Inhibitors of ketohexokinase: Discovery of pyrimidinopyrimidines with specific substitution that complements the ATP-binding site

Article abstract of DOI:10.1021/ml200070g

Attenuation of fructose metabolism by the inhibition of ketohexokinase (KHK; fructokinase) should reduce body weight, free fatty acids, and triglycerides, thereby offering a novel approach to treat diabetes and obesity in response to modern diets. We have

Full text of DOI:10.1021/ml200070g

LETTER
pubs.acs.org/acsmedchemlett
Inhibitors of Ketohexokinase: Discovery of Pyrimidinopyrimidines
with Specific Substitution that Complements the ATP-Binding Site
Bruce E. Maryanoff,*,† John C. O'Neill,* David F. McComsey, Stephen C. Yabut, Diane K. Luci,
Alfonzo D. Jordan, Jr., John A. Masucci, William J. Jones, Marta C. Abad, Alan C. Gibbs, and Ioanna Petrounia
Johnson & Johnson Pharmaceutical Research & Development, Spring House, Pennsylvania 19477-0776, United States
S Supporting Information
b
ABSTRACT: Attenuation of fructose metabolism by the inhibition of ketohexo-
kinase (KHK; fructokinase) should reduce body weight, free fatty acids, and
triglycerides, thereby offering a novel approach to treat diabetes and obesity in
response to modern diets. We have identified potent, selective inhibitors of
human hepatic KHK within a series of pyrimidinopyrimidines (1). For example,
8, 38, and 47 exhibited KHK IC50 values of 12, 7, and 8 nM, respectively, and also
showed potent cellular KHK inhibition (IC50 < 500 nM), which relates to their
intrinsic potency vs KHK and their ability to penetrate cells. X-ray cocrystal
structures of KHK complexes of 3, 8, and 47 revealed the important interactions
within the enzyme's adenosine 5'-triphosphate (ATP)-binding pocket.
KEYWORDS: Kinase, diabetes, obesity, crystal structure, fructose metabolism
oninsulin-dependent diabetes mellitus (NIDDM), which is
characterized by hyperglycemia and insulin resistance, is
body weight, free fatty acids, and triglycerides, thereby offering a
novel approach for treating NIDDM and obesity amidst the modern
diets of postindustrial societies.
N
frequently accompanied by obesity and cardiovascular disease.1ꢀ3
There is a marked positive correlation between the increased
energy intake in the form of highly refined sugars and the
prevalence of NIDDM and obesity.4 Diets high in fructose
promote various metabolic disturbances in animal models,
including weight gain, hyperlipidemia, hypertension, and insulin
resistance.5ꢀ9 Also, in overweight humans, the long-term con-
sumption of fructose is found to increase energy intake, body
weight, fat mass, blood pressure, and plasma triglycerides.10,11
Excessive fructose ingestion stimulates de novo lipogenesis via
upregulation of gene expression,12,13 favors re-esterification of fatty
acids, and increases the production of very low-density lipoprotein
(VLDL) particles.14 Chronic high-fructose diets elevate free fatty
acids and triglycerides, which impair glucose utilization in muscle
tissue and increase the rate of lipolysis in adipose tissue.14 Elevated
triglycerides can impede insulin-signaling pathways,15ꢀ17 support
chronic inflammation,18ꢀ20 and lead to glucolipid toxicity21 with
possible failure of pancreatic β-cells.22
There is human genetic validation of KHK as a therapeutic
target on the basis of mutations that cause essential fructosuria,
an autosomal recessive disorder.27 Individuals with this benign
condition have inactive isoforms of hepatic KHK, such that
ingestion of fructose, sucrose, or sorbitol results in a persistent
rise in blood fructose levels and increased excretion of fructose
into the urine. Thus, a KHK inhibitor would eliminate excess
carbohydrate without a mechanism-based safety issue.
