In-silico Studies of the Antiproliferative Activity of New Anilinoquinazoline Derivatives Against NSCLC Cells.

Article abstract of DOI:10.1016/j.molstruc.2020.129786

The current reversible epidermal growth factor (EGFR) tyrosine kinase inhibitors of non-small cell lung cancer (NSCLC) have been resisted by T790M mutation, while the irreversible inhibitors introduced to overcome the mutation have faced resistance from C979S mutation. In the effort to discover potentially improved reversible EGFR inhibitors, a series of new anilinoquinazoline derivatives with modification on the 2^nd carbon on the aniline ring was synthesized. The derivatives were tested for their antiproliferative activity against NSCLC cell lines with wild-type (A549), exon 19 deletion mutated (H1650) and L858R+T790M (H1975) mutated EGFR kinases. Three derivatives (4-6) performed better than the standard drug, gefitinib, in all cell lines. Derivative 5 recorded the lowest IC₅₀ values in all cell lines (A549: 24.60 ± 0.75 μM, H1650: 14.83 ± 0.54 μM, H1975: 21.72 ± 1.21 μM). A molecular docking study followed by molecular dynamics simulations was performed on derivative 5 and gefitinib using wild-type EGFR kinase (WT-EGFR) and L858R+T790M mutated kinase (LRTM-EGFR) to provide an understanding of their binding mechanisms. In WT-EGFR kinase, derivative 5 recorded better hydrogen bonding occupancy with MET793 (94.16%) and binding energy profile (-35.287 kcal/mol) as compared to gefitinib (91.80%, -26.071 kcal/mol). In LRTM-EGFR kinase, derivative 5 recorded better hydrogen bonding occupancy with MET793 (93.68%) as compared to gefitinib (91.48%). Derivative 5 also recorded additional hydrogen bonding interactions with ASP855, with a total of 60.61% occupancy as well as a better energy profile (-42.867 kcal/mol) as compared to gefitinib (-41.778 kcal/mol).

Full text of DOI:10.1016/j.molstruc.2020.129786

Journal Pre-proof
In-silico Studies of the Antiproliferative Activity of New
Anilinoquinazoline Derivatives Against NSCLC Cells.
Muhammad Kumayl Abdulwahab , Ke Han Tan ,
Rashidi Dzulkeflee , Kok Hoong Leong , Choon Han Heh ,
Azhar Ariffin
PII:
S0022-2860(20)32099-8
DOI:
Reference:
https://doi.org/10.1016/j.molstruc.2020.129786
MOLSTR 129786
To appear in:
Journal of Molecular Structure
Received date:
Revised date:
Accepted date:
12 August 2020
12 December 2020
14 December 2020
Please cite this article as: Muhammad Kumayl Abdulwahab , Ke Han Tan , Rashidi Dzulkeflee ,
Kok Hoong Leong , Choon Han Heh , Azhar Ariffin , In-silico Studies of the Antiproliferative Activity
of New Anilinoquinazoline Derivatives Against NSCLC Cells., Journal of Molecular Structure (2020),
doi: https://doi.org/10.1016/j.molstruc.2020.129786
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier B.V.

1
Highlights
2
3
4
5
6
7
8
9


Reversibly inhibitors have been designed and synthesized in the attempt to overcome the
non-small cell lung cancer (NSCLC) cells that harbor T790M EGFR mutation.
The optimization of 4-anilinoquinazoline scaffold, which is the main core for the first
generation EGFR inhibitors, by the addition of amide group on the 2nd carbon on the
aniline ring, produces derivatives that display better proliferative inhibition of A549,
H1650 and H1975 cells than gefitinib.

