Discovery and Structure-Guided Optimization of Diarylmethanesulfonamide Disrupters of Glucokinase-Glucokinase Regulatory Protein (GK-GKRP) Binding: Strategic Use of a N → S (nN → σ? S-X) Interaction for Conformational Constraint

This is an open access article published under an ACS AuthorChoice License, which permits

copying and redistribution of the article or any adaptations for non-commercial purposes.

Article

pubs.acs.org/jmc

Discovery and Structure-Guided Optimization of

Diarylmethanesulfonamide Disrupters of Glucokinase−Glucokinase

Regulatory Protein (GK−GKRP) Binding: Strategic Use of a N → S

(n_N→ σ*_S−X) Interaction for Conformational Constraint

Lewis D. Pennington,*^,†,#Michael D. Bartberger,*^,‡Michael D. Croghan,^†Kristin L. Andrews,^‡

Kate S. Ashton,^†Matthew P. Bourbeau,^†Jie Chen,^§Samer Chmait,^‡Rod Cupples,^∥Christopher Fotsch,^†

Joan Helmering,^∥Fang-Tsao Hong,^†Randall W. Hungate,^†Steven R. Jordan,^‡Ke Kong,^†Longbin Liu,^†

Klaus Michelsen,^‡Carolyn Moyer,^⊥Nobuko Nishimura,^†Mark H. Norman,^†Andreas Reichelt,^†

Aaron C. Siegmund,^†Glenn Sivits,^∥Seifu Tadesse,^†Christopher M. Tegley,^†Gwyneth Van,^⊥

Kevin C. Yang,^†Guomin Yao,^†Jiandong Zhang,^‡David J. Lloyd,^∥Clarence Hale,^∥

and David J. St. Jean, Jr.^†

‡

§

∥

^†Medicinal Chemistry, Molecular Structure and Characterization, Pharmacokinetics and Drug Metabolism, Metabolic Disorders

⊥

Research, and Pathology, Amgen, One Amgen Center Drive, Thousand Oaks, California 91320-1799, United States

S

* Supporting Information

ABSTRACT: The HTS-based discovery and structure-guided

optimization of a novel series of GKRP-selective GK−GKRP

disrupters are revealed. Diarylmethanesulfonamide hit 6

(hGK−hGKRP IC₅₀= 1.2 μM) was optimized to lead

compound 32 (AMG-0696; hGK−hGKRP IC₅₀= 0.0038

μM). A stabilizing interaction between a nitrogen atom lone

pair and an aromatic sulfur system (n_N→ σ*_S−X) in 32 was

exploited to conformationally constrain a biaryl linkage and

allow contact with key residues in GKRP. Lead compound 32

was shown to induce GK translocation from the nucleus to the

cytoplasm in rats (IHC score = 0; 10 mg/kg po, 6 h) and blood glucose reduction in mice (POC = −45%; 100 mg/kg po, 3 h).

X-ray analyses of 32 and several precursors bound to GKRP were also obtained. This novel disrupter of GK−GKRP binding

enables further exploration of GKRP as a potential therapeutic target for type II diabetes and highlights the value of exploiting

unconventional nonbonded interactions in drug design.

glucose 6-phosphate (2, G6P, Figure 1).⁴GK activity in the

liver mediates glucose storage and disposal processes, as G6P is

INTRODUCTION

■

Type II diabetes mellitus (T2DM) is a grievous glucose

the ﬁrst intermediate in glycogen synthesis and glycolysis, and

utilization disorder aﬄicting nearly 400 million individuals

GK activity in the pancreas modulates insulin secretion.⁵Many

globally at an annual cost of almost 500 billion dollars.¹In

allosteric activators of GK have entered clinical development

coming decades, aging populations in advanced countries and

for the treatment of T2DM;⁶however, observations of

expanding middle classes in developing nations are expected to

hypoglycemia due to GK overactivation have complicated the

increase the number of obese and sedentary people at most risk

progression of these potential drug candidates.⁷A number of

strategies to circumvent this issue are being pursued, such as

liver-selective or partial activators of GK.⁸An alternative

strategy is to identify agents that alter the amount of

physiologically active GK in selected tissues without aﬀecting

its intrinsic catalytic activity.

of suﬀering this progressively debilitating disease.²Although

many oral therapies for T2DM have been developed over the

past 60 years, including agents that stimulate insulin secretion,

improve insulin sensitivity, reduce glucose production, or

increase glucose excretion by the selective modulation of

therapeutic targets in a variety of biological pathways, ultimate

recourse to injectable insulin therapy often has remained

necessary.³Safe and orally bioavailable drugs directed at novel

therapeutic targets are needed to improve the management of

T2DM.

Glucokinase regulatory protein (GKRP) and glucose

cooperatively regulate the cellular location and activity of GK

in hepatocytes.⁹A high intracellular concentration of glucose

Glucokinase (GK) plays a key role in glucose homeostasis

through its physiological function converting glucose (1) to

Received: September 4, 2015

A

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 2 (2.5 and 2.2 Å resolution, respectively) and highlights

the key similarities and diﬀerences in the binding modes of

Figure 1. GKRP-selective GK−GKRP disrupters 3−6 capable of

liberating GK for the conversion of glucose (1) to glucose 6-phosphate

(G6P; 2).

Figure 2. Overlay of the X-ray crystal structures of 3 (transparent

white) and 6 (orange) bound to hGKRP (yellow), PDB codes 4MSU

and 4PX2, respectively. Hydrogen bonding interactions are

represented by dashed lines; residues and interatomic distances (in

Å) of note are labeled in white.

(>5.0 mM) disrupts the inactive GK−GKRP complex that is

sequestered in the nucleus, allows GK translocation to the

cytosol, and stimulates GK activity.¹⁰Seminal studies of the

antidiabetic eﬀects of synthetic ligands that disrupt the GK−

GKRP complex by selectively targeting a formerly unknown

allosteric binding site on GKRP were recently reported (Figure

1).¹¹By use of structure-guided drug design, the original hit 3

(hGK−hGKRP IC₅₀= 1.4 μM) from the ﬁrst HTS campaign

was transformed initially to the ﬁrst orally active tool

compound 4 (AMG-1694; hGK−hGKRP IC₅₀= 0.021

μM)¹²and ultimately to the optimized analogue 5 (AMG-

3969; hGK−hGKRP IC₅₀= 0.0037 μM).¹³In db/db mice, a

single oral dose (100 mg/kg) of 5 was found to eﬀect nearly

complete GK translocation from the nucleus to the cytoplasm

(IHC score = 0) with concomitant blood glucose reduction

(POC = −60%) at the 6 h time point.¹³To further explore the

preclinical pharmacology of this potential therapeutic target for

T2DM, structurally distinct GKRP-selective GK−GKRP

disrupters were sought.¹⁴

these two chemotypes.¹²Notably, 6 was shown to occupy the

same allosteric binding site on GKRP as compounds 3−5 with

negligible diﬀerences in protein secondary structure.^11−13,19

During reﬁnement, the R enantiomer of racemate 6 was

calculated to ﬁt the ligand electron density better than its S-

conﬁgured antipode and to project the sulfonamide NH group

away from a hydrophobic pocket and toward bulk solvent.

Though the phenyl ring of 6 resides in the same region of the

binding pocket as the thiophene ring of 3, these compounds

otherwise have markedly diﬀerent binding modes. The

benzofuran ring system of 6 is shifted over the proline shelf

region^14bcompared to the N-phenylpiperazine core of 3, and

one of the oxygen atoms of the sulfonamide functionality of 6

engages in an intermolecular hydrogen bond with Asn216

rather than Arg215 as observed for 3. Furthermore, while 6

does not engage Arg525, one of the oxygen atoms of its

benzodioxepine ring system was shown to displace the water

molecule present in the X-ray crystal structure of 3¹²and to

form a hydrogen bond with Ile11.

Because the resolution of the X-ray crystal structure depicted

in Figure 2 was insuﬃcient for unambiguous assignment of the

absolute stereochemistry of the bound and presumably

preferred enantiomer of 6, recourse to chemical synthesis and

correlation was made (Scheme 1). First, racemate 6 was

resolved using chiral SFC to the proposed R-conﬁgured

eutomer 7 (hGK−hGKRP IC₅₀= 0.87 μM) and the putatively

S-conﬁgured distomer (hGK−hGKRP IC₅₀= 17 μM). Next,

the commercially available hydrochloride salt of racemic 1-(1-

benzofuran-2-yl)-1-phenylmethanamine (8) was free-based (aq

NaHCO₃) and the resulting neutral amine was resolved using

chiral SFC to the known R-conﬁgured amine 9 and its known

S-conﬁgured antipode,²⁰both of which were sulfonylated

independently using commercially available 3,4-dihydro-2H-

1,5-benzodioxepine-7-sulfonyl chloride (10) to the correspond-

RESULTS AND DISCUSSION

■

A second HTS campaign¹⁵of a separate small-molecule library

of over 126 000 compounds identiﬁed a single potent and

selective hit, N-(1-benzofuran-2-yl(phenyl)methyl)-3,4-dihy-

dro-2H-1,5-benzodioxepine-7-sulfonamide (6, Figure 1).¹⁶In

these luciferase-based luminescence assays measuring ATP

depletion, 6 was found to increase GK activity in the presence

of GKRP (hGK−hGKRP ATP EC₅₀= 0.26 μM) but to have no

eﬀect on GK activity in the absence of GKRP (hGK ATP EC₅₀

> 100 μM).¹⁵Subsequently, HTS hit 6 was demonstrated in

surface plasmon resonance assays to selectively bind GKRP

(hGKRP K_d= 1.1 μM; hGK K_d> 20 μM) and was shown in a

bead-based proximity assay (AlphaScreen) to disrupt the GK−

GKRP binding interaction (hGK−hGKRP IC₅₀= 1.2 μM).¹⁷

The favorable biological activity of 6 coupled with its

reasonable molecular and physicochemical properties (MW =

435, PSA = 78 Å², cLogP = 5.0)¹⁸encouraged the pursuit of

analogues that might exhibit improved pharmacological proﬁles.

An overlay of the X-ray crystal structures of the original HTS

hit 3 and this new HTS hit 6 bound to GKRP is depicted in

B

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

replacing the benzofuran ring of 11 with a benzothiophene ring,

as in 12 (hGK−hGKRP IC₅₀= 0.017 μM; mGK TL EC₅₀= 2.9

μM). It was also not surprising that 12 was shown to retain

selective binding aﬃnity for GKRP (hGKRP K_d= 0.019 μM;

hGK K_d> 10 μM).¹⁷

Scheme 1. Determination of the Absolute Stereochemistry of

the Eutomer 7

Disappointingly, compounds 7, 11, and 12 all suﬀer high

clearance in rat liver microsomes (RLM CL_int> 399 μL min⁻¹

mg⁻¹). A preliminary metabolite identiﬁcation study of 12

using rat liver microsomes indicated that the major metabolite

arose from oxidation and cleavage of the methylene chain of the

benzodioxepine moiety. Examination of the X-ray crystal

structure illustrated in Figure 2 and computational modeling

indicated that substitution of the central carbon atom of the

methylene chain with either a F atom with the S conﬁguration

(axial) or an OH group with the R conﬁguration (equatorial)

might be accommodated in a hydrophobic cleft or hydrophilic

pocket of the binding site, respectively. These strategems were

quickly tested using the R,S-hydroxyl- (13), R,S-ﬂuoro- (14),

and gem-diﬂuoro- (15) analogues, which led to only moderate

biochemical potency loss for 13 and 14 but which resulted in

little to no improvement in microsomal stability for any of the

compounds (hGK−hGKRP IC₅₀= 0.10, 0.074, and 0.81 μM,

respectively; RLM CL_int= 304, >399, and >399 μL min⁻¹mg⁻¹,

respectively). Attempts to improve the microsomal stability of

12 by inserting N atoms into the scaﬀold to lower its

lipophilicity (cLogP of 12 = 6.2) were also unsuccessful, as

illustrated for diaza-analogue 16 (cLogP = 4.0; hGK−hGKRP

IC₅₀= 0.16 μM; RLM CL_int> 399 μL min⁻¹mg⁻¹).

ing R- and S-conﬁgured sulfonamides, respectively. By

comparison of chiral HPLC retention times, the eutomer 7

was conﬁrmed to possess the R conﬁguration and the distomer

was veriﬁed to have the S absolute stereochemistry.

Though benzodioxepine-based diarylmethanesulfonamide 7

exhibits modest biochemical potency toward disrupting the

GK−GKRP binding interaction as reiterated in Table 1 (hGK−

hGKRP IC₅₀= 0.87 μM),¹⁷it does not display measurable

cellular potency in an antibody-based assay of GK translocation

in mouse hepatocytes (mGK TL EC₅₀> 12.5 μM).²¹Analysis

of the X-ray crystal structure depicted in Figure 2 suggested

that a substituent at the ortho position of the phenyl ring of the

ligand, oriented away from the sulfonamide NH group, could

potentially occupy an unﬁlled niche in the binding site of the

protein. Pleasingly, ortho-chloro analogue 11 was found to

exhibit signiﬁcantly improved biochemical and cellular potency

(hGK−hGKRP IC₅₀= 0.052 μM; mGK TL EC₅₀= 6.3

μM).^17,21Further potency enhancement was realized upon

Since modiﬁcation of the benzodioxepine moiety of 12 did

not appear to be a viable approach to improving microsomal

stability, systematic evaluation of the eﬀects of truncating it was

undertaken (Table 2). Surprisingly, excising just one or two

methylene groups from the seven-membered benzodioxepine

ring of 12 to give the six- and ﬁve-membered benzodioxine and

benzodioxole analogues 17 and 18, respectively, was found to

result in a precipitous loss of biochemical and cellular potency

with no gain in microsomal stability (hGK−hGKRP IC₅₀= 2.5

Table 1. Optimization of Benzodioxepine-Based Diarylmethanesulfonamide Disrupters of GK−GKRP Binding

a

b

c

d

compd

R¹

R²

R³

X

Y

Z

hGK−hGKRP IC₅₀(μM) mGK TL EC₅₀(μM)

RLM CL_int(μL min⁻¹mg⁻¹

)

cLogP

7

H

F

H

O

S

CH

N

CH

N

0.87

>12.5

6.3

>399

304

5.0

5.7

6.2

5.1

6.5

7.4

4.0

e

11

12

13

14

15

16

H

Cl

0.052

0.017

0.10

H

2.9

f

OH

S

4.3

f

F

S

0.074

4.1

>399

e

F

S

0.81

>12.5

3.3

H

S

0.16

a

Inhibition of the GK−GKRP binding interaction as measured in a bead-based proximity assay (AlphaScreen) using human GK and GKRP

constructs. Data represent an average of at least three measurements unless otherwise indicated. Standard deviations are reported in the Supporting

Information. GK translocation (TL) as measured in an antibody-based assay using mouse hepatocytes. Data represent an average of at least 3

b

measurements unless otherwise indicated (for EC₅₀> 12.5 μM, data represent 1 measurement). Standard deviations are reported in the Supporting

c

Information. Estimated intrinsic clearance (CL_int) determined by incubation of test compound with rat liver microsomes (RLM) at 37 °C for 30

d

min and measurement of percent turnover. Calculated using Daylight Chemical Information Systems software (Web site: www.daylight.com).

e

f

Data represent an average of two measurements. Standard deviations are reported in the Supporting Information. Prepared and tested as a 1:1

mixture of epimers at this stereocenter.

