Design, synthesis, biological evaluation, common feature pharmacophore model and molecular dynamics simulation studies of ethyl 4-(phenoxymethyl)-2-phenylthiazole-5-carboxylate as Src homology-2 domain containing protein tyrosine phosphatase-2 (SHP2) inhi

Article abstract of DOI:10.1080/07391102.2020.1726817

SHP2 is a non-receptor protein tyrosine phosphatase (PTP) encoded by the PTPN11 gene involved in cell death pathway (PD-1/PD-L1) and cell growth and differentiation pathway (MAPK). Moreover, mutations in SHP2 have been implicated in Leopard syndrome (LS),

Full text of DOI:10.1080/07391102.2020.1726817

Journal of Biomolecular Structure and Dynamics
ISSN: 0739-1102 (Print) 1538-0254 (Online) Journal homepage: https://www.tandfonline.com/loi/tbsd20
Design, synthesis, biological evaluation,
Common feature pharmacophore model and
molecular dynamics simulation studies of ethyl 4-
(phenoxymethyl)-2-phenylthiazole-5-carboxylate
as Src Homology-2 Domain Containing Protein
Tyrosine Phosphatase-2 (SHP2) inhibitors
Jingwei Wu, Weiya Li, Zhihui Zheng, Xinhua Lu, Huan Zhang, Ying Ma &
Runling Wang
To cite this article: Jingwei Wu, Weiya Li, Zhihui Zheng, Xinhua Lu, Huan Zhang, Ying Ma &
Runling Wang (2020): Design, synthesis, biological evaluation, Common feature pharmacophore
model and molecular dynamics simulation studies of ethyl 4-(phenoxymethyl)-2-phenylthiazole-5-
carboxylate as Src Homology-2 Domain Containing Protein Tyrosine Phosphatase-2 (SHP2)
inhibitors, Journal of Biomolecular Structure and Dynamics, DOI: 10.1080/07391102.2020.1726817
To link to this article: https://doi.org/10.1080/07391102.2020.1726817
Accepted author version posted online: 10
Feb 2020.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=tbsd20

Design, synthesis, biological evaluation，Common
feature pharmacophore model and molecular
dynamics simulation studies of ethyl
4-(phenoxymethyl)-2-phenylthiazole-5-carboxylate as
Src Homology-2 Domain Containing Protein Tyrosine
Phosphatase-2 (SHP2) inhibitors
a, ‡
a, ‡
b
b
c
a,
Jingwei Wu , Weiya Li , Zhihui Zheng , Xinhua Lu , Huan Zhang , Ying Ma *,
a,
Runling Wang *
a
Tianjin Key Laboratory on Technologies Enabling Development of Clinical Therapeutics and
Diagnostics (Theranostics), School of Pharmacy, Tianjin Medical University, Tianjin 300070,
China
b
New Drug Research and Development Center of North China Pharmaceutical Group Corporation,
National Microbial Medicine Engineering and Research Center, Hebei Industry Microbial
Metabolic Engineering &Technology Research Center, Key Laboratory for New Drug Screening
Technology of Shijiazhuang City, Shijiazhuang, Hebei, China.
c
Department of Pharmacy, Tianjin Medical University General Hospital, Tianjin 300052, China;
Tel/Fax: +86-22-60362235;
‡
These authors contributed equally to this work
*
Corresponding authors: Ying Ma, E-mail: maying@tmu.edu.cn, Runling Wang ,
E-mail:wangrunling@tmu.edu.cn

Keywords: SHP2; inhibitors; common feature pharmacophore; synthesis; molecular
dynamics simulation
Abstract
SHP2 is a non-receptor protein tyrosine phosphatase (PTP) encoded by the PTPN11
gene involved in cell death pathway (PD-1/PD-L1) and cell growth and differentiation
pathway (MAPK). Moreover, mutations in SHP2 have been implicated in Leopard
syndrome(LS),Noonan syndrome (NS), juvenile myelomonocytic leukemia (JMML)
and several types of cancer and solid tumors. Thus, SHP2 inhibitors are much needed
reagents for evaluation of SHP2 as a therapeutic target. A series of novel ethyl
4
-(phenoxymethyl)-2-phenylthiazole-5-carboxylate derivatives were designed and
synthesized, and their SHP2 inhibitory activities (IC ) were determined. Among the
5
0
desired compounds, 1d shares the highest inhibitory activity (IC = 0.99 μM) against
5
0
SHP2. Additionally, a common feature pharmacophore model was established to
explain the structure activity relationship of the desired compounds. Finally,
molecular dynamics simulation was carried out to explore the most likely binding
mode of compound 1d with SHP2. In brief, the findings reported here may at least
provide a new strategy or useful insights in discovering novel effective SHP2
inhibitors.
List of abbreviations
ANM (Anisotropic Network Mode); DMF (N,N-dimethylformamide); ESI
(electro-spray ionization); JMML (juvenile myelomonocytic leukemia); LS (Leopard

syndrome); MD (molecular dynamics); MM-PBSA (molecular mechanics Poisson
Boltzmann surface area); NMR (nuclear magnetic resonance); NS (Noonan
syndrome); PDB (protein data bank); PTKs (protein tyrosine kinases); PTPs (protein
tyrosine phosphatases); pNPP (para-nitrophenyl phosphate); RMSD (root means
square deviation); RMSF (root mean square fluctuation); SH2 (Src homology-2);
SHP2 (Src homology-2 domain containing protein tyrosine phosphatase-2); SAR
(structure activity relationship); TLC (thin-layer chromatography);
1
Introduction
Lots of fundamental cellular events are regulated by protein dephosphorylation and
phosphorylation controlled by protein-tyrosine phosphatases (PTPs) and
protein-tyrosine kinases (PTKs), respectively(Taute et al., 2019). Not surprisingly,
disturbance of the normal balance between PTK and PTP activity can result in
abnormal tyrosine phosphorylation(Verma & Sharma, 2018), which has been linked to
various human diseases, including cancer, cardiovascular and neurological disorders,
metabolic and autoimmune diseases(Hendriks, Bourgonje, Leenders, & Pulido, 2018;
Khan, Rhett, & O'Bryan, 2019). Because of opposing the action of PTKs, PTPs had
been considered to be prime suspects for potential tumor suppressor genes(Ruckert,
de Andrade, Santos, & Silveira, 2019). Hence, protein tyrosine phosphatases have
been identified as high value therapeutic targets(Neel & Tonks, 2016; Van
Huijsduijnen, Bombrun, & Swinnen, 2002).