Molecular docking and molecular dynamics simulations are performed to understand the
theoretical structure-activity relationship.
10
1

11
12
13
In-silico Studies of the Antiproliferative Activity of New Anilinoquinazoline Derivatives Against NSCLC
Cells.
14
15
Muhammad Kumayl Abdulwahab1, Ke Han Tan2, Rashidi Dzulkeflee3, Kok Hoong Leong2*,
Choon Han Heh2*, Azhar Ariffin1*
16
17
1Department of Chemistry, Faculty of Science, University of Malaya, 50603 Kuala Lumpur,
Malaysia
18
19
2Department of Pharmaceutical Chemistry, Faculty of Pharmacy, 50603 Kuala Lumpur,
Malaysia.
20
21
3Department of Molecular Medicine, Faculty of Medicine, University of Malaya, 50603 Kuala
Lumpur, Malaysia
22
23
24
25
*corresponding author
azhar70@um.edu.my
leongkh@um.edu.my
silverbot@um.edu.my
26
Abstract
27
28
29
30
31
32
33
34
35
The current reversible epidermal growth factor (EGFR) tyrosine kinase inhibitors of non-small
cell lung cancer (NSCLC) have been resisted by T790M mutation, while the irreversible
inhibitors introduced to overcome the mutation have faced resistance from C979S mutation. In
the effort to discover potentially improved reversible EGFR inhibitors, a series of new
anilinoquinazoline derivatives with modification on the 2nd carbon on the aniline ring was
synthesized. The derivatives were tested for their antiproliferative activity against NSCLC cell
lines with wild-type (A549), exon 19 deletion mutated (H1650) and L858R+T790M (H1975)
mutated EGFR kinases. Three derivatives (4-6) performed better than the standard drug,
gefitinib, in all cell lines. Derivative 5 recorded the lowest IC50 values in all cell lines (A549:
2

36
37
38
39
40
41
42
43
44
45
24.60 ± 0.75 µM, H1650: 14.83 ± 0.54 µM, H1975: 21.72 ± 1.21 µM). A molecular docking
study followed by molecular dynamics simulations was performed on derivative 5 and gefitinib
using wild-type EGFR kinase (WT-EGFR) and L858R+T790M mutated kinase (LRTM-EGFR)
to provide an understanding of their binding mechanisms. In WT-EGFR kinase, derivative 5
recorded better hydrogen bonding occupancy with MET793 (94.16%) and binding energy profile
(-35.287 kcal/mol) as compared to gefitinib (91.80%, -26.071 kcal/mol). In LRTM-EGFR
kinase, derivative 5 recorded better hydrogen bonding occupancy with MET793 (93.68%) as
compared to gefitinib (91.48%). Derivative 5 also recorded additional hydrogen bonding
interactions with ASP855, with a total of 60.61% occupancy as well as a better energy profile (-
42.867 kcal/mol) as compared to gefitinib (-41.778 kcal/mol).
46
Keywords
47
48
EGFR kinase inhibitor; anilinoquinazoline; synthesis; antiproliferative activity; molecular
docking; molecular dynamics.
49
1.0 Introduction
50
51
52
53
Epidermal growth factor receptor (EGFR) belongs to a family of four transmembrane
receptors, including EGFR (HER1), HER2, HER3 and HER4 [1]. In a healthy cell,
phosphorylation of EGFR activates multiple downstream pathways that control cell growth and
proliferation [1].
54
55
56
57
Studies on non-small cell lung cancer (NSCLC) have linked the disease with
overexpression of EGFR, which is partly caused by activating mutation in the kinase domain,
including single amino acid substitution, insertion and/or deletion [2, 3]. Among the most
prevalent mutations in EGFR kinase are exon 19 deletion (ex19del), which accounts for 44% of
3

58
59
60
61
62
63
64
all EGFR mutations [3, 4], and exon 21 point mutations (L858R) [5, 6]. Another common
mutation is T790M, which is observed in 51.9% of EGFR inhibitor resistant patients and
appears in the form of ex19del+T790M or L858R+T790M (LRTM) [7, 8]. Irreversible inhibitors
such as osimertinib have shown selective potency against the double mutant EGFR kinase [9],
yet the emergence of another mutation, C797S, has developed resistance against all existing
inhibitors [10]. To overcome this problem, there are numerous studies that have focussed on
discovering new reversible inhibitors against LRTM mutations [11, 12].
65
66
67
68
69
70
71
Anilinoquinazoline pharmacophore is well known and has been continuously studied for
its anticancer and EGFR kinase inhibition activities [13]. Furthermore, there are established
anilinoquinazoline based EGFR kinase inhibitors currently in use for anticancer treatment
(Figure 1). The derivatization of anilinoquinazoline derivatives is commonly performed on
either the quinazoline ring or the aniline ring or both. Derivatization on the aniline ring is more
commonly performed to incorporate the halogen groups at different positions of the rings,
although the incorporation of polar substituents has been reported [14, 15].
72
73
74
75
76
The amide group plays a major role in the biological system, particularly in linking
amino acids together to form proteins. It is also present in a huge number of molecules, including
major marketed drugs [16, 17]. The amide group of ligands, including naquotinib (a non-
anilinoquinazoline irreversible EGFR kinase inhibitor), has also been shown to form
intermolecular hydrogen bonding in many EGFR crystal structures [18, 19].
77
78
79
The rich structural information from the crystal structures of anilinoquinazoline inhibitors
co-crystallized with EGFR kinase has greatly contributed to the drug discovery of EGFR
inhibitors [20, 21]. In these crystal structures, the inhibitors share a common binding mode,
4