Herein, we report potent, selective inhibitors of human hepatic
KHK (KHK-C isoform27ꢀ31), which function by docking within
the ATP-binding site. By using a combination of high-throughput
screening (HTS) and structure-based drug design, we discovered
and optimized a series of pyrimidinopyrimidines (1). We also
obtained X-ray cocrystal structures of KHK, as a pseudohomodimer
(aand bsubunits) with each subunit occupied by an inhibitor ligand,
thereby illustrating the key intermolecular interactions.
Fructose is readily absorbed from the diet and rapidly metabo-
lized in the liver by fructokinase, also known as ketohexokinase
(KHK; EC 2.7.1.3).1 The hepatic enzyme KHK phosphorylates
fructose on position C1 with the aid of adenosine 50-triphosphate
(ATP) to yield fructose-1-phosphate (F1P), which enters normal
metabolic pathways.23ꢀ25 However, in contrast to the tight regula-
tion of glucose pathways, the fructose metabolism lacks robust
control mechanisms.14 KHK has a high KM, is not inhibited by
product, and is not allosterically regulated.26 As a consequence, high
concentrations of fructose can rapidly flow into glycolytic pathways
to provide the components of triglycerides.26 In this context,
targeting fructose metabolism by inhibition of KHK should reduce
Received: March 11, 2011
Accepted: April 18, 2011
Published: April 18, 2011
r
2011 American Chemical Society
538
dx.doi.org/10.1021/ml200070g ACS Med. Chem. Lett. 2011, 2, 538–543
|

ACS Medicinal Chemistry Letters
LETTER
Table 1. Structures and KHK Inhibition Results
for Derivatives of 1 with Variation of R1
Scheme 1. Synthesis of 3
compda
R1
R2
IC50 (nM)b
2
3
2-MeC6H4
2-MeC6H4
Ph
Me
400
210
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
4
3200
2800
100
5
3-MeC6H4
2-MeOC6H4
2-EtOC6H4
2-MeSC6H4
3-MeSC6H4
4-MeSC6H4
2-MeSO2C6H4
2-EtSC6H4
2-CF3SC6H4
2-EtC6H4
2-(i-Pr)C6H4
2-(c-Pr)C6H4
2-FC6H4
6
7
200
8
12
9
2000
>9000
>9000
>9000
3000
130
10
11
12
13
14
15
16
17
18
19
20
21
5000
380
Table 2. Structures and KHK Inhibition Results
for Analogues of 3 with Variation of R2 and R3
1500
540
2-ClC6H4
2-BrC6H4
c-Pr
170
>9000
>9000
c-hexyl
a New compounds were purified by HPLC and characterized by ESI-MS
and 1H NMR (see the Supporting Information). b Inhibition of recom-
binant human hepatic KHK (KHK-C) in terms of IC50 values; >9000
nM relates to <50% inhibition at 9 μM.
compda
N-R2
N-R3
piperazino
IC50 (nM)b
2
22
23
24
25
26
27
28
29
30
31
32
33
34
35
NH-Me
NH-Pr
400
400
piperazino
piperazino
piperazino
piperazino
piperazino
piperazino
piperazino
The activity of KHK can be assessed in vitro with a pyruvate
oxidase-coupled enzyme assay,23 but this approach to screening
for inhibitors is not readily adaptable to a HTS protocol.
Although we developed a ThermoFluor assay32 to assess direct
ligand binding to KHK, we elected to conduct a HTS campaign
by using a fluorescence polarization (FP) assay to measure the
course of the enzymatic reaction. We employed the transcreener
ADP assay, which is a homogeneous, competitive method that
specifically detects adenosine 50-diphosphate (ADP), a primary
KHK reaction product,33 in conjunction with KHK-C that was
recombinantly expressed and purified to homogeneity.27,30,31,34,35
Our HTS campaign involving ca. 800 000 compounds, with a
wide range of structures, yielded several chemotypes as possible
drug discovery starting points. A promising avenue for explora-
tion was the pyrimidino[5,4-d]pyrimidine series, 1, since several
representative compounds were on hand from an earlier chemi-
cal library enhancement project.36 Confirmed hit 2, with an IC50
value of 400 nM, served as a basis for further study. The initial
group of confirmed hits revealed that NH-cyclopropylmethyl
was one of the better groups for NR2, with analogue 3 having an
IC50 value of 210 nM. The corresponding compound that lacks a
2-methyl group (4) had an IC50 of 3200 nM, and 3-methyl
congener 5 had an IC50 of 2800 nM (Table 1).
