Discovery and hit-to-lead optimization of novel allosteric glucokinase activators

Article abstract of DOI:10.1016/j.bmcl.2011.06.128

We report on a hit generation and hit-to-lead program of a novel class of glucokinase activators (GKAs). Hit compounds, activators at low glucose concentration only were identified by vHTS. Scaffold modification reliably afforded activators also at high s

Full text of DOI:10.1016/j.bmcl.2011.06.128

Bioorganic & Medicinal Chemistry Letters 21 (2011) 5417–5422
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery and hit-to-lead optimization of novel allosteric glucokinase activators
Martin Lang a, , Markus H.-J. Seifert a, Kristina K. Wolf a, Andrea Aschenbrenner a, Roland Baumgartner a,
Tanja Wieber a, Viola Trentinaglia a, Marcus Blisse a, Nobumitsu Tajima b, Tokuyuki Yamashita b,
Daniel Vitt a, Hitoshi Noda b
⇑
,
a 4SC AG, Am Klopferspitz 19a, 82152 Planegg-Martinsried, Germany
b Sanwa Kagaku Kenkyusho Co. Ltd, 35 Higashisotobori-cho, Higashi-ku, Nagoya 461-8631, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
We report on a hit generation and hit-to-lead program of a novel class of glucokinase activators (GKAs).
Hit compounds, activators at low glucose concentration only were identified by vHTS. Scaffold modifica-
tion reliably afforded activators also at high substrate level. Potency was increased by introduction of a
hydrogen bond acceptor as proposed by molecular docking. Replacement of the initial alkylene linkers
with a rigid 1,2-phenylene motif followed by further studies eventually furnished a series of potent lead
compounds exhibiting steep SAR.
Received 10 June 2011
Revised 30 June 2011
Accepted 30 June 2011
Available online 18 July 2011
Keywords:
Glucokinase
Ó 2011 Elsevier Ltd. All rights reserved.
Allosteric activators
Hit-to-lead optimization
Molecular modeling
Type 2 diabetes
Glucokinase1 (hexokinase IV, GCK, GK) plays an important role
in the control of whole-body glucose homeostasis by regulating
glucose-stimulated insulin release in pancreatic b-cells and hepatic
glucose metabolism. The enzyme is exceptional within the hexoki-
nase family for its restricted tissue expression and unique kinetics.
An allosteric site was identified in GK about 20 Å distal to the sub-
strate pocket. Direct glucokinase activation by allosteric activators
increases the substrate turnover rate at a given glucose concentra-
tion, effecting augmentation of both insulin release in b-cells and
glucose clearance rate from the bloodstream by the liver, and mit-
igation of hepatic glucose production. The development of small
molecule glucokinase activators (GKAs) is a current approach to-
wards a novel type 2 diabetes treatment.2–6 In this paper, we re-
port on our hit generation and hit-to-lead program leading to a
novel structural class of glucokinase activators.
either 5 or 20 mM glucose concentration. A second alignment ap-
proach was based upon a modified structure 2 (Fig. 1) in order to
de-emphasize local polarity of 1. AstraZeneca’s GKA110 (3) and Ba-
nyu’s 411 and 58 served as templates for further alignment studies,
and a descriptor-based similarity search was carried out to scaffold
5. From the resulting consensus list,12 about 240 selected hit-like
screening compounds were purchased and tested for their GK acti-
vation, resulting in two hit compounds 6 and 7 (Table 1 and
Fig. 2).13
Glucokinase activities were measured by the glucose-6-phos-
phate dehydrogenase (G6PDH) coupled continuous spectrophoto-
metric assay. Briefly, GK catalyzes glucose phosphorylation in the
presence of ATP. The product, glucose-6-phosphate is then oxi-
dized by NADP in the presence of excess G6PDH to produce 6-
phospho-D-gluconate. The rate of increase of absorption of the
Hits were identified by vHTS applying 4SC’s virtual database.7
Our initial approach was a docking experiment into the allosteric
site of active glucokinase (pdb code 1V4S)8 and an alignment study
to Roche’s pioneering activator, Ro-28-16759 (1, Fig. 1). From the
lists of the highest scorers of both experiments, about 300 selected
hit-like screening compounds were purchased and tested for GK
activation. However, none of the compounds of the initial set
met our hit criterion of a P1.2-fold GK activation (c = 10 lM) at
⇑
Corresponding author. Tel.: +49 89 7007630; fax: +49 89 70076329.
E-mail address: mlang1.de@googlemail.com (M. Lang).
New address: Hilti Entwicklungsgesellschaft mbH, Hiltistr. 6, 86916 Kaufering,
Germany.
Figure 1. Templates used for alignment studies during hit generation.
