Synthesis and mechanism of hypoglycemic activity of benzothiazole derivatives

Article abstract of DOI:10.1021/jm4001488

Adenosine 5′-monophosphate activated protein kinase (AMPK) has emerged as a major potential target for novel antidiabetic drugs. We studied the structure of 2-chloro-5-((Z)-((E)-5-((5-(4,5-dimethyl-2-nitrophenyl)furan-2-yl) methylene)-4-oxothiazolidin-2-ylidene)amino)benzoic acid (PT-1), which attenuates the autoinhibition of the enzyme AMPK, for the design and synthesis of different benzothiazoles with potential antidiabetic activity. We synthesized several structurally related benzothiazole derivatives that increased the rate of glucose uptake in L6 myotubes in an AMPK-dependent manner. One compound, 2-(benzo[d]thiazol-2-ylmethylthio)-6-ethoxybenzo[d]thiazole (34), augmented the rate of glucose uptake up to 2.5-fold compared with vehicle-treated cells and up to 1.1-fold compared to PT-1. Concomitantly, it elevated the abundance of GLUT4 in the plasma membrane of the myotubes and activated AMPK. Subcutaneous administration of 34 to hyperglycemic Kuo Kondo rats carrying the Ay-yellow obese gene (KKAy) mice lowered blood glucose levels toward the normoglycemic range. In accord with its activity, compound 34 showed a high fit value to a pharmacophore model derived from the PT-1.

Full text of DOI:10.1021/jm4001488

Subscriber access provided by the Library Service | University of Stellenbosch
Article
Synthesis and Mechanism of Hypoglycemic
Activity of Benzothiazole Derivatives
Ella Meltzer-Mats, Gali Babai-Shani, Lily Pasternak, Neta Uritsky, Tamar Getter, Olga
Viskind, J#rgen Eckel, Erol Cerasi, Hanoch Senderowitz, Shlomo Sasson, and Arie Gruzman
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm4001488 • Publication Date (Web): 10 Jun 2013
Downloaded from http://pubs.acs.org on June 12, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Synthesis and Mechanism of Hypoglycemic Activity of
Benzothiazole Derivatives
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ella MeltzerꢀMats§, Gali BabaiꢀShani∂, Lily Pasternak∂, Neta Uritsky§, Tamar Getter§, Olga Viskind∂,§,
Jürgen Eckel‡, Erol Cerasi†, Hanoch Senderowitz§, Shlomo Sasson∂* and Arie Gruzman, §*
§Division of Medicinal Chemistry, Department of Chemistry, Faculty of Exact Sciences, Bar Ilan
University, Ramat Gan, 52900, Israel
∂Department of Pharmacology, Institute for Drug Research, School of Pharmacy, Faculty of Medicine,
The Hebrew University of Jerusalem, 91120, Jerusalem, Israel
‡German Diabetes Center, Integrative Physiology, PaulꢀLangerhansꢀGroup, Auf’m Hennekamp 65,
40225, Düsseldorf, Germany
†The Endocrinology and Metabolism Service, Department of Medicine, HadassahꢀHebrew University
Medical Center, Jerusalem, 91120, Israel
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 49
1
2
3
4
5
ABSTRACT
6
7
8
9
AMPꢀactivated protein kinase (AMPK) has recently emerged as a major potential target for novel
antidiabetic drugs. We studied the structure of 2ꢀchloroꢀ5ꢀ((Z)ꢀ((E)ꢀ5ꢀ((5ꢀ(4,5ꢀdimethylꢀ2ꢀ
nitrophenyl)furanꢀ2ꢀyl)methylene)ꢀ4ꢀoxothiazolidinꢀ2ꢀylidene)amino)benzoic acid (PTꢀ1), which
attenuates the autoinhibition of the enzyme, for the design and synthesis of different benzothiazoles with
potential antidiabetic activity. We synthesized several structurallyꢀrelated benzothiazole derivatives that
increased the rate of glucose uptake in L6 myotubes in an AMPKꢀdependent manner. One compound, 2ꢀ
(benzo[d]thiazolꢀ2ꢀylmethylthio)ꢀ6ꢀethoxybenzo[d]thiazole (34), augmented the rate of glucose uptake
up to 2.5ꢀfold comparing with vehicleꢀtreated cells and up to 1.1 fold comparing to PTꢀ1.
Concomitantly, it elevated the abundance of GLUT4 in the plasma membrane of the myotubes and
activated AMPK. Subcutaneous administration of 34 to hyperglycemic Kuo Kondo rats carrying the Ayꢀ
yellow obese gene (KKAy) mice lowered blood glucose levels towards the normoglycemic range. In
accord with its activity, compound 34 showed a high fit value to a pharmacophore model derived from
the PTꢀ1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 3 of 49
Journal of Medicinal Chemistry
MANUSCRIPT TEXT
1
2
3
4
Introduction
5
6
7
8
A major pathophysiological defect in T2DM patients is decreased glucose utilization in peripheral
tissues, such as skeletal muscles and fat depots, which aggravates hyperglycemia. 1,2 Clinical
observations in diabetic patients suggest that hyperglycemia per se contributes and even worsens this
phenomenon. 3ꢀ5 Previously, we reported that high glucose levels reduce the rate of glucose transport in
skeletal muscle cells in vivo and in vitro. 5ꢀ9 This autoregulatory mechanism protects the intracellular
environment against deleterious effects of an excessive glucose influx. Nonetheless, this feedback loop
may also worsen the hyperglycemia and contribute further to the development of peripheral insulin
resistance. Agents that increase the rate of glucose entry to skeletal muscles under hyperglycemic
conditions independent of insulin are therefore invaluable for the development of novel antidiabetic
drugs. 10,11
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The enzyme AMPK, which acts as a sensor of cellular energy status, plays a central role in the
regulation of glucose transport in skeletal muscles. 12 Several AMPK isoenzymes have been identified,
of which AMPKα2 is most abundant in skeletal muscles. 13 This enzyme is physiologically activated
when the AMP:ATP ratio is increased in cells. Among other functions, activated AMPK promotes the
translocation of glucose transporter 4 (GLUT4)ꢀcontaining vesicles to the plasma membrane, thus
increasing glucose influx in a nonꢀinsulin dependent manner. 14,15 These properties make AMPK a key
target for the development of antidiabetic drugs. 16 Several classes of direct and indirect activators of
AMPK such as A769662, PTꢀ1, 5ꢀAminoimidazoleꢀ4ꢀcarboxamideꢀ1ꢀβꢀDꢀribofuranoside (AICAR)
(direct activators) and berberine, D942, momordicosides, capsaicinoids, curcumin and lipoic acid
(indirect activators) have been reported. 17,18 Additionally, two large groups of antidiabetic drugs
(biguanides and thiazolidinediones) owe part of their antihyperglycemic properties to the activation of
AMPK. 18 Recently, the American Diabetes Association (ADA) and the European Association for the
Study of Diabetes (EASD) recommended metformin (biguanide AMPK activator) as the first choice for
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 49
the monotherapy treatment of T2DM and suggested to include it in diabetes combination therapy. 19
1
2
3
4
5
However, side effects, such as lactic acidosis and development of tolerance to metformin emphasize the
need for more potent and tissue specific AMPK activators with reduced side effects.
6
7
8
9
In this study we aimed at designing and synthesizing novel and potent activators of AMPK in
skeletal muscles. The known AMPK activator PTꢀ1, which attenuates the autoinhibition of AMPK, 20
was selected as a starting compound in our attempts to design effective AMPK activators (Chart 1).
Based on medicinal chemistry principles and SAR studies of PTꢀ1 we designed and synthesized in
house a series of benzothiazole derivatives that were then screened for their glucose uptake increasing
effects and activation of AMPK in L6 myotubes, and tested in diabetic KKAy mice in terms of
amelioration of hyperglycemia. In order to unveil the molecular determinants responsible for the AMPK
activation of this class of compounds and to direct future synthesis efforts we have used the data
obtained in this study to develop a pharmacophore model for benzothiazoleꢀbased AMPK activators.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Results and Discussion
Pang et al. have reported that PTꢀ1 (Chart 1) activated doseꢀdependently AMPK by relieving the
autoinhibition of the enzyme. 20 In this study also showed that PTꢀ1 increased doseꢀ and timeꢀ
dependently the phosphorylation of AMPK and its downstream substrate acetyl CoA carboxylase
(ACC) in L6 myotubes, without modulating the cellular AMP:ATP ratio. We aimed at elucidating the
structural moieties of PTꢀ1 that were required to activate AMPK and increase the rate of glucose uptake
in L6 myotubes. When tested in in L6 myocytes we found that PTꢀ1 increased the rate of glucose
uptake, reaching maximal effects (1.4 ±0.06ꢀfold) at 5 ꢁM (Supplemental Figure 1). Thus, to identify
the minimal effective moiety of PTꢀ1 required to activate AMPK and stimulate glucose uptake, several
commercially available heterocyclic compounds, which were selected based on structural similarity
with the (2Z,5E)ꢀ5ꢀethylideneꢀ2ꢀ(methylimino)thiazolidinꢀ4ꢀone moiety. A substructure of PTꢀ1: (E)ꢀ2ꢀ
aminoꢀ5ꢀ((E)ꢀ3ꢀ(5ꢀnitrofuranꢀ2ꢀyl)allylidene)thiazolꢀ4(5H)ꢀone (2) (Chart 1) at much higher
concentration then PTꢀ1 (>0.5 mmol/L) only slightly increased the rate of glucose uptake in L6
ACS Paragon Plus Environment