NH-hexyl
1600
2300
1700
300
NH-(c-hexyl)
NEt2
NHꢀCH2CtCH
NHꢀCH2Ph
NHꢀCH2(2-thienyl)
400
300
NHꢀCH2(2-thiazolyl) piperazino
60
NH(CH2-c-Pr)
NH(CH2-c-Pr)
NH(CH2-c-Pr)
NH(CH2-c-Pr)
NH(CH2-c-Pr)
NH(CH2-c-Pr)
homopiperazino
300
N-Me-piperazino
morpholino
1500
>7000
70
4-(NH2CH2)-piperidino
4-(NH2)-piperidino
4-piperidinyl-NH
200
710
a See Table 1. b Inhibition of recombinant human KHK-C; >7000 nM
relates to <50% inhibition at 7 μM.
for 3 (Scheme 1).36 Key trichloro intermediate III, from chlor-
ination of II, was reacted with amine substrates to introduce
groups into the R1, R2, and R3 sites sequentially (III f IV f
V f VI). The Boc protecting group in VI was removed with
CF3CO2H to give final product 3.
To prepare analogues for biological study, including gram
quantities, we used a general synthetic route, which is exemplified
Thus, we prepared a set of close analogues of 3 (Table 1;
6ꢀ21) and obtained particularly impressive potency with
539
dx.doi.org/10.1021/ml200070g |ACS Med. Chem. Lett. 2011, 2, 538–543

ACS Medicinal Chemistry Letters
LETTER
Table 3. Structures and KHK Inhibition Results for
Analogues of 7 with Variation of R2 and R3
compda
R2
N-R3
IC50 (nM)b
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
CH2-c-Bu
piperazino
piperazino
piperazino
piperazino
18
50
CH2CH2-c-Pr
CH2CH2OMe
CH2(2-thienyl)
7.0
30
CH2(2-thiazolyl) piperazino
16
CH2(2-pyridyl)
H
piperazino
piperazino
9.8
7.1
18
Figure 1. Crystal structure of the 8 KHK complex. View of 8 (stick
3
model: C, green; N, blue; and S, yellow) and neighboring KHK residues
(labeled) in subunit a (stick models: C, white; N, blue; O, red; and S,
yellow) and Asp-27 in subunit b (“Asp-27B”) (stick model: C, light blue;
N, blue; and O, red). The conserved water molecule is shown as a red
sphere. The H-bonds between N3 and the water oxygen (3.1 Å), Phe-
245 NR and the water oxygen (2.9 Å), and the piperazine nitrogen and
Asp-27B Oδ (2.8 Å) are denoted by dashed lines.
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
CH2-c-Pr
(R)-3-(NH2)-piperidino
(S)-3-(NH2)-piperidino
4-(NH2CH2)-piperidino
3-(NH2CH2)-azetidino
2,6-diazaspiro[3.3]hept-2-yl
MeNHCH2CH2NMe-
4-(Me2NCH2)-piperidino
N-Me-piperazino
23
10
30
8.0
130
140
110
a See Table 1. b Inhibition of recombinant human KHK-C.
the 2-methylthio group (8; IC50 = 12 nM). Derivatives with
R1 varied (Table 1) served to define a preliminary structureꢀ
activity relationship (SAR). It is evident from this set of com-
pounds that an appropriate ortho substituent is important for
potent KHK inhibition (cf. 3ꢀ8, 8ꢀ10, and 17ꢀ19) and that the
size of this group is crucial, with steric bulk being a serious
limitation (cf. 8, 12; 3, 14, 16; and 17ꢀ19). While 2-SMe (8) was
an ideal substitutent, moderately sized groups, such as Me (3), Br
(19), MeO (6), EtO (7), and Et (14), gave reasonable IC50
values of 100ꢀ200 nM. It also appears that groups with a strong
electrostatic field are less favorable (17 and 18 vs 19).
