Discovery and Rational Design of Natural-Product-Derived 2-Phenyl-3,4-dihydro-2H-benzo[f]chromen-3-amine Analogs as Novel and Potent Dipeptidyl Peptidase 4 (DPP-4) Inhibitors for the Treatment of Type 2 Diabetes

Article abstract of DOI:10.1021/acs.jmedchem.6b00505

Starting from the lead isodaphnetin, a natural product inhibitor of DPP-4 discovered through a target fishing docking based approach, a series of novel 2-phenyl-3,4-dihydro-2H-benzo[f]chromen-3-amine derivatives as potent DPP-4 inhibitors are rationally designed utilizing highly efficient 3D molecular similarity based scaffold hopping as well as electrostatic complementary methods. Those ingenious drug design strategies bring us approximate 7400-fold boost in potency. Compounds 22a and 24a are the most potent ones (IC₅₀ ≈ 2.0 nM) with good pharmacokinetic profiles. Compound 22a demonstrated stable pharmacological effect. A 3 mg/kg oral dose provided >80% inhibition of DPP-4 activity within 24 h, which is comparable to the performance of the long-acting control omarigliptin. Moreover, the efficacy of 22a in improving the glucose tolerance is also comparable with omarigliptin. In this study, not only promising DPP-4 inhibitors as long acting antidiabetic that are clinically on demand are identified, but the target fish docking and medicinal chemistry strategies were successfully implemented.

Full text of DOI:10.1021/acs.jmedchem.6b00505

Subscriber access provided by La Trobe University Library
Article
Discovery and Rational Design of Natural Product-Derived 2-phenyl-3,4-
dihydro-2H-benzo[f]chromen-3-amine Analogs as Novel and Potent Dipeptidyl
Peptidase-4 (DPP-4) Inhibitors for the Treatment of Type 2 Diabetes
Shiliang Li, Hongling Xu, Shichao Cui, Fangshu Wu, Youli Zhang, Ming-Bo Su, Yinghui Gong, Shaobing
Qiu, Qian Jiao, Chun Qin, Jiwei Shan, Ming Zhang, Jiawei Wang, Qiao Yin, Minghao Xu, Xiaofeng
Liu, Rui Wang, Lili Zhu, Jia Li, Yufang Xu, Hualiang Jiang, Zhenjiang Zhao, Jing-Ya Li, and Honglin Li
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00505 • Publication Date (Web): 09 Jul 2016
Downloaded from http://pubs.acs.org on July 9, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pharmacy
Jiang, Hualiang; Shanghai Institute of Materia Medica, Drug Discovery and
Design Center; Shanghai Institute of Materia Medica, Drug Discovery and
Design Center
Zhao, Zhenjiang; East China University of Science and Technology,
Shanghai Key Laboratory of New Drug Design, School of Pharmacy
Li, Jing-Ya; Shanghai Institute of Materia Medica, Chinese Academy of
Sciences
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Li, Honglin; East China University of Science and Technology, Shanghai
Key Laboratory of New Drug Design, School of Pharmacy
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Discovery and Rational Design of Natural Product-Derived
2
-phenyl-3,4-dihydro-2H-benzo[f]chromen-3-amine Analogs as
Novel and Potent Dipeptidyl Peptidaseꢀ4 (DPPꢀ4) Inhibitors for
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
the Treatment of Type 2 Diabetes
1
,†
1,†
†,2
1,†
1
Shiliang Li , Hongling Xu , Shichao Cui , Fangshu Wu , Youli Zhang , Mingbo
2
1
1
1
1
1
Su , Yinghui Gong , Shaobing Qiu , Qian Jiao , Chun Qin , Jiwei Shan , Ming
1
1
1
1
1
1
1
Zhang , Jiawei Wang , Qiao Yin , Minghao Xu , Xiaofeng Liu , Rui Wang , Lili Zhu ,
2
1
2
1,
2,*
1,
Jia Li , Yufang Xu , Hualiang Jiang , Zhenjiang Zhao *, Jingya Li , Honglin Li *
1
State Key Laboratory of Bioreactor Engineering, Shanghai Key Laboratory of New
Drug Design, School of Pharmacy, East China University of Science and Technology,
Shanghai 200237, China.
2
State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica,
Chinese Academy of Sciences. Shanghai 201203, China.
†
Authors contributed equally to this work
*
To whom correspondence should be addressed. Email: zhjzhao@ecust.edu.cn,
jyli@simm.ac.cn, hlli@ecust.edu.cn
ACS Paragon Plus Environment

Page 3 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Abstract
Starting from the lead isoꢀdaphnetin, a natural product inhibitor of DPPꢀ4
discovered through a Target Fishing Docking based approach, a series of novel
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
ꢀphenylꢀ3,4ꢀdihydroꢀ2Hꢀbenzo[f]chromenꢀ3ꢀamine derivatives as potent DPPꢀ4
inhibitors are rationally designed utilizing highly efficient 3D molecular similarity
based scaffold hopping as well as electrostatic complementary methods. Those
ingenious drug design strategies bring us approximate 7400ꢀfold boost in potency.
Compounds 22a and 24a are the most potent ones (IC50: ~2.0 nM) with good
pharmacokinetic profiles. Compound 22a demonstrated stable pharmacological effect.
3
mg/kg oral dose provided > 80% inhibition of DPPꢀ4 activity within 24 h, which is
comparable to the performance of the longꢀacting control Omarigliptin. Moreover, the
efficacy of 22a in improving the glucose tolerance is also comparative with
Omarigliptin. In this study, not only promising DPPꢀ4 inhibitors as long acting
antidiabetic that are clinically on demand are identified, but successfully implemented
the Target Fish docking and medicinal chemistry strategies.
Keywords: Natural product, 3D molecular similarity, Scaffold hopping, DPPꢀ4
inhibitors, OGTT, Diabetes
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Introduction
Diabetes is a fast growing worldwide chronic metabolic disorder. According to
the newest data of the International Diabetes Federation (IDF), about 387 million
adults had diabetes for 2014, and the number will rise to 592 million in 2035. In 2014,
diabetes expenditure reached 612 billion dollars, counting for a ninth of the global
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
healthcare expenditures. It was reported that type 2 diabetes mellitus (T2DM) makes
up approximate 90% of cases of diabetes, which is associated with many micro and
macro vascular complications like coronary artery disease, hypertension, retinopathy,
2
, 3
and stroke, and severely affects patient's quality of life. Current oral antidiabetic
drugs approved by the U.S. FDA are mainly clustered into five categories: biguanides,
4
thiazolidinediones, sulfonylureas, meglitinides and αꢀglucosidase inhibitors.
However, those antidiabetic agents always bring undesirable side effects such as
hypoglycemia, edema, weight gain and gastrointestinal distress, which limit their
5
clinical application. Therefore, it is a necessity to discover novel and potent agents
with minimum side effects for the treatment of T2DM.
Newer approaches for T2DM therapy, such as sodiumꢀglucose coꢀtransporter 2
(SGLT2) inhibitors, glucagonꢀreceptor antagonists and DPPꢀ4 inhibitors, have
6
recently come into focus. Among those possible antidiabetic targets, DPPꢀ4 has been
proved one of the most effective targets for T2DM treatment by the commercial
success of Sitagliptin (1, Figure 1), a firstꢀinꢀclass DPPꢀ4 drug approved by the U.S.
7
FDA in October 2006. DPPꢀ4 enzyme, a serine protease belonging to the
α/βꢀhydrolases (family S9B), is widely distributed and exists in both soluble and
ACS Paragon Plus Environment

