Design, synthesis and biological evaluation of benzothiazepinones (BTZs) as novel non-ATP competitive inhibitors of glycogen synthase kinase-3b (GSK-3b)

European Journal of Medicinal Chemistry 61(2013)95-103 Contents lists available at sciverse ScienceDirect European Journal of Medicinal Chemistry ELSEVIER journalhomepagehttp://www.elsevier.com/locate/ejmech Original article Design, synthesis and biological evaluation of benzothiazepinones (btzs )as nov non-AtP competitive inhibitors of glycogen synthase kinase-3B(GSK-3 Peng Zhang a, Hai-Rong Hu, Shi-Hui Bian Zhao-Hui Huang a, Yong Chua, De-Yong Ye a.* Key Laboratory of State Genetics Engineering, School of life Sciences, Fudan University, 220 Handan Rd, Shanghai 200433, china A INFO ABSTRACT Glycogen synthase kinase-3B(GSK-3B)plays a key role in type ll diabetes and Alzheimers diseases, to eceived 1 March 2012 hich non-ATP competitive inhibitors represent an effectively therapeutical approach due to their good Received in revised form pecificity. Herein, a series of small molecules benzothiazepinones(Btzs)as novel non-ATP competitive 31July2012 Accepted 13 September 2012 inhibitors of GSK-3B have been designed and synthesized. The in vitro evaluation performed by lumi- vailable online 21 September 2012 ent assay showed most biz derivatives have inhibitory effects in micromolar scale. Among them ounds 61, 6t and 6v have the ICso values of 25.0 HM, 27.8 uM and 23.0 uM, respectively. Moreover 6v oid of any inhibitory activity in the assays to other thirteen protein kinases. Besides, SAR is analyzed hypothetical enzymatic binding mode is proposed by molecular docking study, which would be for new candidates design. e 2012 Elsevier Masson SAS. All rights reserved. nzothiazepinone Docking 1. Introduction effects. Normally these inhibitors would be inherently more pecific and usually show a far better true efficacy in cell or in vivo Glycogen synthase kinase-3B(GSK-3B) plays a key role in the assays than in vitro assays 9]. Thus, developing non-ATP compet egulation of many physiological responses in mammalian cells as itive GSK-3B inhibitors with high selectivity has already been a new a multifunctional serine/threonine protein kinase [1]. Its phos- hot spot for medicinal chemists in recent years. phorylation process controls a multitude of cellular processes, To our best knowledge, five kinds of chemical families of GSK-3B including gene transcription, metabolic pathways, cell growth and inhibitor with non-ATP competitive mechanism are already re- differentiation, as well as apoptosis [2,3]. Thus, inhibition of GSK-3B ported(see Fig. 1). It is worth noting that all of their Icso values are may represent a novel approach for the therapy of several human in micromolar level, suggesting poor activity in vitro of such type of diseases, such as type ll diabetes, Alzheimers diseases(AD), chronic inhibitors. The first ones are the thiadiazolidinones(tzds)[10]. inflammatory diseases, bipolar disorders and cancer [4-6] Among them, one compound called tideglusib(NP-12)is currently The search for GSK-3B inhibitors has spanned more than undergoing Phase llb clinical trials both on AD and orphan tauop a decade, and a number of structural diverse molecules that inhibit athy, and it is also the only gsK-3B inhibitor under clinical phase GSK-3B have already been reported in literature [7. However, the far [11. Halomethylketones(HMKs)are the second ones and have great majority of them might offer unfavorable off-target effects recently been viewed as the first irreversible GSK-3B inhibitors nsidering they are almost ATP competitors. Such inhibitors bind [12, 13. The irreversible inhibition is due to the formation of an competitively at the atp binding site of GSK-3B and are likely to act irreversible covalent sulfur-carbon bond between the hMK moiety on other undesirable protein kinases because of the highly and the amino acid residue Cys 199. This key amino acid residue conservation of aTP binding domain among more than 500 protein plays the role as 'gatekeeper' of GSK-3B [14 The small peptide kinases of human kinome [8 Nowadays non-ATP competitive L803-mts acts as substrate competitive inhibitor effectively in vi inhibitors are expected to be promising drugs for reducing adverse for neurological diseases and type ll diabetes [15 Lately, two marine natural products of alkaloid manzamine A and ses- rresponding authors. mailaddresses:cy110@fudan.edu.cn(YChu).dyye@shmueducn(D-YYe). ylation as cell permeable non-ATP competitive inhibitors 5, 16]. 0223-5234s-see front matter o 2012 1 p: //dx. doiorg /10.1016 j. ejmech 2012.0

Original article Design, synthesis and biological evaluation of benzothiazepinones (BTZs) as novel non-ATP competitive inhibitors of glycogen synthase kinase-3b (GSK-3b) Peng Zhang a , Hai-Rong Hu b , Shi-Hui Bian a , Zhao-Hui Huang a , Yong Chu a,*, De-Yong Ye a,* aDepartment of Medicinal Chemistry, School of Pharmacy, Fudan University, 826 Zhangheng Rd, Shanghai 201203, China b Key Laboratory of State Genetics Engineering, School of Life Sciences, Fudan University, 220 Handan Rd, Shanghai 200433, China article info Article history: Received 1 March 2012 Received in revised form 31 July 2012 Accepted 13 September 2012 Available online 21 September 2012 Keywords: GSK-3b Kinase inhibitor Non-ATP competitive Benzothiazepinones Selectivity Docking abstract Glycogen synthase kinase-3b (GSK-3b) plays a key role in type II diabetes and Alzheimer’s diseases, to which non-ATP competitive inhibitors represent an effectively therapeutical approach due to their good specificity. Herein, a series of small molecules benzothiazepinones (BTZs) as novel non-ATP competitive inhibitors of GSK-3b have been designed and synthesized. The in vitro evaluation performed by lumi￾nescent assay showed most BTZ derivatives have inhibitory effects in micromolar scale. Among them compounds 6l, 6t and 6v have the IC50 values of 25.0 mM, 27.8 mM and 23.0 mM, respectively. Moreover 6v is devoid of any inhibitory activity in the assays to other thirteen protein kinases. Besides, SAR is analyzed and a hypothetical enzymatic binding mode is proposed by molecular docking study, which would be useful for new candidates design.