The Src homology-2 domain (SH2) containing protein tyrosine phosphatase-2 (SHP2),
is a non-receptor protein tyrosine phosphatase (PTP) encoded by the PTPN11
gene(Chan & Feng, 2006). It was the first phosphatase to be recognized as a bonafide
oncogene and plays a crucial role in regulation of signaling events such as cell death
pathway (PD-1/PD-L1) and cell growth and differentiation pathway
(MAPK)(Matozaki, Murata, Saito, Okazawa, & Ohnishi, 2009). SHP2 is expressed in
every tissue and contains two N-terminal Src homology-2 (SH2) domains, a
C-terminal domain and a PTP domain. Structural studies have revealed that SHP2
adopts an auto inhibited conformation under basal conditions, which associates the
PTP and SH2 domains and leads to blockage of the catalytic site(Garcia Fortanet et al.,
2
016). Activation of SHP2 occurs via binding of phosphoproteins to the SH2 domain
following cellular stimulation disrupts this interaction, relieving auto-inhibition and
activating cancer dependent phosphatase(X Sun et al., 2018). Mutations in SHP2 that
lead to alterations in the catalytic activity have been implicated in Leopard
syndrome(LS), Noonan syndrome (NS), juvenile myelomonocytic leukemia (JMML)
and several types of cancer and solid tumors(Y.-N. P. Chen et al., 2016; Lawrence et
al., 2008). Moreover, SHP2 has been involved in the pathogenesis of gastric
carcinoma caused by helicobacter pylori(Hatakeyama, 2004). Mutations in PTPN11
associated with JMML are often close to the N-SH2 domain and disrupt the
auto-inhibitory closed conformation, leading to the increase of PTP activity(Xiaojun
Sun et al., 2018). Therefore, down-regulating the PTP activity using SHP2 inhibitors
represents a promising therapeutic method(Zeng et al., 2014).

The organometallic compound sodium stibogluconate(Pathak & Yi, 2001) (Figure 1)
is reported as an irreversible inhibitor of SHP2 in vitro, which induces tyrosine
phosphorylation rapidly and enhances signaling by various cytokines. However, it is
associated with a wide range of serious side effects, including hepato-, cardio-, and
phlebo-toxicity. Screening of compounds in the National Cancer Institute (NCI)
diversity set chemical library led to the discovery of compound NSC-87877(L. Chen
et al., 2006) (Figure 1) with an IC50 of 0.3 μM against SHP2 in vitro. It shows
selectivity over a series of other PTPs, but also inhibits DUSP26. PHPS1 (Hellmuth et
al., 2008)(Figure 1) is identified by a high-throughput virtual screen against the SHP2
catalytic pocket and inhibits SHP2 with an IC of 0.7 μM. Screening of a library
5
0
based on an indole salicylic acid scaffold led to the identification of Compound
II-B08(X. Zhang et al., 2010) (Figure 1), which inhibits SHP2 with an IC of 5.5 μM
5
0
and approximately threefold selectivity over PTP1B and SHP1. Another related
irreversible inhibitor of SHP2 is TPI-1(Kundu et al., 2010), which is an analogue of
cryptotanshinone and has an IC of 0.4 μM. But, it is likely to cause toxicity because
5
0
of generation of reactive oxygen species and the oxidation of the catalytic cysteine
residue of SHP2(Butterworth, Overduin, & Barr, 2014). Duan et al.(Duan et al., 2014)
identified Comp#1 (Figure 1), a benzo[4,5]thiazolo[2,3-c][1,2,4]triazole derivative
with an IC of 4.31 μM against SHP2. Recently, Chen et al.(Y. -N. P. Chen et al., 2016)
5
0
reported a highly potent (IC =0.071 μM) SHP2 inhibitor, SHP099, an allosteric
5
0
modulator that stabilizes the auto-inhibited conformation of SHP2 (Figure 1).
Comp#2, a novel allosteric inhibitor of SHP2 reported by Xie et al.(Xie et al., 2017),

has an IC of 0.7 μM for SHP2 with more than 30-fold selectivity over other PTPs.
5
0
RMC-4550(Koltun et al., 2018), a potent and selective SHP2 allosteric inhibitor
IC =0.538 nM), possesses an overwhelming advantage over SHP099 in terms of in
(
5
0
vitro activity. Patrick Sarver et al.(Sarver et al., 2019) reported a screening method
that led to the identification of SHP394, an orally efficacious inhibitor of SHP2, with
high lipophilic efficiency, improved potency, and enhanced pharmacokinetic
properties.
Comp#3,
,4,6-trihydroxy-5-oxo-5H-benzo[7]annulene reported by Kim et al.(Kim et al.,
020), was obtained as an initial hit with an IC of 0.097 μM from highthroughput
a
novel
potent
SHP2
inhibitor
containing
3
2
5
0
screening (HTS) study. A number of SHP2 inhibitors have been identified in recent
years, however, the challenge is the lacking of a sufficiently well-defined SHP2
inhibitor that was approved to go to the market(Garcia Fortanet et al., 2016).
In this research, we designed the novel classes of SHP2 inhibitors by means of virtual
screening, molecular docking and de novo design. With the aim to discover some
potential SHP2 inhibitors, we evaluated the biological activity of the compounds in
the ZINC chemical library that were obtained by virtual screening. Fortunately,
compound ZINC169503038 was found to have an IC50 value of 4.30 μM against
SHP2 in vitro enzyme activity. Encouraged by the promising finding, we further
designed a series of ethyl 4-(phenoxymethyl)-2-phenylthiazole-5-carboxylate
derivatives by means of scaffold hopping. Then, the target compounds were
synthesized and their SHP2 inhibitory activities (IC ) were determined. In order to
5
0
obtain additional information of the relationship between structure and activity

develop more potent SHP2 inhibitors, common feature pharmacophore studies were
performed. Finally, Molecular dynamic simulation study was carried out to explore
the affinity of SHP2 and the inhibitor.
2
Materials and Methods
The crystal structure of SHP2 protein was downloaded from PDB Bank. Our
calculations were carried out on Dell Precision TM T5500 computer with Discovery
Studio 3.5 software package (http://accelrys.com/) and Gromacs 4.5.5.
2
2
.1 Compounds design
.1 .1 Virtual screening
Virtual screening by molecular docking, using a protein with an experimentally
determined structure as a target, has become an established means for lead discovery
and for enhancing efficiency in structure optimization(Monsen et al., 2017; J.-W. Wu
et al., 2014). In the paper, docking-based virtual screening was carried out through
LibDock module (Accelrys Discovery Studio package), which has been applied to the
GSK validation data set for its efficiency(Rao, Head, Kulkarni, & LaLonde, 2007).
The receptor crystal structure of SHP2 (PDB ID: 3O5X) was downloaded from the
protein data bank (PDB)(Prlić et al., 2016). Using the protocol of “Prepare Protein”,
the 3O5X was prepared by removing alternate conformations, deleting water and
adding the hydrogen atoms. The ligands used for virtual screening were derived from
ZINC database and were prepared by “Prepare Ligand”protocol. The binding sites
sphere was defined using “Find Site From Current Selection” based on the key
residues in the binding pocket, such as P-loop (residues 458-465), the Q loop

(residues 501-507) and the pY recognition loop (residues 277-284) (Zeng et al., 2014).
All prepared ligands were docked to the binding pocket of diverse receptor
conformations through LibDock module. According to the docking scores, the best
candidates were purchased according to the ZINC code and evaluated the
biological activity against SHP2 in vitro enzyme activity. The most potent SHP2
inhibitor was identified as potential scaffold used for the following research of scaf1.2
fold hopping.
2
.1.2 Replace Fragment
Replace Fragment” strategy has been used successfully in the design of inhibitors
“
for many pharmaceutically relevant targets(Scott, Coyne, Hudson, & Abell, 2012).
The aim of “Replace Fragment” is to discover structurally novel compounds starting
from known active compounds by replacement of fragment structure and a lot of
important drug discoveries were made by modifying structures of known drugs, such
as zopiclone, quinpirole and etoricoxib(Böhm, Flohr, & Stahl, 2004; Wang et al.,
2
016). In our studies, repalcement was performed by “Replace Fragment” protocol of
Discovery Studio 3.5. Firstly, build the fragment library that used for “Replace
Fragment”. Secondly, choose the possible cores that would be replaced by the
prepared fragments. Lastly, replace the original fragment and identify the isosteric
molecules with novel structure. When these steps were complete, the molecules were
docked into the receptor pockets by “CDOCKER” module and the molecules perform
higher binding affinities were selected for the following chemical synthesis(Mizwicki
et al., 2012).