C

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 2. Systematic Truncation of the Benzodioxepine

Moiety

a

Inhibition of the GK−GKRP binding interaction as measured in a

bead-based proximity assay (AlphaScreen) using human GK and

GKRP constructs. Data represent an average of at least three

measurements (for EC₅₀> 33 μM, data represent two measurements).

Figure 3. Top-view overlay of the X-ray crystal structures of 4

(transparent white) and 12 (orange) bound to hGKRP (yellow), PDB

codes 4LY9 and 4PX3, respectively. Hydrogen bonding interactions

are represented by dashed lines; residues and interatomic distances (in

Å) of note are labeled in white. Schematic depicts resultant proposed

structure modiﬁcations.

b

Standard deviations are reported in the Supporting Information. GK

translocation (TL) as measured in an antibody-based assay using

mouse hepatocytes. Data represent an average of at least three

measurements (for EC₅₀> 12.5 μM, data represent one measure-

ment). Standard deviations are reported in the Supporting

Information. Estimated intrinsic clearance (CL_int) determined by

incubation of test compound with rat liver microsomes (RLM) at 37

interesting to note that, as with the benzodioxepine moiety of

12, the morpholine ring of 4 was found to be microsomally

labile,¹²requiring replacement with a propynyl group during

the lead optimization eﬀort culminating in the identiﬁcation of

microsomally stable 5.¹³

c

°C for 30 min and measurement of percent turnover.

and 6.5 μM, respectively; mGK TL EC₅₀> 12.5 μM and RLM

CL_int> 399 μL min⁻¹mg⁻¹for both 17 and 18). The para-

methoxyphenyl-, phenyl-, and cyclopropyl analogues 19−21

were also shown to exhibit poor biochemical and cellular

The upper portion of Figure 3 also illustrates the bifurcated

hydrogen bonding interaction between the carbinol group of 4

and the guanidinium side chain of Arg525.¹²Notably, an

Arg525-engaging substituent, whether a carbinol or a suitable

replacement, has served as a potency mainstay in all GKRP

disrupters reported to date.¹¹⁻¹⁴In contrast, this region of the

GKRP binding site is completely unoccupied in the binding

mode of 12. It was therefore proposed that introduction of an

Arg525-engaging moiety onto C7 of the benzothiophene ring

of 12 might improve its potency and allow truncation of the

bulky and microsomally labile benzodioxepine ring system. Due

to the steric constraints imposed by the constricted nature of

the Arg525 region of the GKRP binding site, it was clear that

access to this residue would require a nearly coplanar

arrangement of the benzothiophene moiety and the ring

bearing the guanidine-engaging substituent. Given the prefer-

ence for ≥45° dihedral angles in biphenyl systems,²²the use of

a C7-phenyl ring to bear the 4′-carbinol moiety (Figure 3; X =

CH) was not expected to be optimal.

potency and high microsomal turnover (hGK−hGKRP IC₅₀

=

7.3, >33, and >33 μM, respectively; mGK TL EC₅₀> 12.5 μM

and RLM CL_int> 399 μL min⁻¹mg⁻¹for 19−21). Because

replacement of the microsomally labile benzodioxepine moiety

of 12 did not lead to improved stability for 17−21, it was

speculated that the site of metabolism had shifted to another

part of the scaﬀold in these truncated analogues. At this

juncture, reassessment of the lead optimization direction was

necessary.

A top-view alignment of the X-ray crystal structures of the

biochemically equipotent 4 and benzodioxepine 12 bound to

GKRP is shown in Figure 3 (2.4 Å resolution for both 4 and

12; hGK−hGKRP IC₅₀= 0.021 and 0.017 μM for 4 and 12,

respectively).¹²As previously noted for HTS hits 3 and 6, the

sulfonyl groups of 4 and 12 interact with diﬀerent residues

(Arg215 and Asn216, respectively). In contrast, it was observed

that the benzodioxepine oxygen atom of 12 para to the

sulfonamide group is nearly superimposable with the oxygen

atom of the morpholine ring of 4 and that both of them engage

in intermolecular hydrogen bonds with Ile11.¹²It is also

Intramolecular interactions of sulfur atoms in aromatic

systems with O and N donor atoms (e.g., O → S and N →

S) have been the subject of signiﬁcant attention in the

theoretical, spectroscopic, and crystallographic literature but

have not been exploited to a signiﬁcant extent in drug discovery

D

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

eﬀorts.²³For systems possessing an aromatic sulfur atom and a

nearby heteroatom bearing a lone pair (e.g., carbonyl group,

heteroaromatic ring, etc.) with a favorable orientation with

respect to sulfur, the syn conformation is stabilized and, in fact,

often dominant. This phenomenon has been ascribed to a

stabilizing interaction between the donor lone pair and the σ*

orbital of sulfur and the atom to which it is bonded (n_donor

→

σ*_S−X), outweighing the repulsive interactions between the

sulfur and heteroatom lone pairs. Such a “syn-locked” eﬀect has

been found to be operative in a series of ΔF508-CFTR

correctors for the treatment of cystic ﬁbrosis²⁴and has been

exploited in a recent lead optimization campaign targeting

inhibitors of p38α.²⁵

Density functional theory calculations,²⁶along with natural

bond orbital (NBO) analysis of Weinhold,^27,28demonstrated

that a 2-pyridyl substituent attached to C7 of a benzothiophene

core (Figure 3; X = N) would exhibit a strong coplanarizing

eﬀect due to a nearly geometrically ideal interaction trajectory

(∠N−S−C ≈ 168°) between the lone pair of the pyridine

nitrogen and the σ* orbital of the S−C bond (Figure 4). In the

context of these benzothiophene-containing GKRP disrupters,

this system was predicted to present the C7 pyridyl ring and its

accompanying 4′-carbinol functionality in nearly optimal

alignment for contact with the distal guanidinium side chain

of Arg525. To dramatically lower the molecular weight of the

proposed GK−GKRP disrupters bearing a 4′-substituted C7

pyridyl moiety, truncation of the high molecular weight

benzodioxepine ring system to the low molecular weight, and

presumably more microsomally stable, cyclopropane ring was

chosen (Figure 3). Although cyclopropane analogue 21 had

been found to suﬀer high microsomal turnover, it was unclear

what eﬀect 2-pyridyl moieties attached to C7 of the

benzothiophene core would have on the metabolic stability of

the proposed analogues.

The realization of this unconventional design strategy is

detailed in Table 3. Compared to the 7-hydridobenzothiophene

21 (hGK−hGKRP IC₅₀> 33 μM), the 7-phenylbenzothio-

phene analogues 22 and 23 were found to display improved

potency toward disrupting the GK−GKRP binding interaction

(hGK−hGKRP IC₅₀= 4.4 and 2.7 μM, respectively). Even

more strikingly, and as predicted, the 7-pyridylbenzothiophene

analogues 24 and 25 were shown to exhibit a further 10-fold

Figure 4. (A) B3LYP/6-31+G(d,p) dihedral plot of model 7-(5′-

methyl-2′-pyridyl)benzothiophene and 7-(3′-toluyl)benzothiophene,

demonstrating strong syn-coplanarizing eﬀect of the N → S

interaction. (B) Depiction of the stabilizing interaction between the

occupied nitrogen lone pair (n_N) and unoccupied S−C2 (σ*_S−C

)

NBOs, utilizing the B3LYP/6-31+G(d,p) density.

improvement in biochemical potency (hGK−hGKRP IC₅₀

=

0.30 and 0.19 μM, respectively). These considerable functional

activity improvements for 7-pyridylbenzothiophene analogues

24 and 25 are likely driven by their greater GKRP binding

aﬃnities as measured in surface plasmon resonance assays

(hGKRP K_d= 34, 36, 0.50, and 0.25 μM for 22−25,

respectively), a result of a decreased conformational energy

penalty required for binding. Pleasingly, compounds 24 and 25

gem-dimethylcarbinol moiety is not chemically stable in the

original N-phenylpiperazine series of GKRP disrupters 3−

5,¹¹⁻¹³as this group is para to the amine functionality in these

compounds. Gratifyingly, despite displaying similar biochemical

potency (hGK−hGKRP IC₅₀= 0.11 μM), the gem-

dimethylcarbinol analogue 26 was found to exhibit improved

cellular permeability and potency (P_app= 7.5 × 10⁻⁶cm/s;

mGK TL EC₅₀= 7.7 μM), with only a slight loss in microsomal

stability (RLM CL_int= 34 μL min⁻¹mg⁻¹). However, further

improvement in biochemical and cellular potency was desired.

A front-view alignment of the X-ray crystal structures of 5

and 26 bound to GKRP is depicted in Figure 5 (2.0 and 2.6 Å

resolution, respectively).¹³The crystal structure of 26 with

GKRP clearly reveals the syn-locked orientation of the

benzothiophene sulfur and pyridine nitrogen atoms, as well

as the resultant successful engagement of the 4′-gem-

dimethylcarbinol moiety with the guanidinium side chain of

Arg525. Analysis of the 2-chlorophenyl ring of 26 suggested

that a F atom at the second ortho position might be

also display low clearance in rat liver microsomes (RLM CL_int

<

14 μL min⁻¹mg⁻¹).

Unfortunately, 24 and 25 do not exhibit measurable cell

potency in mouse hepatocytes (mGK TL EC₅₀> 12.5 μM). As

the binding aﬃnities of 24 and 25 to human and mouse GKRP

were found to be similar (mGKRP K_d= 1.8 and 0.85 μM,

respectively), the low cell potency of 24 and 25 was attributed

to their high lipophilicity (cLogP = 5.1; ACDLogP = 6.3;

ACDLogD_7.4= 6.2) and poor cellular permeability (P_app< 1.8

× 10⁻⁶cm/s and P_app= 2.4 × 10⁻⁶cm/s, respectively). The

gem-dimethylcarbinol analogue 26 was designed to simulta-

neously decrease lipophilicity and molecular weight, with the

added beneﬁt of removing a stereocenter. It is notable that a

E

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 3. Modiﬁcation of the C7 Aryl Ring of Cyclopropane-Based Diarylmethanesulfonamide Disrupters of GK−GKRP

Binding

d

e

hGK−hGKRP

h,r,mGKRP

mGK TL

P_app(ER)

RLM CL_int

f

g

compd

R

X

cLogP

ACDLogP (ACDLogD_7.4)

a

b

c

IC₅₀(μM)

K_d(nM)

EC₅₀(μM) (×10⁻⁶cm/s) (μL min⁻¹mg⁻¹

)

h

i

21

>33

4.4

ND

>12.5

7.7

ND

>399

<14

34

4.5

6.4

5.1

4.8

5.7 (5.7)

7.6 (7.6)

6.3 (6.2)

5.6 (5.6)

j

i

22

23

24

25

26

(R)-CF₃

(S)-CF₃

(R)-CF₃

(S)-CF₃

CH₃

CH

N

34000

36000

2.7

0.30

0.19

0.11

500, 430, 1800

250, 800, 850

90, 130, 290

<1.8 (>2.6)

2.4 (2.7)

N

7.5 (2.2)

a

Inhibition of the GK−GKRP binding interaction as measured in a bead-based proximity assay (AlphaScreen) using human GK and GKRP

constructs. Data represent an average of at least three measurements (for EC₅₀> 33 μM, data represent two measurements). Standard deviations are

reported in the Supporting Information. Dissociation constant (K_d) as measured in a surface plasmon resonance assay (Biacore) using human, rat,

b

or mouse GKRP constructs, respectively. Data represent an average of two measurements unless otherwise indicated. Standard deviations are

c

reported in the Supporting Information. GK translocation (TL) as measured in an antibody-based assay using mouse hepatocytes. Data represent an

average of at least three measurements (for EC₅₀> 12.5 μM, data represent at least one measurement). Standard deviations are reported in the

d

Supporting Information. Apparent permeability (P_appA → B) through porcine proximal tubule cells (LLC-PK1 cell line) and eﬄux ratio (ER).

e

Estimated intrinsic clearance (CL_int) determined by incubation of test compound with rat liver microsomes (RLM) at 37 °C for 30 min and

f

g

measurement of percent turnover. Calculated using Daylight Chemical Information Systems software (Web site: www.daylight.com). Calculated

using Advanced Chemistry Development (ACD/Labs) software (Web site: www.acdlabs.com). Compound 21 only bears an H atom at C7 of the

h

i

benzothiophene ring system and is included in Table 3 to facilitate comparison with C7-substituted analogues; see Table 2. Not determined (ND).

j

Dissociation constant (K_d) as measured in a surface plasmon resonance assay (Biacore) using human GKRP constructs only. Data represent one

measurement.

accommodated and engage in a favorable electrostatic

interaction with the sulfonamide NH group (Figure 5; X =

CF),²⁹which could partially mask this hydrogen bond donor,

improve cell permeability, and potentially favor the binding

conformation. The alignment of the X-ray crystal structures of

5 and 26 bound to GKRP clearly shows that the potency-

enhancing, and lipophilicity-lowering, aminopyridine function-

ality present in 5 could be incorporated into 26 in an analogous

fashion, albeit with diﬀering regiochemistry required for

optimal engagement of Gly181 and Met213 (Figure 5; X or

Y = CH or N; R = NH₂).