Page 5 of 49
Journal of Medicinal Chemistry
myotubes and marginally induced Thr172 phosphorylation in AMPK (data not shown). Of the 9
derivatives (all these compounds were randomly selected because of their structure similarity to 2)
tested only 2ꢀaminoꢀ5ꢀethylthiazolꢀ4(5H)ꢀone (3) was active (Table 1). Figure 1 (A and B) shows that
at maximal effective concentration (100 ꢂM) 3 increased glucose uptake in L6 myotubes by 134.9% ±
11.4% (in comparison with the vehicleꢀtreated control) within 5 h of incubation. Above this
concentration 3 exhibited cytotoxic effects. The timeꢀcourse analysis in Figure 1A shows that the
stimulatory effect of 3 peaked at 5 h and then gradually declined to the basal level. Compound 3ꢀ
induced Thr172 phosphorylation in AMPK was maximal at 40 min of incubation and returned to the
basal level after 3 h (Figure 1C). Other, more bulky, compounds structurally related to (2Z,5E)ꢀ5ꢀ
ethylideneꢀ2ꢀ(methylimino)thiazolidinꢀ4ꢀone moiety of PTꢀ1 that include benzene moieties were
purchased and tested (Table 2). Among these compounds, only compounds 4 and 15: 6ꢀ
ethoxybenzo[d]thiazoleꢀ2ꢀthiol and 6ꢀbromobenzo[d]thiazolꢀ2ꢀol, respectively, significantly induced
AMPK phosphorylation and increased the rate of glucose uptake in L6 myotubes (179.1±9.5% and
144.1%±8.8%, respectively, compared to DMSOꢀtreated L6 myotubes; Figure 1B). The maximal
effective concentration of both compounds (100 ꢂM) was similar to that of 3, yet, their stimulatory
effect on glucose uptake persisted for 24 h (Figure 1A and 1B) and was accompanied by a stable Thr172
phosphorylation of AMPK during this period (Figure 1C and 1D). Compounds 3ꢀ18 lacked synergetic
or additive effects to the stimulatory effect of insulin on glucose uptake in L6 myotubes (See
Supplementary Table 1).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Since 4 was more effective than 15 in increasing the rate of glucose transport, it was used as a
starting compound for the design and synthesis of more potent and efficacious derivatives. The first step
was to investigate the role of the ethoxy moiety in 4 on its biological activity. Thus, the ethoxy moiety
was removed from the corresponding phenol alcohol derivative following an overnight reaction with
AlCl3 in dichloroethane at room temperature, to obtain 2ꢀmercaptobenzo[d]thiazolꢀ6ꢀol (19), as
described by Woltersdorf et al. 21 When tested in L6 myotube cultures, 19 failed to stimulate glucose
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 49
uptake, most likely due its substantial low lipophilicity (LogP=1.957±0.420) in comparison with the
parent compound 4 (LogP=3.167±0.436) in resulting poor cellular membrane penetration. 22 We further
synthesized three other novel compounds: 2ꢀ((2,2ꢀdiethoxyethyl)thio)benzo[d]thiazolꢀ6ꢀol (20), 2ꢀ(2ꢀ
ethylbutylthio)benzo[d]thiazolꢀ6ꢀol (21) and 2ꢀ(butylthio)benzo[d]thiazolꢀ6ꢀol (22), which were more
lipophilic than 4 due to the alkylation of the free thiol group. The alkylation was performed in the
presence of an excess of sodium hydride in dry THF (Scheme 1). The alkylation occurred selectively on
the thiol group of 19 and did not involve the unprotected phenol moiety, likely due to lower nucleophilic
property of the phenol group in comparison with the thiol moiety. Compound 20 was not stable during
purification by column chromatography (e.g., silica gel or alumina), as confirmed by MS analysis. Thus,
it was used in cell cultures immediately after synthesis without additional purification. None of these
derivatives (20ꢀ22) increased significantly the rate of glucose uptake in L6 myotubes.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Since the elimination of the ethoxy moiety from the benzene ring in 4 abolished its biological
activity, regardless of the alkylation state of the free thiol group, the alkyl moiety was preserved in
further syntheses. Therefore, the free thiol group was replaced by various lipophilic thioether functional
groups to produce 23ꢀ27 (Scheme 1). In addition, the role of the carboxylic group was investigated
through the synthesis of 2ꢀ(6ꢀethoxybenzo[d]thiazolꢀ2ꢀylthio)acetic acid (23) and 3ꢀ(6ꢀ
ethoxybenzo[d]thiazolꢀ2ꢀylthio)propanoic acid (24). The corresponding iodoꢀ, bromoꢀ alkyl or 3ꢀ
chloroaceticꢀ, 3ꢀchloropropionic acids were reacted with the thiol group of 4 in the presence of an
excess of sodium hydride in dry THF. When tested in L6 myotubes only compound 24 was active.
Figure 2 shows that 50 ꢁM of 24 increased the rate of glucose uptake 1.8ꢀfold within 5 h of incubation,
in comparison with the vehicleꢀtreated cells. Additionally, Thr172 phosphorylation of AMPK peaked at
60 min of incubation.
Next, we changed the thiol group in 4 to sulfonylamide or disulfane moieties, which are less
prone to oxidation. 23 The former was exchanged by sulfanilamide to produce 6ꢀethoxybenzo[d]thiazoleꢀ
2ꢀsulfonamide (28), using a twoꢀstep synthesis according to Schoenwald et al. 24 Briefly, compound 4
ACS Paragon Plus Environment

Page 7 of 49
Journal of Medicinal Chemistry
was treated with ammonium hydroxide (28%) and sodium hypochlorite (5.25%) in an aqueous solution
and the corresponding sulfonamide derivative generated with 84% yield. The latter was oxidized with
KMnO4 in acetone/water (1:1) to yield 28. This compound was inactive in the glucose uptake assay in
L6 myotubes. Further, compounds 1,2ꢀbis(6ꢀethoxybenzo[d]thiazolꢀ2ꢀyl)disulfane (29) and 1,2ꢀdi(1Hꢀ
benzo[d]imidazolꢀ2ꢀyl)disulfane (30) were synthesized, as shown in Scheme 2, using a novel synthetic
method, in which the corresponding thiol was refluxed overnight in the presence of concentrated HCl in
THF to obtain 29. This compound consists of two molecules of 4 covalently bound by an SꢀS bond.
Compound 30 was similarly prepared using two 1Hꢀbenzoimidazoleꢀ2ꢀthiol molecules that were
coupled. In contrast to 29, the sulfur atoms in both benzothiazole moieties in 30 were replaced with
nitrogen atoms and the ethoxy moiety on the benzene ring was also removed. Compound 30 had a weak
stimulatory effect on glucose uptake in L6 myotubes (135± 12.2%). In contrast, 29 substantially
increased the rate of uptake: a maximal effect (189.0 ± 21.8%) was obtained with 150 ꢂM within 5 h of
incubation. It also induced Thr172 phosphorylation in AMPK within 40 min of incubation, an effect that
persisted throughout the 5 h incubation (Figure 2). These effects correlate well with the increased
lipophilicity and enhanced cell permeability of 29 compared with the monomeric form (4) that presents
a free SH group. It is also suggested that the SꢀS bond in 29 is cleaved intracellularly to release free 4.
Nonetheless, despite its marked stimulatory effect on the glucose transport system in L6 myotubes, the
potency of 29 was considered poor (maximal effective concentration above 100 ꢂM).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To improve this low potency, benzo[d]thiazol derivatives were synthesized: conjugation of 6ꢀ
ethoxybenzo[d]thiazolꢀ2ꢀamine or 4ꢀ(6ꢀmethylbenzo[d]thiazolꢀ2ꢀyl)aniline with 2ꢀ
(chloromethyl)benzo[d]thiazole yielded two novel compounds: Nꢀ(benzo[d]thiazolꢀ2ꢀylmethyl)ꢀ6ꢀ
ethoxybenzo[d]thiazolꢀ2ꢀamine (31) and Nꢀ(benzo[d]thiazolꢀ2ꢀylmethyl)ꢀ4ꢀ(6ꢀmethylbenzo[d]thiazolꢀ
2ꢀyl)aniline (32), as shown in Scheme 3. Yet, both 31 and 32 were inactive in the glucose uptake assay
in L6 myotubes (data not shown).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 49
Finally, thioether derivatives of 4 were synthesized, namely, 2ꢀ(6ꢀethoxybenzo[d]thiazolꢀ2ꢀ
ylthio)ꢀNꢀpropylacetamide (33) and 34, as shown in Schemes 1 and 3. Compound 33 was obtained by
coupling propylamine hydrochloride with 23, using 3ꢀ(ethyliminomethyleneamino)ꢀN,Nꢀdimethylꢀ
propanꢀ1ꢀamine (EDC) with 1ꢀhydroxybenzotriazole (HOBt) in chloroform. 25 A twoꢀstep procedure
was employed in the synthesis of 34. First, commercially available 2ꢀmethylbenzo[d]thiazole was
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
transformed into 2ꢀ(chloromethyl)benzo[d]thiazole intermediate by chlorination (thrichloroisocyannuric
acid was used for selective chlorination of the methyl group). The resulting 2ꢀ
(chloromethyl)benzo[d]thiazole was then condensed with 4, using a diꢀisoꢀpropylethyl amine catalysis
in dry THF. Figure 2 (A and B) shows that both compounds markedly increased the rate of glucose
uptake in L6 myotubes doseꢀ and timeꢀdependently. The lipophilic amide derivative 33 was somewhat
less potent and effective than the benzothiazol derivative 34: 100 ꢁM of 33 increased 2ꢀfold the rate of
glucose uptake in L6 myotubes within 5 h of incubation, compared with the 2.5ꢀfold increase induced
by 34 under similar experimental conditions. Moreover, the minimal effective concentration of 34 was
25 ꢁM and its stimulatory effect remained unaltered up to 24 h. The time course of Thr172
phosphorylation of AMPK induced by 33 and 34 varied: the former induced significant Thr172
phosphorylation after 3 h of incubation while the latter caused it already within 1 hour (Figure 2C and
2D).
Compound 34 was the most promising benzothiazole derivative, and its biological activity was
further studied. The compound also has drugable LogP (4.34±0.14) and water solubility (12.09 ±1.71
ꢁM/L); both parameters were experimentally determined as described in the Supplemental Information,
D1. Figure 2E shows that 34 increased doseꢀdependently Thr172 phosphorylation in AMPK in L6
myotubes: already at 5 ꢁM 34 noticeably induced phosphorylation, while at 150 ꢁM it raised it 7.6 fold
in comparison with DMSOꢀtreated control myotubes. Figure 3A shows additivity in glucose transport
stimulatory effects of 34 and insulin in L6 myotubes, suggesting distinct and independent mechanisms
of actions of 34 and insulin.
ACS Paragon Plus Environment

