez-Niño, M.D.; Sanz, A.B.; Martin-Cleary, C.; Ruiz-Ortega, M.; Egido, J.; Navarro-González, J.F.; Ortiz, A.; Fernandez-Fernandez, B. Horizon 2020 in diabetic kidney disease: The clinical trial pipeline for add-on therapies on top of renin angiotensin system blockade. J. Clin. Med. 2015, 4, 1325–1347. [CrossRef] [PubMed]

Int. J. Mol. Sci. 2017, 18, 1408

6.

7. 8.

9.

10. 11.

12. 13.

14. 15. 16. 17. 18.

19.

20. 21. 22.

23.

24. 25.

10 of 11

Cameron, K.O.; Kung, D.W.; Kalgutkar, A.S.; Kurumbail, R.G.; Miller, R.; Salatto, C.T.; Ward, J.; Withka, J.M.; Bhattacharya, S.K.; Boehm, M.; et al. Discovery and preclinical characterization of 6-chloro-5-[4-(1-hydroxycyclobutyl)phenyl]-1H-indole-3-carboxylic Acid (PF-06409577), a direct activator of adenosine monophosphate-activated protein kinase (AMPK), for the potential treatment of diabetic nephropathy. J. Med. Chem. 2016, 59, 8068–8081. [PubMed] Hardie, D.G.; Ross, F.A.; Hawley, S.A. AMPK: A nutrient and energy sensor that maintains energy homeostasis. Nat. Rev. Mol. Cell Biol. 2012, 13, 251–262. [CrossRef] [PubMed] Li, X.D.; Wang, L.L.; Zhou, X.E.; Ke, J.Y.; Waal, P.W.; Gu, X.; Tan, M.H.; Wang, D.; Wu, D.; Xu, H.E.; et al. Structural basis of AMPK regulation by adenine nucleotides and glycogen. Cell Res. 2015, 25, 50–66. [CrossRef] [PubMed] Xiao, B.; Sanders, M.J.; Underwood, E.; Heath, R.; Mayer, F.V.; Carmena, D.; Jing, C.; Walker, P.A.; Eccleston, J.F.; Haire, L.F.; et al. Structure of mammalian AMPK and its regulation by ADP. Nature 2011, 472, 230–233. [CrossRef] [PubMed] Rana, S.; Blowers, E.C.; Natarajan, A. Small molecule adenosine 50 -monophosphate activated protein kinase (AMPK) modulators and human diseases. J. Med. Chem. 2014, 58, 2–29. [CrossRef] [PubMed] Miglianico, M.; Nicolaes, G.A.F.; Neumann, D. Pharmacological targeting of AMP-activated protein kinase and opportunities for computer-aided drug design: Miniperspective. J. Med. Chem. 2016, 59, 2879–2893. [CrossRef] [PubMed] Ross, F.A.; MacKintosh, C.; Hardie, D.G. AMP-activated protein kinase: A cellular energy sensor that comes in 12 flavours. FEBS J. 2016, 283, 2987–3001. [CrossRef] [PubMed] Calabrese, M.F.; Rajamohan, F.; Harris, M.S.; Caspers, N.L.; Magyar, R.; Withka, J.M.; Wang, H.; Borzilleri, K.A.; Sahasrabudhe, P.V.; Hoth, L.R.; et al. Structural basis for AMPK activation: Natural and synthetic ligands regulate kinase activity from opposite poles by different molecular mechanisms. Structure 2014, 22, 1161–1172. [CrossRef] [PubMed] Steinberg, G.R.; Kemp, B.E. AMPK in health and disease. Physiol. Rev. 2009, 89, 1025–1078. [CrossRef] [PubMed] Cameron, K.O.; Kurumbail, R.G. Recent progress in the identification of adenosine monophosphate-activated protein kinase (AMPK) activators. Bioorg. Med. Chem. Lett. 2016, 26, 5139–5148. [CrossRef] [PubMed] Langendorf, C.G.; Kemp, B.E. Choreography of AMPK activation. Cell Res. 2015, 25, 5–6. [CrossRef] [PubMed] Giordanetto, F.; Karis, D. Direct AMP-activated protein kinase activators: A review of evidence from the patent literature. Expert. Opin. Ther. Pat. 2012, 22, 1467–1477. [CrossRef] [PubMed] Cool, B.; Zinker, B.; Chiou, W.; Kifle, L.; Cao, N.; Perham, M.; Dickinson, R.; Adler, A.; Gagne, G.; Iyengar, R.; et al. Identification and characterization of a small molecule AMPK activator that treats key components of type 2 diabetes and the metabolic syndrome. Cell Metab. 2006, 3, 403–416. [CrossRef] [PubMed] Xiao, B.; Sanders, M.J.; Carmena, D.; Bright, N.J.; Haire, L.F.; Underwood, E.; Patel, B.R.; Heath, R.B.; Wlaker, P.A.; Hallen, S.; et al. Structural basis of AMPK regulation by small molecule activators. Nat. Commun. 2013, 4, 3017. [CrossRef] [PubMed] Daina, A.; Michielin, O.; Zoete, V. SwissADME: A free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci. Rep. 2017, 7, 42717. [CrossRef] [PubMed] Zuegg, J.; Cooper, M.A. Drug-Likeness and Increased Hydrophobicity of Commercially Available Compound Libraries for Drug Screening. Curr. Top. Med. Chem. 2012, 12, 1500–1513. [CrossRef] [PubMed] Scott, J.W.; Ling, N.M.; Issa, S.M.A.; Dite, T.A.; O’Brien, M.T.; Chen, Z.P.; Galic, S.; Langendorf, C.G.; Steinberg, G.R.; Kemp, B.E.; et al. Small molecule drug A-769662 and AMP synergistically activate naive AMPK independent of upstream kinase signaling. Chem. Biol. 2014, 21, 619–627. [CrossRef] [PubMed] Sanders, M.J.; Ali, Z.S.; Hegarty, B.D.; Heath, R.; Snowden, M.A.; Carling, D. Defining the mechanism of activation of AMP-activated protein kinase by the small molecule A-769662, a member of the thienopyridone family. J. Biol. Chem. 2007, 282, 32539–32548. [CrossRef] [PubMed] Kaya, M.; Sarhan, A.; Alhajj, R. Multiple sequence alignment with affine gap by using multi-objective genetic algorithm. Comput. Methods Progr. Biomed. 2014, 114, 38–49. [CrossRef] [PubMed] Zou, Q.; Hu, Q.H.; Guo, M.Z.; Wang, G.H. HAlign: Fast multiple similar DNA/RNA sequence alignment based on the centre star strategy. Bioinformatics 2015, 31, 2475–2481. [CrossRef] [PubMed]