Figure 2. Crystal structure of the 3 KHK complex. View of 3 (stick
3
model: C, green; and N, blue) within the KHK ATP-binding pocket of
subunit a (Connelly surface model: gray for subunit a; and light blue for
subunit b). The cyclopropyl group extends beyond the pocket toward
the solvent. The Asp-27 carboxylate of the b subunit (“Asp-27B”)
þ
interacts with the piperazine NH2
.
With the 2-methyl group held constant in R1, we probed the
effects of altering R2 and R3 (Table 2). These results indicate that
the R2 group can vary widely in size and type (2, 3, and 22ꢀ29);
however, large alkyl (23 and 24) and disubstitution (25) are
disfavored. For R3, groups bearing NH2þ or NH3þ are favored
(cf. 3 and 30ꢀ35). Analogues of 8 were then examined to fine-
tune KHK inhibitory potency for advancing to further studies
(Table 3). Several compounds had IC50 values of e50 nM
(36ꢀ47), and besides 8, compounds 36, 38, 40ꢀ43, 45, and 47
had IC50 values of e20 nM. These results support the view, as
noted above, that R2 can vary widely in size and type and that R3
can benefit by having an ammonium group with more than one
proton (NH2þ or NH3þ) (cf. 8 and 50; 45 and 49). Also, R3 can
tolerate conformational constraint (cf. 8 and 47).
by a conserved water molecule, which hydrogen bonds to ring
nitrogen N3 (viz. 1) of the ligand and Phe-245 NR of KHK. The
o-tolyl group occupies a hydrophobic region that is largely
þ
defined by Phe-260, and the piperazine NH2 interacts with
a carboxylate group from Asp-27 in KHK-b (“Asp-27B”)
(Figures 1 and 2). For both structures, the distances for the
two key H-bonds, N3/water oxygen and piperazine nitrogen/
Asp-27 Oδ, were 3.1 and 2.8 Å, respectively (Figure 1). We
suggest that the enhanced potency for 8 may relate to the MeS
group being poised to interact in an optimal manner with a
hydrophobic cleft or patch on the enzyme's surface, proximal to
Phe-260. We also obtained a cocrystal structure for 47 KHK
3
(2.7 Å). The a and b subunits with 47 in the ATP-binding site are
displayed in Figure 3. In subunit a, the terminal nitrogen of the
spiro-bis-azetidine is hydrogen bonded with Asp-27B (3.1 Å) and
Asn-107 (2.9 Å) (Figure S1 in the Supporting Information),41
whereas that interaction is absent in subunit b because of its
“open” conformation (Figure 3).30
We investigated KHKꢀligand complexes via X-ray crystal-
lography to gain an understanding of the key intermolecular
interactions.37ꢀ39 The X-ray cocrystal structures for 3 KHK
3
(2.8 Å resolution) and 8 KHK (2.8 Å) showed that the ligand (3
3
or 8) was present in each ATP-binding site of the KHK
pseudohomodimer (subunits a and b).40 Considering the a
subunit, there is a notable proteinꢀligand interaction mediated
Certain compounds with potent KHK inhibition were studied
in a cellular assay that measured the level of KHK product F1P in
540
dx.doi.org/10.1021/ml200070g |ACS Med. Chem. Lett. 2011, 2, 538–543

ACS Medicinal Chemistry Letters
LETTER
oral bioavailability in rats (F = 34%; oral t1/2 = 4 h), but it had a
high volume of distribution (Vdss = 32 L/kg) and a high rate of
clearance (CL = 160 mL/min/kg).46 Thus, at an oral dose of
10 mg/kg, its plasma Cmax was just 0.16 μM.47
In conclusion, we have identified potent, selective inhibitors of
human ketohexokinase (KHK-C isoform) by virtue of a novel
pyrimidinopyrimidine compound series (viz. 1). An in vitro assay
based on the direct detection of ADP was very useful for
developing the SAR. By appropriate substitution of the hetero-
cyclic nucleus, we obtained low-nanomolar inhibitors that oper-
ate by docking within the ATP-binding pocket of KHK.