The results of incorporating these design elements are

summarized in Table 4. Compared to 26, ortho-ﬂuoro analogue

27 was found to display 2- to 3-fold better binding aﬃnity

(h,r,mGKRP K_d= 0.030, 0.060, and 0.12 μM, respectively) but

only marginally improved biochemical potency (hGK−hGKRP

IC₅₀= 0.056 μM), cell permeability and potency (P_app= 9.5 ×

10⁻⁶cm/s; mGK TL EC₅₀= 5.2 μM), and microsomal stability

(RLM CL_int= 18 μL min⁻¹mg⁻¹). In contrast, aminopyridine

analogue 28 was shown to exhibit 10- to 20-fold better binding

aﬃnity (h,r,mGKRP K_d= 0.0042, 0.013, and 0.024 μM,

nearly 5-fold lower cellular potency (mGK TL EC₅₀= 1.9 μM),

despite having similar cellular permeability (P_app= 7.8 × 10⁻⁶

cm/s). Compared to aminopyridine 28, aminopyridine

regioisomer 30 was shown to possess a 2- to 4-fold decrease

in binding aﬃnity (h,r,mGKRP K_d= 0.019, 0.013, and 0.052

μM, respectively) and biochemical potency (hGK−hGKRP

IC₅₀= 0.025 μM); however, it was found to have improved

microsomal stability (RLM CL_int= 67 μL min⁻¹mg⁻¹) and

cellular permeability (P_app= 19 × 10⁻⁶cm/s), which likely led

to its similar cellular potency (mGK TL EC₅₀= 0.41 μM). The

aminopyrimidine 31 was designed to simultaneously block a

site of metabolism and possess lower lipophilicity. Analogue 31

exhibits excellent binding aﬃnity (h,r,mGKRP K_d= 0.0086,

0.0079, and 0.031 μM, respectively), biochemical potency

(hGK−hGKRP IC₅₀= 0.018 μM), and cellular permeability

and potency (P_app= 19 × 10⁻⁶cm/s; mGK TL EC₅₀= 0.48

μM), as well as low intrinsic clearance (RLM CL_int= 28 μL

min⁻¹mg⁻¹). Finally, ortho-ﬂuoroaminopyridine analogue 32

was found to display the greatest binding aﬃnity (h,r,mGKRP

K_d= 0.0015, 0.0012, and 0.0045 μM, respectively) and

biochemical potency (hGK−hGKRP IC₅₀= 0.0038 μM),

respectively) and functional activity (hGK−hGKRP IC₅₀

=

along with excellent cellular permeability and potency (P_app=

0.0060 μM; mGK TL EC₅₀= 0.43 μM), as well as slightly

improved cellular permeability (P_app= 11 × 10⁻⁶cm/s), but at

the expense of higher microsomal clearance (RLM CL_int= 130

μL min⁻¹mg⁻¹) versus 26. A metabolite identiﬁcation study of

28 using human liver microsomes revealed that the primary

product of metabolism was hydroxylation at the position ortho

to the amino group of the aminopyridine ring. Substitution of

this site with a F atom gave analogue 29,³⁰which was found to

possess signiﬁcantly improved microsomal clearance (RLM

CL_int= 28 μL min⁻¹mg⁻¹) and comparable biochemical

activity (h,r,mGKRP K_d= 0.0066, 0.017, and 0.035 μM,

respectively; hGK−hGKRP IC₅₀= 0.011 μM) but to have

16 × 10⁻⁶cm/s; mGK TL EC₅₀= 0.066 μM) and low intrinsic

clearance (RLM CL_int= 26 μL min⁻¹mg⁻¹). It was also

pleasing that 31 and 32 were found to retain selective binding

aﬃnity for GKRP (hGK K_d> 10 μM for both 31 and 32).

The X-ray crystal structure of 32 bound to GKRP is shown in

Figure 6 (2.2 Å resolution). As expected, the interactions of the

sulfonamide group with Asn216 and the carbinol moiety with

Arg525, as well as the projection of the cyclopropyl

functionality and the coplanar arrangement of the benzothio-

phene core with the C7-pyridyl ring, were all found to be nearly

identical to those observed for 26. It was also gratifying to

observe that the aminopyridine functionality of 32 engaged in

F

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 6. X-ray crystal structure of 32 (orange) bound to hGKRP

(green), PDB code 4PX5. Hydrogen bonding interactions are

represented by dashed lines, and residues and interatomic distances

(in Å) of note are labeled in white.

hydrogen bonds with Gly181 and Met213 as planned. The F

atom of the ortho-ﬂuoroaminopyridine moiety of 32, while

possibly aﬀording a small amount of additional van der Waals

contact with the solvent-exposed protein surface, also may give

rise to a pair of stabilizing electrostatic interactions between

both the sulfonamide NH group and one of the H atoms of the

pendent NH₂functionality. This pair of electrostatically

favorable interactions could serve to mask and reduce the

desolvation penalty of these polar H atoms and stabilize the

observed binding orientation of the ring, as well.

Figure 5. Overlay of the X-ray crystal structures of 5 (transparent

white) and 26 (orange) bound to hGKRP (yellow), PDB codes

4MQU and 4PXS, respectively. Hydrogen bonding interactions are

represented by dashed lines; residues and interatomic distances (in Å)

of note are labeled in white. Schematic depicts resultant proposed

structure modiﬁcations.

Pharmacokinetic proﬁles were obtained for the two

analogues in Table 4 with the best combination of cellular

potency and microsomal stability, 31 (AMG-7549) and 32

Table 4. Modiﬁcation of the 2-Chlorophenyl Ring of Cyclopropane-Based Diarylmethanesulfonamide Disrupters of GK−GKRP

Binding

d

e

hGK−hGKRP

h,r,mGKRP

mGK TL

P_app(ER)

RLM CL_int

f

g

compd

R

X

Y

cLogP ACDLogP (ACDLogD_7.4)

a

b

c

IC₅₀(μM)

K_d(nM)

EC₅₀(μM) (×10⁻⁶cm/s) (μL min⁻¹mg⁻¹

)

26

27

28

29

30

31

32

H

CH

CF

N

CH

CF

N

0.11

90, 130, 290

30, 60, 120

4.2, 13, 24

6.6, 17, 35

19, 13, 52

7.7

7.5 (2.2)

9.5 (2.3)

11 (2.0)

7.8 (1.8)

19 (1.7)

19 (2.3)

16 (2.3)

34

18

130

28

67

28

26

4.8

4.9

3.1

3.2

3.1

2.3

3.2

5.6 (5.6)

5.2 (5.2)

4.5 (4.3)

4.4 (4.0)

4.5 (4.3)

3.4 (3.2)

4.4 (3.9)

H

0.056

0.0060

0.011

0.025

0.018

0.0038

5.2

NH₂

0.43

1.9

N

CH

N

0.41

0.48

0.066

N

8.6, 7.9, 31

1.5, 1.2, 4.5

CF

N

a

Inhibition of the GK−GKRP binding interaction as measured in a bead-based proximity assay (AlphaScreen) using human GK and GKRP

b

constructs. Data represent an average of at least three measurements. Standard deviations are reported in the Supporting Information. Dissociation

constant (K_d) as measured in a surface plasmon resonance assay (Biacore) using human, rat, or mouse GKRP constructs, respectively. Data represent

c

an average of two measurements. Standard deviations are reported in the Supporting Information. GK translocation (TL) as measured in an

antibody-based assay using mouse hepatocytes. Data represent an average of at least three measurements. Standard deviations are reported in the

d

Supporting Information. Apparent permeability (P_appA → B) through porcine proximal tubule cells (LLC-PK1 cell line) and eﬄux ratio (ER).

e

Estimated intrinsic clearance (CL_int) determined by incubation of test compound with rat liver microsomes (RLM) at 37 °C for 30 min and

f

g

measurement of percent turnover. Calculated using Daylight Chemical Information Systems software (Web site: www.daylight.com). Calculated

using Advanced Chemistry Development (ACD/Labs) software (Web site: www.acdlabs.com).

G

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 5. Pharmacokinetic Proﬁles of Key Compounds 31 and 32

b

c

d

e

PPB (%)

rat

CL (L h⁻¹kg⁻¹

2 mg/kg, iv

0.28

)

V_dss(L/kg)

t_1/2(h)

C_max(μM)

t_max(h)

F

a

compd

2 mg/kg, iv

1.3

2 mg/kg, iv

3.0

10 mg/kg, po

(%)

31

32

99

1.9

2.8

4.7

3.3

25

46

f

0.36

1.6

3.1

a

b

In vivo experiments were conducted using male Sprague−Dawley rats (N = 3/group unless otherwise indicated). Percent rat plasma protein

c

binding (PPB) as measured in vitro following rapid equilibrium dialysis. Clearance (CL), volume of distribution (V_dss), and half-life (t_1/2

determined following a single intravenous dose (2.0 mg/kg; vehicle = DMSO). Maximal concentration (C_max) and time at maximal concentration

(t_max) determined following a single oral dose (10 mg/kg; vehicle = 1.0% Tween 80, 2.0% HPMC, 97% H₂O, pH 7.0). Percent bioavailability (F)

calculated using AUC_0→∞values determined from the 2.0 mg/kg (iv) and 10 mg/kg (po) doses. N = 1.

)

d

e

f

(AMG-0696), in preparation for more detailed studies in vivo

(Table 5). Both analogues were found to exhibit high rat

plasma protein binding in an in vitro assay employing rapid

equilibrium dialysis (PPB = 99% for both 31 and 32). In male

Sprague−Dawley rats (N = 3 and 1 for 31 and 32, respectively)

given a single intravenous dose (2.0 mg/kg), these compounds

were shown to display similarly good pharmacokinetic proﬁles

(CL = 0.28 and 0.36 L h⁻¹kg⁻¹, V_dss= 1.3 and 1.6 L/kg, and

t_1/2= 3.0 and 3.1 h for 31 and 32, respectively). In constrast, in

rats (N = 3 for both 31 and 32) given a single oral dose (10

mg/kg), aminopyrimidine 31 was found to exhibit lower peak

concentration, slower absorption, and less bioavailability

compared to ortho-ﬂuoroaminopyridine 32 (C_max= 1.9 and

2.8 μM, t_max= 4.7 and 3.3 h, and F = 25% and 46% for 31 and

32, respectively).

Because 31 and 32 were shown to display good

pharmacokinetic proﬁles, the pharmacodynamic eﬀects of

each compound were evaluated (Figure 7).³¹Six hours after a

single 10 or 100 mg/kg oral dose in Sprague−Dawley rats (N =

6 for both doses), aminopyrimidine 31 was found to exhibit

dose-proportional plasma exposure ([31]_plasma= 2.9 and 23 μM

for the 10 and 100 mg/kg dose, respectively) and robust GK

translocation (Figure 7A; IHC score = 0.33 and 0.0 for the 10

and 100 mg/kg dose, respectively). At the 24 h time point, the

plasma exposure and pharmacodynamic response for both the

10 and 100 mg/kg doses of 31 were shown to be signiﬁcantly

diminished ([31]_plasma= 0.20 and 0.21 μM, IHC score = 2.3

and 1.3 for the 10 and 100 mg/kg dose, respectively). The

pharmacodynamic response for 32 was found to be more

robust and durable, with full GK translocation lasting to the 24

h time point for both the 10 and 100 mg/kg doses (Figure 7B;

[32]_plasma= 0.22 and 6.3 μM for the 10 and 100 mg/kg dose,

Figure 7. Measurement of GK translocation in Sprague−Dawley rats 6

respectively; IHC score = 0.0 for both the 10 and 100 mg/kg

and 24 h after either a 10 or 100 mg/kg oral dose of (A) 31 or (B) 32.

doses).³²This eﬀect on GK translocation in Sprague−Dawley

The statistical signiﬁcance of the GK translocation measurements was

rats is consistent with the GK−GKRP disruption activities of 31

and 32.¹¹⁻¹⁴

based on comparison to the individual vehicle control groups (N = 6/

group; (∗) P < 0.05, (∗∗) P < 0.01, and (∗∗∗) P < 0.001 as calculated

by the ANOVA/Dunnett’s multiple comparison test).

The blood glucose lowering eﬃcacy of 31 and 32 in the

diabetic db/db mouse model was also investigated (Figure 8).³³

Six hours after a single 10, 30, or 100 mg/kg oral dose,

aminopyrimidine 31 was found to exhibit dose-proportional

whole-blood exposure ([31]_blood= 1.2, 4.1, and 19 μM for the

10, 30, and 100 mg/kg dose, respectively), with statistically

signiﬁcant blood glucose lowering achieved for the 100 mg/kg

dose (Figure 8A; POC = −45%).^34,35Similar to the rat GK

translocation response discussed above, the blood glucose

lowering activity for 32 in db/db mice was shown to be better

than that observed for 31, with more robust and statistically

signiﬁcant blood glucose lowering attained for the 100 mg/kg

in db/db mice is also consistent with the GK−GKRP

disruption activities of 31 and 32.¹¹⁻¹⁴

The synthesis of benzodioxepine-based diarylmethane-

sulfonamide 12 is described in Scheme 2 and exempliﬁes the

general approach to compounds 12−21 in Tables 1 and 2.³⁸

Commercially available 2-chlorobenzaldehyde (33) and (S)-2-

methyl-2-propanesulﬁnamide were condensed using Ti(OEt)₄

to aﬀord tert-butanesulﬁnylimine 34 in 88% yield. Regiose-

lective deprotonation of commercially available 1-benzothio-

phene with n-butyllithium followed by addition of 34 gave tert-

butanesulﬁnamide 35 possessing the R absolute stereo-

chemistry at the diarylmethane stereocenter in 65% yield

after removal of the minor undesired diastereomer by ﬂash

dose at both the 3 and 6 h time points (Figure 8B; [32]_blood

=

25 and 21 μM, POC = −45% and −33% for the 3 and 6 h time

points, respectively).^36,37This reduction in blood glucose levels

H

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

provide benzodioxepine-based diarylmethanesulfonamide 12

(66% yield, two steps).

As illustrated in Scheme 3, cyclopropane-based

diarylmethanesulfonamide 25 was accessed by a similar route.

Scheme 3. Chemical Synthesis of Cyclopropane-Based GK−

a

GKRP Disrupter 25

a

Reagents and conditions: (a) n-butyllithium, 1,1,1-triﬂuoroacetone,

THF, −78 °C (65%); (b) chiral SFC puriﬁcation;⁴⁰(c) 38,

Pd(allyl)₂Cl₂, X-Phos, Na₂CO₃·H₂O, 4:1 1,4-dioxane−H₂O, 80 °C

(50%); (d) n-butyllithium (2.2 equiv), 34, THF, −78 °C (51%); (e)

HCl, Et₂O, 23 °C; (f) cyclopropanesulfonyl chloride, i-Pr₂NEt,

DMAP, DMF, 23 °C (70%, two steps).

Figure 8. Measurement of blood glucose lowering in db/db mice 3, 6,

and 24 h after a 10, 30, or 100 mg/kg oral dose of (A) 31 or (B) 32.

The statistical signiﬁcance of the blood glucose measurements was

based on comparison to the individual vehicle control groups (N = 5−

8/group; (∗) P < 0.05 and (∗∗) P < 0.01 as calculated by the

ANOVA/Dunnett’s multiple comparison test).

The synthesis of 25 showcases the standard means to the 7-

arylbenzothiophene disrupters 22−27 in Tables 3 and 4.³⁸

Lithium−iodine exchange of commercially available 2-chloro-4-

iodopyridine (36) using n-butyllithium followed by addition of

1,1,1-triﬂuoroacetone aﬀorded racemic 2-(2-chloro-4-pyridin-

yl)-1,1,1-triﬂuoro-2-propanol (37) in 65% yield, which was

resolved by chiral SFC to aﬀord the enantiomerically pure S-

conﬁgured isomer 38 and its R-conﬁgured antipode.⁴⁰2-

Chloropyridine 38 was coupled with known 2-(1-benzothio-

phen-7-yl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (39)⁴¹using

Pd(allyl)₂Cl₂/X-Phos to give 7-pyridylbenzothiophene 40 in

50% yield. Selective bis-deprotonation of 40 with 2.2 equiv of

n-butyllithium followed by addition of tert-butanesulﬁnylimine

34 gave tert-butanesulﬁnamide 41 possessing the R conﬁg-

uration at the diarylmethane stereocenter in 51% yield after

removal of the minor undesired diastereomer by ﬂash

chromatography on silica gel.³⁹Lastly, the tert-butanesulﬁnyl

group of 41 was removed by treatment with HCl and the

resulting amine was sulfonylated using commercially available

cyclopropanesulfonyl chloride to provide cyclopropane-based

diarylmethanesulfonamide 25 (70% yield, two steps).