80
81
82
83
84
85
86
87
88
89
90
91
where they are oriented in the kinase in such a way that the N-1 of the quinazoline ring forms
hydrogen bonding with the amide of MET793 and the aniline group is sequestered into the
hydrophobic pocket at the back of the ATP binding site [20, 22]. Two amino acids (LYS745 and
ASP855) with the potential of forming hydrogen bonding which have been targeted for the
interaction in previous studies [15, 23], lay closely to the 2nd carbon of the aniline ring of the
inhibitors. Thus, in this study, we synthesized a series of anilinoquinazoline derivatives with the
amide group on the 2nd carbon of the aniline ring; such a study has not been published before.
The derivatives were tested for their antiproliferative activity against NSCLC cell lines with
wild-type (WT), ex19del and LRTM mutated kinases. Based on the antiproliferative activity, we
selected the best performing derivative and performed molecular docking and molecular
dynamics simulations to study the possible binding mechanism of the derivative to the EGFR
kinase.
92
5

93
94
Figure 1: (A)-(D) Structure of anilinoquinazoline EGFR kinase inhibitors in clinical trials and
(E) the general structure of the designed derivatives.
95
96
2.0 Methods and Materials
2.1 Synthesis of the Derivatives
97
98
99
All chemicals were obtained from Merck. The melting points were obtained using a
MELTEMP II, Laboratory Device. NMR spectra were recorded using Jeol ECX-400 MHz. Mass
spectra were recorded using Agilent 6545 LC-MS QTOF.
100
2.1.1 Synthesis of 7-chloroquinazolin-4(3H)-one (2)
101
102
103
104
105
106
107
108
109
2-amino-4-chlorobenzoic acid (0.5 g, 2.91 mmol) was stirred in 7.5 ml of formamide. The
mixture was heated in a microwave oven at 190oC (power 300 kWatt) for 30 minutes. The clear
brown reaction mixture was cooled to room temperature. Iced water was added to the reaction
mixture to produce precipitation. The precipitate was filtered. The crude product was further
purified using column chromatography (EA: hexane, 7:3) to give a white powder. Yield: 0.92 g,
90%. Melting point: 256°C (lit. 251-253°C). 1H NMR (DMSO-D6, 400 MHz), δ, ppm: 7.49 (dd,
1H, J = 8.71, 2.29 Hz), ), 7.66 (d, 1H, 2.29 Hz), 8.05 (d, J = 8.71Hz), 8.09 (s, 1H), 12.35 (s, 1H);
13C NMR (DMSO-D6, 100 MHz), δ, ppm: 121.97, 126.91, 127.54, 128.47, 139.44, 147.43,
150.40, 160.67.
110
2.1.2 Synthesis of 4,7-dichloroquinazoline (3)
111
112
113
Compound 2 (5 g, 27.69 mmol) and DMF (1 ml, 20% w/w) were added to 150 ml of chloroform.
The reaction mixture was cooled in an ice-bath. Oxalyl chloride (4.8 ml, 55.38 mmol) was added
dropwise to the reaction mixture. The reaction mixture was left to equilibrate to room
6