Compound 8 (IC50 = 12 nM) was effective in a cellular functional
assay (IC50 = 400 nM) and was orally bioavailable in rats (F =
34%). Our X-ray cocrystal structures of 3 KHK, 8 KHK, and
Figure 3. Crystal structure of the 47 KHK complex. View of 47 (stick
3
model: C, green; N, blue; and S, yellow) in the ATP-binding pocket of
subunit a (left panel; Connelly surface model, gray; Asp-27B, light blue)
and subunit b (right panel; Connelly surface model, light blue).
3
3
47 KHK display the important intermolecular interactions
3
Table 4. Cellular Functional Results
between the ligand pharmacophore and the enzyme. The avail-
ability of potent, selective KHK inhibitors should facilitate future
studies concerning the role of fructose on metabolic function in
normal animals and in disease models.
’ ASSOCIATED CONTENT
S
Supporting Information. Details for the synthetic pro-
b
cedures; characterization of final products; production and
purification of KHK; biological assay procedures and results;
CEREP profile, kinase profile, X-ray crystallography, and Figure S1.
This material is available free of charge via the Internet at http://
pubs.acs.org.
compd
R2
N-R3
IC50 (nM)a
8
38
40
41
42
46
47
3b
CH2-c-Pr
piperazino
piperazino
piperazino
piperazino
piperazino
400
140
270
270
78
CH2CH2OMe
CH2(2-thiazolyl)
CH2(2-pyridyl)
H
’ AUTHOR INFORMATION
CH2-c-Pr
3-(NH2CH2)-azetidino
2,6-diazaspiro[3.3]hept-2-yl
piperazino
590
360
2400
Corresponding Author
*E-mail: bmaryano@scripps.edu (B.E.M.) or joneill@its.jnj.com
(J.C.O.).
CH2-c-Pr
CH2-c-Pr
a Inhibition of F1P production in HepG2 cell lysates. Number of
experiments (n) and standard deviations are given in the Supporting
Information. b The SMe in 8 is replaced by Me.
Present Addresses
†Department of Chemistry, The Scripps Research Institute,
10550 North Torrey Pines Road, La Jolla, California 92037,
United States. Pennsylvania Drug Discovery Institute, 3805 Old
Easton Road, Doylestown, Pennsylvania 18902, United States.
cell lysates by using LC-MS to quantify F1P (Table 4).42
Compounds 8, 38, 40ꢀ42, and 47 exhibited reasonably potent
cellular KHK inhibition (IC50 < 500 nM), which relates to their
intrinsic potency vs KHK and their ability to enter cells.
Compound 42 is noteworthy given its IC50 value below
100 nM. The much weaker cell inhibition result for 3 vs 8 is
consistent with the relative KHK potencies (IC50 values of 210
and 12 nM, respectively).
Lead compound 8 was examined in a diverse panel of 31
protein kinases, representing different families, for off-target
kinase inhibition at a concentration of 10 μM (Invitrogen; with
100 μM ATP).43 None of the 31 kinases was inhibited >40% at
this high concentration of 8 (IC50 . 10 μM for 31 kinases). For
three off-target metabolic kinases, ribokinase, hexokinase, and
adenosine kinase, the selectivity of 8 was found to be at least
50-fold (relative to KHK inhibition).44,45
Compound 8 has useful solubility (e.g., 1.5 mg/mL in 10%
aqueous Solutol), an acceptable clog D5.0 of 2.9, and favorable
CaCO-2 cell permeability data (ꢁ10ꢀ6 cm/s: A f B, 0.3; B f A,
1.0). It was stable in human and rat liver microsome preparations
(88 and 72% remaining at 10 min) and did not significantly
inhibit cytochrome P450s from human liver microsomes (1A2,
2C19, 2D6, 2C9, and 3A4). Compound 8 exhibited reasonable
Funding Sources
Use of the IMCA-CAT beamline BM-17 at the Advanced Photon
Source was supported by companies of the Industrial Macro-
molecular Crystallography Association by a contract with Haupt-
man-Woodward Medical Research Institute. The Advanced
Photon Source was supported by the U.S. Department of Energy,
Office of Science, Office of Basic Energy Sciences (Contract No.