Scheme 2. Chemical Synthesis of Benzodioxepine-Based

a

GK−GKRP Disrupter 12

The synthetic approach used to acquire phenyl analogue 25

was extended to the more complex aminopyridines 28 and 29

and aminopyrimidine 31, as represented by the assembly of 31

in Scheme 4.³⁸Commercially available 4-chloropyrimidin-2-

amine (42) was regioselectively chlorinated with 1,3,5-

trichloroisocyanuric acid, and then the amine functionality

was protected as its di-tert-butylimidodicarbonate using Boc₂O/

DMAP to provide dichloropyrimidine 43 in 53% yield for the

two-step process. Regioselective palladium-catalyzed cross-

coupling (Pd(AmPhos)₂Cl₂/KOAc) of 43 with commercially

a

Reagents and conditions: (a) (S)-2-methyl-2-propanesulﬁnamide,

Ti(OEt)₄, CH₂Cl₂, 23 °C (88%); (b) 1-benzothiophene, n-

butyllithium, THF, −78 °C (65%); (c) HCl, MeOH, 23 °C; (d) 10,

i-Pr₂NEt, DMAP, CH₂Cl₂, 23 °C (66%, two steps).

chromatography on silica gel.³⁹Finally, the tert-butanesulﬁnyl

group of 35 was removed (HCl/MeOH) and the resulting

amine was sulfonylated using commercially available 3,4-

dihydro-2H-1,5-benzodioxepine-7-sulfonyl chloride (10) to

I

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

The aminopyridines 30 and 32 were prepared by a diﬀerent

route, as shown for 32 in Scheme 5.³⁸Selective deprotonation

Scheme 4. Chemical Synthesis of GK−GKRP Disrupter 31

a

Scheme 5. Chemical Synthesis of GK−GKRP Disrupter 32

a

Reagents and conditions: (a) n-butyllithium, DMF, THF, −78 °C;

(b) (R)-2-methyl-2-propanesulﬁnamide, Ti(OEt)₄, CH₂Cl₂, 23 °C

(87%, two steps); (c) 5-chloro-3-ﬂuoro-2-pyridinamine, n-butyllithium

(1.9 equiv), THF, −78 °C (23%); (d) HCl, MeOH, 23 °C; (e)

cyclopropanesulfonyl chloride, i-Pr₂NEt, DMAP, DMF, 23 °C (83%,

two steps).

a

Reagents and conditions: (a) 1,3,5-trichloroisocyanuric acid, AcOH,

H₂O, 50 °C; (b) Boc₂O, DMAP, THF, 23 °C (53%, two steps); (c)

potassium vinyltriﬂuoroborate, Pd(AmPhos)₂Cl₂, KOAc, 3:1 MeCN−

H₂O, 70 °C (77%); (d) K₂OsO₄·2H₂O, NaIO₄, 4:1 THF−H₂O, 23

°C; (e) (S)-2-methyl-2-propanesulﬁnamide, CuSO₄, CH₂Cl₂, 23 °C

(58%, two steps); (f) n-butyllithium, acetone, THF, −78 °C (61%);

(g) Pd(allyl)₂Cl₂, X-Phos, Na₂CO₃, 6:1 1,4-dioxane−H₂O, 80 °C

(53%); (h) Et₃SiOTf, i-Pr₂NEt, THF, 0 °C (88%); (i) n-butyllithium,

LiCl, 45, THF, −78 °C (quant, crude); (j) HCl, MeOH, 10 → 23 °C;

(k) cyclopropanesulfonyl chloride, i-Pr₂NEt, DMAP, DMF, 13 → 23

°C (49%, three steps); (l) chiral SFC puriﬁcation.⁴²

of 48 with n-butyllithium followed by addition of N,N-

dimethylformamide gave the desired aldehyde, which was

condensed with commercially available (R)-2-methyl-2-pro-

panesulﬁnamide using Ti(OEt)₄to aﬀord tert-butanesulﬁnyli-

mine 50 (87% yield, two steps). Selective bis-deprotonation of

commercially available 5-chloro-3-ﬂuoro-2-pyridinamine with

1.9 equiv of n-butyllithium followed by addition of tert-

butanesulﬁnylimine 50 gave tert-butanesulﬁnamide 51 possess-

ing the R conﬁguration at the diarylmethane stereocenter in

23% yield after removal of the minor undesired diastereomer by

ﬂash chromatography on silica gel.^39,43Global deprotection of

51 by treatment with HCl in methanol followed by selective

sulfonylation of the diarylmethaneamine group with commer-

cially available cyclopropanesulfonyl chloride gave 32 in 83%

yield for the two-step process.

available potassium vinyltriﬂuoroborate gave alkene 44 (77%

yield). Alkene 44 was oxidatively cleaved using K₂OsO₄/NaIO₄

to the desired aldehyde, which was then condensed with

commercially available (S)-2-methyl-2-propanesulﬁnamide

using CuSO₄to aﬀord tert-butanesulﬁnylimine 45 (58% yield,

two steps). Lithium−iodine exchange of commercially available

2-chloro-4-iodopyridine (36) using n-butyllithium followed by

addition of acetone aﬀorded 2-(2-chloro-4-pyridinyl)-2-prop-

anol (46) in 61% yield. 2-Chloropyridine 46 was cross-coupled

with 39 using Pd(allyl)₂Cl₂/X-Phos/Na₂CO₃to give 2-(2-(1-

benzothiophen-7-yl)-4-pyridinyl)-2-propanol (47) in 53%

yield, the gem-dimethylcarbinol moiety of which was then

protected as its triethylsilyl ether (Et₃SiOTf/i-Pr₂NEt) to aﬀord

aﬀord 2-(1-benzothiophen-7-yl)-4-(1-methyl-1-((triethylsilyl)-

oxy)ethyl)pyridine (48) in 88% yield. Selective deprotonation

of 48 (n-butyllithium/LiCl) followed by addition of tert-

butanesulﬁnylimine 45 gave tert-butanesulﬁnamide 49 as a

∼1:1 mixture of diastereomers at the diarylmethane stereo-

center.³⁹The crude material was deprotected by treatment with

HCl in methanol, and the resulting diarylmethaneamine group

was sulfonylated using commercially available cyclopropane-

sulfonyl chloride to give the ﬁnal product as a ∼1:1 mixture of

enantiomers (49% overall yield, three steps). The enantiomeric

mixture was resolved using chiral SFC and the R absolute

stereochemistry of 31 was assigned by comparison of its

quantum mechanically computed and measured optical

rotation.⁴²

CONCLUSIONS

■

A structurally distinct series of GKRP-selective GK−GKRP

disrupters was discovered through an HTS campaign of a

diverse compound library. By leveraging structure-guided drug

design, benzodioxepine-based diarylmethanesulfonamide 6 was

transformed into cyclopropane-based diarylmethane-

sulfonamide 32. This lead optimization eﬀort exploited an

unusual N → S (n_N→ σ*_S−X) interaction to conformationally

constrain a key biaryl linkage. Computational studies revealed

the underpinnings of the N → S (n_N→ σ*_S−X) interaction, and

X-ray analysis of 32 bound to GKRP conﬁrmed its predicted

binding conformation and interactions with the protein. Lead

compound 32 was found to possess improved metabolic

stability, as well as to induce dramatic GK translocation in

Sprague−Dawley rats and robust blood glucose reduction in

db/db mice. This structurally novel disrupter of GK−GKRP

binding allows further exploration of the pharmacology of this

potential therapeutic axis for T2DM and highlights the value of

incorporating unconventional nonbonded interactions in drug

design.

J

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

7.1 Hz, 1 H), 6.15 (d, J = 7.1 Hz, 1 H), 1.18 (s, 9 H). MS m/z 377.9

EXPERIMENTAL SECTION

■

+

[M + H]⁺(378.1 calcd for C₁₉H₂₁ClNOS₂).

Chemistry (General).^38,44Unless otherwise indicated, all

materials were obtained from commercial suppliers and used without

further puriﬁcation. Anhydrous solvents were obtained from Sigma−

Aldrich, Acros, or EM Science and used as received. All reactions

involving air- or moisture-sensitive reagents were performed under a

nitrogen or argon atmosphere. All microwave-assisted reactions were

performed in sealed reaction vials using a Personal Chemistry Emrys

Optimizer microwave synthesizer. Silica gel chromatography was

performed using either glass columns packed with silica gel (200−400

mesh, Sigma-Aldrich) or prepacked silica gel cartridges (Biotage or

RediSep). All ﬁnal compounds were puriﬁed to ≥95% purity unless

otherwise noted, as determined by HPLC−MS obtained on an Agilent

1100 or 1200 spectrometer using the following methods: [A] Agilent

SB-C18 column (50 mm × 3.0 mm, 2.5 μm) at 40 °C with a 1.5 mL/

min ﬂow rate using a gradient of 5−95% [0.1% TFA in acetonitrile] in

[0.1% TFA in water] over 3.5 min; [B] Waters XBridge C18 column

(50 mm × 3.0 mm, 3 μm) at 40 °C with a 1.5 mL/min ﬂow rate using

a gradient of 5−95% [0.1% formic acid in acetonitrile] in [0.1% formic

acid water] over 3.5 min. NMR spectra were obtained with a Bruker

300 MHz or a DRX 400, 500, or 600 MHz spectrometer. Data are

reported as follows: chemical shift in parts per million (ppm, δ units)

from an internal standard, multiplicity (s = singlet, d = doublet, t =

triplet, q = quartet, dd = doublet of doublets, quin = quintet, m =

multiplet, and br s = broad singlet), coupling constant (Hz), and

integration. Low-resolution mass spectrometry (MS) data were

obtained at the same time as the purity determination with the

HPLC−MS instrument using ES ionization (positive mode). High-

resolution mass spectrometry (HRMS) data were obtained with a

Waters Synapt G2 QTOF system operated in either positive or

negative ion mode.

Hydrogen chloride (1.7 mL of a 4.0 M solution with 1,4-dioxane,

6.9 mmol) was added to a solution of 35 (1.3 g, 3.4 mmol) and

methanol (3.4 mL) at room temperature. After 80 min, the reaction

mixture was concentrated under reduced pressure, diethyl ether was

added to the residue, the resulting slurry was subjected to sonication

for 5 min, ﬁltered, and the ﬁlter cake was collected to aﬀord (R)-1-(1-

benzothiophen-2-yl)-1-(2-chlorophenyl)methanamine hydrochloride

(0.90 g, 85% yield) as an oﬀ-white solid. The material was used in

the next step of the synthesis without puriﬁcation.

(R)-1-(1-Benzothiophen-2-yl)-1-(2-chlorophenyl)methanamine hy-

drochloride (0.10 g, 0.32 mmol) was dissolved with dichloromethane

(3.2 mL) and diisopropylethylamine (0.17 mL, 0.97 mmol), and then

3,4-dihydro-2H-1,5-benzodioxepine-7-sulfonyl chloride (10; 0.080 g,

0.32 mmol) was added. After 2 h, 4-dimethylaminopyridine (0.0039 g,

0.032 mmol) was added. After 3 h, more diisopropylethylamine (0.17

mL, 0.97 mmol) was added. After 30 min, more 10 (0.080 g, 0.32

mmol) was added. After 23 h, the reaction mixture was partitioned

between ethyl acetate and saturated aqueous sodium bicarbonate, the

layers were separated, the organic material was washed sequentially

with saturated aqueous sodium bicarbonate and brine, dried (sodium

sulfate), ﬁltered, and the ﬁltrate was concentrated under reduced

pressure. The residue was subjected to ﬂash chromatography on silica

gel (4:1 hexane−ethyl acetate) to give N-((R)-1-benzothiophen-2-

yl(2-chlorophenyl)methyl)-3,4-dihydro-2H-1,5-benzodioxepine-7-sul-

fonamide (12; 0.12 g, 77% yield; 66% yield, two steps) as a colorless

solid. [α]³⁰_D+14.2° (c 1.16, CHCl₃). ¹H NMR (400 MHz, DMSO-d₆)

δ 9.16 (br s, 1 H), 7.93−7.84 (m, 1 H), 7.75−7.66 (m, 1 H), 7.56−

7.49 (m, 1 H), 7.42−7.35 (m, 1 H), 7.35−7.22 (m, 4 H), 7.20 (dd, J =

2.2, 8.5 Hz, 1 H), 7.09 (d, J = 2.3 Hz, 1 H), 6.88 (d, J = 8.4 Hz, 1 H),

6.82 (s, 1 H), 6.16 (br s, 1 H), 4.17−3.96 (m, 4 H), 2.08 (quin, J = 5.6

Hz, 2 H). ¹³C NMR (150 MHz, DMSO-d₆) δ 153.9, 149.9, 144.4,

139.0, 138.7, 136.8, 134.6, 131.5, 129.4, 129.0, 128.6, 127.4, 124.5,

124.4, 123.7, 122.3, 122.2, 121.8, 121.5, 119.9, 70.2, 70.1, 53.6, 30.4.

N-((R)-1-Benzothiophen-2-yl(2-chlorophenyl)methyl)-3,4-di-

hydro-2H-1,5-benzodioxepine-7-sulfonamide (12). Titanium-

(IV) tetraethoxide (220 mL, 1100 mmol) was added to a solution

of 2-chlorobenzaldehyde (33; 30 g, 210 mmol), (S)-2-methyl-2-

propanesulﬁnamide (26 g, 210 mmol), and dichloromethane (430

mL) at room temperature. After 22 h, the reaction mixture was added

to water, dichloromethane was added, the mixture was agitated

vigorously, ﬁltered through a pad of diatomaceous earth, the two layers

comprising the ﬁltrate were separated, the organic material was dried

(sodium sulfate), ﬁltered, and the ﬁltrate was concentrated under

reduced pressure. The residue was subjected to ﬂash chromatography

on silica gel (9:1 hexane−ethyl acetate) to aﬀord N-((S,1E)-(2-

chlorophenyl)methylidene)-2-methyl-2-propanesulﬁnamide (34; 46 g,

HRMS m/z 484.0443 [M

C₂₄H₁₉ClNO₄S₂).

−

H]⁻(484.0444 calcd for

−

N-((R)-(2-Chlorophenyl)(7-(4-((1S)-2,2,2-triﬂuoro-1-hydroxy-

1-methylethyl)-2-pyridinyl)-1-benzothiophen-2-yl)methyl)-

cyclopropanesulfonamide (25). n-Butyllithium (33 mL of a 2.5 M

solution with toluene, 83 mmol) was added to a solution of 2-chloro-

4-iodopyridine (36; 20 g, 82 mmol) and tetrahydrofuran (200 mL) at

−78 °C. After 15 min, 1,1,1-triﬂuoroacetone (28 g, 250 mmol) was

added. After 1 h, saturated aqueous ammonium chloride was added,

the reaction vessel was removed from the cooling bath, the reaction

mixture was allowed to warm to room temperature, partitioned

between more saturated aqueous ammonium chloride and ethyl

acetate, the layers were separated, the aqueous material was washed

with ethyl acetate (2×), the combined organic extract was washed

sequentially with water and brine, dried (sodium sulfate), ﬁltered, and

the ﬁltrate was concentrated under reduced pressure. The residue was

subjected to ﬂash chromatography on silica gel (19:1 hexane−ethyl

acetate) to give racemic 2-(2-chloro-4-pyridinyl)-1,1,1-triﬂuoro-2-

propanol (37; 12 g, 65% yield) as a pale yellow solid. ¹H NMR

(400 MHz, DMSO-d₆) δ 8.63 (d, J = 5.1 Hz, 1 H), 7.91 (s, 1 H), 7.81

(d, J = 5.1 Hz, 1 H), 2.63 (s, 3 H); one exchangeable proton was not

observed. MS m/z 225.9 [M + H]⁺(226.0 calcd for C₈H₈ClF₃NO⁺).