2
.2 ADMET Prediction
The ADMET (absorption, distribution, metabolism, excretion, and toxicity) properties
are a crucial aspect of clinical candidate quality. which were performed using DS v3.5
software to assess the pharmacokinetic properties of the compounds. Some important
ADMET descriptors include log P, the aqueous solubility(Cheng & Merz, 2003),
human intestinal absorption (HIA)(Egan & Lauri, 2002; Egan, Merz, & Baldwin,
2
000),cytochrome P450 (CYP450) 2D6 inhibition(Susnow & Dixon, 2003; Wesson &
Eisenberg, 1992), hepatotoxicity(Cheng & Dixon, 2003; Xia, Maliski, Gallant, &
Rogers, 2004), blood brain barrier penetration (BBB)(Egan, Walters, & Murcko, 2002)
and plasma protein binding(Dixon & Merz, 2001; Votano et al., 2006).
2
2
.3 Chemistry
.3.1 General
All the reagents were purchased from commercial suppliers and were used
without further purification. All the reactions were monitored by thin-layer
chromatography (TLC) on silica gel precoated F254 Merck plates, and spots were
examined under UV light (254 nm or 365 nm). All column chromatography was
performed using 200-300 mesh silica gel. Melting points were measured with an X-6
1
13
micro melting point apparatus and are uncorrected. H NMR spectra and C NMR
spectra were recorded on a Bruker AV400 NMR spectrometer with DMSO-d as
6
solvent and known chemical shifts of residual proton signals of deuterated solvents as
internal standard. High-resolution mass spectra (HR-MS) were determined with an
Agilent Q-TOF 6510 mass spectrometer using direct injection method, and
electro-spray ionization (ESI) was used as an ionization technique in either positive
(ESI+) or negative (ESI-) mode. The synthetic route to the desired products 1a-1j is

illustrated in Schemes 1.
Reagents and conditions: (ⅰ).ethyl 2-chloro-3-oxobutanoate, EtOH, reflux; (ⅱ). NBS, BPO,
CCl , reflux; (ⅲ). 4-substitute phenol, K CO , DMF, 80℃.
4
2
3
2
.3.2 General procedure for the synthesis of compounds 3a-3efrom compounds 2a-2e
The synthetic procedure of compounds 3a-3e: To a stirred solution of 2a-2e (20 mmol)
in EtOH (40 mL) and was added ethyl 2-chloro-3-oxobutanoate (20 mmol, 3.29g), the
reaction mixture was heated at reflux in N atmosphere until the completion of
2
reaction as indicated by TLC analysis (typically within 5 h).
The reaction mixture was poured into ice-water (100 mL) and the mixture thus
obtained was extracted with CH Cl (100 mL × 3). The combined organic phases were
2
2
washed with saturated brine (300 mL), dried over anhydrous Na SO and evaporated
2
4
on a rotary evaporator to afford a residue, which was purified by column
chromatography through a short silica gel column to yield 3a-3e.
Ethyl 2-(2-bromophenyl)-4-methylthiazole-5-carboxylate (3a), white solid, 4.37 g
1
(
67%), m.p. 93.1-94.6°C. H NMR (DMSO-d , 400 MHz) δ: 1.32 (t, J = 7.2 Hz, 3H),
6
2
.73 (s, 3H), 4.33 (q, J = 7.2 Hz, 2H), 7.46-8.16 (m, 4H).
Ethyl 2-(4-methoxyphenyl)-4-methylthiazole-5-carboxylate (3b), white solid, 3.94 g
1
(
71%), m.p. 78.2-78.9°C. H NMR (DMSO-d , 400 MHz) δ: 1.31 (t, J = 7.2Hz, 3H),
6
2
.68 (s, 3H), 3.84 (s，3H), 4.30 (q, J = 7.2 Hz, 2H), 7.08 (d, J = 8.4Hz, 2H), 7.95 (d, J
8.4Hz, 2H).
=
Ethyl 4-methyl-2-(3-phenoxyphenyl)thiazole-5-carboxylate (3c), white solid, 4.68 g
1
(
69%), m.p. 72.1-73.4°C. H NMR (DMSO-d , 400 MHz) δ: 1.31 (t, J=7.2Hz, 3H),
6
2
.67 (s, 3H), 4.31 (q, J=7.2Hz, 2H), 7.10-7.52 (m, 9H).
Ethyl 2-(4-fluorophenyl)-4-methylthiazole-5-carboxylate (3d), white solid, 3.82 g

1
(
72%), m.p. 88.8-89.9°C. H NMR (DMSO-d , 400 MHz) δ: 1.30 (t, J=7.2Hz, 3H),
6
1
2
.88 (s, 3H), 4.30 (q, J=7.2Hz, 2H), 7.60-8.03 (m, 4H). The H NMR data were in
good agreement with those reported(Sierra et al., 2007).
Ethyl 2-(4-bromophenyl)-4-methylthiazole-5-carboxylate (3e), white solid, 4.76 g
1
(
73%), m.p. 74.1-75.2°C. H NMR (DMSO-d , 400 MHz) δ: 1.50 (t, J=7.2Hz, 3H),
6
1
2
.89 (s, 3H), 4.50 (q, J=7.2Hz, 2H), 7.40-8.10 (m, 4H). The H NMR data were in
good agreement with those reported(Tiperciuc et al., 2012).
.3.3 General procedure for the synthesis of compounds 4a-4e from compounds 3a-3e
2
To a stirred solution of 3a-3e (10 mmol) and NBS (1.96 g, 11 mmol) in CCl (50 mL)
4
was added benzoyl peroxide (0.24 g, 1 mmol), the reaction mixture was heated at
reflux in N atmosphere until the completion of reaction as indicated by TLC analysis
2
(typically within 8 h).
On cooling to room temperature, the reaction mixture was poured into ice-water (100
mL) and the mixture thus obtained was extracted with CH Cl (50 mL × 3). The
2
2
combined extracts were washed with 5% brine (100 mL), dried (Na SO ) and
2
4
evaporated on a rotary evaporator to give a residue, which was purified by column
chromatography followed by trituration with n-hexane/EtOAc (1/3 by v/v) to afford
4
a-4e.
Ethyl 4-(bromomethyl)-2-(2-bromophenyl)thiazole-5-carboxylate (4a), white solid,
1
3
4
.16 g (78%), m.p. 127.5-128.4°C. H NMR (DMSO-d , 400 MHz) δ: 1.34 (t, 3H),
6
.38 (q, 2H), 5.05 (s, 2H), 7.52 (m, 1H), 7.60 (m, 1H), 7.88 (m, 1H), 8.17 (m, 1H).
Ethyl 4-(bromomethyl)-2-(4-methoxyphenyl)thiazole-5-carboxylate (4b), white solid,
1
2
2
7
.88 g (81%), m.p. 92.8-93.8°C. H NMR (DMSO-d , 400 MHz) δ: 1.34 (t, J =
6
.4Hz,3H), 3.86 (s, 3H), 4.35 (q, J=2.4 Hz, 2H), 5.00 (s, 2H), 7.10 (d, J=9.2Hz, 2H),
.98 (d, J=9.2Hz, 2H).