2012 Elsevier Masson SAS. All rights reserved. 1. Introduction Glycogen synthase kinase-3b (GSK-3b) plays a key role in the regulation of many physiological responses in mammalian cells as a multifunctional serine/threonine protein kinase [1]. Its phos￾phorylation process controls a multitude of cellular processes, including gene transcription, metabolic pathways, cell growth and differentiation, as well as apoptosis [2,3]. Thus, inhibition of GSK-3b may represent a novel approach for the therapy of several human diseases, such as type II diabetes, Alzheimer’s diseases (AD), chronic inflammatory diseases, bipolar disorders and cancer [4e6]. The search for GSK-3b inhibitors has spanned more than a decade, and a number of structural diverse molecules that inhibit GSK-3b have already been reported in literature [7]. However, the great majority of them might offer unfavorable off-target effects considering they are almost ATP competitors. Such inhibitors bind competitively at the ATP binding site of GSK-3b and are likely to act on other undesirable protein kinases because of the highly conservation of ATP binding domain among more than 500 protein kinases of human kinome [8]. Nowadays non-ATP competitive inhibitors are expected to be promising drugs for reducing adverse effects. Normally these inhibitors would be inherently more specific and usually show a far better true efficacy in cell or in vivo assays than in vitro assays [9]. Thus, developing non-ATP compet￾itive GSK-3b inhibitors with high selectivity has already been a new hot spot for medicinal chemists in recent years. To our best knowledge, five kinds of chemical families of GSK-3b inhibitor with non-ATP competitive mechanism are already re￾ported (see Fig. 1). It is worth noting that all of their IC50 values are in micromolar level, suggesting poor activity in vitro of such type of inhibitors. The first ones are the thiadiazolidinones (TDZDs) [10]. Among them, one compound called tideglusib (NP-12) is currently undergoing Phase IIb clinical trials both on AD and orphan tauop￾athy, and it is also the only GSK-3b inhibitor under clinical phase so far [11]. Halomethylketones (HMKs) are the second ones and have recently been viewed as the first irreversible GSK-3b inhibitors [12,13]. The irreversible inhibition is due to the formation of an irreversible covalent sulfurecarbon bond between the HMK moiety and the amino acid residue Cys199. This key amino acid residue plays the role as ‘gatekeeper’ of GSK-3b [14]. The small peptide L803-mts acts as substrate competitive inhibitor effectively in vivo for neurological diseases and type II diabetes [15]. Lately, two marine natural products of alkaloid manzamine A and ses￾terterpene palinurin are reported to well decrease tau phosphor￾ylation as cell permeable non-ATP competitive inhibitors [5,16]. * Corresponding authors. E-mail addresses: cy110@fudan.edu.cn (Y. Chu), dyye@shmu.edu.cn (D.-Y. Ye). Contents lists available at SciVerse ScienceDirect European Journal of Medicinal Chemistry journal homepage: http://www.elsevier.com/locate/ejmech 0223-5234/$ e see front matter
2012 Elsevier Masson SAS. All rights reserved. http://dx.doi.org/10.1016/j.ejmech.2012.09.021 European Journal of Medicinal Chemistry 61 (2013) 95e103

P Zhang et aL/ European Journal of Medicinal Chemistry 61(2013)95-103 R C50=2-10M IC50=1-5uM Ho、O L803-mts Manzamine a s0=10.2M lCso=4.5μM Fig 1. Structures of reported non-ATP competitive GSK-3B inhibitors. erein, we report another novel small heterocyclic benzothia- cyclization of resulting acids 4a-4j with 2-aminobenzenethiol in nes(Btzs)as reversible non-ATP competitive inhibitors high temp gainst GSK-3B. This scaffold was found by us through a virtual 2, 3-dihydro-1, 5-benzothiazepin-4(5H)-one analogs 5a-5jwere creening for Maybridge database with Autodock 3.0.5. In this successfully obtained in moderate yields from recrystallization [17- docking approach, the X-ray crystal structure of GSK-3B kinase Through the electrophilic substitution of alkyl or acyl halides in PDB ID: 1UV5)was chosen as the model of receptor and a 3D non the nce of NaH at-10C for 30 min, various groups, such as AtP binding pocket composed of Arg 96, Lys 205 and Tyr 216 was alkyl, aromatic alkyl or acyl were introduced to the 5-position of the carefully constructed as screening receptor model. In the results BlZ framework to obtain nalogs 6a-6r, 6u-6y. It is wort several BlZ compounds were top ranked as the effective hits, and noting that unfavorably attacking of C3 position could be avoided wo of them were finally proved against GSK-3B with ICso values of his reaction condition[ 18. As the results, the yields could be high 830 HM and 480 HM respectively(Fig. 2). Inspired by its novel to a range of 80%-100% scaffold new btZ derivatives were designed and successfully Hydrolysis of compound 6m was carried out in a sealed tube synthesized for screening of potent candidates. Finally, some of with concentrated hydrochloric acid at 100C to give compound 6s them demonstrated an in vitro inhibition to GSK-3B in micromolar [19]. which was then treated with thionyl chloride in methanol range, and moreover the compound tested for specificity showed under reflux condition to afford the corresponding methyl ester 6t o activity to a panel of other 13 protein kinases, even the closest(Scheme 2) Cak-1/cyclin B and CK-ll. In this paper, the synthesis, biological The structures of all ew compounds are elu valuation and in silico studies, together with a preliminary their analytical and spectroscopic data(H, C NMR and MS)which structure-activity relationship(SAR)analysis will be described. are collected in the Experimental Section. The results support that the btz derivatives would be promising candidates for further development of pharmacotherapy. 