Racemate 37 was resolved using preparative chiral SFC (two 250

mm × 30 mm, 5 μm Chiralcel OJ-H columns in series), eluting with

90% liquid CO₂in 10% mixture of hexanes/isopropanol (75:25 v/v)

containing 20 mM ammonia in methanol at a ﬂow rate of 120 mL/

min) to give two products in >98% enantiomeric excess. First eluting

peak: (2R)-2-(2-chloro-4-pyridinyl)-1,1,1-triﬂuoro-2-propanol. Second

eluting peak: (2S)-2-(2-chloro-4-pyridinyl)-1,1,1-triﬂuoro-2-propanol

(38).⁴⁰¹H NMR (400 MHz, methanol-d₄) δ 8.40 (d, J = 5.3 Hz, 1 H),

7.69 (s, 1 H), 7.59 (d, J = 5.3 Hz, 1 H), 1.73 (s, 3 H); one

exchangeable proton was not observed. MS m/z 225.9 [M + H]⁺

(226.0 calcd for C₈H₈ClF₃NO⁺).

1

88% yield) as a clear yellow oil. H NMR (400 MHz, CDCl₃) δ 9.05

(s, 1 H), 8.06 (d, J = 7.8 Hz, 1 H), 7.49−7.39 (m, 2 H), 7.38−7.31 (m,

1 H), 1.28 (s, 9 H). MS m/z 243.9 [M + H]⁺(244.0 calcd for

C₁₁H₁₅ClNOS⁺).

n-Butyllithium (2.6 mL of a 2.5 M solution with toluene, 6.4 mmol)

was added to a solution of 1-benzothiophene (0.79 g, 5.9 mmol) and

tetrahydrofuran (26 mL) at −78 °C. After 15 min, a solution of 34

(1.3 g, 5.3 mmol) and tetrahydrofuran (26 mL) was added. After 90

min, saturated aqueous ammonium chloride was added, the reaction

vessel was removed from the cooling bath, and the mixture was

allowed to warm to room temperature. The mixture was partitioned

between ethyl acetate and more saturated aqueous ammonium

chloride, the layers were separated, the organic material was washed

sequentially with saturated aqueous ammonium chloride and brine,

dried (sodium sulfate), ﬁltered, and the ﬁltrate was concentrated under

reduced pressure. The residue was subjected to ﬂash chromatography

on silica gel (gradient elution; 9:1 to 4:1 to 3:1 to 2:1 hexane−ethyl

acetate). The isolated material was resubjected to ﬂash chromatog-

raphy on silica gel (gradient elution; 3:1 to 2:1 hexane−ethyl acetate)

to give (S)-N-((R)-1-benzothiophen-2-yl(2-chlorophenyl)methyl)-2-

methyl-2-propanesulﬁnamide (35; 1.3 g, 65% yield) as a colorless

1

solid. H NMR (400 MHz, DMSO-d₆) δ 7.94−7.86 (m, 1 H), 7.83

(dd, J = 1.5, 7.7 Hz, 1 H), 7.79−7.70 (m, 1 H), 7.53−7.42 (m, 2 H),

7.42−7.35 (m, 1 H), 7.35−7.26 (m, 2 H), 7.02 (s, 1 H), 6.61 (d, J =

A mixture of 38 (1.1 g, 4.7 mmol), 2-(1-benzothiophen-7-yl)-

4,4,5,5-tetramethyl-1,3,2-dioxaborolane (39; 2.1 g, 8.1 mmol),⁴¹

K

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

dicyclohexyl(2′,4′,6′-tris(1-methylethyl)-2-biphenylyl)phosphane

(0.23 g, 0.47 mmol), chloro(2-propen-1-yl)palladium dimer (0.087 g,

0.24 mmol), sodium carbonate monohydrate (1.8 g, 14 mmol), 1,4-

dioxane (8.0 mL), and water (2.0 mL) was stirred at 80 °C for 7 h and

then at room temperature for 14 h. The reaction mixture was

concentrated under reduced pressure, and then the residue was

subjected to ﬂash chromatography on silica gel (gradient elution; 9:1

to 7:3 hexane−ethyl acetate). The isolated material was resubjected to

ﬂash chromatography on silica gel (gradient elution; 9:1 to 7:3

hexane−ethyl acetate) to give (2S)-2-(2-(1-benzothiophen-7-yl)-4-

pyridinyl)-1,1,1-triﬂuoro-2-propanol (40; 0.77 g, 50% yield) as a

colorless oil. ¹H NMR (300 MHz, CDCl₃) δ 8.85 (dd, J = 0.9, 5.3 Hz,

1 H), 8.20−8.16 (m, 1 H), 7.92 (d, J = 7.6 Hz, 2 H), 7.60−7.40 (m, 4

H), 2.72 (s, 1 H), 1.84 (dd, J = 0.3, 0.9 Hz, 3 H). MS m/z 324.0 [M +

H]⁺(324.1 calcd for C₁₆H₁₃F₃NOS⁺).

ethyl)-2-pyridinyl)-1-benzothiophen-2-yl)methyl)cyclopropane-

sulfonamide (25; 0.047 g, 76% yield; 70% yield, two steps) as a

1

colorless solid. [α]³⁰−40.4° (c 0.70, CHCl₃). H NMR (400 MHz,

D

CDCl₃) δ 8.83 (d, J = 5.3 Hz, 1 H), 8.15 (s, 1 H), 7.89 (d, J = 7.4 Hz,

1 H), 7.75 (d, J = 7.8 Hz, 1 H), 7.65 (dd, J = 1.4, 7.6 Hz, 1 H), 7.51−

7.41 (m, 3 H), 7.41−7.30 (m, 2 H), 7.08 (s, 1 H), 6.42 (d, J = 8.0 Hz,

1 H), 5.44 (d, J = 8.0 Hz, 1 H), 2.31−2.24 (m, 1 H), 1.85 (s, 3 H),

1.22−1.07 (m, 2 H), 0.92−0.83 (m, 1 H), 0.82−0.72 (m, 1 H); one

exchangeable proton was not observed. ¹³C NMR (150 MHz, CDCl₃)

δ 156.3, 148.6, 148.2, 146.8, 141.0, 138.0, 137.6, 133.0, 132.7, 130.3,

129.7, 129.2, 127.4, 125.1 (q, J = 285.6 Hz), 125.0, 124.8, 123.1, 122.4,

119.5, 118.1, 74.5 (q, J = 29.5 Hz), 55.7, 31.4, 23.8, 6.0, 5.9. HRMS m/

+

z 567.0784 [M + H]⁺(567.0785 calcd for C₂₆H₂₃ClF₃N₂O₃S₂).

N-((R)-(2-Amino-5-chloro-4-pyrimidinyl)(7-(4-(1-hydroxy-1-

methylethyl)-2-pyridinyl)-1-benzothiophen-2-yl)methyl)-

cyclopropanesulfonamide (31). A solution of 1,3,5-trichloroiso-

cyanuric acid (37 g, 160 mmol), 4-chloropyrimidin-2-amine (42; 42 g,

320 mmol), water (380 mL), and acetic acid (42 mL) was heated at 50

°C. After 15 h, the heating bath was removed and the reaction mixture

was allowed to cool to room temperature, added to crushed ice, and

then the mixture was basiﬁed with 10 M aqueous sodium hydroxide.

After 4 h, the mixture was ﬁltered, and the ﬁlter cake was washed with

water. The solid was suspended in 0.50 M aqueous sodium hydroxide,

and the resulting mixture was stirred for 2 h. The mixture was ﬁltered,

the ﬁlter cake was washed with water and then dried under reduced

pressure at elevated temperature (60 °C) to give 4,5-dichloro-2-

pyrimidinamine (38 g, 73% yield) as a yellow solid. The material was

used in the next step of the synthesis without puriﬁcation.

n-Butyllithium (3.2 mL of a 1.6 M solution with hexanes, 5.1 mmol)

was added to a solution of 40 (0.76 g, 2.3 mmol) and tetrahydrofuran

(2.0 mL) at −78 °C. After 30 min, a solution of 34 (0.74 g, 3.0 mmol)

and tetrahydrofuran (1.0 mL) at −78 °C was added dropwise via a

cannula. After 3 h, saturated aqueous ammonium chloride was added,

the cooling bath was removed, the mixture was allowed to warm to

room temperature, partitioned between more aqueous ammonium

chloride and ethyl acetate, the layers were separated, the aqueous

material was washed with ethyl acetate (2×), the combined organic

material was dried (magnesium sulfate), ﬁltered, and the ﬁltrate was

concentrated under reduced pressure. The residue was subjected to

ﬂash chromatography on silica gel (gradient elution; 9:1 to 2:3

hexane−ethyl acetate). The isolated material was resubjected to ﬂash

chromatography on silica gel (gradient elution; 7:3 to 2:3 hexane−

ethyl acetate) to provide (S)-N-((R)-(2-chlorophenyl)(7-(4-((1S)-

2,2,2-triﬂuoro-1-hydroxy-1-methylethyl)-2-pyridinyl)-1-benzothio-

phen-2-yl)methyl)-2-methyl-2-propanesulﬁnamide (41; 0.67 g, 51%

A solution of 4,5-dichloro-2-pyrimidinamine (38 g, 230 mmol),

tetrahydrofuran (750 mL), and 4-dimethylaminopyridine (1.4 g, 12

mmol) was treated dropwise with a solution of di-tert-butyl

dicarbonate (120 g, 540 mmol) and tetrahydrofuran (250 mL).

After 60 h, the reaction mixture was partitioned between ethyl acetate

and saturated aqueous ammonium chloride, the layers were separated,

the organic material was washed sequentially with water and brine,

dried (magnesium sulfate), ﬁltered, and the ﬁltrate was concentrated

under reduced pressure. The isolated solid was recrystallized (5:1

hexanes−diethyl ether) to give di-tert-butyl (4,5-dichloro-2-

pyrimidinyl)imidodicarbonate (43; 62 g, 53% yield, two steps) as a

yellow solid. ¹H NMR (400 MHz, CDCl₃) δ 8.67 (s, 1 H), 1.48 (s, 18

1

yield) as a light yellow solid. H NMR (400 MHz, CDCl₃) δ 8.89−

8.66 (m, 1 H), 8.14 (s, 1 H), 7.86 (d, J = 7.4 Hz, 1 H), 7.75 (d, J = 7.8

Hz, 1 H), 7.65 (d, J = 7.6 Hz, 1 H), 7.49−7.41 (m, 2 H), 7.38 (d, J =

7.8 Hz, 1 H), 7.34−7.28 (m, 1 H), 7.27−7.19 (m, 2 H), 6.36 (d, J =

5.1 Hz, 1 H), 4.14 (d, J = 5.3 Hz, 1 H), 3.15 (br s, 1 H), 1.81 (s, 3 H),

1.31 (s, 9 H). MS m/z 566.7 [M + H]⁺(567.1 calcd for

+

C₂₇H₂₇ClF₃N₂O₂S₂).

Hydrogen chloride (2.9 mL of a 2.0 M solution with diethyl ether,

5.8 mmol) was added to a solution of 41 (0.66 g, 1.2 mmol) and

diethyl ether (8.0 mL). After 6 h, the reaction mixture was ﬁltered, and

then the ﬁlter cake was washed with diethyl ether and dried under

reduced pressure. The solid was dissolved with 2.0 M ammonia in

methanol and dichloromethane, silica gel was added, and the volatiles

were removed under reduced pressure. The residue was subjected to

ﬂash chromatography on silica gel (gradient elution; 7:3 to 3:7

hexane−ethyl acetate) to yield (2S)-2-(2-(2-((R)-amino(2-

chlorophenyl)methyl)-1-benzothiophen-7-yl)-4-pyridinyl)-1,1,1-tri-

+

H). MS m/z 385.8 [M + Na]⁺(386.1 calcd for C₁₄H₁₉Cl₂N₃NaO₄).

A solution of 43 (32 g, 87 mmol), 1,1-bis[(di-tert-butyl-p-

methylaminophenyl]palladium(II) chloride (2.5 g, 3.5 mmol),

potassium vinyltriﬂuoroborate (15 g, 110 mmol), and potassium

acetate (26 g, 260 mmol), acetonitrile (300 mL), and water (100 mL)

was purged with nitrogen for 5 min and then heated at 70 °C. After 3

h, the heating bath was removed and the reaction mixture was allowed

to cool to room temperature, partitioned between ethyl acetate and

water, the layers were separated, the organic material was washed

sequentially with 1.0 N aqueous HCl and brine, dried (magnesium

sulfate), ﬁltered, and the ﬁltrate was concentrated under reduced

pressure. The isolated material was subjected to ﬂash chromatography

on silica gel (gradient elution; 19:1 to 3:1 hexane−ethyl acetate) to

aﬀord di-tert-butyl (5-chloro-4-ethenyl-2-pyrimidinyl)-

1

ﬂuoro-2-propanol (0.51 g, 92% yield) as a colorless solid. H NMR

(400 MHz, CDCl₃) δ 8.82 (d, J = 5.3 Hz, 1 H), 8.13 (s, 1 H), 7.88−

7.80 (m, 1 H), 7.77−7.70 (m, 1 H), 7.62 (dd, J = 1.7, 7.7 Hz, 1 H),

7.48−7.41 (m, 2 H), 7.38 (dd, J = 1.4, 7.8 Hz, 1 H), 7.31−7.26 (m, 1

H), 7.24−7.19 (m, 1 H), 7.17 (d, J = 0.8 Hz, 1 H), 5.96 (s, 1 H), 1.83

(s, 3 H); two exchangeable protons were not observed. MS m/z 462.8

[M + H]⁺(463.1 calcd for C₂₃H₁₉ClF₃N₂OS⁺).