Ethyl 4-(bromomethyl)-2-(3-phenoxyphenyl)thiazole-5-carboxylate (4c), white solid,
1
3
3
.55 g (85%), m.p. 110.1-111.3°C. H NMR (DMSO-d , 400 MHz) δ: 1.34(t, J=6.8Hz,
6
H), 4.36 (q, J=6.8Hz, 2H), 5.00 (s, 2H), 7.12 (m, 2H), 7.21-7.24 (m, 2H), 7.44-7.48
(m, 2H), 7.55-7.57 (m, 2H), 7.72-7.90 (m, 1H).
Ethyl 4-(bromomethyl)-2-(4-fluorophenyl)thiazole-5-carboxylate (4d), white solid,
1
2
4
.58 g (75%), m.p. 135.8-136.9°C. H NMR (DMSO-d , 400 MHz) δ: 1.33 (t, 3H),
6
.35 (q, J=7.2Hz, 2H), 5.10 (s, 2H), 8.21-8.42 (m, 4H).
Ethyl 4-(bromomethyl)-2-(4-bromophenyl)thiazole-5-carboxylate (4e), white solid,
1
3
4
2
.32 g (82%), m.p. 141.7-143.0°C. H NMR (DMSO-d , 400 MHz) δ: 1.36 (t, 3H),
6
.32 (q, J=7.2Hz, 2H), 5.08 (s, 2H), 8.30-8.51 (m, 4H).
.3.4 General procedure for the synthesis of compounds 1a-1i.
A stirred mixture of 4a-4e (5 mmol), 4-substitute phenol (6 mmol) and K CO (1.38 g,
2
3
1
0 mmol) in DMF (20 mL) was heated at 80°C under N atmosphere until the
2
completion of reaction as indicated by TLC analysis (typically within 10 h).
On cooling to room temperature, the reaction mixture was poured into ice-water (30
mL). The mixture thus obtained was acidified to pH of 7 with concentrated
hydrochloric acid and extracted with CH Cl (30 mL × 3). The combined extracts
2
2
were washed with 5% brine (50 mL), dried and evaporated to afford the crude product,
which was purified by column chromatography through a short silica gel column
followed by trituration with EtOAc/n-hexane to yield 1a-1i.
Ethyl 2-(2-bromophenyl)-4-((4-chlorophenoxy) methyl) thiazole-5-carboxylate (1a),
1
white solid, 1.95 g (85%), m.p. 67.5-68.7°C. H NMR (DMSO-d , 400 MHz) δ: 1.28
6
(t, J=7.2Hz, 3H), 4.34 (q, J=7.2Hz, 2H), 5.49 (s, 2H), 7.07 (d, J=4.8Hz, 2H), 7.09 (d,
1
3
J=4.8Hz, 2H), 7.33-7.36 (m, 2H), 7.46-7.50 (m, 2H). C NMR (100 MHz, DMSO) δ:
1
66.23, 160.52, 157.17, 155.88, 134.26, 132.27, 132.01, 131.39, 129.19, 129.19,

1
28.18, 126.35, 124.67, 121.06, 116.38, 116.38, 63.72, 61.78, 13.94. MS, calcd for
+
C H BrClNO S ([M+Na] ) 475.9522, found 475.9521.
19
15
3
Ethyl
4-((4-methoxyphenoxy)methyl)-2-(4-methoxyphenyl)thiazole-5-carboxylate
1
(
1b), white solid, 1.62 g (81%), m.p. 70.8-72.4°C. H NMR (DMSO-d , 400 MHz) δ:
6
1
6
.28 (t, J=7.2Hz, 3H), 3.71 (s, 3H), 3.85 (s, 3H), 4.32 (q, J=7.2Hz, 2H), 5.38 (s, 2H),
.87 (d, J=9.2Hz, 2H), 6.98 (d, J=9.2Hz, 2H), 7.09 (d, J=8.4Hz, 2H), 7.97 (d,
1
3
J=8.4Hz, 2H). C NMR (100 MHz, DMSO) δ: 169.69, 161.93, 160.64, 157.79,
53.57, 152.39, 152.05, 151.09, 128.38, 124.61, 123.84, 115.64, 115.55, 114.64,
14.49, 114.49, 63.91, 61.52, 55.41, 55.23, 13.92. MS, calcd for C H NO S
1
1
21
21
5
+
(
[M+Na] ) 422.1038, found 422.1033.
Ethyl 2-(4-methoxyphenyl)-4-((p-tolyloxy)methyl)thiazole-5-carboxylate (1c), white
1
solid, 1.69 g (88%), m.p. 97.8-99.2°C. H NMR (DMSO-d , 400 MHz) δ: 1.26 (t,
6
J=7.2Hz, 3H), 2.24 (s, 3H), 3.84 (s, 3H), 4.31 (q, J=7.2Hz, 2H), 5.39 (s, 2H), 6.94 (m,
1
3
2
1
H), 7.09 (m, 4H), 7.97 (m, 2H). C NMR (100 MHz, DMSO) δ 169.72, 161.95,
60.66, 157.72, 156.27, 129.79, 129.79, 129.52, 128.41, 128.41, 124.60, 123.89,
114.69, 114.69, 114.37, 114.37, 63.35, 61.55, 55.45, 20.04, 13.95. MS, calcd for
+
C H NO S ([M+Na] ) 406.1089, found 406.1083.
21
21
4
Ethyl 4-((4-chlorophenoxy)methyl)-2-(4-methoxyphenyl)thiazole-5-carboxylate (1d),
1
white solid, 1.60 g (79%), m.p. 95.3-97.0°C. H NMR (DMSO-d , 400 MHz) δ: 1.26
6
(t, J=7.2Hz, 3H), 3.84 (s, 3H), 4.31 (q, J=7.2Hz, 2H), 5.44 (s, 2H), 7.08 (m, 4H), 7.35
1
3
(
m, 2H), 7.95 (m, 2H). C NMR (100 MHz, DMSO) δ: 169.81, 161.97, 160.61,
57.22, 157.15, 129.22, 129.22, 128.40, 128.40, 124.61, 124.54, 124.04, 116.34,
16.34, 114.68, 114.68, 63.72, 61.58, 55.44, 13.93. MS, calcd for C H1 ClNO S
1
1
2
0
8
4
+
(
[M+ Na] ) 426.0543, found 426.0536.
Ethyl 4-((4-methoxyphenoxy)methyl)-2-(3-phenoxyphenyl)thiazole-5-carboxylate(1e),