2. Results and discussion 2. 1. Chemistry Compounds 6a-6y were synthesized following the general route described in Scheme 1 The synthesis of 3-substituted-acrylic acids 4a-4j was per- formed by the Knoevenagel condensation of various aromatic formaldehyde 3a-3j with propanedioic acid in the presence of 2:lC50=480pM pyridine and piperidine under reflux condition. Followed by the Fig. 2. Hits of 2, 3-dihydro-1,5-benzothiazepin-4(5H -ones from virtual screening

Herein, we report another novel small heterocyclic benzothia￾zepinones (BTZs) as reversible non-ATP competitive inhibitors against GSK-3b. This scaffold was found by us through a virtual screening for Maybridge database with Autodock 3.0.5. In this docking approach, the X-ray crystal structure of GSK-3b kinase (PDB ID: 1UV5) was chosen as the model of receptor and a 3D non￾ATP binding pocket composed of Arg 96, Lys 205 and Tyr 216 was carefully constructed as screening receptor model. In the results several BTZ compounds were top ranked as the effective hits, and two of them were finally proved against GSK-3b with IC50 values of 830 mM and 480 mM respectively (Fig. 2). Inspired by its novel scaffold, more new BTZ derivatives were designed and successfully synthesized for screening of potent candidates. Finally, some of them demonstrated an in vitro inhibition to GSK-3b in micromolar range, and moreover the compound tested for specificity showed no activity to a panel of other 13 protein kinases, even the closest Cdk-1/cyclin B and CK-II. In this paper, the synthesis, biological evaluation and in silico studies, together with a preliminary structureeactivity relationship (SAR) analysis will be described. The results support that the BTZ derivatives would be promising candidates for further development of pharmacotherapy. 2. Results and discussion 2.1. Chemistry Compounds 6ae6y were synthesized following the general route described in Scheme 1. The synthesis of 3-substituted-acrylic acids 4ae4j was per￾formed by the Knoevenagel condensation of various aromatic formaldehyde 3ae3j with propanedioic acid in the presence of pyridine and piperidine under reflux condition. Followed by the cyclization of resulting acids 4ae4j with 2-aminobenzenethiol in high temperature without any solvent, the key intermediates of 2,3-dihydro-1,5-benzothiazepin-4(5H)-one analogs 5ae5j were successfully obtained in moderate yields from recrystallization [17]. Through the electrophilic substitution of alkyl or acyl halides in the presence of NaH at
10 C for 30 min, various groups, such as alkyl, aromatic alkyl or acyl were introduced to the 5-position of the BTZ framework to obtain the analogs 6ae6r, 6ue6y. It is worth noting that unfavorably attacking of C3 position could be avoided in this reaction condition [18]. As the results, the yields could be high to a range of 80%e100%. Hydrolysis of compound 6m was carried out in a sealed tube with concentrated hydrochloric acid at 100 C to give compound 6s [19], which was then treated with thionyl chloride in methanol under reflux condition to afford the corresponding methyl ester 6t (Scheme 2). The structures of all the new compounds are elucidated from their analytical and spectroscopic data (1 H, 13C NMR and MS) which are collected in the Experimental Section. Fig. 1. Structures of reported non-ATP competitive GSK-3b inhibitors. Fig. 2. Hits of 2,3-dihydro-1,5-benzothiazepin-4(5H)-ones from virtual screening. 96 P. Zhang et al. / European Journal of Medicinal Chemistry 61 (2013) 95e103

P Zhang er al/ European Joumal of Medicinal Chemistry 61(2013)95-10 3a-3j 5a与 6a-6r,6u-6y Scheme 1.(a)CH2(COOH), pyridine, piperidine, 2 h, reflux:( b)2-aminobenzenethiol, 4 A molecular sieve, 6h, 180C: (c)RX, X=CI or Br or l,NaH,DMF,0.5h,-10C. 2.2. Biological evaluation and SAR study preincubation time were measured and the results were shown in Fig 3C. Normally, the inhibition effect of reversible inhibitors does 2.2.1. GSK-3B inhibitio not increase at different incubation time. while an irreversible All the newly prepared BTZ derivatives were evaluated for their inhibitor increases the inhibition percentage as it increases the time GSK-3B inhibitory activity by a recently well described Kinase-GloTm of incubation with the enzyme As could be seen in Fig. 3C, the luminescent technique, which is regarded as a safer nonradioactive activities of 6v were almost kept unchanged with the increasing of assay 20. GS-2 is the most utilized substrate of GSK-3B as a small the pre-incubation time, which indicates compound 6v acts as peptide similar to skeletal muscle glycogen synthase. It is a reversible GSK-3B inhibitor. composed of 26-amino acids and contains a prephosphorylated serine residue. In our assay, we used the prephosphorylated 12-2 23. Kinase selectivity studies amino acids