Page 5 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
membrane bound forms. It can selectively degrade peptides or hormones that contain
8,
alanine or proline at the penultimate position, leading to inactivation of the proteins.
9
Glucogonꢀlike peptideꢀ1 (GLPꢀ1), a predominant incretin hormone that regulates
glucose activities in a glucose dependent manner and can inhibit glucagon release,
decrease gastric emptying, and promote the regeneration and differentiation of islet β
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
10ꢀ12
cell,
is exactly the substrate of DPPꢀ4. Therefore, DPPꢀ4 enzyme can quickly
13
inactivate GLPꢀ1, resulting in its short halfꢀtime of only 1ꢀ2 min. Inhibition of
DPPꢀ4 is able to prolong circulating halfꢀtime of GLP, and thus help to maintain blood
glucose homeostasis, improve the function of β cell. Moreover, compared with
2, 14
traditional drugs, DPPꢀ4 inhibitors can avoid the aforementioned side effects,
1
4ꢀ16
making them a promising class of antidiabetic drugs.
Extensive efforts have been made to find DPPꢀ4 inhibitors on the basis of
17
potency, selectivity and safety. More than 4,000 DPPꢀ4 inhibitors are reported,
among which seventeen gliptins are the representatives of those intense researches.
Currently, ten out of those seventeen gliptins are approved (Figure S1) and used for
5
T2DM treatment, while others are in different stages of clinical trials. According to
their structure features as well as the binding mode similarities, the gliptins can be
7
18
divided into several groups: Sitagliptin
and its sister gliptins;
cyanopyrrolidineꢀbased gliptins; pyrimidinedioneꢀbased gliptins; diprolylꢀbased
gliptins; xanthineꢀbased gliptins; and boronic acidꢀbased gliptins. The structures of
1
9
some typical gliptins are shown in Figure 1. Recently, Trelagliptin (3) , and
18, 20
Omarigliptin (MKꢀ3102, 5)
have captured our attention for their advantageous
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
long duration of action, since longꢀacting DPPꢀ4 inhibitors have been clinically
needed for improving patients’ quality of life with a safe, once daily or weekly dosing
regimen.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
With the enthusiasm of discovering structurally different, bestꢀinꢀclass DPPꢀ4
inhibitors, our group previously identified isoꢀdaphnetin (6), a natural product derived
from medicinal plant Daphne odora Thunb. var. marginata, as an inhibitor of DPPꢀ4
2
1ꢀ23
(IC50 = 14.13 ꢁM) through TarFisDock
(a reverseꢀdocking based computational
24
chemical biology approach). Starting with the scaffold of isoꢀdaphnetin, 3D
molecular similarity based scaffold hopping was firstly undertaken, producing a
promising 2ꢀphenylꢀ3,4ꢀdihydroꢀ2Hꢀbenzo[f]chromenꢀ3ꢀamine derivative (compound
10) as the new lead of this paper with IC50 value of 110.10 nM (Figure 2). Subsequent
structural optimization was guided mainly by drug design strategies of ligandꢀreceptor
shape matching and electrostatic complementarity. The combination of those two
characteristics gave birth to two most potent inhibitors (22a and 24a) of this series
with IC50s around 2.0 nM, an approximate 7,400ꢀfold improvement of potency
compared with isoꢀdaphnetin, demonstrating that our drug design strategies are highly
effective. Both 22a and 24a displayed good oral bioavailability (F = 89.2% and
5
4.8% respectively). Moreover, what's particularly exciting is the sustained
pharmacodynamic effect of compound 22a ( > 80% of plasma DPPꢀ4 inhibition over
4 h), which is competitive with the longꢀacting control Omarigliptin in ob/ob mice
2
and the excellent performance of 22a in oral glucose tolerance test (OGTT),
suggesting itself an outstanding candidate as longꢀacting antidiabetic agent targeting
ACS Paragon Plus Environment