1
white solid, 1.79 g (86%), m.p. 116.7-118.4°C. H NMR (DMSO-d , 400 MHz) δ:
6
1
4
.26 (t, J=7.2Hz, 3H), 3.70 (s, 3H), 4.30 (q, J=7.2Hz, 2H), 5.36 (s, 2H), 6.84-6.96 (m,
13
H), 7.10-7.23 (m, 4H), 7.42-7.46 (m, 2H), 7.52-7.58 (m, 2H), 7.75-7.77 (m, 1H). C
NMR (100 MHz, DMSO) δ: 168.77, 160.50, 157.87, 157.63, 155.84, 153.60, 152.28,
33.56, 131.30, 130.25, 130.25, 125.38, 124.18, 121.75, 121.43, 119.23, 119.23,
15.66, 115.66, 115.43, 114.50, 114.50, 63.90, 61.76, 55.27, 13.94. MS, calcd for
1
1
+
C H NO S ([M+ Na] ) 484.1195, found 484.1188.
26
23
5
Ethyl 2-(4-fluorophenyl)-4-((4-methoxyphenoxy)methyl)thiazole-5-carboxylate (1f),
1
white solid, 1.61 g (83%), m.p. 73.5-74.5°C. H NMR (DMSO-d , 400 MHz) δ: 1.26
6
(t, J=7.2Hz, 3H), 3.70 (s, 3H), 4.32 (q, J=7.2Hz, 2H), 5.39 (s, 2H), 6.86-6.99 (m, 4H),
1
3
7
1
.36-7.41 (m, 2H), 8.07-8.10 (m, 2H). C NMR (100 MHz, DMSO) δ: 168.46,
65.19, 162.70, 159.13 (d, J = 269.0 Hz), 153.60, 152.35, 129.03 (2C, d, JC-F = 9.0
C-F
Hz), 128.48 (d, JC-F = 3.0 Hz), 125.02, 116.34 (2C, d, JC-F = 22.3 Hz), 115.56, 115.56,
+
1
14.46, 114.46, 63.88, 61.63, 55.19, 13.85. MS, calcd for C H FNO S ([M+ Na] )
2
0
18
4
4
10.0838, found 410.0835.
Ethyl 4-((4-chlorophenoxy)methyl)-2-(4-fluorophenyl)thiazole-5-carboxylate (1g),
1
white solid, 1.66 g (85%), m.p. 99.0-100.6°C. H NMR (DMSO-d , 400 MHz) δ: 1.27
6
(t, J=7.2Hz, 3H), 4.32 (q, J=7.2Hz, 2H), 5.46 (s, 2H), 7.07-7.09 (m, 2H), 7.34-7.41(m,
1
3
4
2
1
H), 8.06-8.07 (m, 2H). C NMR (100 MHz, DMSO) δ: 168.62, 163.98 (d, J
=
C-F
50.3 Hz), 160.46, 157.17, 157.17, 129.22, 129.22, 129.12 (2C, d, JC-F = 9.0 Hz),
28.45 (d, JC-F = 3.0 Hz), 125.28, 124.66, 116.45 (2C,d, J C-F = 20.9 Hz), 116.34,
+
1
16.34, 63.69, 61.74, 13.90. MS, calcd for C H ClFNO S ([M+ Na] ) 414.0343,
1
9
15
3

found 414.0337.
Ethyl 2-(4-bromophenyl)-4-((p-tolyloxy)methyl)thiazole-5-carboxylate (1h), white
1
solid, 1.90 g (88%), m.p. 91.9-92.8°C. H NMR (DMSO-d , 400 MHz) δ: 1.27 (t,
6
J=7.2Hz, 3H), 2.24 (s, 3H), 4.32 (q, J=7.2Hz, 2H), 5.40 (s, 2H), 6.92 (d, J=8.4Hz,
1
3
2
H), 7.10 (d, J=8.4Hz, 2H), 7.75 (d, J=8.4Hz, 2H), 7.96 (d, J=8.4Hz, 2H). C NMR
(100 MHz, DMSO) δ: 168.45, 160.48, 157.81, 156.21, 132.36, 132.36, 130.96, 129.81,
1
1
29.81, 129.59, 128.49, 128.49, 125.44, 125.17, 114.39, 114.39, 63.33, 61.78, 20.05,
+
3.93. MS, calcd for C H BrNO S ([M+ Na] ) 454.0088, found 454.0083.
2
0
18
3
Ethyl 2-(4-bromophenyl)-4-((4-chlorophenoxy)methyl)thiazole-5-carboxylate (1i),
1
white solid, 1.81 g (80%), m.p. 93.6-95.1°C. H NMR (DMSO-d , 400 MHz) δ: 1.27
6
(t, J=7.2Hz, 3H), 4.32 (q, J=7.2Hz, 2H), 5.45 (s, 2H), 7.08 (d, J=8.8Hz, 2H), 7.35 (d,
13
J=8.8Hz, 2H), 7.74 (d, J=8.4Hz, 2H), 7.95 (d, J=8.4Hz, 2H). C NMR (100 MHz,
DMSO) δ: 166.30, 160.54, 157.17, 155.94, 134.29, 132.36, 132.05, 131.44, 129.22,
1
29.22, 128.26, 126.40, 124.65, 121.07, 116.41, 116.41, 63.73, 61.81, 13.95. MS,
+
calcd for C H BrClNO S ([M+ Na] ) 473.9542, found 473.9537.
1
9
15
3
2
.4 Enzymatic assays
Human recombinant SHP2 (PTP-domain), LAR, SHP1, MEG2, CDC25B or TCPTP
was expressed in E coli and purifed by Ni-NTA affinity chromotagraphy in our
laboratory. The phosphatase activity was assayed using p-nitrophenyl phosphate
(pNPP) as a substrate at 37 °C in 60 μL buffer containing 2 mM EDTA, 2 mM
dithiothreitol and 60 mM calcium citrate. The sample was screened in a 96-well
format at 20 μM compound concentration. The reaction was initiated by the addition

of 0.1μg enzyme. After preincubation for 2 min at 37°C, 50 μL of buffer with 40 μL
pNPP was added and incubated. The reaction was quenched by adding 10μL of 0.2 N
NaOH. The amount of product p-nitrophenol was determined from the absorbance at
4
05 nm detected by a microplate spectrophotometer. The experiments were performed
in triplicate and inhibitory activity was expressed as half maximal inhibitory
concentration IC (J. Wu, Sun, Zhou, Ma, & Wang, 2019).
5
0
2
.5 Pharmacophore generation
The 3D-common feature pharmacophore hypothesis of these SHP2 inhibitors was
identified by applying the HipHop tool embedded in the Accelrys Discovery Studio
3
.5 software (http://accelrys.com/) to analyze the structure-activity relationships (SAR)
of SHP2 protein-ligand interactions.
At first, the 3D structures of chemical compounds were optimized with CHARMm
force field(Brooks et al., 2009) by applying the Full Minimization module embedded
in the Accelrys Discovery Studio 3.5 software. Based on the bioactivity of the
compounds, the molecular property Principal to the molecules was set as a value. “2”
meant that the molecular was active; “1” meant that the molecular was moderately
active; “0” meant that the molecular was inactive. MaxOmitFeat was used to indicate
how many features were allowed to miss for each molecule. Then, Molecular Overlay
was used to automatically align the conformational for each compound by 50% steric
filed and 50% electrostatic filed. In addition, the Input ligands conformation
generation option was set to “BEST”, the Maximum Conformations was set to 200,
the Energy Threshold was set to 10, and the other parameters were kept default values.