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.06.128

5418
M. Lang et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5417–5422
Table 1
Glucokinase activation by Ro-28-1675 (1) and by hit compounds 6–9.a
Compd (10
lM)
Activation at 5 mM glucose
Activation at 20 mM glucose
foldb
relativec
foldb
relativec
1
6
7
8
9
4.67
1.35
1.60
1.88
1.99
1.000
0.049
0.087
0.177
0.215
2.55
1.06
1.02
1.17
1.14
1.000
0.029
–0.051
0.095
0.072
a
Means of triplicates (n = 3, CV 6 10%). Values for 1, 8 and 9 are the means of multiple independent
assays.
b
Glucokinase activity relative to DMSO control (DMSO: 1).
c
Glucokinase activity relative to 10
Supplementary data for details.
lM Ro-28-1675 (1) as a standard activator (DMSO: 0; 1: 1). See
Figure 2. Initial glucokinase activator hits (6, 7). Compounds 8 and 9, selected for
optimization were obtained upon hit explosion.
resulting NADPH at k = 340 nm was used to measure GK activity.14
Compounds (c = 10 lM) were characterized by their fold GK activa-
tion and by their relative GK activation compared with Ro-28-1675
(1) as a standard (Table 1).15 The assay was conducted at two dif-
ferent glucose concentrations, 5 mM and 20 mM which simulated
low (preprandial) and high (postprandial) physiological blood glu-
cose conditions.
Bothhitcompounds6 and7weretakenasa basisforhitexplosion
by a three pillar approach: (i) a numeric, descriptor-based similarity
search for compounds 6 and 7 within 4SC’s virtual database, (ii)
structural alignment of compounds from the virtual database to 6
and 7, and (iii) a conventional chemist’s structural search for similar
scaffolds. The resulting consensus list16 was further restrained tak-
ing patentability aspects into account. Eventually, about 200 screen-
ing compounds were purchased and tested. Six of these, all with an
N-acyl-2-aminothiazole scaffold fulfilled our hit criterion. These
compounds were 1.35-fold to 1.99-fold GK activators at 5 mM glu-
cose. On the other hand, they were generally rather weak GKAs or
weak GK inhibitors at 20 mM glucose. The two most potent candi-
dates, 8 and 9 were selected for optimization (Table 1, Fig. 2).13
Binding of GKAs is to a considerable extent mediated by hydro-
phobic interactions.17 We predicted the binding mode of 9 by its
alignment to the protein scaffold of pdb structure 1V4S by our pro-
prietary docking engine, ProPose.18 The initial model was refined
by short molecular dynamics simulations (t = 5 ps) and subsequent
energy optimization using the MAB force field as implemented in
Moloc.19 Our calculation suggested a hydrogen bond between
backbone amide NH of Arg63 and thiazole nitrogen of 9 and an
H-bond between amide NH of 9 and carbonyl oxygen of Arg63
(Fig. 3), a key residue in hydrogen bond formation with N-acyl-2-
aminothiazoles and other classes of GKAs.3f,17 The formation of a
weak20 H-bond between OH of Tyr215 and furan oxygen of 9 cer-
tainly remains debatable. The binding modes of known GKAs were
predicted similarly (data not shown).21 Based on our analysis, we
defined the following rationales for further optimization of com-
pounds 8 and 9: (i) keep both H-bonds between the ligand and
backbone of Arg63, (ii) replace furan with a better H-bond acceptor
group to increase the strength of a H-bond with the hydroxy group
of Tyr215, and (iii) introduce an appropriate H-bond acceptor
group at C-5 for interaction with the side chain OH of Ser64.
Figure 3. Calculated binding mode of hit compound 9. The GK structure is
displayed as cartoon in steel blue. Selected residues are shown as sticks (oxygens in
red, nitrogen in blue). Coloring for the ligand is as follows: carbons in yellow,
oxygens in red, nitrogens in blue, sulfur in orange. Predicted H-bonds are shown as
red dashes.22
In parallel to this binding mode analysis, chemistry-driven scaf-
fold investigations were carried out (Table 2 and Supplementary
data). The fused cyclohexane ring was opened (10), replaced with
cyclopentane (11), and steric bulk was increased (12). Most
remarkably, the replacement of the tetrahydrobenzothiazole scaf-
fold with cyclopentane annelated thiazole (11) furnished signifi-
cant GK activation at 20 mM glucose, an attribute not exhibited
by our previous compounds. Direct GK activators at low glucose
levels which behave more or less indifferently at higher glucose
concentration decrease S0.5 with no or little effect on vmax. Activa-
tors over a larger range of substrate concentration decrease S0.5 and
3f,23
additionally increase vmax
.