Page 7 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
DPPꢀ4 for onceꢀdaily oral administration. This study not only provides structurally
distinct skeleton that differs from other known inhibitors in later stages of clinical
development or launched for DPPꢀ4 inhibition, but also underlines that computational
methods are highly efficient toolkits to accelerate the process of drug discovery and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
5
development.
Figure 1
Chemistry
Starting with the lead compound 6, a series of novel potent DPPꢀ4 inhibitors
were designed and synthesized. Compounds 7aꢀb, 8aꢀb, 9 containing a scaffold of
dihydropyrano[2,3ꢀf]chromenꢀ2(8H)ꢀone were synthesized as shown in Scheme 1.
The synthetic sequence started from demethylation of 2,6ꢀdimethoxybenzaldehyde 30
with
,6ꢀdihydroxybenzaldehyde 31. Subsequently, cyclization of intermediate 31 with
E)ꢀ1,2,4ꢀtrifluoroꢀ5ꢀ(2ꢀnitrovinyl)benzene 38, which was obtained from the
condensation of 2,4,5ꢀtrifluorobenzaldehyde 37 and nitromethane in NaOH aqueous
anhydrous
aluminum
chloride
in
dichloromethane
to
give
2
(
26
solution, brought chromene 32 via cascade reaction of Michael addition and aldol
2
7, 28
condensation.
4
Reduction of chromene 32 with NaBH gave dihydrochroman 33,
followed by reduction using metal Zn in the presence of 6 N hydrochloric acid and
ethanol to achieve a mixture of cisꢀracemic 34 and transꢀracemic 35 that could be
isolated by column chromatography. Then, cisꢀracemic 34 and transꢀracemic 35
o
reacted with ethyl acetoacetate 36 in 60% sulfuric acid at 55 C respectively. The four
resulting isomers were separated by chiral column chromatography to get 7aꢀb, 8aꢀb.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Transꢀracemic 9 was made by cyclization of intermediate 35 and ethyl propiolate 39
1
8, 29ꢀ31
in the presence of TFA and Palladium acetate. According to the literatures
and
the results of the in vitro DPPꢀ4 inhibition assay, we were advised that only one
isomer of the four resulting stereoisomers is the most active against DPPꢀ4. To
identify the absolute configuration of the most effective chiral isomer as well as
validate our hypothesis of the active configuration, we did circular dichroism (CD)
spectra for 8a and 8b. As a result, 8a was the most active isomer, which is consistent
with our hypothesis. Meanwhile, the active configuration of our compounds was also
confirmed by the Xꢀray crystal structure of DPPꢀ4 in complex with a representative
compound 22a. Based on the absolute configuration of the final products 8a and 8b,
we could infer the configuration of the intermediates and match the chemical
differences to the corresponding isomers. During our characterization of 34, 35, 7a,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
7b, 8a and 8b, it was found that the H NMR is similar between transꢀisomer and
cisꢀisomer except for a common remarkable difference: there is a doublet around 5
1
ppm in H NMR for transꢀisomer, while only a singlet at about 5 ppm for cisꢀisomer,
32
which is consistent with the article reported by Varma et al. Also, transꢀisomer was
18
more polar than cisꢀisomer after reduction by Zn and hydrochloric acid. These
differences of physical property were applied to initially distinguish the
configurations of the other compounds during synthesis.
Scheme 1
Scheme 2
Scheme 2 describes the general synthesis of dihydro benzo[f]chromen amine
ACS Paragon Plus Environment

Page 9 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
analogues. Different substitutes (E)ꢀnitrovinylbenzene 41a-e were obtained using the
same synthetic method as compound 38. Naphthaldehyde 43a-e, 43g, 43h could be
prepared by formylation of corresponding starting material 42a, 42b, 42d-f, 42h, 42i
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
in the presence of 1,1ꢀdichlorodimethyl ether and TiCl or in the presence of POCl3
4
and DMF. Starting compounds 42a, 42b, 42e were commercially available while 42d
was given by reacting commercial reagent 42c with methanesulfonyl chloride in the
presence of trimethylamine. Intermediates 42f, 42i were obtained by reaction of 42e
or 42g with dimethyldisulfide in the presence of nꢀBuLi followed by oxidation with
3
3
H O , and 42h was obtained by coupling of 42g with methylsulfonamid using CuI
2
2
as catalyst. Bromonaphthaldehyde 43d was protected by glycol, which then coupled
with morpholine followed by deprotection with HCl solution to yeild 43f.
Intermediates 43a-c, 43e-h could be converted into 45b, 45e, 45f, 45i-l by
3
demethylation in the presence of AlCl while 45c, 45d, 45g, 45h, 45m were directly
afforded by formylation of starting material 44a-e. Heating hydroxyl naphthaldehyde
45a-m with nitrovinylbenzene 41a-e or 38 selectively obtained a series of racemic
benzo[f]chromene 46a-d, 46f-s, meanwhile, 46e was easily got by Suzuki coupling of
3 4
46d with (3ꢀMethylsulfonylphenyl)boronic acid using Pd(PPh ) as the catalyst.
Subsequently, 46a-s were reduced to corresponding dihydro benzo[f]chromene by
NaBH followed by reduction with Zn in the presence of HCl solution to yield
4
resulting transꢀracemic 10-29 via column chromatography isolation. Moreover, we
selectively separated some transꢀracemic compounds with high activity like 11, 14, 15,
1
7, 20-24 by chiral column chromatography to get corresponding single isomer.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Results and Discussion
Design and Modification Strategy
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2
Lead compound 6 (isoꢀdaphnetin), a natural product derived from medicinal
plant Daphne odora Thunb. var. marginata, was previously identified as an inhibitor
of DPPꢀ4 through a reverseꢀdocking based computational chemical biology
22
approach. The low structural similarities between compound 6 and other classical
DPPꢀ4 inhibitors (both 2D and 3D similarities, Figure S2ꢀS3) are the most direct
evidence to demonstrate the novelty of our lead compound. Therefore, based on this
novel scaffold, we conducted structural optimization to enhance its potency. When
comparing the scaffold of 6 with Sitagliptin (Figure 2), it was found that besides the
common hydrophobic multiꢀsubstituted benzene moiety, both of them also harbor a
chiral carbon atom connected with a polar hydrogen donor (ꢀCH OH and ꢀNH ,
3
2
respectively). Those two similar pharmacophores were supposed to be key structural
determinants for the inhibition of DPPꢀ4 enzyme. After aligning all the DPPꢀ4
3
4
i
inhibitors deposited in the ChEMBL database with IC50 or K value lower than 50
nM (about 2000 compounds) to the biological active conformation of Sitagliptin
extracted from its complex crystal structure with DPPꢀ4 (PDB code: 1X7O) by
3
5ꢀ37
SHAFTS (an efficient 3D molecular similarity evaluation program),
it could be
clearly observed that no matter how diverse the structures are, most of them share two
identical wellꢀsuperimposed pharmacophores, an ellipsoidal hydrophobic head and an
ACS Paragon Plus Environment