4-Dimethylaminopyridine (0.0013 g, 0.011 mmol) was added to a

mixture of (2S)-2-(2-(2-((R)-amino(2-chlorophenyl)methyl)-1-benzo-

thiophen-7-yl)-4-pyridinyl)-1,1,1-triﬂuoro-2-propanol (0.049 g, 0.11

mmol), cyclopropanesulfonyl chloride (13 μL, 0.13 mmol),

diisopropylethylamine (55 μL, 0.32 mmol), and N,N-dimethylforma-

mide (0.70 mL). After 24 h, the mixture was diluted with methanol

(1.0 mL), the solution was ﬁltered, and the ﬁltrate was subjected to

reverse-phase HPLC (Phenomenex Gemini-NX C18 110 Å, 100 mm

× 50 mm, 10 μm, 10−95% H₂O/MeCN with 0.1% TFA). The

product-containing fractions were combined, neutralized with solid

sodium bicarbonate, and extracted with dichloromethane (2×). The

combined organic material was dried (magnesium sulfate), ﬁltered,

and the ﬁltrate was concentrated under reduced pressure to give N-

((R)-(2-chlorophenyl)(7-(4-((1S)-2,2,2-triﬂuoro-1-hydroxy-1-methyl-

1

imidodicarbonate (44; 24 g, 77% yield) as a yellow solid. H NMR

(400 MHz, CDCl₃) δ 8.65 (s, 1 H), 7.15 (dd, J = 10.6, 16.8 Hz, 1 H),

6.76 (dd, J = 1.8, 16.8 Hz, 1 H), 5.84 (dd, J = 1.8, 10.6 Hz, 1 H), 1.49

(s, 18 H). MS m/z 378.1 [M + Na]⁺(378.1 calcd for

+

C₁₆H₂₂ClN₃NaO₄).

Potassium osmate dihydrate (0.099 g, 0.27 mmol) was added to a

solution of 44 (24 g, 67 mmol), sodium periodate (40 g, 190 mmol),

tetrahydrofuran (400 mL), and water (100 mL) at room temperature.

After 18 h, more potassium osmate dihydrate (0.050 g, 0.14 mmol)

was added. After 4 h, Na₂SO₃(30 g) and water (400 mL) were added.

After 45 min, ethyl acetate was added, the layers were separated, the

aqueous material was washed with ethyl acetate, the combined organic

material was washed sequentially with water and brine, dried

(magnesium sulfate), ﬁltered, and the ﬁltrate was concentrated

under reduced pressure to aﬀord di-tert-butyl (5-chloro-4-formyl-2-

L

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

pyrimidinyl)imidodicarbonate (24 g, quant yield) as a thick oil. The

material was used in the next step of the synthesis without puriﬁcation.

A mixture of di-tert-butyl (5-chloro-4-formyl-2-pyrimidinyl)-

imidodicarbonate (24 g, 67 mmol), (S)-2-methyl-2-propanesulﬁna-

mide (9.8 g, 80 mmol), copper(II) sulfate (32 g, 200 mmol), and

dichloromethane (200 mL) was stirred at room temperature. After 15

h, the reaction mixture was ﬁltered through a plug of diatomaceous

earth, the ﬁlter cake was washed with dichloromethane, and the ﬁltrate

was concentrated under reduced pressure. The residue was subjected

to ﬂash chromatography on silica gel (gradient elution; 19:1 to 7:3

hexane−ethyl acetate) to aﬀord di-tert-butyl (4-((E)-(((S)-tert-

butylsulﬁnyl)imino)methyl)-5-chloro-2-pyrimidinyl)imidodicarbonate

(45; 18 g, 58% yield, two steps) as a yellow solid. ¹H NMR (400 MHz,

CDCl₃) δ 8.85 (s, 1 H), 8.83 (s, 1 H), 1.47 (s, 18 H), 1.31 (s, 9 H).

MS m/z 483.0 [M + Na]⁺(483.2 calcd for C₁₉H₂₉ClN₄NaO₅S⁺).

n-Butyllithium (8.4 mL of a 2.5 M solution with toluene, 21 mmol)

was added to a solution of 2-chloro-4-iodopyridine (36; 5.0 g, 21

mmol) and tetrahydrofuran (50 mL) at −78 °C. After 15 min, acetone

(4.6 mL, 63 mmol) was added. After 1 h, saturated aqueous

ammonium chloride was added, the cooling bath was removed, the

mixture was allowed to warm to room temperature, partitioned

between more saturated aqueous ammonium chloride and ethyl

acetate, the layers were separated, the aqueous material was washed

with ethyl acetate (2×), the combined organic material was washed

sequentially with water and brine, dried (sodium sulfate), ﬁltered, and

the ﬁltrate was concentrated under reduced pressure. The residue was

subjected to ﬂash chromatography on silica gel (6:1 hexane−ethyl

acetate) to give 2-(2-chloro-4-pyridinyl)-2-propanol (46; 2.2 g, 61%

yield) as a colorless solid. ¹H NMR (400 MHz, DMSO-d₆) δ 8.33 (d, J

= 5.2 Hz, 1 H), 7.53 (s, 1 H), 7.47 (dd, J = 1.2, 4.8 Hz, 1 H), 5.42 (s, 1

H), 1.41 (s, 6 H). MS m/z 172.1 [M + H]⁺(172.0 calcd for

C₈H₁₁ClNO⁺).

(200 mL of a 0.50 M solution with tetrahydrofuran, 99 mmol) at −78

°C. After 30 min, the mixture was added via cannula over 10 min to a

solution of 45 (23 g, 50 mmol) and tetrahydrofuran (200 mL) at −78

°C. After 20 min, saturated aqueous ammonium chloride was added

slowly, the mixture was partitioned between ethyl acetate and 5.0%

aqueous sodium bicarbonate, the layers were separated, the aqueous

material was washed with ethyl acetate, the combined organic material

was washed with brine (2×), dried (magnesium sulfate), ﬁltered, and

the ﬁltrate was concentrated under reduced pressure to aﬀord a ∼1:1

diastereomeric mixture of di-tert-butyl (4-((((S)-tert-butylsulﬁnyl)-

amino)(7-(4-(1-methyl-1-((triethylsilyl)oxy)ethyl)-2-pyridinyl)-1-ben-

zothiophen-2-yl)methyl)-5-chloro-2-pyrimidinyl)imidodicarbonate

(49; 42 g, quant yield) as an amber oil. The material was used in the

next step of the synthesis without puriﬁcation.

Hydrogen chloride (200 mL of a 4.0 M solution with 1,4-dioxane,

790 mmol) was added to a stirring suspension of 49 (42 g, 50 mmol)

and methanol (200 mL) at 10 °C, and then the cooling bath was

removed and the reaction mixture was allowed to warm to room

temperature. After 72 h, diethyl ether was added over a 90 min period.

The mixture was ﬁltered and the ﬁlter cake was washed with diethyl

ether to aﬀord ∼1:1 enantiomeric mixture of 2-(2-(2-(amino(2-amino-

5-chloro-4-pyrimidinyl)methyl)-1-benzothiophen-7-yl)-4-pyridinyl)-2-

propanol trihydrochloride (17 g, 65% yield) as a pale yellow solid. The

material was used in the next step of the synthesis without puriﬁcation.

Cyclopropanesulfonyl chloride (4.9 mL, 49 mmol) was added to a

stirring solution of a ∼1:1 enantiomeric mixture of 2-(2-(2-(amino(2-

amino-5-chloro-4-pyrimidinyl)methyl)-1-benzothiophen-7-yl)-4-pyri-

dinyl)-2-propanol trihydrochloride (17 g, 32 mmol), 4-dimethylami-

nopyridine (0.40 g, 3.2 mmol), diisopropylethylamine (23 mL, 130

mmol), and N,N-dimethylformamide (100 mL) at 13 °C, and then the

cooling bath was removed and the reaction mixture was allowed to

warm to room temperature. After 18 h, the reaction mixture was

partitioned between ethyl acetate and 5.0% aqueous sodium

bicarbonate, the layers were separated, the organic material was

washed sequentially with water (2×) and brine, dried (magnesium

sulfate), ﬁltered, and the ﬁltrate was concentrated under reduced

pressure. The residue was suspended in 9:1 dichloromethane−

isopropanol, the suspension was ﬁltered, and then the ﬁlter cake was

washed with diethyl ether and collected to aﬀord the desired product

(2.1 g) as white solid. Silica gel was added to the ﬁltrate, the volatiles

were removed under reduced pressure, and the residue was subjected

to ﬂash chromatography on silica gel (gradient elution; 49:1 to 19:1

dichloromethane−2.0 M NH₃in methanol). The isolated materials

were combined to aﬀord a ∼1:1 enantiomeric mixture of N-((2-amino-

5-chloro-4-pyrimidinyl)(7-(4-(1-hydroxy-1-methylethyl)-2-pyridinyl)-

1-benzothiophen-2-yl)methyl)cyclopropanesulfonamide (13 g, 49%

yield, three steps) as a tan solid. The combined material was subjected

to preparative chiral SFC (Chiralpak AD-H column; 25 cm × 30 mm,

5 μm) eluting with 60% liquid CO₂in 40% isopropanol (with 20 mM

NH₃) at a ﬂow rate of 120 mL/min) to aﬀord two products in >99%

enantiomeric excess. First eluting peak: N-((S)-(2-amino-5-chloro-4-

pyrimidinyl)(7-(4-(1-hydroxy-1-methylethyl)-2-pyridinyl)-1-benzo-

thiophen-2-yl)methyl)cyclopropanesulfonamide. Second eluting peak:

N-((R)-(2-amino-5-chloro-4-pyrimidinyl)(7-(4-(1-hydroxy-1-methyl-

ethyl)-2-pyridinyl)-1-benzothiophen-2-yl)methyl)cyclopropane-

A stirring mixture of 46 (24 g, 0.14 mol), 2-(1-benzothiophen-7-yl)-

4,4,5,5-tetramethyl-1,3,2-dioxaborolane (39; 40 g, 0.15 mol), 1,4-

dioxane (190 mL), water (30 mL), allylpalladium(II) chloride dimer

(2.5 g, 7.0 mmol), 2-(dicyclohexylphosphino)-2′,4′,6′,-triisopropyl-

1,1′-biphenyl (6.7 g, 14 mmol), and sodium carbonate (45 g, 0.042

mol) was heated at 80 °C. After 12 h, the heating bath was removed,

the reaction mixture was allowed to cool to room temperature, ﬁltered

through a pad of diatomaceous earth, and the ﬁltrate was concentrated

under reduced pressure. The residue was partitioned between water

and ethyl acetate, the layers were separated, the aqueous material was

washed with ethyl acetate (2×), the combined organic extract was

washed sequentially with water and brine, dried (sodium sulfate),

ﬁltered, and the ﬁltrate was concentrated under reduced pressure. The

residue was subjected to ﬂash chromatography on silica gel (9:1

hexane−ethyl acetate) to give 2-(2-(1-benzothiophen-7-yl)-4-pyridin-

1

yl)-2-propanol (47; 20 g, 53% yield) as a pale yellow solid. H NMR

(400 MHz, DMSO-d₆) δ 8.71 (d, J = 5.2 Hz, 1 H), 8.23 (s, 1 H), 8.12

(d, J = 7.6 Hz, 1 H), 7.99 (d, J = 8.0 Hz, 1 H), 7.82 (d, J = 5.6 Hz, 1

H), 7.60−7.47 (m, 3 H), 5.39 (s, 1 H), 1.51 (s, 6 H). MS m/z 270.0

[M + H]⁺(270.1 calcd for C₁₆H₁₆NOS⁺).

A solution of triethylsilyl triﬂuoromethanesulfonate (32 mL, 140

mmol) and tetrahydrofuran (50 mL) was added to a solution of 47 (30

g, 110 mmol), diisopropylethylamine (29 mL, 170 mmol), and

tetrahydrofuran (100 mL) at 0 °C. After 3 h, the reaction mixture was

partitioned between diethyl ether and 1.0 M aqueous KH₂PO₄, the

layers were separated, the organic material was washed with brine,

dried (magnesium sulfate), ﬁltered, and the ﬁltrate was concentrated

under reduced pressure. The residue was subjected to ﬂash

chromatography on silica gel (gradient elution; hexane to 9:1

hexane−ethyl acetate) to aﬀord 2-(1-benzothiophen-7-yl)-4-(1-meth-

yl-1-((triethylsilyl)oxy)ethyl)pyridine (48; 37 g, 88% yield) as a

colorless oil. ¹H NMR (400 MHz, CDCl₃) δ 8.77−8.72 (m, 1 H), 8.08

(d, J = 0.8 Hz, 1 H), 7.89 (d, J = 7.6 Hz, 2 H), 7.56 (d, J = 5.7 Hz, 1

H), 7.50 (t, J = 7.6 Hz, 1 H), 7.41 (d, J = 5.7 Hz, 1 H), 7.34 (dd, J =

1.7, 5.2 Hz, 1 H), 1.63 (s, 6 H), 1.00 (t, J = 7.9 Hz, 9 H), 0.71−0.64

(m, 6 H). MS m/z 384.1 [M + H]⁺(384.2 calcd for C₂₂H₃₀NOSSi⁺).

n-Butyllithium (22 mL of a 2.5 M solution with hexanes, 55 mmol)

was added to a solution of 48 (19 g, 50 mmol) and lithium chloride

sulfonamide (31)⁴²as a tan solid. [α]²⁰+48.3° (c 0.39, MeOH). H

1

D

NMR (400 MHz, DMSO-d₆) δ 8.71 (d, J = 5.1 Hz, 1 H) 8.34 (s, 1 H)

8.21 (s, 1 H) 8.10 (d, J = 8.6 Hz, 2 H) 7.94−7.85 (m, 1 H) 7.58−7.44

(m, 2 H) 7.33 (d, J = 1.0 Hz, 1 H) 6.98 (br s, 2 H) 6.09 (dd, J = 0.9,

8.9 Hz, 1 H) 5.35 (br s, 1 H) 2.45−2.33 (m, 1 H) 1.50 (s, 6 H) 0.97−

0.79 (m, 3 H) 0.78−0.68 (m, 1 H). ¹³C NMR (100 MHz, DMSO-d₆)

δ 163.0, 162.0, 160.4, 158.0, 154.6, 148.0, 145.0, 140.5, 136.4, 132.6,

124.8, 124.8, 122.7, 122.3, 119.0, 116.6, 114.8, 70.6, 54.0, 31.2, 30.8,

5.3, 4.9. HRMS m/z 530.1086 [M + H]⁺(530.1087 calcd for

+

C₂₄H₂₅ClN₅O₃S₂).

N-((R)-(2-Amino-5-chloro-3-ﬂuoro-4-pyridinyl)(7-(4-(1-hy-

droxy-1-methylethyl)-2-pyridinyl)-1-benzothiophen-2-yl)-

methyl)cyclopropanesulfonamide (32). n-Butyllithium (23 mL of

a 2.5 M solution with toluene, 57 mmol) was added to a solution of 2-

(1-benzothiophen-7-yl)-4-(1-methyl-1-((triethylsilyl)oxy)ethyl)-

M

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

pyridine (48; 22 g, 57 mmol) and tetrahydrofuran (120 mL) at −78

°C. After 30 min, N,N-dimethylformamide (22 mL, 290 mmol) was

added. After 40 min, saturated aqueous sodium bicarbonate and ethyl

acetate were added, the layers were separated, the organic material was

washed sequentially with saturated aqueous sodium bicarbonate and

brine, dried (sodium sulfate), ﬁltered, and the ﬁltrate was concentrated

under reduced pressure to give 7-(4-(1-methyl-1-((triethylsilyl)oxy)-

ethyl)-2-pyridinyl)-1-benzothiophene-2-carbaldehyde (24 g, quant

yield) as an oﬀ-white solid. The material was used in the next step

of the synthesis without puriﬁcation.