The trend observed with 9 and 11 also applied for their thio-
phene analogs 13 and 14, corroborating a correlation between en-
zyme activation at high glucose level (effect on vmax) and the
annulation partner of the thiazole ring, uncoupled from glucoki-
nase activation at low glucose concentration (effect on S0.5). Po-
tency of cyclopentane fused thiazoles 11 and 14 was in the range
of 20–30% of Ro-28-1675 at both 5 and 20 mM glucose. The corre-
sponding tetrahydrobenzothiazoles 9 and 13 were equally potent
activators at 5 mM glucose but significantly less activating at
20 mM glucose. Interestingly, within a recently reported series of
substituted N-(thiazol-2-yl)-2-aryl-3-(tetrahydropyranyl)propana-
mides, cyclohexane and cyclopentane fused thiazoles were among
the most potent GKAs.24
The weak H-bond acceptors furan and thiophene were replaced
by pyridine, tetrahydrofuran, thiazole and isoxazole all of which

M. Lang et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5417–5422
5419
Table 2
Table 3
SAR of scaffold modifications of hit compounds 8 and 9a
SAR of H-bond acceptor groups attached to C-5 of 11a
Activation at
5 mM glucose
Activation at
20 mM glucose
Activation at
20 mM
glucose
5
5
Activation at
5 mM glucose
H
N
X
2
H
N
Compd (10
lM)
O
2
X
fold
relative
fold
relative
Compd
(10 lM)
N
3
3
n
fold relative fold relative
O
N
S
O
n
10
11
12
13
14
O
O
O
S
1.09
1.88
1.25
2.12
2.22
0.022
0.219
0.059
0.287
0.312
0.96
1.42
1.02
1.06
1.37
–0.022
0.247
0.012
0.038
0.231
OH
S
15
16
1
1
1.30
1.92
0.069
0.212
1.08
1.41
0.050
0.246
OMe
N
S
17
18
19
20
21
1
1
1
1
1
1.98
2.07
1.56
1.12
1.62
0.282
0.290
0.151
0.032
0.178
1.51
1.44
1.23
1.04
1.29
0.331
0.287
0.154
0.027
0.187
OMe
N
N
S
S
O
O
O
OMe
NHMe
NMe2
N
S
S
O
a
Means of triplicates (n = 3, CV 6 10%). Values for 11,13 and 14 are the means of
NH
multiple independent assays. See also footnotes of Table 1.
22
23
1
1
1.40
2.35
0.105
0.312
1.11
1.51
0.081
0.305
O
O
exhibit stronger H-bond basicities.20,25 However, this approach
was initially unsuccessful in terms of GKA potency (Supplementary
data) and not pursued further in favor of the attachment of a third
H-bond acceptor.
O
2'
An sp2 or an sp3 oxygen (Table 3) or a pyridine-type nitrogen
(Supplementary data) was connected by different linkers to C-5
of 11. Compounds 17, 18 and 23 displayed slightly better potencies
compared with 11 of about 30% GK activation relative to 1 at both
low and high glucose levels, whereas 16 retained the potency of 11.
The remaining compounds of this series 15 and 19–22 were less
potent GKAs. In view of the improved potency of 17, 18, and 23,
we further modified these compounds by replacement of the eth-
ylene and propylene linkers with a rigid phenylene motif.
Whereas our tested phenylene bridged analogs of 16 and 17
carrying an sp2 oxygen were inactive (Supplementary data), the
ortho-acetophenone 24 seemed a promising candidate for further
optimization, overmatching GK activation of the parent com-
pounds 18 and 23 at both 5 mM and 20 mM glucose. In a sharp
contrast, both the meta-acetophenone 25 and the naked 5-phenyl-
furan 26 were no activators (Table 3). The proposed binding mode
of 24, calculated by alignment to the protein scaffold of 1V4S and
subsequent energy minimization is shown in Fig. 4. Retainment of
both characteristic H-bonds between the ligand and backbone car-
bonyl and amide NH of Arg63 is predicted. Our calculation suggests
the participation of the carbonyl oxygen of the acetophenone part
in an extensive H-bond network with donors of the side chains of
Arg250 and Ser64 and possibly the backbone amide NH of Thr65.
1'
1'
24
1
2.43
1.00
0.329
0.000
1.63
0.384
O
3'
25
26
1
2
0.93 –0.039
0.89 –0.062
O
0.99 –0.002
a
Means of triplicates (n = 3, CV 6 10%). Values for GKAs with >1.1-fold activation
are the means of at least two independent assays. See also footnotes of Table 1.
20 mM glucose compared with Ro-28-1675 (Table 4) and was
therefore the first candidate to meet our lead criterion of a P0.5-
fold relative GK activation.
We next investigated the influence of the 2-methyl group (Ta-
ble 4) and found that this alkyl residue had a major impact on
the biological activity. Omission of the 2-methyl substituent (28)
greatly reduced the GK activation, and its extension by one carbon
(29) had a similar, though less pronounced effect.
We were interested in whether the initially discovered correla-
tion of thiazole scaffold and GK activation at high glucose concen-
tration still applied for the phenylene substituted series. We
synthesized tetrahydrobenzothiazoles 30–32 to find that all three
compounds were less potent GKAs compared with their cyclopen-
tane fused counterparts 27–29. With the exception of the weak
activator 31, the activity loss of cyclohexane annulated thiazoles
was most prominent at high glucose concentration. This trend is
in agreement with the SAR of our early generation GKAs.