114
115
116
117
118
119
120
121
122
temperature by stirring for two hours. The reaction mixture was then heated to 60oC and stirred
overnight. After the reaction was completed, it was cooled in an ice-bath. Then, 4.95 g of sodium
hydrogen carbonate in 50 ml of cold water was added dropwise to the reaction mixture. The
mixture was separated using a separating funnel, and the organic layer was collected and dried
using anhydrous sodium sulphate. The organic layer was concentrated under vacuum. The crude
product was purified using column chromatography (ethyl acetate: hexane, 1:1). Yield: 4.41 g,
1
80%. Melting point: 140°C. H NMR (CDCl3, 400 MHz), δ, ppm: 7.67 (dd, 1H, J = 8.70, 1.84
13
Hz), 8.07 (d, 1H, 1.84 Hz), 8.21 (d, J = 8.70Hz), 9.04 (s, 1H); C NMR (CDCl3, 100 MHz), δ,
ppm: 122.60, 127.35, 127.98, 130.36, 141.62, 151.66, 154.81, 162.52
123
2.1.3. General Procedure for the Synthesis of Derivatives 4-11
124
125
126
127
128
129
130
2-amino-4-chlorobenzoic acid (0.30 g, 1.75 mmol) was dissolved in 9 ml of THF. Then, CDI
(0.34 g, 2.10 mmol) was added to the solution. The solution was stirred for 1 hour, after which
the respective amines were added to the solution. The solution was stirred further overnight. The
completion of the reaction was observed using TLC. Then, the reaction mixture was run through
silica, and the solvent was concentrated using a rotary evaporator. The crude intermediate was
reacted with Compound 3 (0.35 g, 1.75 mmol) in THF at room temperature and stirred overnight.
7

131
132
The precipitate formed was filtered. The crude product was stirred in isopropanol for one hour
before being filtered to give the pure product.
133
2.1.3.1. 4-chloro-2-((7-chloroquinazolin-4-yl)amino)-N-propylbenzamide, 4
134
135
136
137
138
139
140
141
142
143
Propan-1-amine (0.14 ml, 1.75 mmol) was used as the reactant. A yellowish powder was
1
obtained. Yield: 0.36 g, 54%. Melting point: 232°C. H NMR (DMSO-D6, 400 MHz), δ, ppm:
0.80 (t, 3H, J = 7.34 Hz, H-22), 1.44 (m, 2H, H-21), 3.15 (m, 2H, H-20), 7.40 (dd, 1H, J = 8.47,
2.26 Hz, H-15), 7.85 (d, 1H, J = 8.47 Hz, H-14), 7.93 (dd, 1H, J = 9.03, 2.26 Hz, H-7), 7.99 (d,
1H, J = 2.26 Hz, H-9), 8.34 (d, 1H, J = 9.03 Hz, H-6), 8.57 (s, 1H, H-17), 8.94 (s, 1H, H-2), 9.96
(t, 1H, J = 5.65 Hz, H-19), 13.11 (br, 1H, H-11); 13C NMR (DMSO-D6, 100 MHz), δ, ppm: 12.0
(CH3, C-22), 22.5 (CH2, C-21), 41.7 (CH2, C-20), 113.7 (C, C-5), 122.1 (CH, C-9), 124.0 (CH,
C-17), 124.2 (C, C-13), 125.5 (CH, C-15), 125.8 (CH, C-6), 129.8 (CH, C-7), 130.8 (CH, C-14),
136.5 (C, C-16), 138.9 (C, C-12), 140.7 (C, C-8), 143.2 (C, C-4), 153.4 (CH, C-2), 159.0 (C, C-
10), 167.4 (C, C-18); HREIMS m/z 397.0604 [M+Na]+ (calc. for C18H16Cl2N4ONa 397.0594).
144
2.1.3.2. 4-chloro-2-((7-chloroquinazolin-4-yl)amino)-N-butylbenzamide, 5
145
146
147
148
149
150
151
152
N-butylamine (0.17 ml, 1.75 mmol) was used as the reactant. A yellowish powder was obtained.
Yield: 0.35 g, 52%. Melting point: 230°C. 1H NMR (DMSO-D6, 400 MHz), δ, ppm: 0.84 (t, 3H,
J = 7.43 Hz, H-23), 1.28 (m, 2H, H-22), 1.45 (m, 2H, H-21), 2.26 (m, 2H, H-20), 7.43 (dd, 1H, J
= 8.47, 2.26 Hz, H-15), 7.88 (d, 1H, J = 8.47 Hz, H-14), 7.96 (dd, 1H, J = 8.47, 2.26 Hz, H-7),
8.00 (d, 1H, J = 2.26 Hz, H-9), 8.36 (d, 1H, J = 8.47 Hz, H-6), 8.63 (s, 1H, H-17), 8.95 (s, 2H,
H-2, H-19), 13.06 (br, 1H, H-11); 13C NMR (DMSO-D6, 100 MHz), δ, ppm: 14.19 (CH3, C-23),
20.15 (CH2, C-22), 31.34 (CH2, C-21), 40.66 (CH2, C-20), 113.69 (C, C-5), 122.28 (CH, C-9),
124.04 (CH, C-17), 124.35 (C, C-13), 125.49 (CH, C-15), 125.84 (CH, C-6), 129.75 (CH, C-7),
8