Page 9 of 49
Journal of Medicinal Chemistry
Further, AMPKꢀinduced phosphorylation of ACC and AMPK substrate 160 kDa (ASꢀ160) was
1
2
determined. Figure 3B shows that 34 significantly induced Ser79 and Thr642 phosphorylation of ACC and
ASꢀ160, respectively. The classical AMPK activators: AICAR (4 mmol/L) and Dꢀsorbitol
(hyperosmolar shock inducer at 0.25 M) served as positive controls in this experiment. Additionally,
Compound C, a selective inhibitor of AMPK 26 was used to analyze the role of AMPK in mediating the
action of 34. When added (5 ꢁM) to myotube cultures, it significantly impeded the ability of 34 to
augment the phosphorylation of both AMPK substrates in L6 myotubes. The activation of ACC leads to
increased production of malonylꢀCoA, 27 which further inhibits carnitine acylꢀCoA transferaseꢀ1, the
enzyme that transports fatty acids into mitochondria and leads to the inhibition of mitochondrial fatty
acid degradation. 28 To exclude the possibility that the activation of ACC inhibited fatty acid metabolism
and caused cytotoxicity, we assessed potential cytotoxic effects of 34 in the (3ꢀ(4,5ꢀDimethylthiazolꢀ2ꢀ
yl)ꢀ2,5ꢀdiphenyltetrazolium bromide test (MTT). This assay measures the reduction of a tetrazolium salt
into a formazan product by dehydrogenase enzymes in functional mitochondria. 29 The data in
Supplemental Figure 2 show that L6 myotube viability remained intact upon exposure to high
concentrations of 34 during prolonged incubation periods. Thapsigarginꢀinduced apoptosis served as a
positive control in this assay. 30 Reduced cellular ATP levels and increased ATP:AMP ratio are
associated with upregulation of glucose transport and AMPK activation, respectively. 31,32 Therefore, we
tested whether 34 induced such effects; Table 3 shows that the cellular concentration of ATP was
unaltered in the presence of 200 ꢁM 34 in myotube cultures. Thus, it appears that the stimulatory effect
of 34 on AMPK was unrelated to cellular AMP level.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The ability of compound 34 to mobilize GLUT4ꢀrich vesicles to the plasma membrane of
myotubes was then investigated using the mycꢀtaggedꢀGLUT4 L6 or mycꢀtaggedꢀGLUT1 L6 cell lines.
In this assay the abundance of the myc epitope embedded in the first extracellular loop of the transporter
is determined by a colorimetric immunodetection assay. 33 Figure 3C shows that insulin and 34
independently increased the abundance of GLUT4myc, but not GLUT1myc, in the plasma membrane of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 49
L6 myotubes. The effect of insulin on the GLUT4myc translocation to the plasma membrane of
myotubes was fast (30 min) whereas 34 exerted its effects following a 5ꢀh incubation period. Neither
1
2
3
4
5
6
7
8
agent altered the total content of GLUT4 or GLUT1 in myotubes, as determined in whole cell lysates of
treated L6 myotubes (insert, Figure 3C).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The fact that 34 increased GLUT4 content in the plasma membrane raised the possibility
that 34 could employ the insulin transduction pathway to mobilize GLUT4ꢀcontaining vesicles to the
plasma membrane. To this end, the phosphatidylinositide 3ꢀkinase (PI3K) inhibitor wortmannin and an
Akt/PKB inhibitor (1,3ꢀDihydroꢀ1ꢀ(1ꢀ((4ꢀ(6ꢀphenylꢀ1Hꢀimidazo[4,5ꢀg]quinoxalinꢀ7ꢀyl)phenyl)methyl)ꢀ
4ꢀpiperidinyl)ꢀ2Hꢀbenzimidazolꢀ2ꢀone triꢀfluoroacetate salt hydrate) were used in 34ꢀtreated cells. 34 In
control myotubes, both inhibitors suppressed insulinꢀstimulated glucose uptake; however, the
stimulatory effect of 34 on glucose uptake remained intact in the presence of each inhibitors (Figure
3D), suggesting that the insulin transduction mechanism was not involved in mediating 34ꢀinduced
recruitment of GLUT4 to the surface of L6 myotubes. Conversely, the AMPK inhibitor compound C
completely blocked 34ꢀmediated stimulation of the rate of glucose transport in L6 myotubes.
Additionally, we used a commercial available kit to determine the AMPK activity in vitro, as described
in Figure 3E. In this assay PTꢀ1 or 34 markedly increased the kinase activity of recombinant human full
length AMPK by 2.04.0±0.04 and 4.30±0.05ꢀfold, respectively, comparing with the DMSO treated
control. These findings suggest that 34 also interacts with AMPK and modifies it catalytic activity.
KKAy mice develop diabetic traits similar to those characterizing T2DM patients, including
hyperinsulinemia, hyperglycemia, obesity, and peripheral resistance to insulin, and several studies
indicate they are an excellent animal model for in vivo evaluation of antidiabetic drugs. 35 Therefore,
we tested the antihyperglycemic potential of 34 by subcutaneously injecting it to diabetic KKAy mice
(30 or 75 mg/kg body weight, suspension in sesame oil, twice daily), while the control group received
the vehicle only. Oral (gavage) administration of 34 to these mice failed to lower significantly blood
glucose levels, possibly due to a low biostability of 34 (Supplemental Information, D2). Indeed when
ACS Paragon Plus Environment

Page 11 of 49
Journal of Medicinal Chemistry
incubated with mouse hepatic microsomes 34 was subjected to a rapid Phase I metabolism with an in
vitro estimated halfꢀlife time of 2.4 min and rapid clearance (692 ꢁl/min/mg). In addition, 34
significantly inhibited CytP450 2D6 and 3A4 activity by 64.4% ±4.9% and 77.1% ±11.5%,
respectively, an in vitro assay (Supplemental Information D3). The activity of 1A2, 2C9 and 2C19
isoforms of P450 was not affected in the presence of 34. Figure 4A shows a substantial blood glucose
lowering effect of 34: it reduced blood glucose by nearly 30%, 45% and 50% on days 1, 2 and 3 of
treatment, respectively, whereas the oil treatment (vehicle) had no effect. It should also be noted that
despite the marked antihyperglycemic effects of 34, normoglycemia was not attained. In addition, food
intake of 34ꢀtreated mice was not significantly different from the that measured in the cages of control
mice. Furthermore, ipꢀGTT was performed in KKAy mice treated with 34 for two days after an
overnight fast, as described. 6 Figure 4B shows improved tolerance to glucose in comparison with the
vehicleꢀtreated group. In the treated mice the AUC were 1523.9 ± 132.9 vs. 1150.6 ± 125.8 or 753.6 ±
142.2 mmol glucose/l/min for oilꢀtreated and 30 or 75 mg/kg 34ꢀtreated KKAy mice, respectively. The
daily food intake was not affected by 34, as shown in Supplemental Table 6. These results show that 34
improved total body glucose clearance in diabetic KKAy mice by 40%ꢀ50% following a twoꢀday
treatment.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While the design strategy of the compounds synthesized and tested in this work was based on
medicinal chemistry considerations, we have performed a postꢀsynthesis pharmacophore analysis of the
results with three main goals in mind: (1) Rationalize the design strategy and experimental (tested
compounds activity) data (2) Elucidate the structural determinant responsible for the biological activity;
(3) Direct future synthesis efforts. With this in mind, we have first generated a 3D pharmacophore
model from the lowest energy conformation of the starting compound (PTꢀ1). The resulting
pharmacophore (Figure 5A) consists of two aromatic moieties corresponding to the furan ring and the 5ꢀ
aminoꢀchloro benzoic acid ring, three hydrophobic centers located on the chlorine atom of the 5ꢀaminoꢀ
chloro benzoic acid ring and on the benzene moiety and the methyl group of the terminal 1,2ꢀdimethylꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 49
4ꢀnitrobenzene ring and four Hꢀbond acceptors located on the carboxylic group, the sulfur atom of the
central thiazolidine ring, the carbonyl oxygen of the same ring and on the nitro moiety of the 1,2ꢀ
dimethylꢀ4ꢀnitrobenzene ring. The fit of PTꢀ1 to this model is, not surprisingly, good (fit value of 9; see
Figure 5A and Supplemental Table 4). Next, we have screened compounds 2ꢀ4 (Chart 1) against this
model and in accord with the biological results (Table 2), and found compound 4 to have the best fit
values among the three. Moreover, this compound fits almost equally well into two, largely nonꢀ
overlapping parts of the pharmacophore model (fit values of 2.49 and 2.23, Figure 5B), thereby
justifying the overall design strategy, which was largely based on tethering together two analogs of this
compound. Finally, all 31 compounds tested in this work were screened into the pharmacophore model.
The results are presented in Supplemental Table 4 and confirm that, overall; the more active compounds
have better fit values. Using a cutoff of 3, 5 out of 6 active compounds (83% success) have a fit value
above this number whereas 22 out of 25 inactive compounds (88% success) have fit values below it or
do not map to the pharmacophore model at all. In particular our lead compound (34), fits into the
pharmacophore model with a fit value of 4.63 (Figure 5C).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While our preliminary pharmacophore model gave reasonable results, it was developed from a
single compound and consequently does not take advantage of the new data obtained in this work. With
this in mind, we have developed a new pharmacophore model based on the three most active
compounds (29, 33, 34) and further refined it based on data on 2 inactive compounds (25 and 31, see
method section for more details). The resulting pharmacophore (Figure 6A) consists of a single
aromatic moiety located on the central phenyl ring, five hydrogen bond acceptors located on the sulfur
and on the nitrogen atoms of the benzothiazole group, on the sulfur atom of the linker between the two
benzothiazole rings, on the oxygen of the ethoxy group that is connected to the benzothiazole moiety
and on the second ("right hand side") benzothiazole moiety and two hydrophobic moieties located on
the ethoxy group and on the phenyl group of the second benzothiazole moiety. The fit values resulting
from the mapping of the compounds considered in this work into this pharmacophore model are
ACS Paragon Plus Environment