Int. J. Mol. Sci. 2017, 18, 1408

26. 27.

28. 29.

30.

31.

32. 33.

11 of 11

Zou, Q.; Li, X.B.; Jiang, W.R.; Lin, Z.Y.; Li, G.L.; Chen, K. Survey of MapReduce frame operation in bioinformatics. Brief. Bioinform. 2014, 15, 637–647. [CrossRef] [PubMed] Gladue, D.P.; Baker-Bransetter, R.; Holinka, L.G.; Fernandez-Sainz, I.J.; O’Donnell, V.; Fletcher, P.; Lu, Z.Q.; Borca, M.V. Interaction of CSFV E2 protein with swine host factors as detected by yeast two-hybrid system. PLoS ONE 2014, 9, e85324. [CrossRef] [PubMed] Seeliger, D.; de Groot, B.L. Ligand docking and binding site analysis with PyMOL and Autodock/Vina. J. Comput. Aided Mol. Des. 2010, 24, 417–422. [CrossRef] [PubMed] Asokan, R.; Nagesha, S.N.; Manamohan, M.; Krishnakumar, N.K.; Mahadevaswamy, H.M.; Rebijith, K.B.; Prakash, M.N.; Sharath Chandra, G. Molecular diversity of Helicoverpa armigera Hubner (Noctuidae: Lepidoptera) in India. Orient. Insects 2012, 46, 130–143. [CrossRef] Dammganamet, K.L.; Bembenek, S.D.; Venable, J.W.; Castro, G.G.; Mangelschots, L.; Peeterst, D.C.G.; Mcallister, H.M.; Edwards, J.P.; Disepio, D.; Mirzadegan, T. A prospective virtual screening study: Enriching hit rates and designing focus libraries to find inhibitors of PI3Kδ and PI3Kγ. J. Med. Chem. 2016, 59, 4302–4313. [CrossRef] [PubMed] Lionta, E.; Spyrou, G.; Vassilatis, D.K.; Cournia, Z. Structure-Based Virtual Screening for Drug Discovery: Principles, Applications and Recent Advances. Curr. Top. Med. Chem. 2014, 14, 1923–1938. [CrossRef] [PubMed] Drug-Likeness and Molecular Property Prediction. Available online: http://www.molsoft.com/mprop/ (accessed on 9 June 2017). Kashem, M.A.; Nelson, R.M.; Yingling, J.D.; Pullen, S.S.; Prokopowicz, A.S., III; Jones, J.W.; Wolak, J.P.; Rogers, G.R.; Morelock, M.M.; Snow, R.J.; et al. Three Mechanistically Distinct Kinase Assays Compared: Measurement of Intrinsic ATPase Activity Identified the Most Comprehensive Set of ITK Inhibitors. J. Biomol. Screen. 2007, 12, 70–83. [CrossRef] [PubMed] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Identification of Direct Activator of Adenosine Monophosphate-Activated Protein Kinase (AMPK) by Structure-Based Virtual Screening and Molecular Docking Approach.

Adenosine monophosphate-activated protein kinase (AMPK) plays a critical role in the regulation of energy metabolism and has been targeted for drug de...
3MB Sizes 0 Downloads 9 Views