153
154
155
130.81 (CH, C-14), 136.42 (C, C-16), 138.92 (C, C-12), 140.58 (C, C-8), 143.50 (C, C-4),
153.52 (CH, C-2), 158.95 (C, C-10), 167.42 (C, C-18); HREIMS m/z 389.0923 [M+H]+ (calc. for
C19H19Cl2N4O 389.0901).
156
2.1.3.3. 4-chloro-2-((7-chloroquinazolin-4-yl)amino)-N-(2-methoxyethyl)benzamide, 6
157
158
159
160
161
162
163
164
165
166
2-methoxyethanamine (0.15 ml, 1.75 mmol) was used as the reactant. A yellowish powder was
1
obtained. Yield: 0.36 g, 52%. Melting point: 224°C. H NMR (DMSO-D6, 400 MHz), δ, ppm:
3.14 (s, 3H, H-22), 3.37 (m, 4H, H-20,21), 7.40 (dd, 1H, J = 8.47, 1.69 Hz, H-14), 7.83 (d, 1H, J
= 8.47 Hz, H-14), 7.91 (dd, 1H, J = 9.03, 2.26 Hz, H-7), 7.98 (s, 1H, H-9), 8.37 (d, 1H, J = 9.03
Hz, H-6), 8.52 (s, 1H, H-17), 8.90 (s, 1H, H-2), 8.97 (s, 1H, H-19), 12.95 (br, 1H, H-11); 13C
NMR (DMSO-D6, 100 MHz), δ, ppm: 58.37 (CH3, C-22), 70.52 (2CH2, C-20,21), 113.72 (C, C-
5), 122.29 (CH, C-9), 124.06 (CH, C-17), 124.34 (C, C-13), 125.48 (CH, C-15), 125.91 (CH, C-
6), 129.63 (CH, C-7), 130.87 (CH, C-14), 136.51 (C, C-16), 138.97 (C, C-12), 140.50 (C, C-8),
143.64 (C, C-4), 153.51 (CH, C-2), 158.84 (C, C-10), 167.62 (C, C-18); HREIMS m/z 391.0714
[M+H]+ (calc. for C18H17Cl2N4O2 391.0724).
167
2.1.3.4. 4-chloro-2-((7-chloroquinazolin-4-yl)amino)-N-(2-hydroxyethyl)benzamide, 7
168
169
170
171
172
173
174
2-aminoethanol (0.10 ml, 1.75 mmol) was used as the reactant. A white powder was obtained.
1
Yield: 0.36 g, 54%. Melting point: 218°C. H NMR (DMSO-D6, 400 MHz), δ, ppm: 3.27 (m,
2H, H-20), 3.44 (t, 2H, J = 5.65, H-21), 7.39 (dd, 1H, J = 8.47, 2.26 Hz, H-15), 7.88 (d, 1H, J =
8.47 Hz, H-14), 7.92 (dd, 1H, J = 9.03, 1.69 Hz, H-7), 7.97 (d, 1H, J = 1.69 Hz, H-9), 8.32 (d,
1H, J = 9.03 Hz, H-6), 8.62 (s, 1H, H-17), 8.92 (s, 1H, H-2), 8.94 (t, 1H, J = 5.08 Hz, H-19),
13.00 (br, 1H, H-11); 13C NMR (DMSO-D6, 100 MHz), δ, ppm: 42.84 (CH2, C-20), 59.76 (CH2,
C-21), 113.95 (C, C-5), 123.27 (CH, C-9), 123.39 (CH, C-17), 124.92 (CH, C-15), 125.55 (CH,
9