Page 11 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
adjacent chiral carbon atom attached with an amino. Consequently, the two common
pharmacophores must play pivotal role in the inhibition process of DPPꢀ4. Due to the
3
unflavored larger size of ꢀOCH groups substituted on the phenyl moiety of 6 and the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
lack of amino group on the chiral carbon atom, lead 6 displayed moderate biological
activity against DPPꢀ4. Given that Sitagliptin is now the most popular marketed
DPPꢀ4 inhibitor, the above mentioned two pharmacophores of Sitagliptin were grafted
onto the scaffold of 6, resulting in compound 8a. Our hypothesis was borne out when
8
a was tested to have an IC50 of 154.67 nM against DPPꢀ4, an approximately 91ꢀfold
improvement in potency (Table 1).
Table 1
Molecular docking simulation was conducted to analyze the binding mode of 8a.
The 2,4,5ꢀtrifluorobenzene group penetrates into the hydrophobic subpocket of DPPꢀ4
formed by residues Val656, Tyr631, Tyr662, Trp659, Tyr666, and Val711, while the
tetrahydroꢀ2Hꢀpyranꢀ3ꢀamine nitrogen atom positions appropriately to generate
hydrogen bonding interactions with the side chains of Glu205, Glu206 and Tyr662
(Figure 3). In addition, the tetrahydroꢀ2Hꢀpyranꢀ3ꢀamine oxygen atom is hydrogen
bonded to the side chain of Tyr547. With this binding orientation, the catalytic activity
related residues of DPPꢀ4 (Catalytic triad: Ser630, Asp708 and His740; related
38, 39
residues Tyr547, Glu205 and Glu206)
are occupied to interact with 8a, leading to
a decreased catalytic activity of DPPꢀ4 enzyme against its substrates. Moreover, we
noticed that the coumarin fragment of 8a could form πꢀπ stacking interaction with
Phe547 at the opening entrance of the binding site. In order to improve the binding
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
affinity of our inhibitors, we decided to conduct a scaffold hopping by replacing the
coumarin fragment with a naphthalene ring, which has a much stronger πꢀconjugated
system. As a result, even the transꢀracemic compound 10 displayed improved
biological activity than 8a (absolute configuration), demonstrating the beneficial πꢀπ
stacking interaction brought by scaffold hopping. At this point, a novel scaffold
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(2ꢀ(2,4,5ꢀtrifluorophenyl)ꢀ3,4ꢀdihydroꢀ2Hꢀbenzo[f]chromenꢀ3ꢀamine)
was
successfully designed through the 3D molecular similarity based scaffold hopping
approach.
Figure 3
Beginning with the new scaffold of compound 10, structural optimization was
undertaken to further enhance the binding affinity of the inhibitors. First of all, the
electrostatic potentials of the DPPꢀ4 enzyme were calculated using the Adaptive
40ꢀ42
PoissonꢀBoltzmann Solver (APBS) software package.
The electrostatic potentials
on the surface of the DPPꢀ4 substrate binding site are presented in Figure 4A.
Proposed binding mode of 10 was simulated by molecular docking with PDB 4PNZ
as the receptor. We could clearly observe from Figure 4A that the entrance of the
binding site (S2 extended subsite) is widely open and covered with the relatively
strong positive electrostatic potential. Meanwhile, the large space on the right side of
the entrance, near Ser209 and Arg358, also captured our attention. In order to probe
this distinctive entrance region, three substituents with increasing
1
2
3
electronꢀwithdrawing ability (ꢀOCH , ꢀBr, and ꢀCN) were introduced to R and R
positions of 10, respectively. The inhibitory activities of compounds 11ꢀ13 and 15ꢀ17
ACS Paragon Plus Environment

Page 13 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
are listed in Table 2. For the R position, compound 11 with ꢀOCH
3
group performed
much better than 12 (with ꢀBr), while compound 12 performed better than 13 (with
1
ꢀ
CN), suggesting that electronꢀdonating substituents are more suitable for R .
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
Interestingly, when it came to R position, the situation was just the opposite:
compound 17 with ꢀCN was the most potent one (IC50 = 8.42 nM), while compound
1
5 with ꢀOCH
3
was the least potent one (IC50 = 49.73 nM) among the three, indicating
2
that electronꢀwithdrawing groups are preferred at R position. This parallel
optimization strategy was considered to be a highly efficient approach to provide
primary exploration information and accelerate the subsequent structural optimization
process.
Figure 4
On the one hand, compound 11 was docked to the substrate binding site of
DPPꢀ4 to help explore the binding patterns of the right side of the pocket. Besides
almost the same hydrophobic interaction, hydrogen bond and salt bridge around the
catalytic triad as those of 8a, the naphthalene ring of 11 contacts Phe357 with
3
favorable πꢀπ interactions. Moreover, the ꢀOCH group could insert into the concave
surface formed by the backbones of Glu206, Val207 and Phe208 with the oxygen
atom hydrogen bonded to the sidechain of Ser209 (Figure 4B). The increased binding
affinity of 11 compared with 10, 12-13 informed us that shape complementarity is an
1
important factor for the optimization at R position. In order to generate electrostatic
1
interactions with Arg358 through R position, the substituent should contain an
electronꢀwithdrawing tail and a proper rigid linker to ensure that it could extent along
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the concave surface. Then compound 14, which possesses a 3ꢀ(methylsulfonyl)phenyl
substituent, was designed and synthesized, displaying approximate 4ꢀfold
improvement of potency compared with 11 (14.60 nM vs 60.61 nM). The predicted
binding mode of 14 (Figure 4C) presents that the phenyl group fits well with the
concave surface with one methylsulfonyl oxygen atom hydrogen bonded to Arg358
and another oxygen atom hydrogen bonded to water molecule W985, which appears
frequently in most of the DPPꢀ4 crystal structures, and thus was remained in the
molecular docking simulation. The beneficial van der Waals (vdW) interactions
resulting from shape matching as well as the anticipated hydrogen bond with Arg358
contribute to the increased binding affinity of 14, demonstrating the rationality of our
optimization hypothesis.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
On the other hand, electrostatic surface potential (ESP) of the substituents at R
in compounds 15ꢀ17 with the relative orientations corresponding to bound
43
conformations predicted by molecular docking was calculated by Jaguar with DFT
at B3LYP/ccꢀpVTZ(ꢀf) level. There is an increasing trend of negative electrostatic
3
potential around the groups ꢀOCH , ꢀBr, and ꢀCN as the red color of the surfaces turns
darker (Figure S4). Intriguingly, the binding affinities of 15ꢀ17 increased as the
negative electrostatic potential increased, suggesting us that electrostatic
2
complementarity is a crucial consideration for the optimization at R since the
cliffꢀlike entrance of the binding site is covered with relatively strong positive
electrostatic potential which is attributable to the backbones of residues Arg356,
Phe357 and Arg358. Perhaps a ꢀCOOH, which is able to generate stronger negative
ACS Paragon Plus Environment