Titanium tetraethoxide (61 mL, 290 mmol) was added to a stirring

solution of 7-(4-(1-methyl-1-((triethylsilyl)oxy)ethyl)-2-pyridinyl)-1-

benzothiophene-2-carbaldehyde (24 g, 58 mmol), (R)-2-methyl-2-

propanesulﬁnamide (7.1 g, 58 mmol), and dichloromethane (120 mL)

at room temperature. After 13 h, the reaction mixture was partitioned

between water and dichloromethane, the layers were separated, the

aqueous material was washed with dichloromethane, the combined

organic material was dried (sodium sulfate), ﬁltered, and the ﬁltrate

was concentrated. The residue was subjected to ﬂash chromatography

on silica gel (5:1 hexane−ethyl acetate) to give (R)-2-methyl-N-((1E)-

(7-(4-(1-methyl-1-((triethylsilyl)oxy)ethyl)-2-pyridinyl)-1-benzothio-

phen-2-yl)methylidene)-2-propanesulﬁnamide (50; 26 g, 87% yield,

two steps) as a yellow solid. ¹H NMR (400 MHz, CDCl₃) δ 8.85 (s, 1

H), 8.81−8.73 (m, 1 H), 8.08 (s, 1 H), 8.02−7.96 (m, 1 H), 7.96−7.90

(m, 1 H), 7.84 (s, 1 H), 7.54 (t, J = 7.7 Hz, 1 H), 7.35 (dd, J = 1.7, 5.2

Hz, 1 H), 1.63 (s, 6 H), 1.30 (s, 9 H), 1.00 (t, J = 7.9 Hz, 9 H), 0.67

(q, J = 7.9 Hz, 6 H). MS m/z 515.1 [M + H]⁺(515.2 calcd for

C₂₇H₃₉N₂O₂S₂Si⁺).

stirred for 20 min, partitioned between more ethyl acetate and

saturated aqueous sodium bicarbonate, the layers were separated, the

organic material was washed sequentially with saturated aqueous

sodium bicarbonate (4×) and brine (2×), dried (sodium sulfate),

ﬁltered, and the ﬁltrate was concentrated under reduced pressure. The

residue was subjected to ﬂash chromatography on silica gel (49:1

dichloromethane−methanol) to give N-((R)-(2-amino-5-chloro-3-

ﬂuoro-4-pyridinyl)(7-(4-(1-hydroxy-1-methylethyl)-2-pyridinyl)-1-

benzothiophen-2-yl)methyl)cyclopropanesulfonamide (32; 2.0 g, 83%

1

yield) as an oﬀ-white solid. [α]³⁰+27.6° (c 1.42, CHCl₃). H NMR

D

(400 MHz, DMSO-d₆) δ 8.70 (d, J = 5.1 Hz, 1 H), 8.67 (d, J = 8.8 Hz,

1 H), 8.22 (s, 1 H), 8.11 (d, J = 7.2 Hz, 1 H), 7.92 (s, 1 H), 7.90 (d, J

= 7.4 Hz, 1 H), 7.53 (t, J = 7.6 Hz, 1 H), 7.48 (dd, J = 1.5, 5.2 Hz, 1

H), 7.36−7.29 (m, 1 H), 6.56 (s, 2 H), 6.31 (d, J = 8.8 Hz, 1 H), 5.35

(s, 1 H), 2.48−2.39 (m, 1 H), 1.50 (s, 6 H), 1.02−0.76 (m, 4 H). ¹³C

NMR (125 MHz, DMSO-d₆) δ 160.3, 154.4, 149.2 (d, J = 12.7 Hz),

147.9, 145.1, 143.7 (d, J = 260.7 Hz), 141.8 (br d, J = 6.4 Hz), 140.5,

136.1, 132.4, 132.4 (br d, J = 10.0 Hz), 124.8, 124.7, 122.5, 121.3,

118.9, 116.5, 115.0, 70.5, 50.9, 31.1, 31.1, 30.7, 5.0, 5.0. HRMS m/z

+

547.1034 [M + H]⁺(547.1035 calcd for C₂₅H₂₅ClFN₄O₃S₂).

ASSOCIATED CONTENT

* Supporting Information

■

S

The Supporting Information is available free of charge on the

ACS Publications website at DOI: 10.1021/acs.jmed-

chem.5b01367.

Experimental procedures and characterization data for

compounds 7, 9, 11, 13−24, and 26−30,³⁸absolute

stereochemistry determinations for compound 31 and

intermediate 38, energetic parameters from the quantum

mechanical analysis of the model systems in Figure 4, as

well as statistical analyses of compound activities in key

biological assays (PDF)

n-Butyllithium (32 mL of a 2.5 M solution with toluene, 80 mmol)

was added slowly to a solution of 5-chloro-3-ﬂuoro-2-pyridinamine

(6.3 g, 43 mmol) and tetrahydrofuran (80 mL) at −78 °C. After 2 h, a

solution of 50 (10 g, 19 mmol) and tetrahydrofuran (17 mL) was

added slowly. After 80 min, saturated aqueous sodium bicarbonate was

added, the cooling bath was removed, the mixture was allowed to

warm to room temperature, partitioned between ethyl acetate and

more saturated aqueous sodium bicarbonate, the layers were separated,

the organic material was washed with brine, dried (sodium sulfate),

ﬁltered, and the ﬁltrate was concentrated under reduced pressure. The

residue was subjected to ﬂash chromatography on silica gel (gradient

elution; 5:1 to 3:1 hexane−acetone). The isolated material was

resubjected to ﬂash chromatography on silica gel (1:1 hexane−ethyl

acetate) to give (R)-N-((R)-(2-amino-5-chloro-3-ﬂuoro-4-pyridinyl)-

(7-(4-(1-methyl-1-((triethylsilyl)oxy)ethyl)-2-pyridinyl)-1-benzothio-

phen-2-yl)methyl)-2-methyl-2-propanesulﬁnamide (51; 2.9 g, 23%

Molecular formula strings (CSV)

Accession Codes

Atomic coordinates for the X-ray crystal structures of

compounds 6, 12, 26, and 32 bound to hGKRP (2.2, 2.4,

2.6, and 2.2 Å resolution for 6, 12, 26, and 32, respectively)

have been deposited in the RCSB (PDB codes 4PX2, 4PX3,

4PXS, and 4PX5 for 6, 12, 26, and 32, respectively).

1

yield) as a pale yellow solid. H NMR (400 MHz, DMSO-d₆) δ 8.70

AUTHOR INFORMATION

Corresponding Authors

*L.D.P.: phone, +1-781-609-6868; e-mail, lewis.pennington@

alkermes.com.

■

(d, J = 5.3 Hz, 1 H), 8.17 (s, 1 H), 8.04 (d, J = 6.9 Hz, 1 H), 7.93 (d, J

= 7.2 Hz, 1 H), 7.89 (s, 1 H), 7.54 (t, J = 7.6 Hz, 1 H), 7.50−7.43 (m,

1 H), 6.51 (s, 2 H), 6.38 (d, J = 7.4 Hz, 1 H), 6.22 (d, J = 7.6 Hz, 1 H),

1.61 (s, 6 H), 1.16 (s, 9 H), 0.99−0.90 (m, 9 H), 0.68−0.55 (m, 6 H);

one exchangeable proton was not observed. MS m/z 661.0 [M + H]⁺

(661.2 calcd for C₃₂H₄₃ClFN₄O₂S₂Si⁺).

*M.D.B.: phone, +1 805-447-3446; e-mail, mbartber@amgen.

com.

Hydrogen chloride (11 mL of a 4.0 M solution with 1,4-dioxane, 44

mmol) was added to a solution of 51 (2.9 g, 4.4 mmol) and methanol

(44 mL) at room temperature. After 20 min, the reaction mixture was

concentrated under reduced pressure, the residue was partitioned

between ethyl acetate and saturated aqueous sodium bicarbonate, the

layers were separated, the organic material was washed sequentially

with saturated aqueous sodium bicarbonate and brine, dried (sodium

sulfate), ﬁltered, and the ﬁltrate was concentrated under reduced

pressure to give 2-(2-(2-((R)-amino(2-amino-5-chloro-3-ﬂuoro-4-

pyridinyl)methyl)-1-benzothiophen-7-yl)-4-pyridinyl)-2-propanol (1.9

g, quant yield) as a pale yellow solid. The material was used in the next

step of the synthesis without puriﬁcation.

4-Dimethylaminopyridine (0.052 g, 0.43 mmol) was added to a

solution of 2-(2-(2-((R)-amino(2-amino-5-chloro-3-ﬂuoro-4-

pyridinyl)methyl)-1-benzothiophen-7-yl)-4-pyridinyl)-2-propanol (1.9

g, 4.3 mmol), N,N-dimethylformamide (4.3 mL), diisopropylethyl-

amine (2.2 mL, 13 mmol), and cyclopropanesulfonyl chloride (0.66

mL, 6.4 mmol) at room temperature. After 21 h, saturated aqueous

sodium bicarbonate and ethyl acetate were added, the mixture was

Present Address

^#L.D.P.: Alkermes, 852 Winter Street, Waltham, MA 02451.

Notes

The authors declare the following competing ﬁnancial

interest(s): All authors are current or former employees of

Amgen.

ACKNOWLEDGMENTS

■

We thank Christopher Wilde (Molecular Structure and

Characterization, Amgen) and Iain Campuzano (Molecular

Structure and Characterization, Amgen) for conﬁrmation of the

structure assignments of 12, 25, 31, and 32 using NMR and

HRMS methods, respectively; Kyung-Hyun Gahm, Samuel

Thomas, Wesley Barnhart, and Zheng Hua (Molecular

Structure and Characterization, Amgen) for conducting

compound separations by chiral SFC; Robert Kurzeja (Protein

N

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

B.; Doliba, N.; Li, C.; Vanderkooi, J. M.; Naji, A.; Sarabu, R.; Grimsby,

J. Glucokinase activators for diabetes therapy (May 2010 status

report). Diabetes Care 2011, 34 (Suppl. 2), S236−S243.

(7) (a) Rees, M. G.; Gloyn, A. L. Small molecular glucokinase

activators: has another new anti-diabetic therapeutic lost favor? Br. J.

Pharmacol. 2013, 168, 335−338. (b) Matschinsky, F. M. GKAs for

diabetes therapy: why no clinically useful drug after two decades of

trying? Trends Pharmacol. Sci. 2013, 34, 90−99.

(8) Filipski, K. J.; Pfefferkorn, J. A. A patent review of glucokinase

activators and disruptors of the glucokinase − glucokinase regulatory

protein interaction: 2011−2014. Expert Opin. Ther. Pat. 2014, 24,

875−891.

(9) Anderka, O.; Boyken, J.; Aschenbach, U.; Batzer, A.; Boscheinen,

O.; Schmoll, D. Biophysical characterization of the interaction between

hepatic glucokinase and its regulatory protein: impact of physiological

and pharmacological effectors. J. Biol. Chem. 2008, 283, 31333−31340.

(10) Vandercammen, A.; Van Schaftingen, E. The mechanism by

which rat liver glucokinase is inhibited by the regulatory protein. Eur. J.

Biochem. 1990, 191, 483−489.

(11) Lloyd, D. J.; St. Jean, D. J., Jr.; Kurzeja, R. J. M.; Wahl, R. C.;

Michelsen, K.; Cupples, R.; Chen, M.; Wu, J.; Sivits, G.; Helmering, J.;

Ashton, K. S.; Pennington, L. D.; Fotsch, C. H.; Vazir, M.; Chen, K.;

Chmait, S.; Zhang, J.; Liu, L.; Norman, M. H.; Andrews, K. A.;

Bartberger, M. D.; Van, G.; Galbreath, E. J.; Vonderfecht, S. L.; Wang,

M.; Jordan, S. R.; Veniant, M. M.; Hale, C. Antidiabetic effects of

glucokinase regulatory protein small-molecule disruptors. Nature 2013,

504, 437−440.

(12) Ashton, K. S.; Andrews, K. L.; Bryan, M. C.; Chen, J.; Chen, K.;

Chen, M.; Chmait, S.; Croghan, M.; Cupples, R.; Fotsch, C.;

Helmering, J.; Jordan, S. R.; Kurzeja, R. J. M.; Michelsen, K.;

Pennington, L. D.; Poon, S. F.; Sivits, G.; Van, G.; Vonderfecht, S. L.;

Wahl, R. C.; Zhang, J.; Lloyd, D. J.; Hale, C.; St. Jean, D. J., Jr. Small

molecule disruptors of the glucokinase−glucokinase regulatory protein

interaction: 1. Discovery of a novel tool compound for in vivo proof-

of-concept. J. Med. Chem. 2014, 57, 309−324.

Technologies, Amgen) for providing puriﬁed proteins; Robert

Wahl and Kui Chen (Molecular Structure and Characterization,

Amgen) for developing biochemical assays; David Cronk,

Andrew Southan, Scott Pollack, and Kevin Nash (Compound

Libraries and Assay Development and Screening, BioFocus) for

executing the HTS campaign; Ramesh Senaiar and co-workers

(Medicinal Chemistry, Aurigene) for preparing synthetic

́

intermediates; and Scott Simonet and Murielle Veniant

(Metabolic Disorders Research, Amgen), Dean Hickman

(Pharmacokinetics and Metabolism, Amgen), and Philip Tagari

(Therapeutic Discovery, Amgen) for endorsing this research.

The Advanced Light Source for X-ray analyses was supported

by the U.S. Department of Energy under Contract DE-AC02-

05CH11231. All live animal studies were conducted in an

AAALAC-accredited facility, and husbandry procedures met all

of the recommendations of the Guide for the Care and Use of

Laboratory Animals. All work using research animals was

conducted under Institutional Animal Care and Use

Committee (IACUC) approved protocols.

ABBREVIATIONS USED

■

T2DM, type II diabetes mellitus; GK, glucokinase; GKRP,

glucokinase regulatory protein; G6P, glucose 6-phosphate; TL,

translocation; RLM, rat liver microsome; CL_int, intrinsic

clearance; cLogP, calculated log P (logarithm of the octanol/

water partition coeﬃcient) using the Daylight Chemical

Information Systems algorithm; ACDLogP, calculated log P

using the Advanced Chemistry Development (ACD/Labs)

algorithm; ACDLogD_7.4, calculated log D using the Advanced

Chemistry Development (ACD/Labs) algorithm; K_d, dissoci-

ation constant; P_app, apparent permeability; HPMC, hydrox-

ypropyl methylcellulose; PPB, plasma protein binding; CL,

clearance of compound from plasma; V_dss, volume of

distribution at steady state; t_1/2, half-life of compound in

plasma; C_max, maximum compound concentration in plasma;

(13) St. Jean, D. J., Jr.; Ashton, K. S.; Bartberger, M. D.; Chen, J.;

Chmait, S.; Cupples, R.; Galbreath, E.; Helmering, J.; Hong, F.-T.;

Jordan, S. R.; Liu, L.; Kunz, R. K.; Michelsen, K.; Nishimura, N.;

Pennington, L. D.; Poon, S. F.; Reid, D.; Sivits, G.; Stec, M. M.;

Tadesse, S.; Tamayo, N.; Van, G.; Yang, K.; Zhang, J.; Norman, M. H.;

Fotsch, C.; Lloyd, D. J.; Hale, C. Small molecule disruptors of the

glucokinase−glucokinase regulatory protein interaction: 2. Leveraging

structure-based drug design to identify analogs with improved

pharmacokinetic profiles. J. Med. Chem. 2014, 57, 325−338.