175
176
177
C-6), 129.55 (CH, C-7), 130.87 (CH, C-14), 136.59 (C, C-16), 139.52 (C, C-12), 140.25 (C, C-
8), 144.73 (C, C-4), 154.02 (CH, C-2), 158.64 (C, C-10), 167.92 (C, C-18); HREIMS m/z
377.0552 [M+H]+ (calc. for C17H15Cl2N4O2 377.0567).
178
2.1.3.5. 4-chloro-2-((7-chloroquinazolin-4-yl)amino)-N-(3-hydroxypropyl)benzamide, 8
179
180
181
182
183
184
185
186
187
188
3-aminopropan-1-ol (0.14 ml, 1.75 mmol) was used as the reactant. A white powder was
1
obtained. Yield: 0.36 g, 52%. Melting point: 220°C. H NMR (DMSO-D6, 400 MHz), δ, ppm:
1.58 (m, 2H, H-21), 3.24 (m, 2H, H-20), 3.37 (t, 2H, J = 6.21, H-22), 7.40 (dd, 1H, J = 8.47, 2.26
Hz, H-15), 7.84 (d, 1H, J = 8.47 Hz, H-14), 7.92 (dd, 1H, J = 8.47, 2.26 Hz, H-7), 7.99 (s, 1H,
H-9), 8.35 (d, 1H, J = 8.47 Hz, H-6), 8.50 (s, 1H, H-17), 8.95 (s, 2H, H-2,19), 13.16 (br, 1H, H-
11); 13C NMR (DMSO-D6, 100 MHz), δ, ppm: 32.44 (CH2, C-21), 37.35 (CH2, C-20), 58.98
(CH2, C-22), 113.69 (C, C-5), 112.13 (CH, C-9), 124.03 (CH, C-17), 124.16 (C, C-13), 125.54
(CH, C-15), 125.83 (CH, C-6), 129.84 (CH, C-7), 130.81 (CH, C-14), 136.48 (C, C-16), 138.91
(C, C-12), 140.65 (C, C-8), 143.14 (C, C-4), 153.45 (CH, C-2), 159.00 (C, C-10), 167.51 (C, C-
18); HREIMS m/z 391.0726 [M+H]+ (calc. for C18H17Cl2N4O2 391.0724).
189
2.1.3.6. (4-chloro-2-((7-chloroquinazolin-4-yl)amino)phenyl)(morpholino)methanone, 9
190
191
192
193
194
195
196
Morpholine (0.16 ml, 1.75 mmol) was used as the reactant. A white powder was obtained. Yield:
0.42 g, 59%. Melting point: 230°C. 1H NMR (DMSO, 400 MHz), δ, ppm: 3.29 (m, 4H, H-19),
3.45 (m, 4H, H-20), 7.50 (m, 2H, H-14,17), 7.65 (d, 1H, J = 1.13 Hz, H-15), 7.90 (dd, 1H, J =
9.03, 2.26 Hz, H-7), 8.04 (d, 1H, J = 2.26 Hz, H-9), 8.82 (d, 1H, J = 9.03 Hz, H-6), 8.88 (s, 1H,
H-2), 11.97 (br, 1H, H-11); 13C NMR (DMSO, 400 MHz), δ, ppm: 42.33 (CH2, C-20), 47.88
(CH2, C-19), 66.34 (2CH2, C-21), 112. 69 (C, C-5), 120.52 (CH, C-9), 127.56 (CH, C-6), 127.88
(CH, C-17), 128.11 (CH, C-15), 129.63 (CH, C-7), 130.38 (CH, C-14), 131.71 (C, C-13), 134.62
10