W-31-109-Eng-38).
’ ACKNOWLEDGMENT
We are grateful to the spectroscopy group for analytical
characterization data for new compounds; to Keli Dzordzorme,
Shariff Bayoumy, and Robyn Williams for cloning and expressing
KHK; to Alexander Barnakov and Ludmila Barnakova for
purifying KHK; to Cynthia Milligan for KHK crystallization; to
Edward Lawson and Gregory Leo for assistance in preparing the
Supporting Information; and to Prof. Erick Carriera (ETH,
Z€urich, CH) for a generous gift of N-Boc-2,6-diazaspiro-
[3.3]heptane, used in the synthesis of 47. We thank the staff at
IMCA-CAT for their support during the collection of X-ray
diffraction data.
541
dx.doi.org/10.1021/ml200070g |ACS Med. Chem. Lett. 2011, 2, 538–543

ACS Medicinal Chemistry Letters
LETTER
’ REFERENCES
(1) Basciano, H.; Federico, L.; Adeli, K. Fructose, insulin resistance,
and metabolic dyslipidemia. Nutr. Metab. 2005, 2, 5.
(2) Zimmet, P.; Alberi, K. G.; Shaw, J. Global and societal implica-
tions of the diabetes epidemic. Nature 2001, 414, 782–787.
(3) Songer, T. J. The economic cost of NIDDM. Diabetes Metab. Rev.
1992, 8, 389–404.
(4) Gross, L. S.; Li, L.; Ford, E. S.; Liu, S. Increased consumption of
refined carbohydrates and the epidemic of type 2 diabetes in the United
States: An ecological assessment. Am. J. Clin. Nutr. 2004, 79, 774–779.
(5) Kasim-Karakas, S. E.; Vriend, H.; Almario, R.; Chow, L. C.;
Goodman, M. N. Effects of dietary carbohydrates on glucose and lipid
metabolism in golden Syrian hamsters. J. Lab. Clin. Med. 1996,
128, 208–213.
(6) Hwang, I. S.; Ho, H.; Hoffman, B. B.; Reaven, G. M. Fructose-
induced insulin resistance and hypertension in rats. Hypertension 1987,
10, 512–516.
(7) Reiser, S.; Hallfrisch, J. Insulin sensitivity and adipose tissue
weight of rats fed starch or sucrose diets ad libitum or in meals. J. Nutr.
1977, 107, 147–155.
(8) Zavaroni, I.; Sander, S.; Scott, S.; Reaven, G. M. Effect of
fructose-feeding on insulin secretion and insulin action in the rat.
Metabolism 1980, 29, 970–973.
(9) Martinez, F. J.; Rizza, R. A.; Romero, J. C. High-fructose feeding
elicits insulin resistance, hyperinsulinemia, and hypertension in normal
mongrel dogs. Hypertension 1994, 23, 456–463.
(10) Raben, A.; Vasilaras, T. H.; Moller, A. C.; Astrup, A. Sucrose
compared with artificial sweeteners, different effects on ad libitum food
intake and body weight after 10 wk of supplementation in overweight
subjects. Am. J. Clin. Nutr. 2002, 76, 721–729.
(11) Teff, K. L.; Elliott, S. S.; Tschop, M.; Kieffer, T. J.; Rader, D.;
Heiman, M.; Townsens, R. R.; Keim, N. L.; D'Alessio; Havel, P. J.
Dietary fructose reduces circulating insulin and leptin, attenuates
postprandial suppression of ghrelin, and increases triglycerides in
women. J. Clin. Endocrinol. Metab. 2004, 89, 2963–2972.
(12) Kok, N.; Roberfroid, M.; Delzenne, M. Dietary oligofructose
modifies the impact of fructose on hepatic triacylglycerol metabolism.
Metabolism 1996, 45, 1547–1550.