Page 15 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
electrostatic potential than ꢀCN (ESPmin: ꢀ146.48 vs ꢀ39.19 kcal/mol, Figure 4D), shall
perform better. However, compound 18 (IC50 = 17.65 nM) was less potent than 17
(
IC50 = 8.42 nM). We conjectured that the negative electrostatic potential of the
COOH group may be too strong for the positive electrostatic potential of the subsite.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ
Aiming to seek the optimal matching of the electrostatic potential, some substituents
like morpholine, ꢀNHSO CH and ꢀSO CH were selected for the ESP computation.
As a result, the minimum ESPs of those three fragments were in a descending order:
26.88, ꢀ31.86 and ꢀ37.69 kcal/mol, respectively. Just as expected, the biological
2
3
2
3
ꢀ
activity of compound 19 was less potent than 20 and 21. However, it was arresting
that the IC50 value of 19 (IC50 = 70.42 nM) is larger than compound 15 (IC50 = 49.73
3
nM), which contains a ꢀOCH group and has a less negative electrostatic potential
(ESPmin: ꢀ21.10 kcal/mol). The diminished activity of 19 was likely attributable to the
farther distance between the point with the minimum ESP (around the morpholine
oxygen atom) and the positive electrostatic potential surface of the subsite compared
with that of 15, informing us that distance is a crucial element that influence the
electrostatic interaction. In addition, the fact that compounds 17, 20 and 21 exhibited
approximately equal potent activities against DPPꢀ4 despite the different minimum
ESPs of their fragments had also invited us to think that the minimum ESP for the
inhibitor to generate optimal complementarity of electrostatic potential with the
receptor is not a fixed value but in a range, taking distance into consideration.
Table 2
1
The above discussion of the biological activities of inhibitors optimized at R and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
R positions were all concentrated on the transꢀracemic molecules. In order to
accelerate the structural optimization pipeline and improve working efficiency, only
transꢀracemic molecules with superior potency and favorable physicochemical
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
properties were selected for single isomer separation. For R position, compound 11
was chosen other than 14 because of the disadvantageous lipophilicity of 14 (clogP:
4
4
2
4
.83) calculated by QikProp. For R positon, compounds 15, 17, 20 and 21 were
selected considering that 16 and 19 exhibited higher lipophilicity (clogP: 4.96 and
.66, respectively), and that 18 displayed unfavorable higher pKa value, which may
4
have more potential to block the human potassium channel humanꢀEtherꢀaꢀgoꢀgo
18, 45
Related Gene (hERG)
. The biological data of the absolute configurations of those
selected transꢀracemic molecules are listed in Table 2. Comparisons of the activities
of 11a with 11b and 15a with 15b indicated that the (2R, 3S)ꢀisomer of the analog in
this series is the most potent isomer. Additionally, just as the almost equipotent
activities of transꢀracemic 17, 20 and 21, the absolute configurations 17a, 20a and
21a also displayed similar potency against DPPꢀ4 at single digital nanomolar range.
Encouraged by the successful improvement of potency that is attributable to
1
2
either shape matching at R or electrostatic complementarity at R , the combination of
those two beneficial substituting positions were undertaken. The ꢀOCH
3
group was
1
2
fixed as R , while ꢀOCH , ꢀNHSO CH and ꢀSO CH groups were selected for R .
3
2
3
2
3
Then, compounds 22aꢀ24a were synthesized and separated as (2R, 3S)ꢀisomers. As
shown in Table 2, all of the three doubleꢀsubstituted analogs exhibited increased
potency against DPPꢀ4 compared with the singleꢀsubstituted ones, once again proving
ACS Paragon Plus Environment

Page 17 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the rationality and validity of our structural optimization strategy. It was interesting
that 22a displayed equipotent activity to 24a (IC50: 2.06 nM vs 1.90 nM). Besides the
vdW interactions and hydrogen bond between the methoxyl oxygen atom and Ser209
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
at R (Figure 5A), the strengthened negative electrostatic potential (ESPmin: ꢀ36.41
kcal/mol, Figure S5) produced by the coꢀoccurrence of the two methoxyl oxygen
atoms was the major contributor to the substantially improved potency of 22a in
comparison with 11a and 15a (7ꢀfold and 10ꢀfold improvement, respectively). As the
negative electrostatic potential of 24a increased (ESPmin: ꢀ42.14 kcal/mol, Figure S5),
so did the biological activity, compared with 21a. The DPPꢀ4 inhibitory activities of
2
the compounds optimized at R correlate well (R = 0.84) with their calculated ESPmin
(Figure S6). Moreover, the binding mode of 24a informed us that one of the
methylsulfonyl oxygen atom could participate into the hydrogen bond network
bridged by the versatile water W985 (Figure 5B). Compound 23a was less
competitive than 22a and 24a probably because of its less strong negative electrostatic
potential. Therefore, 22a and 24a, with highly potent bioactivities, were picked out
for pharmacokinetic studies.
Figure 5
Last but not least, the hydrophobic subpocket of DPPꢀ4 formed by residues
Val656, Tyr631, Tyr662, Trp659, Tyr666, and Val711 was revisited with the aim of
verifying the necessity of the ellipsoidal hydrophobic pharmacophore. Firstly, the
fluorine atom at 5 position of the 2,4,5ꢀtrifluorobenzene group in compound 11 was
removed, leading to a decreased activity (25, IC50 = 78.33 nM). When we continued
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to cut off the fluorine atom at 4 position, compound 26 displayed an approximate
9
ꢀfold decrease in potency. Then, compound 27, with a fluorine atom at 2 and 5
positions simultaneously, was synthesized, which showed improved activity than 25
and 26. Also, we continued to cut off the fluorine atom at 2 position, compound 28
displayed approximate 4ꢀfold decrease in potency compared with compound 27.
Moreover, the replacement of the 2 position fluorine atom by a chlorine atom in
compound 29 brought us comparable potency to 26. Combined with the efficacy of 11,
conclusion was drawn that coupled with one fluorine atom at 2 position of the phenyl
group, another fluorine atom substituted at 5 position is more preferred, making the
hydrophobic head more like a rugby.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Crystal Structure of DPP-4 in Complex with Compound 22a
In order to investigate the actual binding patterns of the inhibitors in this series
and prove the rationale of our computational design strategies, we selected one of the
most potent compound 22a to coꢀcrystallize with DPPꢀ4. The crystal structure of
DPPꢀ4 in complex with compound 22a was refined to 2.75 Å and the coordinates
have been deposited in the Protein Data Bank (PDB) as the entry 5J3J (Table 3). The
binding orientation of 22a in the crystal structure (Figure 6) is almost the same as
46
predicted above in Figure 5A. However, due to the flexibility of the residue Arg358,
it was attracted by the strengthened negative electrostatic potential (ESPmin: ꢀ36.41
kcal/mol) produced by the coꢀoccurrence of the two methoxyl oxygen atoms in 22a,
o
and rotated its guanidine group with approximate 70 to simultaneously generate a
ACS Paragon Plus Environment