Page 13 of 49
Journal of Medicinal Chemistry
presented in Supplemental Table 5 and show that all of the active compounds are characterized by high
fit values (above cutoff of 19; 100% success) whereas 24 out of the 25 inactive compounds are
characterized by values below this threshold (96% success). In particular, this pharmacophore model
well explained the activity difference between 34 (active) and its close, albeit inactive analog 31. These
two compounds differ by a single moiety, namely, a sulfur atom in 34 that maps into an Hꢀbond
acceptor pharmacophoric feature versus an NH group (an Hꢀbond donor) in 31 (see arrows in Figure
6BꢀC).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While the two pharmacophores were built based on different structures they share many
common features, in particular in the region corresponding to the ethoxybenzothiazole moiety of
compound 34 (see Figures 5 and 6 where the pharmacophoric features corresponding to the
ethoxybenzothiazole moiety are circled ). Yet the two pharmacophores differ in the regions
corresponding to the second ("right hand side") benzothiazole moiety and the linker between the two
ring systems of 34. Importantly, the Hꢀbond acceptor feature responsible for reproducing the activity
differences between compounds 34 and 31 (see above) is missing in the first pharmacophore.
Accordingly, the secondꢀgeneration pharmacophore outperforms the first one in terms of its ability to
differentiate between active and inactive compounds (see above and Tables 4 and 5). This is not
surprising since this pharmacophore was built on a larger collection of compounds taking into account
both active and inactive ones. This new pharmacophore model will direct our future synthesis efforts
aimed at developing yet more potent benzothiazoleꢀbased AMPK activators.
Validation of our in vitro results, that PTꢀ1 and 34 directly activate AMPK, was conducted using
the binding mode suggested by Pang. 20 Our docking simulations identified favorable interactions
between PT1 and specific binding site residues. The lead compound of this study, 34 together with the
reference compound PT1 were docked into the binding site of AMPK to gain additional insight into
their potential interactions with the kinase. In particular, PT1 forms a salt bridge with K154 which was
shown by mutation studies to be important for the binding of this compounds .20 Additional interactions
are formed with S97 and A149 (Hꢀbonds), K78 (πꢀcation interaction) and P74 (πꢀπ interaction; see
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 49
Figure 7A). Only one of these interactions is preserved for analogue 34 (Hꢀbond with K154). However,
this compound is stabilized by an additional Hꢀbond with K41 (figure 7B). These differences in the
interactions pattern between the three compounds are reflected in their overall docking scores (ꢀ5.24, ꢀ
3.65 for PT1 and 34, respectively). Overall, our docking simulations suggest that PT1 is the best binder
followed by analogue 34. However, these results do not indicate the degree to which these compounds
could activate the protein.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Conclusion
Various substituted benzothiazol derivatives were synthesized and screened for their potential
hypoglycemic (antihyperglycemic) activity. The ethoxybenzothiazole moiety in 34 was found to be
critical for increasing glucose transport in L6 myotubes and for the activation of AMPK. In accord with
this observation, this moiety fits three of the pharmacophoric features found to be important for the
biologic activity (hydrophobic, aromatic and Hꢀbond acceptor, located on the ethoxy, phenyl and
thiazolidin ring, respectively; see Figure 6B). Compound 34 significantly increases the rate of glucose
uptake in L6 myotubes at pharmacologically relevant concentrations (25 ꢂM). Moreover, already at
concentration of 5 ꢁM it increased phosphorylation of AMPK in L6 myotubes.
The activity of 34 is not mediated by the insulin transduction pathway, since the inhibition of
both key insulin sensitive enzymes PI3K and AKT/PKB did not alter 34ꢀinduced stimulation of glucose
uptake. The protein ASꢀ160, shared by the insulinꢀ and AMPKꢀinduced GLUT4 translocation
mechanisms, was markedly phosphorylated by 34. These results emphasize an important property of
34, namely, increasing glucose uptake in a skeletal muscle in vitro model (L6 myotubes) in the absence
of insulin. The stimulatory effect of 34 was blocked by an AMPK inhibitor (compound C), pointing to
the essential role of AMPK activation in the biological effect of 34. The enzyme ACC, which is also
activated by AMPK, is not directly related to glucose uptake regulation in skeletal muscles. However,
this downstream target of AMPK was phosphorylated and activated by 34 in L6 myotubes, thereby
ACS Paragon Plus Environment

Page 15 of 49
Journal of Medicinal Chemistry
supporting our main conclusion that compound 34 leads to physiological phosphorylation of AMPK
and activation of all its downstream pathways. Our docking simulations and direct effect of 34 on
recombinant human AMPK suggest that such activation may result from a direct binding of compound
34 to AMPK. The biological relevance of these in vitro findings have been demonstrated by the effect
of compound 34 in decreasing the blood glucose level in diabetic KKAy mice. Metabolic instability of
the 34 can be improved by replacing of methylꢀthio linker by more metabolic stable linkers. Thus, also
may lead to oral activity of those new 34 derivatives.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Experimental Section
Materials and Methods
ATP Assay Kit (Colorimetric) and rabbit polyclonal antiꢀglucose transporter GLUT1 and GLUT4
antibodies were obtained from Abcam (Cambridge, MA). Human insulin (Actrapid) was purchased
from Novo Nordisk (Bagsvaerd, Denmark). AICAR, BSA, MT, compound 4, compound C, 2ꢀdeoxyꢀDꢀ
glucose (dGlc), Dꢀglucose, Oꢀphenylenediamine (OPD), and the protease inhibitor cocktail were
purchased from SigmaꢀAldrich Chemicals (Rehovot, Israel). Compound 2 was from SynthonꢀLab Ltd.
(St. Petersburg, Russia). Chemical Block Ltd. (Moscow, Russia) supplied compound 3. Glycerol and
sodium
fluoride
were
from
Merck
(Whitehouse
Station,
NJ).
Mercaptoethanol,
phenylmethanesulfonylfluoride (PMSF), sodium orthovanadate, sodiumꢀβꢀglycerophosphate, sodium
pyrophosphate and SDS were purchased from Alfa Aesar (Ward Hill, MA). PTꢀ1 was supplied by
Tocris (Bristol, UK). ADPꢀGlotm AMPK A1/B1/G1 kinase kit was purchased from SignalChem
(Richmond, Canada). American Radiolabeled Chemicals (St. Louis, MO) supplied [3H]dGlc [2.22
TBq/mmol (60 Ci/mmol)]. Antibodies against AMPKα, ASꢀ160, pThr642ꢀASꢀ160 and pThr172ꢀAMPK
were purchased from Cell Signaling Technology (Beverly, MD). ACC, pSer79ꢀACC and αꢀtubuline were
from Millipore (Billerica, MA). The AntiꢀcꢀMyc (Aꢀ14) antibody was from Santa Cruz Biotechnology
(Santa Cruz, CA), horseradish peroxidase (HRP)ꢀconjugated antiꢀrabbit IgG and EZꢀECL
chemoluminescence detection kit were from Jackson ImmunoResearch (West Grove, PA). Goat serum,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 49
fetal calf serum (FCS), Lꢀglutamine, αꢀminimal essential medium (MEM) and antibiotics were
purchased from Biological Industries (BethꢀHaemek, Israel). Organic solvents (HPLC grade) were from
Frutarom Ltd. (Haifa, Israel). Dry THF was obtained using distillation from a boiled blue color mix with
sodium/benzophenone. The melting points were determined with FisherꢀJohns melting point apparatus
1
2
3
4
5
6
7
8
9
1
13
(Palmerton, PA). The H NMR and C NMR spectra were recorded at room temperature on a Bruker
Advance NMR spectrometer (Vernon Hills, IL) operating at 200 and 300 MHz and were in accord with
the assigned structures. Chemical shift values were reported relative to TMS that was used as an internal
standard. The samples were prepared by dissolving the synthesized compounds in DMSOꢀd6 (δH = 2.50
ppm, δC = 39.52 ppm) or in CDCl3 (δH = 7.26 ppm, δC = 77.16 ppm). 36 Chemical shifts were expressed
in δ (ppm) and coupling constants (J) in hertz. The splitting pattern abbreviations are as follows: s,
singlet; d, doublet; t, triplet; q, quartet; quint, quintet; m, unresolved multiplet due to the field strength
of the instrument; dd, doublet of doublet. A QTof micro spectrometer (Micromass, Milford, MA) in the
positive ion mode was used for mass spectrometry. Data were processed using massLynX v.4.1
calculation and deconvolution software (Waters Corporation, Milford, MA). Column chromatography
was performed on Merck Silica gel 60 (230ꢀ400 mesh; Merck, Darmstadt, Germany). Analytical and
preparative HPLC (Young Lin Instruments, Anyang, Korea) were performed on LUNA C18(2)
preparative (10 ꢁm, 100 x 30 mm) or analytical (5 ꢁm, 250 x 4.6 mm) columns, both from Phenomenex
Inc. (Torrance, CA). Acetonitrile and doubly distilled water were used as an eluent in different ratios.
Analytical thin layer chromatography was carried out on preꢀcoated Merck Silica gel 60F254 (Merck)
sheets using UV absorption and iodine physical adsorption for visualization. The ≥95% purity of the
final compounds 19, 22ꢀ34 and the <95% purity of the final compounds 20, 21 was confirmed using
HPLC analysis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General procedure for the synthesis of Sꢀalkyl benzothiazoles (20ꢀ22).
To an iceꢀcold stirred solution of 19 (0.20 g, 1.09 mmol) in dry THF (15 mL) sodium hydride (0.11 g,
4.58 mmol) was added in three portions. Then, the appropriate alkylhalide (2.73 mmol in dry THF, 2
ACS Paragon Plus Environment