(13) Miyazaki, M.; Dobrzyn, A.; Man, W. C.; Chu, K.; Sampath, H.;
Kim, H. J.; Ntambi, J. M. Stearoyl-CoA desaturase I gene expression is
necessary for fructose-mediated induction of lipogenic gene expression
by sterol regulatory element-binding protein-1c-dependent and -inde-
pendent mechanisms. J. Biol. Chem. 2004, 279, 25164–25171.
(14) Mayes, P. A. Intermediary metabolism of fructose. Am. J. Clin.
Nutr. 1993, 58, 754S–765S.
(15) Dresner, A.; Laurent, D.; Marcucci, M.; Griffin, M. E.; Dufour,
S.; Cline, G. W.; Slezak, L. A.; Andersen, D. K.; Hundal, R. S.; Rothman,
D. L.; Petersen, K. F.; Shulman, G. I. Effects of free fatty acids on glucose
transport and IRS-1-associated phosphatidylinositol 3-kinase activity.
J. Clin. Invest. 1999, 103, 253–259.
(16) Ueno, M.; Bezerra, R. M.; Silva, M. S.; Tavares, D. Q.; Carvalho,
C. R.; Saad, M. J. A high-fructose diet induces changes in pp185
phosphorylation in muscle and liver of rats. Braz. J. Med. Biol. Res.
2000, 33, 1421–1427.
(21) Swanson, J. E.; Laine, D. C.; Thomas, W.; Bantle, J. P. Meta-
bolic effects of dietary fructose in healthy subjects. Am. J. Clin. Nutr.
1992, 58, 851–856.
(22) Lewis, G. F.; Carpentier, A.; Adeli, K.; Giacca, A. Disordered fat
storage and mobilization in the pathogenesis of insulin resistance and
type 2 diabetes. Endocr. Rev. 2002, 23, 201–229.
(23) Raushel, F. M.; Cleland, W. W. Bovine liver fructokinase:
Purification and kinetic properties. Biochemistry 1977, 16, 2169–2175.
(24) Raushel, F. M.; Cleland, W. W. Determination of the rate-
limiting steps and chemical mechanism of fructokinase by isotope
exchange, isotope partitioning, and pH studies. Biochemistry 1977,
16, 2176–2181.
(25) Raushel, F. M.; Cleland, W. W. Substrate and anomeric
specificity of fructokinase. J. Biol. Chem. 1973, 248, 8174–8177.
(26) Elliott, S. S.; Keim, N. L.; Stern, J. S.; Teff, K.; Havel, P. J.
Fructose, weight gain, and the insulin resistance syndrome. Am. J. Clin.
Nutr. 2002, 76, 911–922.
(27) Asipu, A.; Hayward, B. E.; O'Reilly, J.; Bonthron, D. T. Proper-
ties of normal and mutant recombinant human ketohexokinases and
implications for the pathogenensis of essential fructosuria. Diabetes
2003, 52, 2426–2432.
(28) Bonthron, D. T.; Brady, N.; Donaldson, I. A.; Steinmann, B.
Molecular basis of essential fructosuria: Molecular cloning and muta-
tional analysis of human ketohexokinase (fructokinase). Hum. Mol.
Genet. 1994, 3, 1627–1631.
(29) Trinh, C. H.; Asipu, A.; Bonthron, D. T.; Phillips, S. E. V.
Structures of alternatively spliced isoforms of human ketohexokinase.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 2009, D65, 201–211.
(30) Gibbs, A. C.; Abad, M. C.; Zhang, X.; Tounge, B. A.;
Lewandowski, F. A.; Struble, G. T.; Sun, W.; Sui, Z.; Kuo, L. C. Electron
density guided fragment-based lead discovery of ketohexokinase inhibi-
tors. J. Med. Chem. 2010, 53, 7979–7991.
(31) Details for KHK-C cloning and expression are given in the
Supporting Information (see the paragraph at the end of this paper).
(32) Matulis, D.; Kranz, J. K.; Salemme, F. R.; Todd, M. J. Thermo-
dynamic stability of carbonic anhydrase: measurements of binding
affinity and stoichiometry using ThermoFluor. Biochemistry 2005,
44, 5258–5266.