Page 19 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hydrogen bond with each of the methoxyl oxygen atom (Figure 6B). This
phenomenon may also happen to the inhibitor 24a, which has a methylsulfonyl group
acting as Hꢀbond receptor and has a comparable ESPmin to 22a. Therefore, future
structureꢀbased drug design targeting DPPꢀ4 should take the flexibility of the residue
Arg358 into consideration.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Besides, we also performed surface plasmon resonance (SPR) experiments to
evaluate the kinetic parameters of 22a binding to DPPꢀ4 enzyme (Table 4 and Figure
D
S7). As expected, compounds 22a and 5 are equipotent in terms of K values (2.38
nM vs 2.75 nM). Both compounds have quite long dissociation profile (k ) than
off
Sitagliptin, which may be a favorable factor for longꢀlasting in vivo efficacy.
ꢀ1 ꢀ1
6
5
Moreover, 22a associated faster than compound 5 (kon (M s ): 1.44×10 vs 2.58×10 )
to DPPꢀ4.
Figure 6
Table 3
Table 4
Pharmacokinetic Study of Compounds 22a and 24a
Pharmacokinetic studies of the most potent inhibitors, 22a and 24a, were
performed by administering SpragueꢀDawley (SD) rats with a 1 mg/kg intravenous
(IV) dose and a 2 mg/kg oral dose (PO) of the compounds. In general, both of
compounds displayed high oral bioavailability, good exposures (AUCs) and almost
the same time to reach the maximum concentration (Tmax) (Table 5). Compared with
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2a, compound 24a had a larger IV exposure, a bigger maximum plasma
concentration Cmax (2317.99 ꢁg/L) and a lower plasma clearance (0.04 L/h/kg),
resulting in a longer terminal halfꢀlife of 5.35 h. However, 24a was less competitive
than 22a in terms of oral bioavailability, probably due to the first pass effect, which
could greatly reduce the concentration of 24a before it enters the systemic circulation,
since the AUC0ꢀ∞ of 24a after PO dosing was relatively low in comparison with that of
the IV dosing.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 5
Effect on in vivo DPP-4 Activity
On the basis of the superior properties of compound 22a including potent
inhibitory potency in vitro, and favorable pharmacokinetic characteristics, it was
advanced into further efficacy study. Firstly, the plasma DPPꢀ4 inhibition study of 22a
was primarily evaluated in ICR mice at dose of 5 mg/kg. As shown in Figure 7A,
compound 22a displayed sustained up to 80% inhibitory effect of DPPꢀ4 activity in
2
4 h after single dose administration, which is comparable to the performance of the
longꢀacting control 5. Those results inspired us to conduct another experiment to
further validate the duration of action and in vivo DPPꢀ4 inhibition effect of 22a in
type 2 diabetic ob/ob mice. In the subsequent efficacy study, multiple doses of 22a (1,
3
and 10 mg/kg) were assigned. As a result, compound 22a significantly inhibited the
DPPꢀ4 activity in a doseꢀdependent manner (Figure 7B and S8, P < 0.001). More than
8
0% of the DPPꢀ4 activity was inhibited at all dose levels of 22a (1, 3 and 10 mg/kg)
ACS Paragon Plus Environment

Page 21 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
within 24 h. Particularly, the sustained pharmacodynamics effect of 22a was
comparable to the positive control 5 at the dose of 3 mg/kg. The above in vivo DPPꢀ4
inhibition study indicated that compound 22a is qualified to be further developed as
onceꢀdaily antidiabetic agent.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7
Effect on OGTT in ICR mice
Considering the favorable pharmacokinetic profiles and good in vivo DPPꢀ4
inhibition activity of compound 22a, it was further advanced to an OGTT to evaluate
its ability of improving the glucose clearance capacity. For this study, 8hrsꢀfasted
mice were administered orally with a single dose of vehicle or 22a (1, 3, 10mg/kg) 60
minutes before the glucose challenge (3.0 g/kg, orally loaded), and blood glucose was
measured at time points of 0 (baseline), 15, 30, 60, and 120 min after the glucose
challenge. As shown in Figure 8A, compound 22a significantly reduced glucose
excursion in a doseꢀdependent manner at the time points from 30 min to 60 min
(Table S1 and Figure S9, P < 0.05). Although the efficacy of 1 mg/kg 22a (Reduction
percentage of AUC0ꢀ120: 11.1%) was less competitive than 5 of 10 mg/kg (15.47%),
both 3 and 10 mg/kg 22a were comparable to 5 with the reduction percentage of
AUC0ꢀ120 around 16%. Collectively, the AUC0ꢀ120 of OGTT demonstrated that the
efficacy of 22a in improving the glucose tolerance is as good as, if not better than 5
(Figure 8B).
Figure 8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Conclusion
In summary, starting from the natural product isoꢀdaphnetin, a series of novel
ꢀphenylꢀ3,4ꢀdihydroꢀ2Hꢀbenzo[f]chromenꢀ3ꢀamine derivatives with highly potent
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
DPPꢀ4 inhibitory activities were designed and synthesized through utilizing highly
efficient 3D molecular similarity based scaffold hopping and electrostatic matching
strategies. The customized design strategies brought us approximately 7400ꢀfold
improvement in potency compared with the lead. Compounds 22a and 24a were the
most potent inhibitors with good pharmacokinetic profiles. The in vivo efficacy
studies demonstrated that 22a displayed sustained pharmacodynamics effect on the
inhibition of DPPꢀ4 activity in ob/ob mice, and also presented comparable
performance to the longꢀacting control 5 in improving the glucose tolerance. Inꢀdepth
preclinical characterization of compound 22a is undertaken and will be reported in
due course. Finally, this study provides a promising candidate as longꢀacting
antidiabetic agent targeting DPPꢀ4 for onceꢀdaily oral administration, which is mainly
attributed to the extremely useful and efficient computational methods.
ACS Paragon Plus Environment