polypeptide substrate 65 HSSPHQ(pS)EDEEE,which High selectivity of protein kinase inhibitors is critical to avoid is as same as the one used by Andrea Baki in a high throughput despread effects in a potential therapy. Thus for evaluating the GSK-3B enzyme was incubated with ATP and GS-2 in the presence sentative compound 6v was assayed against a panel of other 13 or absence of the tested compound, and then the amount of ATP kinases. These kinases include four serine/threonine kinases as emaining in solution, which inversely correlates to kinase activity, Cdk-1/cyclin B, CK-ll, PKA and PKCa which are highly close to GSK was quantified following the kinase reaction. The ICso values are 3B, and nine tyrosine kinases as Flt-1, KDR, PDGFR-B, EPH-A2, EGFR, listed in Table 1 and it can be seen that some of compounds really ErbB2, ErbB4, RON and Abl, which play important role in cancer inhibit GSK-3B in micromolar scale. Among them the best three signaling pathways. All of the inhibitory assays were carried out at compounds 61, 6t and 6v have the ICso values of 25.0 uM, 27. 8 uM a 100 uM of 6v concentration. The serine/ threonine kinases inhi- and 23.0 uM, respectively bition were performed at Millipore Corporation(Dundee, UK) by KinaseProfilerTM service, and the tyrosine kinases screen was 2. 2. 2. Mode of inhibition assayed utilizing an ELISA approach as described in Experimental In order to explore the biological mechanism of BTZ derivatives, section. The results listed in Table 2 showed that compound 6v compound 6v was used to study the kinetic features. We firstly tested almost displayed no inhibitory activity against the whole set of whether a competition effect exists between 6v and ATP. That is, kinases, which to some extent demonstrates that BTZs might have eping the concentration ofGS-2 unchanged the enzyme inhibitor good specificity with respect to GSK-3B activities of compound 6v were measured at its two differer concentrations separately while AtP concentrations varied. The 2. 2. 4. Structure-activity relationships (SAR)analysis results are showed in Fig 3A and it can be seen from the double In order to explore the features between the chemical reciprocal plotting of the data, in which compound 6v acts as a non structures and their GSK-3B inhibition, a preliminary structure ATP competitive inhibitor. Then we explored that whether there is activity relationship of BTzs was analyzed based on the preceding a competitive relationship between 6v and GS-2. In these experi- in vitro data, which were listed in Table 1. The substituents ments, GS-2 concentrations varied and atP concentration was kept attached to the n5 of the btz ring seemed to be important for unchanged as the activities of 6v were measured at concentrations of keeping the activities, of which the benzyl group was the best one 25 HM and 50 uM separately. The results were showed in Fig. 3B and(6f vs 6a-e)in the test set. Furthermore, ortho-nitro group and the double reciprocal plotting of the data in this figure indicates that meta-carbomethoxy group could dramatically contribute to the compound 6v acts as a noncompetitive inhibitor of GS-2 inhibitory potency as attached to the benzyl moiety( 61, 6t). To further investigate the interaction feature of Blzs to the On the other hand. the size and nature of the substituents r enzyme, the inhibitory activities of compound 6v at different attached to the c2 should also be crucial for inhibition. The b COOMe Scheme 2. Reagents and conditions: (a)conc. Ha1, 100oC,7 h; (b)SoCl, MeOH, reflux

2.2. Biological evaluation and SAR study 2.2.1. GSK-3b inhibition All the newly prepared BTZ derivatives were evaluated for their GSK-3b inhibitory activity by a recently well described Kinase-Glo luminescent technique, which is regarded as a safer nonradioactive assay [20]. GS-2 is the most utilized substrate of GSK-3b as a small peptide similar to skeletal muscle glycogen synthase. It is composed of 26-amino acids and contains a prephosphorylated serine residue. In our assay, we used the prephosphorylated 12- amino acids polypeptide substrate 650HSSPHQ (pS)EDEEE, which is as same as the one used by Andrea Baki in a high throughput luminescent assay for looking for GSK-3b inhibitors [20]. Briefly, GSK-3b enzyme was incubated with ATP and GS-2 in the presence or absence of the tested compound, and then the amount of ATP remaining in solution, which inversely correlates to kinase activity, was quantified following the kinase reaction. The IC50 values are listed in Table 1 and it can be seen that some of compounds really inhibit GSK-3b in micromolar scale. Among them the best three compounds 6l, 6t and 6v have the IC50 values of 25.0 mM, 27.8 mM and 23.0 mM, respectively. 