Page 17 of 49
Journal of Medicinal Chemistry
mL) was added dropꢀwise to the reaction solution. Stirring at 0 °C was continued for 1 hour followed by
additional 1 hour incubation at room temperature. The reaction progress was followed by TLC (100%
CH2Cl2). The reaction was quenched by the addition of ice (about 15 g) followed by 100 mL of
chloroform. The organic layer was separated, washed twice in water, dried over sodium sulfate, filtered
and evaporated to yield the solid product. The compounds were purified by column chromatography
using 100% dichloromethane as an eluent.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2ꢀ(2,2ꢀdimethoxyethylthio)benzo[d]thiazolꢀ6ꢀol (20). Yield ~70% colorless syrup. The purity of the
compound was 60% we did not succeed purifying 20 (See supplemental information). 1H NMR (CDCl3,
200 MHz): δ 1.43 (s, 6H, ꢀSCH2CH(OCH3)2), 3.53 (d, 2H, J=5.4, ꢀSCH2CH(OCH3)2), 4.71 (t, 1H,
J=5.4 Hz, ꢀSCH2CH(OCH3)2), 6.90ꢀ6.98 (m, 2H, HOꢀPhꢀ), 7.20 (d, 1H, J=2.4 Hz, HOꢀPhꢀ), 7.68 (d,
13
1H, J=9.0 Hz, HOꢀPhꢀ). C NMR was not conducted because the compound disintegrated during the
night NMR testing. HPLC (gradient water/CH3CN from 100 % of water in 60 min, flow = 1 mL/min, λ
= 254 nm): tR = 40.6 min.
2ꢀ(2ꢀethylbutylthio)benzo[d]thiazolꢀ6ꢀol (21). Yield 50% white powder, mp 105ꢀ110 °C. The purity of
the compound was 86% we did not succeed purifying 21 more than that (See supplemental information).
1H NMR (CDCl3, 300 MHz): δ 0.90 (t, 6H, J=9.0 Hz, ꢀSCH2CH(CH2CH3)2), 1.44ꢀ1.52 (m, 4H, ꢀ
SCH2CH(CH2CH3)2), 1.65ꢀ1.70 (m, 1H, ꢀSCH2CH(CH2CH3)2), 3.32 (d, 2H, J=6.3 Hz, ꢀ
SCH2CH(CH2CH3)2), 6.92 (dd, 1H, J=2.5, 6.3 Hz, HOꢀPhꢀ), 7.20 (d, 1H, J=2.4 Hz, HOꢀPhꢀ), 7.71 (d,
1H, J=8.7 Hz, HOꢀPhꢀ). 13C NMR (CDCl3, 50 MHz): δ 10.7 (ꢀSCH2CH(CH2CH3)2), 25.1 (ꢀ
SCH2CH(CH2CH3)2), 37.6 (ꢀSCH2CH(CH2CH3)2), 40.5 (ꢀSCH2CH(CH2CH3)2), 106.5 (HOꢀPhꢀ), 115.0
(HOꢀPhꢀ), 121.8 (HOꢀPhꢀ), 129.7 (HOꢀPhꢀ), 141.0 (HOꢀPhꢀ), 143.1 (HOꢀPhꢀ), 154.6 (HOꢀPhꢀ
[d]thiazoleꢀSꢀ). MS (ES+): 267 (MH+). HRMS (CI+) 267.0770 (M+), 268.076 (MH+). HPLC (gradient
water/CH3CN from 100 % of water in 60 min, flow = 1 mL/min, λ = 254 nm): tR = 48.2 min.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 49
2ꢀ(butylthio)benzo[d]thiazolꢀ6ꢀol (22). Yield 60% light brown crystals, mp 80ꢀ85 °C. 1H NMR (CDCl3,
300 MHz): δ 0.90 (t, 3H, J=7.3 Hz, ꢀSCH2CH2CH2CH3), 1.43ꢀ1.55 (m, 2H, ꢀSCH2CH2CH2CH3), 1.73ꢀ
1.83 (m, 2H, ꢀSCH2CH2CH2CH3), 3.30 (t, 2H, J=7.3 Hz, ꢀSCH2CH2CH2CH3), 6.92 (dd, 1H, J=2.5, 6.3
1
2
3
4
5
6
7
8
13
Hz, HOꢀPhꢀ), 7.21 (d, 1H, J=2.4 Hz, HOꢀPhꢀ), 7.71 (d, 1H, J=8.7 Hz, HOꢀPhꢀ). C NMR (CDCl3, 50
9
MHz): δ 13.4 (ꢀSCH2CH2CH2CH3), 21.7 (ꢀSCH2CH2CH2CH3), 31.1 (ꢀSCH2CH2CH2CH3), 33.7 (ꢀ
SCH2CH2CH2CH3), 106.7 (HOꢀPhꢀ), 115.6 (HOꢀPhꢀ), 121.6 (HOꢀPhꢀ), 136.1 (HOꢀPhꢀ), 147.0 (HOꢀPhꢀ
), 153.8 (HOꢀPhꢀ), 165.0 (HOꢀPhꢀ[d]thiazoleꢀSꢀ). MS (ES+): 240 (MH+). HRMS (CI+) 239.0462 (M+),
240.050 (MH+). HPLC (gradient water/CH3CN from 100 % of water in 60 min, flow = 1 mL/min, λ =
254 nm): tR = 41.7 min.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General procedure for the synthesis of Sꢀalkyl 6ꢀethoxybenzo[d]thiazoles (25ꢀ27).
These compounds were synthesized according to the synthetic procedure for Sꢀalkyl benzo[d]thiazoles,
using was 4 as the starting molecule. 37
6ꢀethoxyꢀ2ꢀ(propylthio)benzo[d]thiazole (25). The remaining yellow syrup was further purified by silica
gel column chromatography (eluent: 100% dichloromethane). Attempts to crystallize the syrup were
1
unsuccessful. Yield 10% colorless syrup. H NMR (CDCl3
,
200 MHz): δ 1.05 (t, 3H, J=7.4 Hz, ꢀ
SCH2CH2CH3), 1.41 (t, 3H, J=7.0 Hz, ꢀPhꢀOCH2CH3), 1.82 (m, 2H, ꢀSCH2CH2CH3), 3.26 (t, 2H, J=7.2
Hz, ꢀSCH2CH2CH3), 4.01 (q, 2H, J=6.9 Hz, ꢀPhꢀOCH2CH3), 6.97 (dd, 1H, J=2.5, 6.4 Hz, EtOꢀPhꢀ),
7.18 (d, 1H, J=2.2 Hz, EtOꢀPhꢀ), 7.73 (d, 1H, J=9.0 Hz, EtOꢀPhꢀ). 13C NMR (CDCl3, 50 MHz): δ 13.2
(ꢀSCH2CH2CH3), 14.7 (ꢀPhꢀOCH2CH3), 22.6 (ꢀSCH2CH2CH3), 35.5 (ꢀSCH2CH2CH3), 63.9 (ꢀPhꢀ
OCH2CH3), 104.6 (EtOꢀPhꢀ), 115.0 (EtOꢀPhꢀ), 121.7 (EtOꢀPhꢀ), 136.3 (EtOꢀPhꢀ), 147.7 (EtOꢀPhꢀ),
156.2 (EtOꢀPhꢀ), 163.8 (EtOꢀPhꢀ[d]thiazoleꢀSꢀ). MS (ES+): 254 (MH+). HRMS (CI+) 253.0583 (M+),
254.064 (MH+). HPLC (gradient water/CH3CN from 100 % of water in 60 min, flow = 1 mL/min, λ =
254 nm): tR = 52.1 min.
ACS Paragon Plus Environment