Furans 33–35 followed similar trends as their thiophene ana-
logs 28, 29 and 32. Again, both omission of the 2-methyl residue
(33) and its replacement by ethyl (34) proved detrimental for bio-
logical activity. In fact, tetrahydrobenzothiazole 35 exhibited
greater potency compared with parent 34 at 5 mM glucose but as
Furthermore, the model proposes a face-to-face
p-p-stacking
interaction between the phenyl ring of 24 and Tyr214. Key residues
supposably involved in hydrophobic interaction of the 5-aryl-2-
methyl-3-furoyl moiety of 24 are Pro66, Met210, Ile211 and
Met235 (not shown in Fig. 4). Val62, Pro66, Ile159, Val452,
Val455 and Ala456 are predicted to mediate hydrophobic interac-
tion of the cyclopentane fused thiazole part. A hydrogen bond be-
tween the hydroxy group of Tyr215 and furan oxygen as discussed
for 9 is not supported by the docking experiment (O–O distance:
4.85 Å).
Compound 27, the thiophene analog of 24 displayed a surpris-
ingly high relative GK activation of 66% and 63% at 5 mM resp.

5420
M. Lang et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5417–5422
Shifting the carbonyl oxygen by one bond (36 and 37 vs 24 and
25) resulted in a complete loss of enzyme activation (Table 4). The
ortho-benzoate ester 38, like the ortho-benzonitrile26 40 was a
weak GKA whereas both the ortho-benzoic acid 39 and meta-ben-
zonitrile 41 were inactive. The ortho-nitrobenzene 4227 was a rea-
sonable GK activator.
Ortho-(methylsulfonyl)benzenes 43–47 were strong activators.
Again, the 2-ethyl substituted furan (44) was less potent compared
with its 2-methyl analog (43). In contrast to 44, the corresponding
tetrahydrobenzothiazole (45) was inactive at 20 mM glucose. The
SAR between thiophenes (43, 44) and furans (46, 47), though less
pronounced, complied reasonably with that of the acetophenone
series.
Ortho-benzamides 48 and 49 were the strongest glucokinase
activators identified in the course of this hit-to-lead project with
a potency between 69% and 74% of that of Ro-28-1675 at both 5
and 20 mM glucose. Again, furans (50 and 51) were less potent
than their thiophene counterparts. Whereas in the furan series,
replacement of the 2-methyl group with an ethyl residue was
advantageous (51 vs 50), both 2-methyl and 2-ethyl substituents
were equally well tolerated in the thiophene series (49 vs 48).
The ortho-dimethylbenzamide 52, by contrast, was not a glucoki-
nase activator.
Figure 4. Calculated binding mode of 24. Only selected residues are shown. The
proposed
p–p-stacking interaction is highlighted in green, H-bonds are shown as
red dashes.
Compounds 10–14 were synthesized by standard coupling reac-
tion28 of commercially available 2,5-dimethyl-3-furoic acid and
2,5-dimethyl-3-thenoic acid, respectively with 2-aminothiazoles
expected, was substantially less potent at high substrate concen-
tration. Generally, within the ortho-acetophenone series, furans
(24, 33–35) were weaker GKAs than their corresponding thio-
phenes (27–29, 32) with the only exception of equipotent com-
pounds 35 and 32 at 20 mM glucose.
A–D. The latter were prepared by Hantzsch type reaction of a-leav-
ing group substituted ketones with thiourea,29 under Gewald con-
ditions,30 or by reaction of cycloalkanones with iodine and
Table 4
SAR of optimization of compound 24a
R3
R2
H
X
N
N
R1
S
O
n
Compd (10
lM)
R1
R2
R3
X
n
Activation at 5 mM glucose
Activation at 20 mM glucose
fold
relative
fold
relative
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
Me
H
Et
Me
H
Et
H
Et
Et
Me
Me
Me
Me
Me
Me
Me
Me
Et
Et
Me
Et
Me
Et
Me
Et
Ac
Ac
Ac
Ac
Ac
Ac
Ac
Ac
Ac
OAc
H
CO2Me
CO2H
CN
H
NO2
SO2Me
SO2Me
SO2Me
SO2Me
SO2Me
CONH2
CONH2
CONH2
CONH2
CONMe2
H
H
H
H
H
H
H
H
H
H
OAc
H
H
H
CN
H
H
H
H
H
H
H
H
H
H
H
S
S
S
S
S
S
1
1
1
2
2
2
1
1
2
1
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
3.52
2.02
2.46
3.17
1.57
2.22
1.77
1.64
1.89
0.95
0.94
1.36
0.92
1.25
0.99
1.89
3.11
2.76
2.16
2.81
2.92
3.34
3.70
2.29
2.61
0.96
0.660
0.274
0.446
0.554
0.145
0.312
0.207
0.193
0.278
À0.012
À0.019
0.098
À0.023
0.067
À0.002
0.244
0.642
0.537
0.297
0.504
0.586
0.714
0.690
0.340
0.453
À0.011
1.71
1.45
1.50
1.49
1.26
1.23
1.42
1.41
1.25
0.88
0.92
1.10
0.94
1.10
0.85
1.46
1.56
1.41
0.95
1.72
1.39
1.80
1.92
1.58
1.61
0.96
0.627
0.296
0.446
0.392
0.214
0.182
0.275
0.361
0.194
À0.072
À0.047
0.054
À0.034
0.064
À0.090
0.267
0.495
0.363
À0.041
0.470
0.343
0.713
0.736
0.418
0.509
À0.026
O
O
O
O
O
O
O
O
O
O
S
S
S
O
O
S
S
O
O
O
Me
a
Means of triplicates (n = 3, CV 6 10%). Values for GKAs with >1.1-fold activation are the means of at least two independent assays. See also footnotes of Table 1.