2.2.2. Mode of inhibition In order to explore the biological mechanism of BTZ derivatives, compound 6v was used to study the kinetic features.We firstly tested whether a competition effect exists between 6v and ATP. That is, keeping the concentration of GS-2 unchanged, the enzyme inhibitory activities of compound 6v were measured at its two different concentrations separately while ATP concentrations varied. The results are showed in Fig. 3A and it can be seen from the double reciprocal plotting of the data, in which compound 6v acts as a non￾ATP competitive inhibitor. Then we explored that whether there is a competitive relationship between 6v and GS-2. In these experi￾ments, GS-2 concentrations varied and ATP concentration was kept unchanged as the activities of 6v were measured at concentrations of 25 mM and 50 mM separately. The results were showed in Fig. 3B and the double reciprocal plotting of the data in this figure indicates that compound 6v acts as a noncompetitive inhibitor of GS-2. To further investigate the interaction feature of BTZs to the enzyme, the inhibitory activities of compound 6v at different preincubation time were measured and the results were shown in Fig. 3C. Normally, the inhibition effect of reversible inhibitors does not increase at different incubation time, while an irreversible inhibitor increases the inhibition percentage as it increases the time of incubation with the enzyme. As could be seen in Fig. 3C, the activities of 6v were almost kept unchanged with the increasing of the pre-incubation time, which indicates compound 6v acts as a reversible GSK-3b inhibitor. 2.2.3. Kinase selectivity studies High selectivity of protein kinase inhibitors is critical to avoid widespread effects in a potential therapy. Thus for evaluating the selectivity of BTZ compounds as potential inhibitors, the repre￾sentative compound 6v was assayed against a panel of other 13 kinases. These kinases include four serine/threonine kinases as Cdk-1/cyclin B, CK-II, PKA and PKCa which are highly close to GSK- 3b, and nine tyrosine kinases as Flt-1, KDR, PDGFR-b, EPH-A2, EGFR, ErbB2, ErbB4, RON and Abl, which play important role in cancer signaling pathways. All of the inhibitory assays were carried out at a 100 mM of 6v concentration. The serine/threonine kinases inhi￾bition were performed at Millipore Corporation (Dundee, UK) by KinaseProfiler service, and the tyrosine kinases screen was assayed utilizing an ELISA approach as described in Experimental section. The results listed in Table 2 showed that compound 6v almost displayed no inhibitory activity against the whole set of kinases, which to some extent demonstrates that BTZs might have good specificity with respect to GSK-3b. 2.2.4. Structureeactivity relationships (SAR) analysis In order to explore the main features between the chemical structures and their GSK-3b inhibition, a preliminary structuree activity relationship of BTZs was analyzed based on the preceding in vitro data, which were listed in Table 1. The substituents R2 attached to the N5 of the BTZ ring seemed to be important for keeping the activities, of which the benzyl group was the best one (6f vs 6aee) in the test set. Furthermore, ortho-nitro group and meta-carbomethoxy group could dramatically contribute to the inhibitory potency as attached to the benzyl moiety (6l, 6t). On the other hand, the size and nature of the substituents R1 attached to the C2 should also be crucial for inhibition. The a bc Scheme 1. (a) CH2(COOH)2, pyridine, piperidine, 2 h, reflux; (b) 2-aminobenzenethiol, 4 A molecular sieve, 6 h, 180 C; (c) R2 X, X ¼ Cl or Br or I, NaH, DMF, 0.5 h,
10 C. a b Scheme 2. Reagents and conditions: (a) conc. HCl, 100 C, 7 h; (b) SOCl2, MeOH, reflux. P. Zhang et al. / European Journal of Medicinal Chemistry 61 (2013) 95e103 97

P Zhang et aL/ European Journal of medicinal Chemistry 61(2013)95-103 Table 2 GSK-3B inhibitory activities of compounds 6a-6y. ry activity ( inhibition) of compound 6v(100 HM)against severa R rosine loh Cdk-1/cyclin B 1.5 EPH-A2 -24.3 GFR-Ba-81.8 PKC 12.0 Erbb Su11248 as reference inhibitor for Flt-1(87.1% inhibition), KDR(89.7% inhibi for EGFR(86.9% inhibition). ErbB2(79.4% 2-Thienyl hibition)and ErbB4(82.8% for EPH-A2(83. 1% inhibition), Abl (90.0% d pD173074 as reference inhibitor for RON(93.4% inhibition). 3-Pyridyl 2.2.5. Molecular docking In order to gain an in-depth understanding on the interaction 2-F-Bn 2-CI-Bn mechanism for BTZs within the non-AtP binding pocket of GSK-3B, docking study of compound 6v was performed utilizing the gold 5.0 [21] software. The GSK-3B crystal structure as PDB ID of 1PYX 4-CH,0- was chosen as the model of receptor because it was lately proven to 3-COOMe-Bn be the best one for the non- atP binding in a docking study with three GSK-3B crystal structures(PDB ID: 1PYX, 1Q41, and 1Q4L [22]. The ligand was prepared by minimizing the energy of compound 6v using Sybyl 6.9 31 with the MMFF94 force, and the CI-Ph 2-NOz-B binding site was defined as a sphere of 10 A radius around the 4-Br-Ph 2-NOz-B 78 TDZD-8 residue arg 209, which is suggested to be a key residue for GSK-3B binding process [22]. The suggested binding mode was shown in value of at least two separate determinations, each determina- Fig. 4 and several key interactions were observed. Compound 6v TDZD-8, the first reported non-ATP competitive GSK-3B inhibitor, was used as was located between residues Arg 209 and Ser 236, and its BTZ ring reference compound in this study. ound with Arg 209 by cation-T interaction In addition, there were two hydrogen bonds formed between 6v and the binding pocket. compounds with methyl moiety or without substituent at C2 did One hydrogen bond was attributed to the oxygen of carbonyl on not show any activities(6h, 6i). However, the inhibition potency BIZ ring with Arg 209, and the other one interacted between nitro as enhanced