With the aid of Edit and Cluster Features tool, the list of pharmacophore feature types
for the compounds was represented. Four features, namely hydrogen bond acceptors
(
HBA), hydrogen bond doners (HBD), Ring_Aromatic (RA) and hydrophobe area
HYP), were used to build the SHP2 inhibitors hypotheses. The minimum numbers of
(
HBA, HBD, RA and HYP were set to 1 and the maximum numbers of them were set
to 2. Moreover, pharmacophore model computation was initialized using the HipHop
algorithm. Finally, pharmacophoric hypothesis was generated from these aligned
structures.
2
.6 Molecular Dynamics Simulations
The mechanism of how the compound inhibited the SHP2 protein at molecular level
were studied by applying the “GROMACS 4.5.5 package(Loeffler & Winn, 2012;
Pandey, Roy, & Doerksen, 2020)”with AMBER99B force field(Smith, Rao, Segelken,
&
Cruz, 2015) for 20 ns. The topology files were firstly generated. The protein
topology file was produced automatically and the ligand topology file were generated
by utilizing ACPYPE Server(W Sousa da Silva & Vranken, 2012) with the atom type
setting as default GAFF(Generalised Amber Force Field). Secondly, the box and
solvate was defined(Sadeghi-Kaji, Shareghi, Saboury, & Farhadian, 2020; H. Zhang
et al., 2020). The dodecahedron box was defined as the unit cell and was filled with
explicit single-point charge (SPC) water molecules, which were more than 3000 and
+
-
less than 4000. Thirdly, the Na /Cl ions were added to neutralize the system by
replacing some solvent molecules randomly. Fourthly, the system was minimized by
using the steepest decent method. The minimization was stopped when the maximum

force < 10.0 kJ/mol.The system was then heated to 310 K during a 200 ps NVT
simulation with 1 fs time step. The pressure was then equilibrated to 1 atm during a
5
00 ps NPT simulation with 2 fs time step. In both simulations, all heavy atoms were
2
position restrained with the force constant of 1000 kJ/(mol nm )(Mohammadi,
Shareghi, Akbar Saboury, & Farhadian, 2020). The position restraint was gradually
removed during a 1 ns simulation with a time step of 2 fs. After that, molecular
dynamics study was initiated by using the Particle Mesh Ewald algorithm
(PME)(Nocito & Beran, 2018) and (Uusitalo, Ingólfsson, Marrink, & Faustino, 2017)
with a time step of 2 fs. The short-range electrostatic cutoff and the short-range van
der Waals cutoff were 1.0 nm, respectively(Maroli & Kolandaivel, 2020).
Free energy were calculated on the MD trajectory using g_mmpbsa(Kumari, Kumar,
Open Source Drug Discovery, & Lynn, 2014). The calculation formula is as
follows:(Saiz-Urra, Cabrera, & Froeyen, 2011):
훥Gbind = Gcomplex– Greceptor − Gligand (1)
훥Gbind = Egas + G – T훥S (2)
sol
Egas = E + E + Eele (3)
int
vdw
Gsol = G + G (4)
GB
SA
GSA = 훾SASA (5)
where T is for temperature and S is for the total solute entropy; Egas signifies
gas-phase energy, which is the sum of internal energy (E ), electrostatic (E ), and
int
ele
van der Waals contributions (Evdw). The solvation free energy (G ) can be further
sol
decomposed into polar and nonpolar solvation states. The polar solvation contribution

(
GGB and G ) is determined by solving Poisson Boltzmann (PB) and Generalized
PB
Borne (GB) equations. The nonpolar solvation contribution (G ) is estimated using
SA
-1
-2
0
.0072 kcal mol A as the value for constant γ and the solvent-accessible surface
area (SASA) determined using a water probe radius of 1.4 Å. Dielectric constant
values for solute and solvent were set to 1 and 80, respectively(Mehla & Ramana,
2
3
3
016).
. Results and discussion
.1 Compounds design
As shown in Figure 2. The high-throughput virtual screen of approximately 3 million
compounds in ZINC database against the SHP2 catalytic domain led to the
identification of several candidates. These compounds were then purchased from
reagent companies and evaluated the biological activity against SHP2. Fortunately, we
discovered a potent SHP2 inhibitor, ZINC169503038, which inhibits SHP2 with an
IC of 4.30 μM. In the paper, scaffold hopping was performed by “Replace Fragment”
5
0
protocol of Discovery Studio 3.5. The first step was to replace scaffold A (colored in
red, Figure 2), which generated 5 scaffolds. The second step was to replace scaffold
B (colored in blue, Figure 2), which generated 4 scaffolds. Therefore, we have a total
of 20 different compounds for the ZINC169503038 derivatives thus generated. These
compounds were then re-docked into the binding pocket of diverse receptor (3O5X)
conformations through “CDOCKER” module. The “CDOCKER interaction energy”
was an important index for evaluating the binding affinities. The docking score of top

1
0 compounds (1a-1i) and ZINC169503038 were listed in Table 1, the compound 1d
(- CDOCKER interaction energy = 50.82 Kcal/moL) had the higher docking score
than ZINC169503038 (CDOCKER interaction energy = 30.95 Kcal/moL).
The two-dimensional (2D) diagram of compound 1d-3O5X interaction was shown in
Figure 3. Pink boxes such as Tyr279, Lys366, Trp423, Gly427, Arg465 and Gln510
represented hydrogen bonding, charge or polar interactions, while green boxes like
Val428, Cys459, Gly464, Ala509, His426, Thr507, Arg362, Lys364, Ser460, Gln506,
Ala461and Thr357 were involved in van der waals interactions. The 4-chlorophenyl
of compound 1d formed cation-Pi interaction with Lys366. Moreover, the N and O
atoms of compound 1d formed seven potential H-bonding with the side chains of
Glu510, Trp423 Arg465 and Lys366 (see blue dashed arrow directed towards the
electron donor in Figure 3). In short, compound 1d not only formed the cation-Pi
interaction but also more hydrogen bonds with the SHP2 protein, resulting in the
strong binding affinities.
Furthermore, the binding modes of 3O5X and compound 1d obtained from the
docking simulation was shown in Figure 4C, it was obvious that the key residues
(Arg362, Lys364, Lys366, His426, Arg465 and Gln510) for the binding interactions
between the ligand and the receptor were fully consistent with the previous results. As
shown in Figure 4C, the three O atoms of compound 1d forms six hydrogen bonds
with Gln510, Arg465 and Lys364 while the N atom of compound 1d interacts with
residue Gln510 via a strong H-bond. These H-bonds significantly enhance the

inhibitor-receptor
-(phenoxymethyl)-2-phenylthiazole-5-carboxylate derivatives complex within the
active site to a great extent.
interaction
and
stabilize
the
ethyl
4
3
.2 ADMET Pridiction
There are two important factors that affect oral bioavailability: one is Human
Intestinal Absorption (HIA), the other is solubility. Solubility plays a key role in the
pharmacological activity of a compound. HIA has a pronounced effect on the
therapeutic effect of drugs. All compounds (1a-1i) showed good human intestinal
absorption due to appropriate aqueous solubility and LogP. Plasma protein plays an
important role in drug distribution. All compounds (1a-1i) were found to be highly
bound with plasma protein. CYP2D6 is one of the important enzymes involved in
drug metabolism, which is associated with the metabolism and elimination of
approximately 25% of clinically used drugs. Nine compounds were predicted to be
non-inhibitors of cytochrome P450 2D6 (CYP2D6). For hepatotoxicity, all
compounds were predicted non-toxic. For brain/blood barrier (BBB), compound 1g,
1
h and 1i had a very good penetrant level, compound 1a, 1b, 1c and 1e had a good
penetrant level, while compound 1d and 1f showed a moderate penetrant level.
Therefore, as mentioned above, the values for the ADMET properties of all
compounds listed in Table 1 are within the acceptable range for human beings,
indicating these compounds found in this study can be utilized as candidates for the
purpose of developing new drugs.
3
.3 Chemistry