(33) Purchased from BellBrook Laboratories, Madison, NJ.
(34) Kozak, M.; Hayward, B.; Borek, D.; Bonthron, D. T.; Jaskolski,
M. Expression, purification and preliminary crystallographic studies of
human ketohexokinase. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2001,
D57, 586–588.
(35) Details on the screening assay method are given in the
Supporting Information (see the paragraph at the end of this paper).
(36) El-Araby, M.; Pottorf, R. S.; Player, M. R. Synthesis of a 2,4,8-
trisubstituted pyrimidino[5,4-d]pyrimidine library via sequential SNAr
reactions on solid-phase. Combin. Chem. High Throughput Screening
2004, 7, 413–421.
(37) Structural information on KHK from X-ray crystallography has
been reported (ref 29 and 30).
(38) Details for our X-ray crystallography are given in the Support-
ing Information (see the paragraph at the end of this paper).
(39) The atomic coordinates and structure factors for KHK com-
plexes with 3, 8, and 47 were deposited in the Protein Data Bank, with
the accession codes 3QA2, 3Q92, and 3QAI, respectively [Protein Data
Bank, Research Collaboratory for Structural Bioinformatics (http://
www.rcsb.org)].
(40) Native KHK is a functional dimer with two subunits that are
basically the same (homodimer) (refs 23, 28, and 30) except for their
spatial disposition due to conformational differences (hence, our use of
the term “pseudohomodimer”) (refs 29, 30, and this work). The
subunits are designated a and b herein.
(17) Koteish, A.; Diehl, A. M. Animal models of steatosis. Semin.
Liver Dis. 2001, 21, 89–104.
(18) Wu, T.; Giovannucci, E.; Pishon, T.; Hankinson, S. E.; Ma, J.;
Rifai, N.; Rimm, E. B. Fructose, glycemic load, and quantity and quality
of carbohydrate in relation to plasma C-peptide concentrations in US
women. Am. J. Clin. Nutr. 2004, 80, 1043–1049.
(19) Oron-Herman, M.; Rosenthal, T.; Sela, B. A. Hyperhomocys-
teinemia as
a component of syndrome X. Metabolism 2003,
52, 1491–1495.
(41) See the paragraph about Supporting Information at the end of
this paper.
(20) Holven, K. B.; Aukrust, P.; Retterstol, K.; Hagve, T. A.;
Morkrid, L.; Ose, L.; Nenseter, M. S. Increased levels of C-reactive
protein and interleukin-6 in hyperhomocysteinemic subjects. Scand. J.
Clin. Lab. Invest. 2006, 66, 45–54.
(42) Details on the cellular assay procedure, including a negative
control, are given in the Supporting Information.41
542
dx.doi.org/10.1021/ml200070g |ACS Med. Chem. Lett. 2011, 2, 538–543

ACS Medicinal Chemistry Letters
LETTER
(43) The compound kinase specificity was assessed against a panel of
31 kinases (see the Supporting Information) with a FRET assay platform
by Invitrogen (http://www.invitrogen.com).41 Representative kinases
across the various kinase families were tested with 10 μM compound and
100 μM ATP. Inhibition activities were ranked based on percent
inhibition at 10 μM. Compounds with inhibition of less than 25% at
10 μM were classified as selective for KHK.
(44) Compounds of interest were tested for inhibition of metabolic
kinase activity via ribokinase (EC 2.7.1.15), hexokinase (EC 2.7.1.1),
and adenosine kinase (EC 2.7.1.20). These kinases were selected
because of their structural or sequence similarity to KHK.
(45) A CEREP panel for receptors and ion channels was performed
(http://www.cerep.fr/cerep/users/index.asp).41
(46) Experimental details are given in Supporting Information.41
(47) After iv dosing to rats at 2 mg/kg, 8 had a C0 of 0.6 μM and a t1/2
of 2 h.
543
dx.doi.org/10.1021/ml200070g |ACS Med. Chem. Lett. 2011, 2, 538–543