Page 19 of 49
Journal of Medicinal Chemistry
6ꢀethoxyꢀ2ꢀ(2ꢀethylbutylthio)benzo[d]thiazole (26). The compound was obtained as a yellow syrup that
was further purified by silica gel column chromatography (eluent: 100% dichloromethane). Attempts to
crystallize the syrup were unsuccessful. Yield 10% colorless syrup. 1H NMR (CDCl3, 300 MHz): δ 0.93
(t, 6H, J=7.4 Hz, ꢀSCH2CH(CH2CH3)2), 1.41ꢀ1.50 (m, 4H, ꢀSCH2CH(CH2CH3)2), 1.63ꢀ1.70 (m, 1H, ꢀ
SCH2CH(CH2CH3)2), 3.33 (d, 2H, J=6.3 Hz, ꢀSCH2CH(CH2CH3)2), 4.05 (q, 2H, J=7.0 Hz, ꢀPhꢀ
OCH2CH3), 6.98 (dd, 1H, J=2.4, 6.6 Hz, EtOꢀPhꢀ), 7.20 (d, 1H, J=2.4 Hz, EtOꢀPhꢀ), 7.73 (d, 1H, J=9.0
Hz, EtOꢀPhꢀ). 13C NMR (CDCl3, 75 MHz): δ 10.9 (ꢀSCH2CH(CH2CH3)2), 14.8 (ꢀPhꢀOCH2CH3), 25.2 (ꢀ
SCH2CH(CH2CH3)2), 37.6 (ꢀSCH2CH(CH2CH3)2), 40.7 (ꢀSCH2CH(CH2CH3)2), 64.1 (ꢀPhꢀOCH2CH3),
104.8 (EtOꢀPhꢀ), 115.1 (EtOꢀPhꢀ), 121.8 (EtOꢀPhꢀ), 136.4 (EtOꢀPhꢀ), 147.8 (EtOꢀPhꢀ), 156.3 (EtOꢀPhꢀ
[d]thiazoleꢀSꢀ). MS (ES+): 296 (MH+), 318 (MNa+). HRMS (CI+) 295.1020 (M+), 296.106 (MH+).
HPLC (gradient water/CH3CN from 100 % of water in 60 min, flow = 1 mL/min, λ = 254 nm): tR =
61.1 min.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2ꢀ(butylthio)ꢀ6ꢀethoxybenzo[d]thiazole (27). The compound was purified by silica gel column
chromatography (eluent: 100% dichloromethane). Yield 30% colorless crystalline needles, mp 30ꢀ35
1
°C. H NMR (CDCl3, 200 MHz): δ 1.12 (t, 3H, J=7.0 Hz, ꢀSCH2CH2CH2CH3), 1.56ꢀ1.74 (m, 5H, ꢀ
SCH2CH2CH2CH3, ꢀPhꢀOCH2CH3), 1.87ꢀ2.02 (m, 2H, ꢀSCH2CH2CH2CH3), 3.46 (t, 2H, J=7.3 Hz, ꢀ
SCH2CH2CH2CH3), 4.21 (q, 2H, J=7.0 Hz, ꢀPhꢀOCH2CH3), 7.15 (dd, 1H, J=2.4, 6.6 Hz, EtOꢀPhꢀ), 7.37
13
(d, 1H, J=2.6 Hz, EtOꢀPhꢀ), 7.90 (d, 1H, J=9.0 Hz, EtOꢀPhꢀ). C NMR (CDCl3, 50 MHz): δ 13.4(ꢀ
SCH2CH2CH2CH3), 14.7 (ꢀPhꢀOCH2CH3), 21.8 (ꢀSCH2CH2CH2CH3), 31.2 (ꢀSCH2CH2CH2CH3), 33.4
(ꢀSCH2CH2CH2CH3), 64.0 (ꢀPhꢀOCH2CH3), 104.7 (EtOꢀPhꢀ), 115.0 (EtOꢀPhꢀ), 121.8 (EtOꢀPhꢀ), 136.2
(EtOꢀPhꢀ), 156.1 (EtOꢀPhꢀ), 163.9 (EtOꢀPhꢀ[d]thiazoleꢀSꢀ). MS (ES+): 268 (MH+), 318 (MNa+). HRMS
(MALDI) 268.082 (MH+). HPLC (gradient water/CH3CN from 100 % of water in 60 min, flow = 1
mL/min, λ = 254 nm): tR = 44.8 min.
General procedure for the synthesis of (6ꢀethoxybenzo[d]thiazolꢀ2ꢀyl)methanamine derivatives (31ꢀ32).
To a stirred solution of commercially available 6ꢀethoxybenzo[d]thiazolꢀ2ꢀamine (1.00 g, 4.16 mmol) or
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 49
2ꢀ(4ꢀaminophenyl)ꢀ6ꢀmethylbenzo[d]thiazole (1.00 g, 5.15 mmol) with triethylamine (1.08 mL, 7.72
mmol or 0.87 mL, 6.24 mmol, respectively) and 4ꢀdimethylaminopyridine (0.06 g, 0.49 mmol or 0.05 g,
0.41 mmol, respectively) in THF (35 or 50 mL, respectively) was added previously synthesized 2ꢀ
(chloromethyl)benzo[d]thiazole according to Russell et al. 38 (0.94 g, 5.15 mmol or 0.76 g, 4.14 mmol).
The reaction mixture stirred in room temperature or refluxed overnight, respectively. Then, the solution
was cooled to room temperature, and water (100 mL) followed by chloroform (100 mL) were added.
The organic layer was separated, washed twice in water, dried over sodium sulfate, filtered and
evaporated. The crude solid was purified by HPLC.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nꢀ(benzo[d]thiazolꢀ2ꢀylmethyl)ꢀ6ꢀethoxybenzo[d]thiazolꢀ2ꢀamine (31). Yield 20% brown solid, mp 150ꢀ
1
156 °C. H NMR (DMSOꢀd6, 200 MHz): δ 1.33 (t, 3H, J=6.9 Hz, ꢀPhꢀOCH2CH3), 4.04 (q, 2H, J=6.9
Hz, ꢀPhꢀOCH2CH3), 5.22 (s, 2H, ꢀNHꢀCH2ꢀPhꢀ), 7.03 (d, 2H, J=8.8 Hz, EtOꢀPhꢀ), 7.22 (d, 2H, J=8.6
Hz, EtOꢀPhꢀ), 7.53 (m, 2H, EtOꢀPhꢀ), 8.09 (dd, 1H, J=9.0 Hz, EtOꢀPhꢀ). 13C NMR (DMSOꢀd6, 50
MHz): δ 14.6 (ꢀPhꢀOCH2CH3), 32.6 (ꢀNHꢀCH2ꢀPhꢀ), 63.7 (ꢀPhꢀOCH2CH3), 115.1 (ꢀPhꢀOEt), 121.8 (ꢀ
PhꢀOEt), 123.1 (ꢀbenzo[d]thiazole), 126.4 (ꢀbenzo[d]thiazole), 126.6 (ꢀPhꢀOEt), 127.1 (ꢀ
benzo[d]thiazole), 128.0 (ꢀSCH2ꢀbenzo[d]thiazole), 151.8 (ꢀPhꢀOEt, ꢀbenzo[d]thiazole), 159.3 (ꢀPhꢀOEt,
ꢀbenzo[d]thiazole), 172.0 (ꢀbenzo[d]thiazole). MS (ES+): 342 (MH+). HPLC (gradient water/CH3CN
from 100 % of water in 60 min, flow = 1 mL/min, λ = 254 nm): tR = 26.8 min.
Nꢀ(benzo[d]thiazolꢀ2ꢀylmethyl)ꢀ4ꢀ(6ꢀmethylbenzo[d]thiazolꢀ2ꢀyl)aniline (32). Yield 40% yellow
1
solid, mp 143ꢀ145 °C. H NMR (CDCl3, 400 MHz): δ 2.46 (s, 3H, CH3ꢀPhꢀ), 4.85 (s, 2H, ꢀNHꢀCH2ꢀ),
6.77 (dd, 2H, J=2.0, 4.8 Hz, ꢀPhꢀNHꢀ), 7.24 (dd, 1H, J=1.4, 7.2 Hz, CH3ꢀbenzo[d]thiazole), 7.38 (t, 1H,
J=7.6 Hz, ꢀCH2ꢀbenzo[d]thiazole), 7.49 (t, 1H, J=7.6 Hz, ꢀCH2ꢀbenzo[d]thiazole), 7.62 (s, 1H, CH3ꢀ
benzo[d]thiazole), 7.83ꢀ7.91 (m, 4H, CH3ꢀbenzo[d]thiazole, ꢀCH2ꢀbenzo[d]thiazole), 8.01 (d,1H, J=8.4
Hz, CH3ꢀbenzo[d]thiazole). 13C NMR (CDCl3, 150 MHz): δ 21.5 (CH3ꢀPhꢀ), 46.6 (ꢀNHꢀCH2ꢀ), 113.1 (ꢀ
PhꢀNHꢀ), 121.2 (ꢀPhꢀNHꢀ), 121.8 (CH3ꢀPhꢀ), 122.0 (CH3ꢀPhꢀ), 122.9 (ꢀbenzo[d]thiazole), 124.4 (ꢀCH2ꢀ
benzo[d]thiazole), 125.2 (ꢀbenzo[d]thiazole), 126.2 (ꢀbenzo[d]thiazole), 127.6 (CH3ꢀPhꢀ), 129.0 (ꢀPhꢀ
ACS Paragon Plus Environment