considerably when aromatic groups, such as thienyl, group of the compound and Ser 236. Finally,, C2-subsitituted group furyl, phenyl and benzyl(6f, 6g 6k and 6v) were introduced to O2, of the inhibitor was extended to the hydrophobic region consisting suggesting favorable hydrophobic interactions with the enzyme. of three amino residues (Leu 169, Pro 331 and Thr 330). This In a summary the presence of both a bulky hydrophobic docking mode can well explain the inhibitory activity of compound substituent at N5 position and an aromatic group at C2 position in 6v to GSK-3B. It is worth to note that this result is also consistent the BTZ scaffold(61, 6t, 6u and 6v)seems favorable to increase the with the recently proposed hypothesis by Martinez et al. 221. GSK-3B inhibitory activity. which assumed that three key residues(Arg 209, Thr 235, Ser 236) A2.5 25μM 1.5 25μM control 1[Gs-2] pre-incubation time E+l (min) Fig 3. Kinetic data determined for the compound 6v. (A)ATP concentrations varied from 0.5 uM to 8 uM in the reaction mixture: GS-2 concentration was kept constant at 6.25 HM epicted in the plot. (B)GS-2 concentrations varied from 0.78 uM to 12.5 M in the reaction mixture: ATP concentration was kept constant at 2 HM; ompound concentrations were depicted in the plot. (C)Time dependent GSK-3B inhibition of 6v. Each point was the mean of two separate experiments and each experiment

compounds with methyl moiety or without substituent at C2 did not show any activities (6h, 6i). However, the inhibition potency was enhanced considerably when aromatic groups, such as thienyl, furyl, phenyl and benzyl (6f, 6g, 6k and 6v) were introduced to C2, suggesting favorable hydrophobic interactions with the enzyme. In a summary, the presence of both a bulky hydrophobic substituent at N5 position and an aromatic group at C2 position in the BTZ scaffold (6l, 6t, 6u and 6v) seems favorable to increase the GSK-3b inhibitory activity. 2.2.5. Molecular docking In order to gain an in-depth understanding on the interaction mechanism for BTZs within the non-ATP binding pocket of GSK-3b, a docking study of compound 6v was performed utilizing the GOLD 5.0 [21] software. The GSK-3b crystal structure as PDB ID of 1PYX was chosen as the model of receptor because it was lately proven to be the best one for the non-ATP binding in a docking study with three GSK-3b crystal structures (PDB ID: 1PYX, 1Q41, and 1Q4L) [22]. The ligand was prepared by minimizing the energy of compound 6v using Sybyl 6.9 [31] with the MMFF94 force, and the binding site was defined as a sphere of 10 A radius around the residue Arg 209, which is suggested to be a key residue for GSK-3b binding process [22]. The suggested binding mode was shown in Fig. 4 and several key interactions were observed. Compound 6v was located between residues Arg 209 and Ser 236, and its BTZ ring bound with Arg 209 by cation-p interaction. In addition, there were two hydrogen bonds formed between 6v and the binding pocket. One hydrogen bond was attributed to the oxygen of carbonyl on BTZ ring with Arg 209, and the other one interacted between nitro group of the compound and Ser 236. Finally, C2-subsitituted group of the inhibitor was extended to the hydrophobic region consisting of three amino residues (Leu 169, Pro 331 and Thr 330). This docking mode can well explain the inhibitory activity of compound 6v to GSK-3b. It is worth to note that this result is also consistent with the recently proposed hypothesis by Martinez et al. [22], which assumed that three key residues (Arg 209, Thr 235, Ser 236) Table 1 GSK-3b inhibitory activities of compounds 6ae6y. Compound R1 R2 IC50 (mM)a 6a 2-Thienyl Et >100 6b 2-Thienyl i Pr >100 6c 2-Thienyl n Bu >100 6d 2-Thienyl Cyclohexthylmethyl >100 6e 2-Thienyl Benzoyl >100 6f 2-Thienyl Bn 47.5 6g 2-Furyl Bn 77.2 6h H Bn >100 6i Me Bn >100 6j 3-Pyridyl Bn >100 6k Ph Bn 42.7 6l Ph 2-NO2eBn 25.0 6m Ph 2-CNeBn >100 6n Ph 2-F-Bn >100 6o Ph 2-Cl-Bn >100 6p Ph 2-Br-Bn 73.9 6q Ph 2-CH3eBn 76.1 6r Ph 4-CH3OeBn 73.7 6s Ph 3-COOHeBn >100 6t Ph 3-COOMeeBn 27.8 6u Ph 3-ClePhCOCH2 37.8 6v PhCH2 2-NO2eBn 23.0 6w 4-FePh 2-NO2eBn 81.5 6x 4-ClePh 2-NO2eBn 71.3 6y 4-BrePh 2-NO2eBn 67.8 TDZD-8b e e 1.4 a IC50, the mean value of at least two separate determinations, each determina￾tion was mean of triplicate experiments. b TDZD-8, the first reported non-ATP competitive GSK-3b inhibitor, was used as reference compound in this study. Fig. 3. Kinetic data determined for the compound 6v. (A) ATP concentrations varied from 0.5 mM to 8 mM in the reaction mixture; GS-2 concentration was kept constant at 6.25 mM; compound concentrations were depicted in the plot. (B) GS-2 concentrations varied from 0.78 mM to 12.5 mM in the reaction mixture; ATP concentration was kept constant at 2 mM; compound concentrations were depicted in the plot. (C) Time dependent GSK-3b inhibition of 6v. Each point was the mean of two separate experiments and each experiment performed in triplicate. Table 2 Inhibitory activity (% inhibition) of compound 6v (100 mM) against several protein kinases. Serine/ threonine kinases % Inhibition Tyrosine kinases % Inhibition Tyrosine kinases % Inhibition Cdk-1/cyclin B 1.5 Flt-1a 0.6 ErbB4b 0 CK-II 0 KDRa 5.3 EPH-A2c
24.3 PKA 0 PDGFR-ba