197
198
(C, C-16), 136.17 (C, C-12), 141.27 (C, C-8), 141.49 (C, C-4), 152.53 (CH, C-2), 160.45 (C, C-
10), 166.50 (C, C-18); HREIMS m/z 425.0557 [M+Na]+ (calc. for C19H16Cl2N4O2Na 425.0543).
199
2.1.3.7. 4-chloro-2-((7-chloroquinazolin-4-yl)amino)-N-morpholinobenzamide, 10
200
201
202
203
204
205
206
207
208
209
Morpholin-4-amine (0.17 ml, 1.75 mmol) was used as the reactant. A white powder was
1
obtained. Yield: 0.36 g, 49%. Melting point: 208°C. H NMR (DMSO-D6, 400 MHz), δ, ppm:
2.76 (m, 4H, H-20), 7.45 (dd, 1H, J = 8.47, 2.26 Hz, H-15), 7.74 (d, 1H, J = 8.47 Hz, H-14), 7.93
(dd, 1H, J = 8.47, 2.26 Hz, H-7), 7.98 (d, 1H, J = 2.26 Hz, H-9), 8.26 (d, 1H, J = 2.26 Hz, H-6),
8.50 (d, 1H, J = 8.47 Hz, H-17), 8.90 (s, 1H, H-2), 9.93 (br, 1H, H-19), 12.50 (br, 1H, H-11); 13C
NMR (DMSO-D6, 100 MHz), δ, ppm: 54.83 (2CH2, C-20), 66.33 (2CH2, C-21), 113.40 (1C, C-
5), 112.09 (CH, C-9), 125.19 (CH, C-17), 125.28 (C, C-13), 125.82 (CH, C-15), 126.28 (CH, C-
6), 129.68 (CH, C-7), 131.07 (CH, C-14), 136.28 (C, C-16), 138.31 (C, C-12), 140.71 (C, C-8),
143.27 (C, C-4), 153.38 (CH, C-2), 159.34 (C, C-10), 164.72 (C, C-18); HREIMS m/z 440.0681
[M+Na]+ (calc. for C19H17Cl2N5O2Na 440.0652).
210
2.2 In-vitro Antiproliferative Study
211
212
213
214
215
216
217
218
Antiproliferation activity of the synthesized derivatives was evaluated using an A549 cell
line (human lung cancer with WT-EGFR), an H1650 cell line (an inherent ex19del mutated
EGFR human lung cancer), and an H1975 cell line (human lung cancer with LRTM-EGFR),
which were obtained from AddexBio Ltd. USA. DMEM was used as the media for the A549 cell
line, while RPMI 1640 was used as the media for the H1650 and H1975 cell lines. The cells were
cultured in the respective media, in a cell incubator maintained at 37°C in a humidified
atmosphere of 5% CO2. The cell proliferation activities of the compounds against the cell lines
were determined using the standard MTT assay [24]. The IC50 values of the samples and positive
11

219
220
221
control were determined from dose-response curves using Prism 8.0.2 software for Windows
(GraphPad Software Inc., La Jolla, CA, USA). The results were expressed as mean ± standard
deviation of triplicates [24].
222
2.3 Molecular Docking
223
224
A workstation with four Intel Core 2 Duo E6850 3.00 GHz microprocessors, 8 GigaBytes
of Random-Access Memory (RAM) and an Ubuntu 18.04 Linux operating system was used.
225
226
227
228
229
230
231
232
233
234
235
236
237
238
Blind docking that covered the whole EGFR kinase crystal structure was performed using
AutoDock Vina software [25]. The models for the WT-EGFR and LRTM-EGFR kinases were
obtained from the Protein Data Bank (PDB ID: 1M17 and 5EDR, respectively) [11, 20]. The
molecular docking for exon 19 deletion mutated kinase was not performed because its crystal
structure has yet to be established in the Protein Data Bank, and the region of the deletions was
not a part of the ligand binding site [26]. Then, the present ligand and water molecules were
removed from the crystal structure using Biovia Discovery Studio Visualiser 4.5 software [27].
Subsequently, AutoDockTools 1.5.6 software was used to add all hydrogen atoms, to merge
nonpolar hydrogen atoms, to check and repair missing atoms, to add Gasteiger charges, and to
assign AD4 atom types to the protein structure. Finally, the protein structure was saved in .pdbqt
format for further docking [28, 29]. A docking grid box of the kinase structure was then set using
AutoDockTools 1.5.6 software with a grid spacing of 1.0 Å, dimensions of 126 x 126 x 126
points along the x, y and z axes, and centered on the kinase (x = -57.6, y = -5.1, z = -20.5),
covering the whole kinase for blind docking [30].
239
240
The most potent derivative based on the in-vitro antiproliferative study was selected for
mechanism study via molecular docking followed by molecular dynamics simulations, with
12