Page 21 of 49
Journal of Medicinal Chemistry
NHꢀ), 134.6 (CH3ꢀPhꢀ), 134.8 (ꢀbenzo[d]thiazole), 134.9 (CH3ꢀPhꢀ), 149.1 (ꢀPhꢀNHꢀ), 152.4 (CH3ꢀPhꢀ),
153.4 (ꢀbenzo[d]thiazole), 167.3 (ꢀbenzo[d]thiazole), 171.3 (ꢀbenzo[d]thiazole). MS (ES+): 388 (MH+).
HRMS (MALDI) 388.096 (MH+). HPLC (gradient water/CH3CN from 100 % of water in 60 min, flow
= 1 mL/min, λ = 254 nm): tR =51.4 min.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2ꢀ(benzo[d]thiazolꢀ2ꢀylmethylthio)ꢀ6ꢀethoxybenzo[d]thiazole
(34).
The
compound
2ꢀ
(chloromethyl)benzo[d]thiazole (1.00 g, 5.45 mmol), dissolved in THF (50 mL), was added dropꢀwise
to a stirred solution of 4 (1.15 g, 5.45 mmol), Nꢀethyldiisopropylethylamine (1.35 mL, 8.17 mmol) and
4ꢀdimethylaminopyridine (0.08 g, 0.54 mmol) in THF (50 mL). The reaction mixture was refluxed
overnight. Then, the solution was cooled to room temperature, and water (100 mL) followed by
chloroform (100 mL) were added. The organic layer was separated, washed twice in water, dried over
sodium sulfate, filtered and evaporated. The crude solid was purified by column chromatography using
100% dichloromethane as an eluting solvent. The compound was obtained as dark brown solid. Yield
1
35% brown solid, mp 78ꢀ82 °C. H NMR (DMSOꢀd6, 200 MHz): δ 1.34 (t, 3H, J=6.9 Hz, ꢀPhꢀ
OCH2CH3), 4.05 (q, 2H, J=7.0 Hz, ꢀPhꢀOCH2CH3), 5.09 (s, 2H, ꢀSCH2ꢀbenzo[d]thiazole), 7.04 (dd, 1H,
J=2.5, 6.4 Hz, ꢀSCH2ꢀbenzo[d]thiazole), 7.38ꢀ7.54 (m, 2H, ꢀSCH2ꢀbenzo[d]thiazole), 7.60 (d, 1H, J=2.4
13
Hz, ꢀPhꢀOEt), 7.76 (d, 1H, J=9.0 Hz, ꢀPhꢀOEt), 7.95ꢀ8.06 (m, 2H, ꢀPhꢀOEt). C NMR (DMSOꢀd6, 50
MHz): δ 14.9 (ꢀPhꢀOCH2CH3), 34.9 (ꢀSCH2ꢀbenzo[d]thiazole), 64.0 (ꢀPhꢀOCH2CH3), 105.8 (ꢀPhꢀOEt),
115.9 (ꢀPhꢀOEt), 122.2 (ꢀSCH2ꢀbenzo[d]thiazole), 122.6 (ꢀSCH2ꢀbenzo[d]thiazole), 122.8 (ꢀPhꢀOEt),
125.6 (ꢀSCH2ꢀbenzo[d]thiazole), 126.6 (ꢀSCH2ꢀbenzo[d]thiazole), 136.8 (ꢀPhꢀOEt, ꢀSCH2ꢀ
benzo[d]thiazole), 147.0 (ꢀPhꢀOEt, ꢀSCH2ꢀbenzo[d]thiazole), 156.4 (ꢀPhꢀOEt), 161.5 (OEtꢀ
benzo[d]thiazole), 168.4 (ꢀCH2ꢀbenzo[d]thiazole). MS (ES+): 359 (MH+). HPLC (gradient
water/CH3CN from 100 % of water in 60 min, flow = 1 mL/min, λ = 254 nm): tR =60.0 min.
Molecular modeling
Pharmacophore models
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 49
The first pharmacophore model was generated from compound PTꢀ1 using a Common Feature
Pharmacophore Generation Procedure as implemented in Catalyst/HipHop which is part of the
Discovery Studio (DS) V3.5 suit of programs (Accelrys Software Inc., Discovery Studio Modeling
Environment, Release 3.5, San Diego: Accelrys Software Inc., 2012). Prior to pharmacophore
generation, compound PTꢀ1 was subjected to a conformational search procedure using the CHARmm
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
force field and the poling algorithm as implemented in the Catalyst/BEST procedure. 255 conformations
were generated within an energy cutoff of 20.0 Kcal/mol and the lowest energy one was used as input to
the pharmacophore generation procedure. For hypothesis generation, the following chemical functions
were selected in the feature directory of Catalyst: hydrogen bond acceptor, hydrogen bond donor,
aromatic ring and hydrophobic groups. Compounds were fitted to this model by using the Ligand
Pharmacophore mapping protocol as implemented in DS. Briefly, compounds were subjected to a
conformational search procedure as described above and the resulting conformations were subsequently
fitted into the pharmacophore model using a flexible fitting method while setting the maximum omitted
features parameter of 3. Fit values were obtained according to equation 1 and compared to the
biological activity of these compounds.

2 

D
T




(1)
FitValue = W 1−

∑
∑






Where W is a weight factor, D is the displacement of the feature from the center of the location
constraint and T is the radius of the location constraint sphere for the feature.
and compared to the biological activity of these compounds.
The second pharmacophore model was generated from the three most active compounds (29, 33
and 34) using a Common Feature Pharmacophore Generation procedure as implemented in DS V3.5.
Catalyst/Best method with energy threshold of 20.0 Kcal/mol was chosen for conformation generation.
Two inactive compounds (25 and 31) were selected for refinement of the pharmacophore hypothesis.
Refinement was performed by setting higher weights (10) to the features that are missing in the inactive
compounds but which are present in the active ones. These features include a hydrogen bond acceptor
ACS Paragon Plus Environment

Page 23 of 49
Journal of Medicinal Chemistry
located on the linker between two benzothiazole moieties and a hydrogen bond acceptor located on the
second benzothiazole moiety. The ligand pharmacophore mapping protocol was used for mapping the
set of 31 ligands considered in this work to the pharmacophore model described above while employing
the flexible fitting method and setting a maximum of two omitted features. The resulting fit values were
compared to the biological activity of the ligands.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Docking simulations
The structure of AMPK was downloaded from the PDB (2Y94) and prepared for docking simulations
using the prepare protein module implemented in DS. Structures of PT1 and of compound 34 were
drawn in ChemDraw and prepared for docking using the LigPrep procedure as implemented in Maestro
(Maestro, version 9.3, Schrodinger, LLC, New York, NY, 2012). Docking was performed using the
induced fit procedure (IFD) as implemented in Maestro which combines docking by Glide with
sidechain optimization performed by Prime. In the present study we used extra precision (XP) docking
with the OPLS2005 force field. All other IFD parameters were kept at their default values. The docking
box was positioned at the vicinity of residues Glu 94 and Lys 154 in accord with the position of PT1
within the AMPK binding site as previously suggested by Pang et al. 20
Animals
Diabetic male KKAy mice (8 to 12ꢀweek old, 20ꢀ25 g) were purchased from The Jackson Laboratory
(Bar Harbor, ME). The animals were routinely kept in 12ꢀh light/dark cycles and provided with food
and water ad libitum. Blood glucose levels at sacrifice ranged between 17ꢀ25 mmol/l. The ethics
committee of the Hebrew University of Jerusalem approved the study protocol for animal welfare. The
Hebrew University of Jerusalem is an Association for Assessment and Accreditation of Laboratory
Animal Care (AAALAC) international accredited institution.
Intraperitoneal Glucose Tolerance Test (ipꢀGTT)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 49
A standard ipꢀGTT test was performed on mice after an overnight fast. Glucose in saline was injected ip
(1.5 g/kg body weight). Venous blood samples from small tail clips were taken for glucose
determination using a glucometer (FreeStyle Freedom, Abbott Diabetes Care, Alamada, CA).
1
2
3
4
5
6
7
8
9
L6 myotube cultures
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
L6 myotubes were maintained as previously described.39 All experiments were conducted on fully
differentiated myotubes.
2ꢀ[3H,1]ꢀDeoxyꢀDꢀglucose uptake assay
The rate of [3H]dGlc uptake in myotubes, in the absence or presence of insulin, was determined as
previously described. 7 Briefly, L6 myotube cultures were preincubated in αꢀMEM supplemented with
2% (v/v) FCS and 23.0 mmol/L. Dꢀglucose for 24 h, and then 4 h before the experiment the medium
was changed to serumꢀfree medium with 0.5% BSA. The myotubes were then treated as described in
Fig. 1. The insulin effect was measured after its addition (200 nmol/L) to cultures for the last 30 min of
treatment. The cultures were then rinsed 3 times with PBS at room temperature and incubated with PBS,
pH 7.4, containing 0.1 mmol/L dGlc and 1.3 ꢁCi/mL [3H]dGlc for 5 min at room temperature. At the
end of the assay, the myotubes were lysed in 0.1% (w/v) SDS in water and taken for liquid scintillation.
Preparation of cell lysates and WB (Western blot) analysis
Whole cell lysates were prepared as previously described 7 with some minor modifications: the lysis
buffer contained 50 mmol/L TrisꢀHCl, pH 7.5, 1 mmol/L EDTA, 1mmol/L EGTA, 1 mmol/L Na3VO4,
150 mmol/L NaCl, 50 mmol/L NaF, 10 mmol/L sodiumꢀglycerophosphate, 5 mmol/L sodium
pyrophosphate, and 1 mmol/L PMSF, supplemented with 0.1% (v/v) NPꢀ40, 0.1% (v/v) 2ꢀβꢀ
mercaptoethanol, and protease inhibitor cocktail (1:100 dilution). The cells were washed with iceꢀcold
PBS, and 1 mL of lysis buffer was then added and incubated at 4 °C for 40 min. The resulting cell
lysates were centrifuged at 8,700 g x 30 min at 4 °C and the resulting supernatant fractions were
ACS Paragon Plus Environment