M. Lang et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5417–5422
5421
thiourea31 (Scheme 1). Compounds 16–19, 23–38, and 41–51 were
synthesized by standard coupling of suitable 3-furoic acid and 3-
thenoic acid precursors E–J with 2-aminothiazoles
B or D
(Scheme 2; see Supplementary data for the synthesis of 22 and
E–J). Hydroxymethylfuran 15 was obtained by acetate cleavage
from 23. Amides 20 and 21 were synthesized from ester 19. Methyl
ester 38 served as a precursor for both 39 and 52. Nitrile 40 was
prepared from the corresponding amide 48 by treatment with Bur-
gess reagent32 (Scheme 2).
In conclusion, novel allosteric glucokinase activator hit com-
pounds were identified and developed into a series of potent leads
by molecular modeling-supported medicinal chemistry. The
replacement of the initial tetrahydrobenzothiazole moiety (9 and
13) with cyclopentane fused thiazole (11 and 14) was crucial for
GK activation at high glucose level (effect on vmax), a robust struc-
ture-activity relationship valid with our final 1,2-phenylene com-
pounds, too (30, 32, 35, 45 vs 27, 29, 34, 44). Potency was
increased by attachment of H-bond acceptor groups to 11 via eth-
ylene and propylene spacers (17, 18, 23), and further improved by
the integration of a rigid 1,2-phenylene linker instead of the initial
alkylene units (24). Substitution at C-2’ was essential for the 5-aryl
substituted compounds (24–52), and SAR was steep with regard to
R2 (Fig. 5). We believe that both hydrogen bond acceptor strength
and steric demands of R2, favoring a preferred conformation33 of
the ligand are crucial for biological activity. In our lead series, thio-
phenes were generally more potent than their isosteric furan ana-
logs. In most cases the 2-methyl residue was superior to ethyl and
H. Thienylbenzamides 48 and 49 were the strongest glucokinase
activators identified in the course of this study. Compounds 27,
43, 44, 46–49 and 51 were fed into a lead optimization program.
Scheme 1. Synthesis of aminothiazoles A–D and screening compounds 10–14.
Reagents and conditions: (a) TosCl (1.1 equiv), pyridine, CHCl3, rt, 18 h; (b) thiourea
(1.0 equiv), EtOH, rfx, 4 h; (c) sulfur (1.1 equiv), H2NCN (1.1 equiv), EtOH (ꢀ1 mL/
mmol), rt, then slowly add NHEt2 (ꢀ1 mL/mmol), then stir at 40 °C for 2.5–4 h; (d)
NBS (1.05 equiv), NH4OAc (0.1 equiv), Et2O, rt, 0.5 h (no purification); (e) thiourea
(2 equiv), NEt3 (1.5 equiv), EtOH, rfx, 2 h; (f) I2 (1 equiv), thiourea (2 equiv), neat,
110 °C, 2.5 h; (g) HBTU, cat. DMAP, DIEA, CH2Cl2, rt, 12–24 h. Yields were not
optimized.
Acknowledgments
The authors are grateful to Ms Heike Fischer and Ms Ilka Rudat
for assistance with screening compound logistics.
Supplementary data
Supplementary data (assay details, correlation of fold activation
and GK enzyme concentration for 1, additional GKA data, synthesis
of building blocks E–J, synthesis of 22, and 1H NMR spectra of lead
compounds 27, 43, 44, 46–49 and 51) associated with this article
can be found, in the online version, at doi:10.1016/j.bmcl.
2011.06.128.
References and notes
1. Printz, R. L.; Magnuson, M. A.; Granner, D. K. Annu. Rev. Nutr. 1993, 13, 463.
2. Matschinsky, F.M.; Magnuson, M.A. (Eds.), Glucokinase and Glycemic Disease,
Vol. 16, Karger, 2004.
Scheme 2. Synthesis of 15–21 and 23–52. Reagents and conditions: (a) HBTU, cat.
DMAP, DIEA, CH2Cl2, rt, 12–24 h; (b) 0.5
N LiOH, MeOH-H2O (3:1), rfx, 1–3 h; (c)
Ac2O, cat. DMAP, NEt3, THF, rt, 2 h; (d) Burgess reagent (1.5 + 0.5 equiv), THF, 60 °C,
2 h + 12 h, rt. See Supplementary data for the synthesis of 22 and building blocks E–J.