81.8 Ablc 18.1 PKCa 12.0 EGFRb 0 RONd 0 e e ErbB2b 2.7 e e a Su11248 as reference inhibitor for Flt-1 (87.1% inhibition), KDR (89.7% inhibi￾tion) and PDGFR-b (82.1% inhibition). b BIBW2992 as reference inhibitor for EGFR (86.9% inhibition), ErbB2 (79.4% inhibition) and ErbB4 (82.8% inhibition). c Dasatinib as reference inhibitor for EPH-A2 (83.1% inhibition), Abl (90.0% inhibition). d PD173074 as reference inhibitor for RON (93.4% inhibition). 98 P. Zhang et al. / European Journal of Medicinal Chemistry 61 (2013) 95e103

P Zhang er al/ European Joumal of Medicinal Chemistry 61(2013)95-103 ADME properties are difficult to predict accurately just by in silico and so the predicted properties should be used with caution. For our study, more experiments should be done to explore the real 3. Conclusion ARG 209 SER 26 n summary, we have disclosed a novel series of BTZ compounds as non-ATP competitive GSK-3B inhibitors Based on the structure modification of the hits found from a structure-based docking screening, twenty five btz derivatives were successfully synthe- sized and the in vitro luminescent bioassay showed that about half of them can inhibit GSK-3B with ICso values in micromolar level Among them, three compounds 61, 6t and 6v have ICso values of 25.0 uM, 27.8 HM and 23.0 HM respectively. As the representative compound 6v is proved to reversibly inhibit GSK-3B as non-ATP competitive mechanism through kinetic analysis. Moreover, it almost has not shown any inhibitory activity to a whole panel of other 13 protein kinases, which to some extent suggests good Fig 4. Suggested binding mode for compound 6v in GSK-3B(PDB ID: IPYX) specificity of BIZ compounds with respect to GSK-3B. The results of our primary SAR study also remind us that the nature of BTZ moiety In potentially related to the activ and certain substituents would be important for retaining GSK-3B ore, it is hoped that the prese inhibition. The presence of both a bulky hydrophobic substituent at study of more potent Gsk-3B seem crucial to increase the inhibitory activity. Finally, the binding mode of compound 6v to a non-atP binding pocket of GSK-3B was established by molecular docking and several key interactions were Finally, the adme descriptors module available in Discovery bserved. In this binding mode the affinity seems mainly attrib- Studio(DS)3.0[23 was used to predict a range of drug-like ited to the hydrogen-bonds interaction by the carbonyl and nitro properties for the compound 6v. This protocol uses the qsar groups of inhibitor 6v with Arg 209 and Ser 236 respectively, which models and is suitable to estimate the adme related properties for let the compound more closely bind to the pocket of enzyme and small molecules. The following properties, and classes of proper- hoped that the inhibitory potency and specificity of BTZs toward ties, can be computed such as human intestinal absorption( HIA), GSK-3B shou queous solubility, blood-brain barrier penetration(BBB), cyto low them worth to further study as potenti therapeutic candidates for severe unmet human diseases where chrome P450(CYP450)2D6 inhibition and plasma protein binding. GSK-3B is up-regulated. The resulted data were listed in Table 3. It showed compound 6v might have good intestinal absorption after oral administration and nedium ability to cross the blood-brain barrier (BBB), but its ity predicted was not so good. In addition, 4. Experi compound 6v was likely to inhibit CYP2D6 enzyme, which belongs 4.1. Chemistry to CYP450 enzyme family. The results also showed that the binding between inhibitor and plasma protein is less than 90%, which Reagents were purchased from commercial sources and then proteins in the blood. However, it is a common experience that chromatography was carried out at medium pressure using silica Co. Ltd. All the reactions were monitored by thin layer chroma- 3 tography(TLC)on silica gel. H NMR andC NMR spectra were ADME properties of compound 6v predicted. obtained on Mercury Plus 400 spectrometers working at 400 MHz nd 100 MHz, respectively. Chemical shifts(O)are reported in parts per million(ppm) relative to internal tetramethylsilane(TMs)andJ values are reported in Hertz. MS spectra were recorded on al 00 Agilent LC-MS 1100 instrument with an ESI mass selective detector. Melting points were determined by an SGW X-4 thermometer and AlogP98 were uncorrected(slide method). Level 0 means inhibitor has good human intestinal absorption(HlA)after o 4.1.1. General procedure for the synthesis of 3-substituted acrylic eous solubility of inhibitor is not very good and its drug- A mixture of substituted carbaldehyde(200 mmol), propen dioic acid (20.8 g. 200 mmol) in a solution of pyridin e(10 ml d Level 2 means inhibitor has medium ability to cross the blood-brain barrier 120 mmol )and piperidine(1 ml)was warmed at reflux for 2 h. The e Level 1 means inhibitor likely to inhibit CYP2D6 enzyme. resultant solution was poured into 2 M HCl ag. and then cooled to f Levelo means the binding between inhibitor and plasma protein is less than 90% room temperature. The present solid was collected by filtration, (No markers flagged and Alog P98 4.0). washed with water and recrystallized from ethanol/ water