Page 25 of 49
Journal of Medicinal Chemistry
separated and kept at ꢀ20 °C until used. Protein content in the supernatant was determined according to
Bradford 40, using a BSA standard dissolved in the same buffer. Aliquots (5ꢀ60 ꢁg of protein) were
mixed with the sample buffer [62.5 mmol/L TrisꢀHCl, pH 6.8, 2% (w/v) SDS, 10% (v/v) glycerol, 50
mmol/L DTT, and 0.01% (w/v) bromophenol blue], heated at 95°C for 5 min. Samples for WB analyses
of GLUT1 and GLUT4 were prepared as previously described. 41 The proteins were separated on 10%
SDSꢀPAGE and WB analyses were performed using antibodies according to our previously established
protocols.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Colorimetric determination of surface GLUT1myc and GLUT4myc in L6 myotubes
The colorimetric detection of surface GLUT4myc or GLUT1myc in L6 myotubes was performed as
described. 33 Briefly, cultured myotubes were incubated with rabbit antiꢀcꢀMyc antibody (1:200
dilution), washed and fixed with 3% formaldehyde, and further interacted with goat HRPꢀconjugated
antiꢀrabbit IgG (1:2000 dilution). Following the washes, a solution of OPD was added, and the culture
plates were taken for absorbance measurement at 492 nm to estimate the relative abundance of
GLUT1myc or GLUT4myc on the plasma membrane of the myotubes. The GLUT1myc and
GLUT4myc L6 cells were the courtesy of Dr. A. Klip, Hospital for Sick Children (Toronto, ON,
Canada).
Colorimetric determination of cellular ATP level in L6 myotubes
ATP content in lysates of L6 myotubes was measured with the ATP assay kit. A 34, DMSO and known
uncoupler dinitrophenol (DNP) as a positive control were added to L6 myotubes for indicate din the
Table 3 times. Myotubes in 6ꢀwell plates were washed 3 times with cold 0.15 M Tris HCl, pH 7.75. The
wells were then treated with 200 ꢂL of 0.5% (w/v) iceꢀcold TCA, followed by the addition of 400 ꢂL of
TrisꢀHCl buffer. The mixture was centrifuged in Eppendorf tubes (12,000 rpm, 30 min at 4 °C) and the
supernatant taken for ATP determination according to the manufacturer’s instructions. The assay was
performed in 96ꢀmultiꢀwell plates (Nunc, Roskilde, Denmark) in an EXx800 spectrophotometer (Biotec
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 49
Instruments, Winooski, VT) at room temperature. ATP standard curve was constructed according to the
manufacturer’s protocol, using the ATP supplied in the kit.
1
2
3
4
5
LogP calculations
6
7
8
9
Theoretical LogP values were calculated using Advanced Chemistry Development (ACD/Labs)
Software V11.02 (© 1994ꢀ2012 ACD/Labs). The experimental LogP/D for 34 was determined as
described in SI.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Statistical analysis
Results are given as Mean ± SEM. Statistical significance (p<0.05) was calculated among experimental
groups using the twoꢀtailed Student’s t test. The http://www.graphpad.com/quickcalcs/ttest1.cfm
online service was used for statistical evaluations.
ACKNOWLEDGMENTS
This study was supported partly by a BarꢀIlan University new faculty grant and the EFSD/DꢀCure
Program for Young Investigator Awards for Collaborative Diabetes Research between Israel and Europe
(to A.G and J.E). This study was also supported by Diab R&D (France) and grants from the Applied
Research Program A of the Hebrew University, and the Israel Science Foundation of The Israel
Academy of Sciences and Humanities (to S.S). S. Sasson is member of the David R. Bloom Center for
Pharmacy and the Dr. Adolf and Klara Brettler Center for Research in Molecular Pharmacology and
Therapeutics, and the Adolf D. and Horty Storch Chair in Pharmaceutical Sciences, Faculty of
Medicine, The Hebrew University of Jerusalem. We thank Ms. Duha Fahham and Mr. Yosef Avrahami
for their technical help in conducting glucose uptake assays. We are thankful to Dr. Rachel Persky for
the MS analysis, to biochemistry department of Enamine Ltd. for performing the biostability studies and
to Mr. Yosef Mackler for editing the manuscript.
ACS Paragon Plus Environment

Page 27 of 49
Journal of Medicinal Chemistry
SUPPORTING INFORMATION
1
2
3
4
5
6
7
8
9
Supporting Information Available: supportive biological data, results of biostability of 34, analytical
data of synthesized compounds and molecular modeling fit values/calculations. This material is
available free of charge via the Internet at http://pubs.asc.org.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Corresponding Authors: *A.G.: email: gruzmaa@biu.ac.il, Phone: +972ꢀ3ꢀ7384597, Fax: +972ꢀ3ꢀ738ꢀ
4053; *S.S. shlomo.sasson@mail.huji.ac.il, Phone +972ꢀ2ꢀ576ꢀ8798, Fax:+972ꢀ2ꢀ675ꢀ8741
ABBREVIATION LIST
AAALAC, Association for Assessment and Accreditation of Laboratory Animal Care;
ACC, Acetyl CoA Carboxylase;
AICAR, 5ꢀAminoimidazoleꢀ4ꢀcarboxamideꢀ1ꢀβꢀDꢀribofuranoside;
ADA, American Diabetes Association;
AMPK, AMPꢀactivated protein kinase;
ASꢀ160, AMPK substrate 160 kDa;
GTT, glucose tolerance test;
dGlc, 2ꢀdeoxyꢀDꢀglucose;
DNP, Dinitrophenol;
DS, Discovery Studio;
EDC, 3ꢀ(ethyliminomethyleneamino)ꢀN,Nꢀdimethylꢀpropanꢀ1ꢀamine;
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 49
EASD, European Association for the Study of Diabetes;
1
2
3
4
FCS, fetal calf serum;
5
6
7
GLUT, glucose transporter;
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HCGM, High concetration glucose medium: 23.0 mmol/L DꢀGlucose;
HOBt, 1ꢀhydroxybenzotriazole;
HRP, horseradish peroxidase;
KKAy, Kuo Kondo rats carrying the Ay yellow obese gene;
αMEM, alphaꢀMinimal essential medium;
MTT, (3ꢀ(4,5ꢀDimethylthiazolꢀ2ꢀyl)ꢀ2,5ꢀdiphenyltetrazolium bromide test
OPD, Oꢀphenylenediamine;
PI3K, Phosphatidylinositide 3ꢀKinase;
PMSF, phenylmethanesulfonylfluoride;
WB, Western Blot
ACS Paragon Plus Environment

Page 29 of 49
Journal of Medicinal Chemistry
1
2
3
FIGURE CAPTIONS
4
5
6
7
Figure 1. Compounds 3, 4 and 15 increase the rate of glucose uptake in L6 myotubes by activating
8
9
AMPK.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Compounds 24, 29, 33 and 34 increase the rate of glucose uptake in L6 myotubes by
activating AMPK.
Figure 3. Compound 34 induces GLUT4 translocation to the plasma membrane of L6 myotubes in a
nonꢀinsulinꢀdependent manner and stimulates phosphorylation of AMPK downstream targets.
Figure 4. Antihyperglycemic effects of 34 in diabetic KKAy mice.
Figure 5. PTꢀ1 based pharmacophore model.
Figure 6. The pharmacophore model generated based on the compounds synthesized in the present
work.
Figure 7. Binding modes of PT1 and compound 34 to AMPK.
SCHEME TITLES
Scheme 1. Synthesis of 20ꢀ27 and 33.
Scheme 2. Synthesis of 29 and 30.
Scheme 3. Synthesis of 31, 32 and 34.
CHART TITLES
Chart 1. Determination of possible minimal active moiety of PTꢀ1.
ACS Paragon Plus Environment