3. General Reviews on the topic: (a) Matschinsky, F. M.; Zelent, B.; Doliba, N.; Li,
C.; Vanderkooi, J. M.; Naji, A.; Sarabu, R.; Grimsby, J. Diabetes Care 2011, 34,
S236; (b) Matschinsky, F. M.; Zelent, B.; Doliba, N. M.; Kaestner, K. H.;
Vanderkooi, J. M.; Grimsby, J.; Berthel, S. J.; Sarabu, R. in Diabetes – Perspectives
in Drug Therapy, Handb. Exp. Pharmacol., Vol. 203, Springer, 2011, p. 3575; (c)
Matschinsky, F. M.; Porte, D. F1000 Med. Rep. 2010, 2, 43; (d) Matschinsky, F. M.
Nat. Rev. Drug Discov. 2009, 8, 399; (e) Manojit, P. Drug Discovery Today 2009,
14, 784; (f) Grimsby, J.; Berthel, S. J.; Sarabu, R. Curr. Top. Med. Chem. 2008, 8,
1524; (g) Zelent, D.; Najafi, H.; Odili, S.; Buettger, C.; Weik-Collins, H.; Li, C.;
Doliba, N.; Grimsby, J.; Matschinsky, F. M. Biochem. Soc. Trans. 2005, 33, 306; (h)
Kietzmann, T.; Ganjam, G. K. Expert Opin. Ther. Patents 2005, 15, 705.
4. Reviews rather focused on structures of GKAs: (a) Sarabu, R.; Berthel, S. J.;
Kester, R. F.; Tilley, J. W. Expert Opin. Ther. Patents 2011, 21, 13; (b) Manojit, P.
Curr. Med. Chem. 2009, 16, 3858; (c) Fyfe, M. C. T.; Procter, M. J. Drugs Fut. 2009,
34, 641; (d) Sarabu, R.; Berthel, S. J.; Kester, R. F.; Tilley, J. W. Expert Opin. Ther.
Patents 2008, 18, 759; (e) Johnson, T. O.; Humphries, P. S. Annu. Rep. Med. Chem.
2006, 41, 141; (f) Guertin, K. R.; Grimsby, J. Curr. Med. Chem. 2006, 13, 1839; (g)
Sarabu, R.; Grimsby, J. Curr. Opin. Drug Discov. Devel. 2005, 8, 631.
5. Review on molecular physiology of mammalian GK: Iynedjian, P. B. Cell. Mol.
Life Sci. 2009, 66, 27.
Figure 5. SAR trends in and around lead series.

5422
M. Lang et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5417–5422
6. Review on direct small-molecule kinase activation: Simpson, G. L.; Hughes, J.
A.; Washio, Y.; Bertrand, S. M. Curr. Opin. Drug Discov. Devel. 2009, 12, 585.
7. (a) Seifert, M. H. J.; Kraus, J.; Kramer, B. Curr. Opin. Drug Discov. Devel. 2007, 10,
298; (b) Seifert, M. H. J.; Wolf, K.; Vitt, D. Biosilico 2003, 1, 143.
8. Kamata, K.; Mitsuya, M.; Nishimura, T.; Eiki, J.; Nagata, Y. Structure 2004, 12,
429.
RSC Drug Discovery Series No. 4, Royal Society of Chemistry, 2011, p. 51; (b)
Takahashi, K.; Hashimoto, N.; Nakama, C.; Kamata, K.; Sasaki, K.; Yoshimoto, R.;
Ohyama, S.; Hosaka, H.; Maruki, H.; Nagata, Y.; Eiki, J.; Nishimura, T. Bioorg. Med.
Chem. 2009, 17, 7042; (c) Ishikawa, M.; Nonoshita, K.; Ogino, Y.; Nagae, Y.;
Tsukahara, D.; Hosaka, H.; Maruki, H.; Ohyama, S.; Yoshimoto, R.; Sasaki, K.;
Nagata, Y.; Eiki, J.; Nishimura, T. Bioorg. Med. Chem. Lett. 2009, 19, 4450; (d)
Mitsuya, M.; Kamata, K.; Bamba, M.; Watanabe, H.; Sasaki, Y.; Sasaki, K.;
Ohyama, S.; Hosaka, H.; Nagata, Y.; Eiki, J.; Nishimura, T. Bioorg. Med. Chem. Lett.
2009, 19, 2718; (e) Ref. 11; (f) Bebernitz, G. R.; Beaulieu, V.; Dale, B. A.; Deacon,
R.; Duttaroy, A.; Gao, J.; Grondine, M. S.; Gupta, R. C.; Kakmak, M.; Kavana, M.;
Kirman, L. C.; Liang, J.; Maniara, W. M.; Munshi, S.; Nadkarni, S. S.; Schuster, H.
F.; Stams, T.; St. Denny, I.; Taslimi, P. M.; Vash, B.; Caplan, S. L. J. Med. Chem.