in an allosteric binding site are potentially related to the active conformation of GSK-3b. Therefore, it is hoped that the present work will be helpful for further study of more potent GSK-3b inhibitors. 2.2.6. Predicting ADME properties Finally, the ADME descriptors module available in Discovery Studio (DS) 3.0 [23] was used to predict a range of drug-like properties for the compound 6v. This protocol uses the QSAR models and is suitable to estimate the ADME related properties for small molecules. The following properties, and classes of proper￾ties, can be computed such as human intestinal absorption (HIA), aqueous solubility, bloodebrain barrier penetration (BBB), cyto￾chrome P450 (CYP450) 2D6 inhibition and plasma protein binding. The resulted data were listed in Table 3. It showed compound 6v might have good intestinal absorption after oral administration and medium ability to cross the bloodebrain barrier (BBB), but its aqueous solubility predicted was not so good. In addition, compound 6v was likely to inhibit CYP2D6 enzyme, which belongs to CYP450 enzyme family. The results also showed that the binding between inhibitor and plasma protein is less than 90%, which means the compound is unlikely to be highly bound to carrier proteins in the blood. However, it is a common experience that ADME properties are difficult to predict accurately just by in silico and so the predicted properties should be used with caution. For our study, more experiments should be done to explore the real properties of these compounds. 3. Conclusion In summary, we have disclosed a novel series of BTZ compounds as non-ATP competitive GSK-3b inhibitors. Based on the structure modification of the hits found from a structure-based docking screening, twenty five BTZ derivatives were successfully synthe￾sized and the in vitro luminescent bioassay showed that about half of them can inhibit GSK-3b with IC50 values in micromolar level. Among them, three compounds 6l, 6t and 6v have IC50 values of 25.0 mM, 27.8 mM and 23.0 mM respectively. As the representative, compound 6v is proved to reversibly inhibit GSK-3b as non-ATP competitive mechanism through kinetic analysis. Moreover, it almost has not shown any inhibitory activity to a whole panel of other 13 protein kinases, which to some extent suggests good specificity of BTZ compounds with respect to GSK-3b. The results of our primary SAR study also remind us that the nature of BTZ moiety and certain substituents would be important for retaining GSK-3b inhibition. The presence of both a bulky hydrophobic substituent at N5 position and an aromatic group at C2 position in BTZ scaffold seem crucial to increase the inhibitory activity. Finally, the binding mode of compound 6v to a non-ATP binding pocket of GSK-3b was established by molecular docking and several key interactions were observed. In this binding mode, the affinity seems mainly attrib￾uted to the hydrogen-bonds interaction by the carbonyl and nitro groups of inhibitor 6v with Arg 209 and Ser 236 respectively, which let the compound more closely bind to the pocket of enzyme and contributed to the inhibitory activity. According to our work, it is hoped that the inhibitory potency and specificity of BTZs toward GSK-3b should allow them worth to further study as potential therapeutic candidates for severe unmet human diseases where GSK-3b is up-regulated. 4. Experimental 4.1. Chemistry Reagents were purchased from commercial sources and then used without further purification except special case. Flash column chromatography was carried out at medium pressure using silica gel (200e300 mesh) purchased from Qingdao Haiyang Chemical Co. Ltd. All the reactions were monitored by thin layer chroma￾tography (TLC) on silica gel. 1 H NMR and 13C NMR spectra were obtained on Mercury Plus 400 spectrometers working at 400 MHz and 100 MHz, respectively. Chemical shifts (d) are reported in parts per million (ppm) relative to internal tetramethylsilane (TMS) and J values are reported in Hertz. MS spectra were recorded on an Agilent LC-MS 1100 instrument with an ESI mass selective detector. Melting points were determined by an SGW X-4 thermometer and were uncorrected (slide method). 4.1.1. General procedure for the synthesis of 3-substituted acrylic acid A mixture of substituted carbaldehyde (200 mmol), propene￾dioic acid (20.8 g, 200 mmol) in a solution of pyridine (10 ml, 120 mmol) and piperidine (1 ml) was warmed at reflux for 2 h. The resultant solution was poured into 2 M HCl aq. and then cooled to room temperature. The present solid was collected by filtration, washed with water and recrystallized from ethanol/water. Fig. 4. Suggested binding mode for compound 6v in GSK-3b (PDB ID: 1PYX). Table 3 ADME properties of compound 6v predicted. Value Level Absorptiona NVb 0 Solubilityc
4.619 2 BBBd
0.06 2 CYP2D6e 0.722 1 PPBf NV 0 AlogP98f 3.733 0 a Level 0 means inhibitor has good human intestinal absorption (HIA) after oral administration. b NV means no value was given. c Level 2 means the aqueous solubility of inhibitor is not very good and its drug￾likeness properties are low. d Level 2 means inhibitor has medium ability to cross the blood
brain barrier (BBB). e Level 1 means inhibitor likely to inhibit CYP2D6 enzyme. f Level 0 means the binding between inhibitor and plasma protein is less than 90% (No markers flagged and AlogP98 < 4.0). P. Zhang et al. / European Journal of Medicinal Chemistry 61 (2013) 95e103 99