Firstly, the synthetic route to desired products 1a-1i was shown in Scheme 1.
Cyclization of benzothioamide derivatives (2a-2e) with ethyl 2-chloro-3-
oxobutanoate in EtOH at reflux smoothly afforded 3a-3e according to known
procedures. Secondly, bromination of ethyl 4-methyl-2-phenylthiazole-5- carboxylate
derivatives (3a-3e) with N-Bromosuccinimide in the presence of benzoyl peroxid as
catalyst in carbon tetrachloride at reflux afforded 4a-4e. Thirdly, O-alkylation of ethyl
4
-(bromomethyl) -2-phenylthiazole-5-carboxylate derivatives with substituted phenol
in the presence of K CO in DMF at 80°C produced the final desired products
2
3
1
a-1i(J.-w. Wu et al., 2019).
3
.4 In vitro SHP2 inhibitory activity
The results of in vitro inhibitory assay of 10 synthesized compounds (1a-1i) as well as
comp#1 as positive control against SHP2 were summarized in Table 2. Comp#1 was
identified by Duan et al. against the SHP2 catalytic pocket and inhibits SHP2 with an
IC of 4.31 μM (Figure 1). Most of the synthesized compounds (1a-1i) were more
5
0
active than the positive control except compound 1a and 1g. It was obvious that half
of the desired compounds (1b-1e) showed improved SHP2 inhibitory activity (IC50
values were range from 0.99 μM to 3.90 μM) as compared with ZINC169503038
(
IC =4.30 μM). The rest of the desired compounds (1f-1i) were only slightly less
50
active (IC values were range from 4.60 μM to 11.48 μM) than ZINC169503038
5
0
except 1a, displayed an IC50 of 35.56 µ M, which is 8-fold less potent than
ZINC169503038. As shown in Table 2, compounds 1b-1d showed higher and similar
inhibitory activity (IC values were range from 0.99 μM to 1.25 μM), leading to the
5
0

identification of 4-methoxyphenyl as the optimal scaffold A. It should be noted that
the compounds 1a (scaffold A= 2-bromophenyl, IC =35.56 μM) and 1f-1i (scaffold
5
0
A= 4-bromophenyl or 4-fluorophenyl, IC =4.6-11.48 μM) with halogen substituted
5
0
benzene exhibited lower IC50 values than ZINC169503038. These observations
strongly indicated that the scaffold A is very important t for the significant impacts on
inhibitory activity. Among the four scaffold B groups, 4-methoxyphenyl seemed the
worse one (1g < 1f, 1i < 1h).
As shown in Table 3, we also explored the selectivity of compound 1a-1i for other
phosphatases, such as src homologous phosphatase-1 (SHP1), homogeneous T cell
protein tyrosine phosphatase (TCPTP), maternal-effect germ-cell defective 2 (MEG2),
cell division cycle 25 homolog B (CDC25B) and leukocyte antigen-related
phosphatase (LAR). It could be seen from the Table 3 that all compounds (1a-1i)
seem to be inactive (IC > 100 μM) against SHP1, MEG2, CDC25B and LAR.
5
0
Compound 1c and 1d showed strong selectivity (more than 33-fold) for SHP2 over
TCPTP, compound 1f, 1g and 1h showed moderate selectivity (more than 3-fold)
while compound 1i showed poor selectivity (less than 2-fold). In short, most of
compounds (1a-1h) have some selective inhibition for SHP2 over SHP1, TCPTP,
MEG2, CDC25B and LAR. However, compound 1i shows no selectivity for SHP2
over TCPTP.
In summary, systematic SAR exploration of a ZINC169503038-based hit led to the
discovery of ethyl 4-((4-chlorophenoxy)methyl)-2-(4-methoxyphenyl)thiazole
-5-carboxylate (1d) as highly active SHP2 inhibitors, which was 4-fold more active

than parent ZINC169503038 in SHP2 inhibitory assay (IC values for 1d was 0.96
5
0
0
.99 μM, against SHP2 vs 4.30μM for ZINC169503038). The hypothesis was tested
in the following 3D-QSAR study.
.5 Pharmacophore hypotheses generated by the HipHop method
3
The 3D-common feature pharmacophore hypothesis of these SHP2 inhibitors was
identified by applying the HipHop tool to analyze the SAR of SHP2 protein-ligand
interactions. The reliability of the generated pharmacophore model was assessed by
aligning the docked pose of the reference compound 1d to the conformation generated
in Hiphop. Comparing the two conformations with compound 1d in the top ten
hypotheses and SHP2, the RMSD was found to be 0.38 Å, 0.67 Å, 0.93Å, 1.24 Å,
1
.57 Å, 1.89 Å, 1.91 Å, 2.10 Å, 2.10 Å, 2.46 Å, respectively. As shown in Figure 4,
comparing the conformations of compound 1d in the Hypo-1 and in the active sites of
SHP2, the RMSD was only 0.38 Å, indicating that the Hypo-1 was a highly reliable
pharmacophore model of SHP2 inhibitors. Thus, the top one hypothesis-Hypo-1 had
the least RMSD value and possessed the highest ranking score was chosen for the
further study. The generated 3D common feature pharmacophore hypothesis had two
HBAs, one RA, two HYPs. Fit value could provide information in whether the
molecule could fit into the pharmacophore hypothesis. As shown in Table 4, the fit
value of the most active compound 1d was 4.96, whereas, the least active compound
1
a had the fit value of 1.60. The most active compound 1d with IC 0.99 μM and
50
least active compound 1a with IC 50 35.56 μM were aligned with Hypo-1 to verify
whether the hypothesis could distinguish the active compounds from the inactive