2009, 52, 6142; (g) Efanov, A. M.; Barrett, D. G.; Brenner, M. B.; Briggs, S. L.;
Delaunois, A.; Durbin, J. D.; Giese, U.; Guo, H.; Radloff, M.; Sanz Gil, G.; Sewing,
S.; Wang, Y.; Weichert, A.; Zaliani, A.; Gromada, J. Endocrinology 2005, 146,
3696; (h) Dunten, P.; Swain, A.; Kammlott, U.; Crowther, R.; Lukacs, C. M.; Levin,
W.; Reik, L.; Grimsby, J.; Corbett, W. L.; Magnuson, M. A.; Matschinsky, F. M.;
Grippo, J. F. in Glucokinase and Glycemic Disease, Frontiers in Diabetes, Vol. 16,
Karger, 2004, p. 145; (i) Ref. 8.
9. (a) Haynes, N.-E.; Corbett, W. L.; Bizzarro, F. T.; Guertin, K. R.; Hilliard, D. W.;
Holland, G. W.; Kester, R. F.; Mahaney, P. E.; Qi, L.; Spence, C. L.; Tengi, J.;
Dvorozniak, M. T.; Railkar, A.; Matschinsky, F. M.; Grippo, J. F.; Grimsby, J.;
Sarabu, R. J. Med. Chem. 2010, 53, 3618; (b) Grimsby, J.; Sarabu, R.; Corbett, W.
L.; Haynes, N.-E.; Bizzarro, F. T.; Coffey, J. W.; Guertin, K. R.; Hilliard, D. W.;
Kester, R. F.; Mahaney, P. E.; Marcus, L.; Qi, L.; Spence, C. L.; Tengi, J.; Magnuson,
M. A.; An Chu, C.; Dvorozniak, M. T.; Matschinsky, F. M.; Grippo, J. F. Science
2003, 301, 370.
10. (a) McKerrecher, D.; Allen, J. V.; Bowker, S. S.; Boyd, S.; Caulkett, P. W. R.;
Currie, G. S.; Davies, C. D.; Fenwick, M. L.; Gaskin, H.; Grange, E.; Hargreaves, R.
B.; Hayter, B. R.; James, R.; Johnson, K. M.; Johnstone, C.; Jones, C. D.; Lackie, S.;
Rayner, J. W.; Walker, R. P. Bioorg. Med. Chem. Lett. 2005, 15, 2103; (b)
Brocklehurst, K. J.; Payne, V. A.; Davies, R. A.; Carroll, D.; Vertigan, H. L.;
Wightman, H. J.; Aiston, S.; Waddell, I. D.; Leighton, B.; Coghlan, M. P.; Agius, L.
Diabetes 2004, 53, 535.
22. Illustration generated with the PyMOL Molecular Graphics System, DeLano
Scientific, San Carlos, CA, USA. http://www.pymol.org.
11. Nishimura, T.; Iino, T.; Mitsuya, M.; Bamba, M.; Watanabe, H.; Tsukahara, D.;
Kamata, K.; Sasaki, K.; Ohyama, S.; Hosaka, H.; Futamura, M.; Nagata, Y.; Eiki, J.
Bioorg. Med. Chem. Lett. 2009, 19, 1357.
23. Ralph, E. C.; Thomson, J.; Almaden, J.; Sun, S. Biochemistry 2008, 47, 5028.
24. Li, F.; Zhu, Q.; Zhang, Y.; Feng, Y.; Leng, Y.; Zhang, A. Bioorg. Med. Chem. 2010,
18, 3875.
12. Filter criteria were (i) high alignment score; (ii) molecular weight (MW)
6550 gmol–1; (iii) 20 Å2 6 topological polar surface area (TPSA) 6 180 Å2; (iv)
H-bond donor count (Ndon) 66; (v) H-bond acceptor count (Nacc) 610; (vi)
number of rotatable bonds (Nrot) 610; (vii) absence of reactive groups; (viii)
presence of 2 or 3 small aromatic/lipophilic rings; (ix) visual inspection. About
15% of the compounds tested were N-acyl-2-aminothiazoles.
25. Relative H-bond basicities: thiophene < furan < tetrahydrofuran < pyridine. See
Refs. 20 and additionally, pyridine and THF vs furan: (a) Nobeli, I.; Price, S. L.;
Lommerse, J. P. M.; Taylor, R. J. Comput. Chem. 1997, 18, 2060; Pyridine vs
furan: (b) Lee, H. M.; Singh, N. J.; Kim, K. S. In Hydrogen Bonding – New Insights
Challenges and Advances in Computational Chemistry and Physicsm; Grabowski,
S. J., Ed.; Springer, 2005; Vol. 3, p 149; THF vs furan: (c) Kaur, D.; Khanna, S.
Comput. Theor. Chem. 2011, 963, 71; Thiophene vs furan: (d) Cooke, S. A.;
Corlett, G. K.; Legon, A. C. J. Chem. Soc., Faraday Trans. 1998, 94, 1565.