Page 23 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Experimental Section
1
. In vitro Enzyme Assay.
The expression and purification of DPPꢀ4 enzyme was conducted using a
baculovirus system in insect cells (Sf9, a clonal isolate of Spodoptera frugiperda Sf21
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
47
cells (IPLBꢀSF21ꢀAE)) according to the literature. Some details are listed below.
The cell culture consisting of the secreted DPPꢀ4 was harvested at 72 h after infection.
The supernatant was loaded to a 5ꢀml HisTrap FF nickelꢀchelate column (GE
Healthcare) and equilibrated with binding buffer (50 mM Tris pH 7.9, 400 mM NaCl).
The column was washed with ten column volumes of washing buffer (50 mM
TrisꢀHCl pH 7.9, 400 mM NaCl, 20 mM imidazole and 0.25 mM
tris(2ꢀcarboxyethyl)phosphine (TCEP). DPPꢀ4 was eluted with 20 column volumes of
a linear gradient of imidazole (20ꢀ500 mM). DPPꢀ4 was further purified using a
sizeꢀexclusion column (Superdex 200 10/300 column, GE Healthcare) equilibrated
with 25 mM Tris pH 7.9, 150 mM NaCl, 0.25 mM TCEP and 1 mM EDTA.
The DPPꢀ4 activity was determined by measuring the rate of hydrolysis of a
substrate GlyꢀProꢀAMC (sigma), the hydrolyzed fluorescent product
amidomethylcoumarin (AMC) was continuously monitored using an excitation
wavelength of 360 nm and an emission wavelength of 460 nm every 60 seconds for
30 minutes using a BioTek microplate reader. A typical reaction contained 15 ng/ml
enzyme, 150 ꢁM GlyꢀProꢀAMC, different concentrations of the test compounds and
assay buffer (25 mM HEPES, pH 7.5, 150 mM NaCl, 0.12 mg/ml BSA) in total
reaction volume of 100 ꢁl. The highꢀthroughput screening of the DPPꢀ4 inhibitors was
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
carried out in triplicate. And the IC50 data were calculated using the software Origin
8
.0.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
. Co-crystallization of DPP-4 with Compound 22a and Structure
Determination
Purified DPPꢀ4 was concentrated to 15 mg/mL for crystallization and the purity
of the crystallization sample was confirmed by SDSꢀPAGE. DPPꢀ4 and 22a were
o
mixed and incubated at 4 C overnight. Crystals were grown by vapour diffusion at
2
93 K using the hangingꢀdrop method by mixing equal volumes of protein and
reservoir solution containing 18% polythylene glycol 3350, 200 mM NaCl and 100
mM TrisꢀHCl buffer pH 8. Prior to data collection, the crystals were soaked in the
reservoir solution including 25% glycerol, and then flash frozen in liquid nitrogen.
Xꢀray diffraction data were collected at 100 K on beam line BL17U at the
Shanghai Synchrotron Radiation Facility (SSRF). The raw data were processed using
48, 49
MOSFLM and SCALA program from CCP4 suite.
The complex structure was
determined by molecular replacement using PDB entry 4N8D (without ligands and
water molecules) as the search model. Structure refinement was conducted with
50
51
REFMAC 5. After cycles of refinement and model building with COOT, the
molecular graphics were prepared with PYMOL (DeLano,2002) to generate the
figures. Statistics of data collection, processing, and refinement were listed in Table 3.
Bound inhibitors are clearly visible in the electron density map.
ACS Paragon Plus Environment

Page 25 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
. Surface Plasmon Resonance (SPR)
BIAcore T200 was used to conduct the surface plasmon resonance (SPR)
o
experiments at the temperature of 25 C. The general procedures were identical to that
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
2
reported by Wang et al.. In this study, the purified DPPꢀ4 is diluted in sodium acetate
solution (pH 4.5) to a final concentration of 50 ꢁg/mL. The immobilization level was
1
3,000 RU resonance units (RU). Compounds 22a, 24a, Sitagliptin and MKꢀ3102
with concentrations of 0 nM, 0.19 nM, 0.39 nM, 0.78 nM, 1.56 nM, 3.125 nM, 6.25
nM, 12.5 nM and 25 nM were injected respectively for 480 s at a low speed of 30
ꢁ
L/min. Data processing and analysis were performed using BIAevaluation 1.1
software.
4
. Pharmacokinetic Assay in SD rats.
Single dose pharmacokinetic studies on the two most active compounds 22a and
4a were performed using male SD rats in full compliance with the Guide for the
2
Care and Use of Laboratory Animals. After intravenous (1 mg/kg) and oral
administration (2 mg/kg), blood samples were collected at specified intervals (5 min,
15 min, 30 min, 1 h, 2 h, 4 h, 6 h, 8 h and 24 h) with heparin sodium as antiꢀcoagulate.
Centrifugation of blood samples gave the plasma fraction. PK parameters were
calculated with nonꢀcompartment model using WinNonlin5.2 based on the
LCꢀMS/MS quantitation data.
5
. In vivo DPP-4 inhibition study.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Male ICR mice (18ꢀ25 g) from Shanghai SLAC laboratory Animal Co., Ltd.,
were housed 10 per cage and given access to normal diet rodent chow and water ad
libitum. Mice (n=2/group) were randomly assigned to treatment groups. Baseline (t =
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ60 min) blood was collected into a tube containing heparin sodium from tail nick
blood. Animals were then treated orally with vehicle (0.3% sodium
carboxymethylcellulose, 150 ꢁL), 22a (5 mg/kg, 150 ꢁL) or 5 (5 mg/kg, 150 ꢁL).
Subsequently, blood per animal was collected at 1, 2, 3, 4, 6, 8, 10, 12 and 24 h. All
samples were centrifuged at 1000×g for 10 min and the plasma was harvested.
o
Aliquots of plasma samples were stored at ꢀ70 C until analysis.
ob
ob
Female B6.VꢀLep /Lep (ob/ob) mice (from Jackson Laboratory, Bar Harbor,
ME, USA) were bred at the Shanghai Institute of Materia Medica, Chinese Academy
o
of Sciences. The animals were housed in a temperature controlled room (22 ± 2 C)
with a 12ꢀh light/dark, and allowed access to normal chow diet and water. Animal
experiments were approved by the Animal Care and Use Committee, Shanghai
Institute of Materia Medica. Mice (n=6/group) were randomly assigned to treatment
groups and fasted 2 h. Baseline blood samples were collected from retroorbital plexus
of mice. Animals were then administrated orally with vehicle (0.5% sodium
carboxymethylcellulose, 10 ml/kg), 22a (1, 3 and 10 mg/kg) or 5 (3 mg/kg).
Subsequently, blood samples were collected at specified intervals: 1, 2, 4, 6, 8 and 24
h. All samples were centrifuged at 1000×g for 10 min and the serum was harvested.
o
Aliquots of serum samples were stored at ꢀ20 C until analysis.
For the measurement of in vivo DPPꢀ4 activity, a continuous fluorometric assay
ACS Paragon Plus Environment