compounds. It could be seen from Figure 5 the most active compound 1d mapped all
the features in Hypo-1, whereas the least active compound 1a only mapped the two
HBA and RA features and failed to map the two HYP features. In order to verify the
reliability of the Hypo-1, we compared the pharmcophore model with the active site
of SHP2. The two HBAs were pointed to the Arg465 and Gln 510, respectively, and
the RA was pointed to the Lys366, while one HYP was oriented towards the His426,
and the other HYP was pointed to the Arg362 and Lys364. Thus, the chemical features
of Hypo-1 could interact with key residues in the active site of SHP2. The generated
pharmacophore model revealed that the two HBAs, one RA, and two HYPs features
might play an important role in binding to the active site of SHP2. According to these,
two HBA atoms or groups at the specific positions were the essentials in the molecule
to bind to the SHP2 protein, while bulky aromatic moieties with appropriate active
shape in the molecule were the necessary to fit into the one RA and two HYPs
features.
The results of the compounds mapping onto features were given in Table 4. the 1b, 1c,
and
1d
compounds
were
discovered
as
the
most
4
-(phenoxymethyl)-2-phenylthiazole-5-carboxylate derivatives with the best fit value
of 4.80, 4.95, and 4.96, respectively, for the generated pharmacophore hypothesis,
fitting to all the mapped common features in the anticipated model with a specified
“Pharmprint” value of “11111”. The fit values and “Pharmprint” values of these SHP2
inhibitors based on the pharmacophoric features in the Hypo-1 model were given in
Table 4. The compounds having 4-methoxy group at the benzene ring in scaffold A,

for example compound 1b, compound 1c and compound 1d rather than holding a
-nitro, 3-phenoxy, and 4-bromol group at the benzene ring in scaffold A, such as
4
compond 1e and compound 1h possessed a higher fit values, meaning that they
possessed a better match with all the features in the pharmacophoric features. This
explains why compounds with 4-methoxy group at the benzene ring in scaffold A,
such as compound 1b, compound 1c and compound 1d had more significant
bioactivity than compounds holding a 4-nitro, 3-phenoxy, and 4-bromol group at the
benzene ring in scaffold A. Owing to matching with all the features except the HYP
features, the compounds with 2-bromol and 4-fluoro group at the benzene ring in
scaffold A showed a lower fit value than compounds 4-methoxy, 4-nitro, 3-phenoxy,
and 4-bromol group at the benzene ring in scaffold A, such as compound
1
a,compound 1f and compound 1g, which was accordance with the SHP2 inhibitory
activity results. This observation indicated why structure with the 4-methoxy group at
the benzene ring in scaffold A was more favourable than the other group, including
2
-bromol,4-nitro, 3-phenoxy, 4-fluoro and 4-bromol groups, for increasing potency in
this set of compounds.
.6 Molecular Dynamics Simulations
3
To explore the mechanism of how compound 1d inhibit SHP2, compound 1d in
complex with SHP2 and SHP2 without ligand were subjected to 20 ns MD
simulations, respectively. As shown in Figure 6, the RMSD values of
SHP2-compound 1d and SHP2 without ligand systems were stable from 2 ns to 20 ns.
The RMSD value was a key criterion applied to evaluate the stability of the

protein-complex and the protein systems. During the simulation, the RMSD values of
SHP2 with ligand compound 1d and without ligand systems were found to be
relatively stable about 0.21 Å and 0.26 Å, respectively. As we could see from the
Figure 6, the RMSD value of SHP2 with ligand compound 1d was smaller than that
of SHP2, indicating that the complexes of SHP2 with ligand compound 1d was more
stable than SHP2. Thus, the compound 1d could make the SHP2 protein stable.
As shown in panel B of Figure 6, the root mean square fluctuations for all the
side-chain atoms of the receptors were also computed within 20 ns molecular
dynamics simulations, to perform an in-depth study of the interactions of the SHP2
binding domain with the inhibitor. The larger the RMSF fluctuated, the more unstable
the residues were, whereas the smaller the RMSF fluctuated, the more stable the
residues were. In the active site-pTyr recognition loop (residues 271-284), the
WPD-loop (residues 421-431) and the Q-loop ((residues 501-507), the order of the
RMSF values was 3O5X-compound 1d system < 3O5X without ligand system. The
results suggested that the compound 1d had effect on the three regions (pTyr
recognition loop, WPD-loop and Q-loop) with decreasing the interactions of the
residues.
To provide further understanding of compound 1d in inhibiting the fluctuation of
loops of SHP2 (Figure 7), the Anisotropic Network Mode (ANM) was performed by
the VMD software(Kotamarthi, Yadav, & Ainavarapu, 2015; Sarkar, 2018) The arrow
represented the direction and amplitude of the motion. The motions of the pTyr

recognition loop (residues residues 271-284 colored in red), the WPD-loop (residues
21-431 colored in blue) and the Q-loop (residues 501-507 colored in purple) in the
4
SHP2-compound 1d system were all smaller than that in the SHP2 without ligand
system, showing that compound7d could inhibit the fluctuations of the pTyr
recognition loop, the WPD-loop and the Q-loop. So, the ligand (compound 1d)
binding to the receptor (3O5X) might increase the stability of the residues in the
active site, causing the inhibition activities. The analysis of ANM was in accordance
with RMSF analysis.
For better understanding of the specific interactions between protein systems, the
binding free energy was calculated using MM/PBSA method, which is consisting of
four terms, including the polar solvation free energy, the non-polar solvation free
energy, the van der waals interaction energy and the electrostatic energy. As we could
see from the Table 5, The binding energy of compound 1d with SHP2 is -76.737
kcal/mol, while the binding energy of compound 1a with SHP2 is -66.949 kcal/mol,
indicating that the system of SHP2-compound 1d seems more stable than
SHP2-compound 1a, which was in accord with the bioactivity results. Polar solvation
free energy made a positive contribution to the total free binding energy while
non-polar solvation free energy, electrostatic interactions and Van der Waals made a
negative contribution to the total interaction energy, which meant that Van der Waals,
electrostatic interactions and non-polar solvation energy together are in favor of the
stability of the complex system. We could clearly see that the Van der Waal

interactions and the non-polar solvation free energy were the major favorable
contributions to the binding energies between SHP2 and compound 1d/1a.
For the purpose of discerning the contributions of binding free energy of each residue,
the interaction energies are decomposed into individual residue contributions using
energy decomposition scheme(Baker, Sept, Joseph, Holst, & McCammon, 2001). The
residues with the top 10 interaction energies for compound 1d with SHP2,
respectively, are shown in Figure 8. The binding energy is defined as the sum of van
der Waals interactions and short range electrostatic interactions between the ligand
and the receptor. The key residues were found in three significantly different regions:
the active site-pTyr recognition loop (residues 271-284), WPD-loop (residues
4
21-431), P-loop (residues 501-507), which was in accordance with the RMSF results
in Figure 5. As shown in Figure 8, Ile282, Pro284, Tyr279 and Phe285 in pTyr
recognition loop made large contributions to the binding affinity. Their total energy
contributions are -6.20 kcal/mol, -3.07 kcal/mol, -1.99 kcal/mol and -1.69 kcal/mol,
respectively. Gly503, Ser502 and Met504 in Q-loop makes remarkable contributions
to the binding affinity, their total energy contributions are -4.29 kcal/mol, -2.52
kcal/mol and -1.65 kcal/mol, respectively. In a word, all of these are very important
reasons to explain why compound 1d showed good activities.
4
Conclusion
In conclusion, a series of novel ethyl 4-(phenoxymethyl)-2-phenylthiazole-5-
carboxylate derivatives targeting SHP2 were designed with the aid of scaffold
hopping approach. The designed compounds were synthesized, and their activities