26. The nitrile group as an H-bond acceptor: (a) Fleming, F. F.; Yao, L.; Ravikumar,
P. C.; Funk, L.; Shook, B. C. J. Med. Chem. 2010, 53, 7902; (b) Desiraju, G. R.;
Steiner, T. The Weak Hydrogen Bond in Structural Chemistry and Biology; Oxford
Science Publications, 2001, p. 198; (c) Le Questel, J.-Y.; Berthelot, M.; Laurence,
C. J. Phys. Org. Chem. 2000, 13, 347; (d) Allerhand, A.; Schleyer, P. v. R. J. Am.
Chem. Soc. 1963, 85, 866.
13. Compunds 6 and 9 were purchased from Enamine Ltd, 7 was from 4SC’s
proprietary screening compound library, and
ChemBrigde Corp.
8 was purchased from
14. Davidson, A. L.; Arion, W. J. Arch. Biochem. Biophys. 1987, 253, 156.
15. The fold activation of given GKA increases with reduced enzyme
a
concentration (Fig. S1). Relative activation of GKAs displays negligible
enzyme batch-to-batch variation.
16. Filter criteria were (i) MW 6550 gmol–1; (ii) 20 Å2 6 TPSA 6 180 Å2; (iii) Ndon
66; (iv) Nacc 610; (v) Nrot 610; (vi) absence of reactive groups; (vii) visual
inspection. About 53% of the compounds tested were N-acyl-2-aminothiazoles.
17. For a docking assessment of GKAs, see: Kumari, V.; Li, C. Curr. Chem. Genom.
2008, 2, 76.
27. The nitro group as an H-bond acceptor: (a) Johnstone, R. A. W.; Loureiro, R. M.
S.; Cristiano, M. L. S.; Labat, G. ARKIVOC 2010, 142; (b) Allen, F. H.; Baalham, C.
A.; Lommerse, J. P. M.; Raithby, P. R.; Sparr, E. Acta Cryst. 1997, B53, 1017; (c)
Baitinger, W. F.; Schleyer, P. v. R.; Murty, T. S. S. R.; Robinson, L. Tetrahedron
1964, 20, 1635.
18. (a) Seifert, M. H. J. J. Chem. Inf. Model 2005, 45, 449; (b) Seifert, M. H. J.; Schmitt,
F.; Herz, T.; Kramer, B. J. Mol. Model. 2004, 10, 342.
19. (a) Gerber, P. R.; Müller, K. J. Comput. Aided Mol. Des. 1995, 9, 251; (b) Gerber
Molecular Design (http://www.moloc.ch).
20. Measures of H-bond basicity of acceptors and hydrogen-bond acidity of
donors: Overview: (a) Laurence, C.; Brameld, K. A.; Graton, J.; Le Questel, J.-Y.;
Renault, E. J. Med. Chem. 2009, 52, 4073; Overview: (b) Todeschini, R.;
Consonni, V. Molecular Descriptors for Chemoinformatics, Vol. 1: Alphabetical
Listing in Methods and Principles in Medicinal Chemistry, Vol. 41, Wiley, 2009, p.
448; pKa slide rule: (c) Gilli, P.; Pretto, L.; Bertolasi, V.; Gilli, G. Acc. Chem. Res.
2009, 42, 33; Prediction of Hacc strength: (d) Schwöbel, J.; Ebert, R.-U.; Kühne,
R.; Schüürmann, G. J. Phys. Chem. A 2009, 113, 10104.
28. HBTU: (a) Dourtoglou, V.; Ziegler, J.-C.; Gross, B. Tetrahedron Lett. 1978, 15,
1269; DMAP: (b) Steglich, W.; Höfle, G. Angew. Chem., Int. Ed. Engl. 1969, 8, 981.
29.
a
-Bromination of ketones with NBS/NH4OAc: (a) Tanemura, K.; Suzuki, T.;
Nishida, Y.; Satsumabayashi, K.; Horaguchi, T. Chem. Commun. 2004, 470;
Reaction of -halogenoketones with thiourea: (b) Traumann, V. Liebigs Ann.
Chem. 1888, 249, 31.
a
30. Gewald, K.; Böttcher, H.; Mayer, R. J. Prakt. Chem. 1964, 23, 298.
31. (a) King, L. C.; Hlavacek, R. J. J. Am. Chem. Soc. 1950, 72, 3722; (b) Dodson, R. M.;
King, L. C. J. Am. Chem. Soc. 1945, 67, 2242.
32. Claremon, D. A.; Phillips, B. T. Tetrahedron Lett. 1988, 29, 2155.
33. Example of biphenyl-monitored effective size of ortho substituents: Ruzziconi,
R.; Spizzichino, S.; Mazzanti, A.; Lunazzi, L.; Schlosser, M. Org. Biomol. Chem.
2010, 8, 4463.
21. (a) More recently, a number of crystal structure X-ray data based analyses of GK
activator binding modes have been published, see: (a) Sarabu, R.; Tilley, J. W.;
Grimsby, J. in Accounts in Drug Discovery: Case Studies in Medicinal Chemistry,