(14) Structurally related variations of the sulfonylpiperazine scaﬀold

present in 3−5 have been subsequently reported by our laboratories:

(a) Nishimura, N.; Norman, M. H.; Liu, L.; Yang, K. C.; Ashton, K. S.;

Bartberger, M. D.; Chmait, S.; Chen, J.; Cupples, R.; Fotsch, C.;

Helmering, J.; Jordan, S. R.; Kunz, R. K.; Pennington, L. D.; Poon, S.

F.; Siegmund, A.; Sivits, G.; Lloyd, D. J.; Hale, C.; St. Jean, D. J., Jr.

Small molecule disruptors of the glucokinase−glucokinase regulatory

protein interaction: 3. Structure−activity relationships within the aryl

carbinol region of the N-arylsulfonamido-N′-aryl-piperazine series. J.

Med. Chem. 2014, 57, 3094−3116. (b) Hong, F.-T.; Norman, M. H.;

Ashton, K. S.; Bartberger, M. D.; Chen, J.; Chmait, S.; Cupples, R.;

Fotsch, C.; Jordan, S. R.; Lloyd, D. J.; Sivits, G.; Tadesse, S.; Hale, C.;

St. Jean, D. J., Jr. Small molecule disruptors of the glucokinase−

glucokinase regulatory protein interaction: 4. Exploration of a novel

binding pocket. J. Med. Chem. 2014, 57, 5949−5964. (c) Tamayo, N.

A.; Norman, M. H.; Bartberger, M. D.; Hong, F.-T.; Bo, Y.; Liu, L.;

Nishimura, N.; Yang, K. C.; Tadesse, S.; Fotsch, C.; Chen, J.; Chmait,

S.; Cupples, R.; Hale, C.; Jordan, S. R.; Lloyd, D. J.; Sivits, G.; St. Jean,

D. J., Jr. Small molecule disruptors of the glucokinase−glucokinase

regulatory protein interaction: 5. A novel aryl sulfone series,

optimization through conformational analysis. J. Med. Chem. 2015,

58, 4462−4482.

t_max, time at which maximum compound concentration in

plasma was achieved; AUC, area under the plasma concen-

tration−time curve; F, bioavailability; POC, percent-of-control;

NBO, natural bond orbital; B3LYP, Becke, three-parameter,

Lee−Yang−Parr hybrid density functional

REFERENCES

■

(1) International Diabetes Federation. IDF Diabetes Atlas. http://

www.idf.org/ (accessed May 26, 2015).

(2) (a) Ettaro, L.; Songer, T. J.; Zhang, P.; Engelgau, M. M. Cost-of-

illness studies in diabetes mellitus. PharmacoEconomics 2004, 22, 149−

164. (b) Seuring, T.; Archangelidi, O.; Suhrcke, M. The economic

costs of type 2 diabetes: a global systematic review. PharmacoEco-

nomics 2015, 33, 811−831.

(3) Verspohl, E. J. Novel pharmacological approaches to the

treatment of type 2 diabetes. Pharmacol. Rev. 2012, 64, 188−237.

(4) (a) Agius, L. Glucokinase and molecular aspects of liver glycogen

metabolism. Biochem. J. 2008, 414, 1−18. (b) Matschinsky, F. M.

Glucokinase as glucose sensor and metabolic signal generator in

pancreatic β-cells and hepatocytes. Diabetes 1990, 39, 647−652.

(5) Matschinsky, F. M.; Glaser, B.; Magnuson, M. A. Pancreatic beta-

cell glucokinase: closing the gap between theoretical concepts and

experimental realities. Diabetes 1998, 47, 307−315.

(6) (a) Sarabu, R.; Grimsby, J. Targeting glucokinase activation for

the treatment of type 2 diabetesa status review. Curr. Opin. Drug

Discovery Dev. 2005, 8, 631−637. (b) Meininger, G. E.; Scott, R.; Alba,

M.; Shentu, Y.; Luo, E.; Amin, H.; Davies, M. J.; Kaufman, K. D.;

Goldstein, B. J. Effects of MK-0941, a novel glucokinase activator, on

glycemic control in insulin-treated patients with type 2 diabetes.

Diabetes Care 2011, 34, 2560−2566. (c) Matschinsky, F. M.; Zelent,

(15) For the original design and execution of this GKRP HTS

strategy, see the following: Chen, K.; Michelsen, K.; Kurzeja, R. J. M.;

Han, J.; Vazir, M.; St. Jean, D. J., Jr.; Hale, C.; Wahl, R. C. Discovery of

O

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

small-molecule glucokinase regulatory protein modulators that restore

glucokinase activity. J. Biomol. Screening 2014, 19, 1014−1023.

(16) The ﬁrst HTS campaign of nearly 2 million compounds also

resulted in only a single potent, selective, and structurally distinct hit 3.

(17) For detailed descriptions of the surface plasmon resonance assay

(Biacore) using human GK or human, rat, or mouse GKRP constructs,

as well as the bead-based proximity assay (AlphaScreen) using human

GK and GKRP constructs, see refs 11−14.

(18) (a) cLogP and PSA were calculated with the Daylight Suite,

version 4.81, Daylight Chemical Information Systems, Inc. (b)

ACDLogP and ACDLogD_7.4were calculated with ACD 2012,

Advanced Chemistry Development, Inc. (ACD/Labs).

(19) Very little protein backbone or side chain movement was found

to occur upon the binding of 3 versus 6 with GKRP: the rmsd (both

backbone and side chain atoms) of common residues within a 4.5 Å

radius of either 3 or 6 is only ∼0.76 Å.

(20) Botta, M.; Corelli, F.; Gasparrini, F.; Messina, F.; Mugnaini, C.

Chiral azole derivatives. 4. Enantiomers of bifonazole and related

antifungal agents: synthesis, configuration assignment, and biological

evaluation. J. Org. Chem. 2000, 65, 4736−4739 (see the Supporting

Information for details).

(21) For detailed descriptions of this cell assay using mouse or rat

primary hepatocytes, see refs 11−14.

(22) Johansson, M. P.; Olsen, J. Torsional barriers and equilibrium

angle of biphenyl: reconciling theory with experiment. J. Chem. Theory

Comput. 2008, 4, 1460−1471.

(23) For a recent comprehensive review, see the following: Beno, B.

R.; Yeung, K.-S.; Bartberger, M. D.; Pennington, L. D.; Meanwell, N.

A. A survey of the role of noncovalent sulfur interactions in drug

design. J. Med. Chem. 2015, 58, 4383−4438.

(29) For a recent review, see the following: Gillis, E. P.; Eastman, K.

J.; Hill, M. D.; Donnelly, D. J.; Meanwell, N. A. Applications of

fluorine in medicinal chemistry. J. Med. Chem. 2015, 58, 8315−8359.

(30) St. Jean, D. J., Jr.; Fotsch, C. Mitigating heterocycle metabolism

in drug discovery. J. Med. Chem. 2012, 55, 6002−6020.

(31) For detailed descriptions of this GK translocation assay using

Sprague−Dawley rats, see refs 11−14.

(32) Interestingly, the total plasma concentration of 31 (2.9 μM)

coinciding with statistically signiﬁcant GK translocation (TL) at the 6

h time point for the 10 mg/kg dose translates into an unbound plasma

concentration (0.029 μM; PPB = 99%; Table 5) that is well below the

cellular potency of 31 (rGK TL EC₅₀= 0.14 for 31).^21,31Similarly, the

total plasma concentration of 31 or 32 (∼0.20 μM) coinciding with

statistically signiﬁcant GK translocation (TL) at the 24 h time point

for the 10 mg/kg dose translates into an unbound plasma

concentration (∼0.0020 μM; PPB = 99%; Table 5) that is even

further below the cellular potency of 31 or 32 (rGK TL EC₅₀= 0.037

μM for 32). It should be noted that for in vivo experiments GK TL

was measured using an immunohistochemistry readout with visual

scoring, and for in vitro experiments GK TL was measured using a

diﬀerential ﬂuorescence intensity readout with confocal microscopy

detection.^21,31One explanation for this exposure−eﬀect disconnect

may be that compounds 31 and 32 achieve higher concentrations in

liver tissue than in plasma.

(33) For detailed descriptions of this blood glucose assay using db/

db mice, see refs 11−14.

(34) Statistically signiﬁcant blood glucose lowering was not achieved

at the 3 h time point for the 100 mg/kg dose of 31 but was achieved at

this time point for the 10 mg/kg dose of 31; this discrepancy is

attributed to the variability in the measured blood glucose levels for

the 100 mg/kg dose of 31 at the 3 h time point.

(24) Yu, G. J.; Yoo, C. L.; Yang, B.; Lodewyk, M. W.; Meng, L.; El-

Idreesy, T. T.; Fettinger, J. C.; Tantillo, D. J.; Verkman, A. S.; Kurth,

M. J. Potent s-cis-locked bithiazole correctors of ΔF508 cystic fibrosis

transmembrane conductance regulator cellular processing for cystic

fibrosis therapy. J. Med. Chem. 2008, 51, 6044−6054.

(25) Lin, S.; Wrobleski, S. T.; Hynes, J., Jr.; Pitt, S.; Zhang, R.; Fan,

Y.; Doweyko, A. M.; Kish, K. F.; Sack, J. S.; Malley, M. F.; Kiefer, S. E.;

Newitt, J. A.; McKinnon, M.; Trzaskos, J.; Barrish, J. C.; Dodd, J. H.;

Schieven, G. L.; Leftheris, K. Utilization of a nitrogen−sulfur

nonbonding interaction in the design of new 2-aminothiazol-5-

ylpyrimidines as p38α MAP kinase inhibitors. Bioorg. Med. Chem.

Lett. 2010, 20, 5864−5868.

(26) All quantum mechanical calculations were performed within the

Gaussian 09 program system: Frisch, M. J.; Trucks, G. W.; Schlegel, H.

B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.;

Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato,

M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.;

Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.;

Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai,

H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.;

Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.;

Keith, T.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.;

Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J.

M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.;

Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;

Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.;

Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.;

Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.;

Cioslowski, J.; Fox, D. J. Gaussian 09, revision D.01; Gaussian, Inc.:

Wallingford, CT, U.S., 2009.

(35) The whole-blood concentration of 31 (19 μM) coinciding with

statistically signiﬁcant blood glucose lowering at the 6 h time point for

the 100 mg/kg dose is ∼40-fold higher than the cellular potency of 31

(mGK TL EC₅₀= 0.48 μM).³³It should be noted that 31 is likely

highly protein bound in mouse (rat PPB = 99%; Table 5).

(36) Although the variability in the measured blood glucose levels

confounds clear interpretation, it was also encouraging to observe

statistically signiﬁcant blood glucose lowering at the 3 and 6 h time

points for the 10 and 30 mg/kg doses of 32, respectively.

(37) The whole-blood concentration of 32 (21−25 μM) coinciding

with statistically signiﬁcant blood glucose lowering at the 3 and 6 h

time points for the 100 mg/kg dose is ∼320- to 380-fold higher than

the cellular potency of 32 (mGK TL EC₅₀= 0.066 μM).³³It should be

noted that 32 is likely highly protein bound in mouse (rat PPB = 99%;

Table 5).

(38) For the initial disclosure of the chemical synthesis of these

compounds, see the following: (a) Bartberger, M. D.; Croghan, M. D.;

Fotsch, C. H.; Norman, M. H.; Pennington, L. D.; Reichelt, A.; St.

Jean, D. J., Jr.; Tegley, C. M. Benzodioxepine and benzodioxine

compounds that interact with glucokinase regulatory protein for the

treatment of diabetes. WIPO/PCT Pat. Appl. WO 2012138776 A1,

2012. (b) Ashton, K.; Bartberger, M. D.; Bourbeau, M. P.; Croghan,

M. D.; Fotsch, C. H.; Hungate, R. W.; Kong, K.; Nishimura, N.;

Norman, M. H.; Pennington, L. D.; Reichelt, A.; Siegmund, A. C.;

Tadesse, S.; St. Jean, D. J., Jr.; Yang, K. C.; Yao, G. Benzothiophene

sulfonamides and other compounds that interact with glucokinase

regulatory protein. WIPO/PCT Pat. Appl. WO 2013173382 A1, 2013.

(39) For a comprehensive review of the synthetic and stereochemical

aspects of tert-butanesulﬁnylimine chemistry, see the following: Robak,

M. T.; Herbage, M. A.; Ellman, J. A. Synthesis and applications of tert-

butanesulfinamide. Chem. Rev. 2010, 110, 3600−3740.

(27) NBO 6.0. Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.;

Carpenter, J. E.; Bohmann, J. A.; Morales, C. M.; Landis, C. R.;

Weinhold, F. Theoretical Chemistry Institute, University of Wisconsin,

Madison, 2013.

(28) Weinhold, F.; Landis, C. R. Natural bond orbitals and

extensions of localized bonding concepts. Chem. Educ. Res. Pract.

2001, 2, 91−104.

(40) Assignments of the absolute conﬁguration of 38 and its antipode

were determined by comparison of their computed and measured

vibrational circular dichroism (VCD) spectra (Stephens, P. J.; Devlin,

F. J.; Pan, J.-J. Chirality 2008, 20, 643−663) and optical rotations; see

the Supporting Information for details.

(41) Akay, S.; Yang, W.; Wang, J.; Lin, L.; Wang, B. Synthesis and

evaluation of dual wavelength fluorescent benzo[b]thiophene boronic

P

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

acid derivatives for sugar sensing. Chem. Biol. Drug Des. 2007, 70,

279−289.

(42) Assignments of the absolute conﬁguration of 31 and its antipode

were determined by comparison of their computed and measured

optical rotations; see the Supporting Information for details.

(43) Schlosser, M.; Bobbio, C. Creating structural manifolds from a

common precursor: basicity gradient-driven isomerization of halopyr-

idines. Eur. J. Org. Chem. 2002, 4174−4180.

(44) For the experimental details of key biological assays, see refs

11−14.

Q

DOI: 10.1021/acs.jmedchem.5b01367

J. Med. Chem. XXXX, XXX, XXX−XXX

Article Doi

Discovery and Structure-Guided Optimization of Diarylmethanesulfonamide Disrupters of Glucokinase-Glucokinase Regulatory Protein (GK-GKRP) Binding: Strategic Use of a N → S (nN → σ? S-X) Interaction for Conformational Constraint

DOI: 10.1021/acs.jmedchem.5b01367

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.jmedchem.5b01367

Full text of DOI:10.1021/acs.jmedchem.5b01367

Products guided by the article

R&D Labs maybe for 1402597-53-8

Relevant to this article

Hot Product