Page 27 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
using GlyꢀProꢀAMC as substrates was applied, which can release the fluorescent
AMC group after being cleaved by DPPꢀ4. For the plasma samples, the reaction
solution was made up of 50% plasma samples, 125 ꢁM GlyꢀProꢀAMC, and buffer
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(
100 mM 4ꢀ(2ꢀhydroxyethyl)ꢀ1ꢀpiperazineethanesulfonic acid (HEPES), pH 7.5, 0.1
mg/mL BSA) in a total reaction volume of 20 ꢁL. For the serum samples, the reaction
solution was made up of 10 ꢁL serum samples, 10 ꢁM GlyꢀProꢀAMC, and buffer (100
mM HEPES, pH 7.5, 0.1 mg/mL BSA) in a total reaction volume of 20 ꢁL. Liberation
o
of AMC was monitored continuously in a 96ꢀwell plate in 5 min at 37 C using an
excitation wavelength of 360 nm and an emission wavelength of 460 nm. The data are
reported as % activity calculated as follows: %activity =Vts/Vcs*100%, where Vts is the
rate of reaction of treated sample and Vcs is the rate of reaction of control sample.
6. OGTT in ICR Mice
Male ICR mice (20ꢀ25 g) from Shanghai SLAC laboratory Animal Co., Ltd.,
were housed 5 per cage and given access to normal diet rodent chow and water ad
libitum. Animals were divided into 5 groups with 10 in each and assigned as follows:
Group I (served as control) received 0.5% sodium carboxymethyl cellulose (CMCꢀNa)
solution, Group II (served as positive control) received 5 (10 mg/kg) orally, Group III
(served as experimental group) received the indicated doses of 22a (1, 3, 10 mg/kg)
orally. During the experiment, 8 h fasted mice were administered with vehicle or 5 or
22a (1, 3, 10 mg/kg) after the fasting blood glucose was estimated. After 60 minutes,
animals were orally loaded with glucose (3 g/kg), and tail vein blood was collected at
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 88
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
(before glucose challenge), 15, 30, 60, 90 and 120 min after the glucose challenge.
Blood glucose levels were measured using a glucometer (ACCUꢀCHEK ActiveꢀBlood
Glucose Meter, Roche). Area under the curve (AUC) for the glucose tolerance test
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
3
was calculated using the trapezoidal method. Percentage reduction in AUC glucose
in treated animals was determined relative to the vehicleꢀtreated group. Students’s
tꢀtest was used for comparisons between two groups. Analysis of variance (ANOVA)
test was performed for comparisons among more than two groups. P < 0.05 was
considered to be statistically significant.
7
. Chemistry.
Chemistry: Reagents and Methods. All commercial chemicals and solvents are
reagent grade and were used without further purification unless otherwise noted. All
1
reactions are stirred magnetically. H (400 MHz) NMR spectra are recorded on a
1
Bruker AVꢀ400 instrument in CDCl or DMSOꢀd . Some H (500 MHz) NMR spectra
3
6
are recorded on a Varian InNovaꢀ500 spectrometer in DMSOꢀd
6
. Chemical shifts are
given in parts per million (ppm) downfield from internal reference Me
4
Si in δ units,
and coupling constants (J) are given in hertz (Hz). Selected data are reported in the
following manner: chemical shift, multiplicity, coupling constant. Lowꢀresolution
mass spectra are performed on a liquid chromatographyꢀmass spectrometer (LCꢀMS),
o
using an Agilent 1200 series with an Eclipse XDBꢀC18 column at 25 C. The gradient
solvent system is from 80% A and 20% B to 10% A and 90% B (A = 10 mM
2 3
HCOONH aqueous solution, B = CH CN) for 15 min with a flow rate of 0.4 mL/min.
ACS Paragon Plus Environment

Page 29 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
High resolution mass (HRMS, ESI) spectral data are acquired by use of a bruker
daltonics 7T mass spectrometer in positive ion mode. All final products are purified
to > 95% purity by HPLC analysis, using an Essentia LCsolution 15C with a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
o
Diamonsil C column at 35 C. The gradient solvent system is from 80% A and 20%
1
8
B to 100% A and 0% B (20 min) or 20% A and 80% B to 80% A and 20% B (15 min)
to 100% A and 0% B (10 min) (A = CH CN, B = H O) with a flow rate of 0.5 mL/min.
3
2
Melting point was tested using a WRSꢀ1B digital melting point apparatus. Flash
chromatographic separations are performed using silica gel (particle size 0.040ꢀ0.045
mm). Solvents are removed by rotary evaporation under vacuum using a standard
rotovap equipped with a cold alcohol condenser. All filtration are performed with a
vacuum unless otherwise noted.
26
General procedure for compounds 38 and 41a-e. To a solution of MeOH, water,
and 2.5N NaOH (1.2 equiv., VMeOH/Vwater/VNaOH = 1/1/1) was added a solution of
benzyaldehyde 37 or 40a-e (1.0 equiv.) and MeNO
2
(1.2 equiv.) in MeOH (60 mL)
drop wise over 30ꢀ60 min, while the internal temperature was maintained between 5
o
o
and 10 C. Then, the reaction solution was agitated another 30 min at 0ꢀ10 C and
added dropwise to a solution of ZnCl (5.0 equiv.) in concd HCl (13 mL) and water
2
(17 mL) for 2ꢀ4 h. After addition, the slurry was allowed to warm to ambient
temperature and stirred for 1 h before filtration. The filter cake was washed with 40%
MeOH aqueous solution and vacuum dried at ambient temperature to give light
yellow products with 75.0%ꢀ81.7% yield.
ACS Paragon Plus Environment