241
242
243
244
245
246
gefitinib as the standard ligand. The derivative and the standard ligand, gefitinib, were
constructed using ChemBio 3D Ultra 12.0, and their energy was minimized using the MM2
energy minimization calculation [31]. Then, AutoDockTools 1.5.6 software was used to add
flexibility to flexible bonds on the ligands, after which the ligands were saved in .pdbqt format as
input files for docking [28]. The results were analysed using Biovia Discovery Studio Visualiser
4.5 software.
247
2.4 Molecular Dynamics Study
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
Molecular dynamics (MD) simulations were performed to obtain further insight into the
binding mechanism of the most potent derivative and gefitinib against WT-EGFR and LRTM-
EGFR. The MD simulations were carried out using GROMACS 2019 [32]. The receptor files for
the simulations were prepared using the GROMACS ‘pdb2gmx’ module with the AMBER force
field ff14SB [33]. The ligand files were prepared using Antechamber Python Parser Interface
(ACPYPE) [34] software, after which the Generalized AMBER Force Field (GAFF) [35, 36] and
AM1-BCC charge models were applied to the files. The simulations were set up by suspending
the protein-ligand complex at the center of a TIP3P [37] water box, where the edge of the box
was at least 10 Å away from the complex. All systems were solvated with an explicit solvent
model and were neutralized by adding counter ions prior to the energy minimization stage. The
simulations were started by minimizing the solvated complex structure with the steepest descent
algorithm followed by the conjugated gradient algorithm. After energy minimization,
equilibrium simulation was carried out with a constant temperature and pressure at 300 K and 1
atm, respectively. All MD simulations during the production stage were carried out for 50 ns,
while a trajectory was captured every 20 ps. For RMSD and RMSF calculations, the analyses of
the collected trajectories were carried out using the GROMACS ‘rms’ and ‘rmsf’ modules. The
13

264
265
266
267
268
269
270
MM-PBSA binding free energy calculations and per-residue energy decomposition analysis were
performed using g_mmpbsa [38] software. The water molecules and counter ions were removed
from the captured trajectory frames before MM-PBSA calculations were carried out. For the
hydrogen bond occupancy calculations, the trajectories collected from GROMACS MD
simulation were converted into AMBER netcdf format (.nc) with the cpptraj [39] module in
AmberTools. The converted trajectories were then used for hydrogen bond occupancy analysis
with cpptraj’s hbond module.
271
2.5 In-silico ADME Evaluation
272
273
The pharmacokinetic and druglikeness properties of the derivatives were
predicted in-silico using the SwissADME computational tool [51].
274
275
3.0 Result and Discussion
3.1 Synthesis of Derivatives
276
277
278
279
280
281
282
The derivatives were synthesized according to the route of synthesis by Baumann, et. al.
[40] with modification, as outlined in Scheme 1. One of the common methods of synthesizing
quinazolinone is through conventional heating of 2-aminobenzoate and formamide [40-42]. Our
employment of this method on 2-amino-4-chlorobenzoic acid was unfavorable and took 16 hours
to complete. On the other hand, the utilization of microware irradiation, as reported by
Glowacka, et. al. and Ouahrouch, et. al., cut the reaction time to 30 minutes and produced a high
yield [43, 44].
283
284
The reaction of intermediate 2 with oxalyl chloride and catalytic DMF, as outlined in
Scheme 1, yielded 4,7-dichloroquinazoline, 3; a good yield [45-47].
14

285
286
287
288
289
290
291
To obtain derivatives 4-10, 2-amino-4-benzoic acid, 1 was first reacted with 1,1-
carbonyldiimidazole in THF, followed by the addition of the respective amine as reported by
Mizutani, et. al. [48]. From the TLC monitoring (ethyl acetate:hexane, 4:6), we observed that the
excess starting material formed salt with the by-product imidazole and stayed at the baseline of
the TLC plate. We decided to utilize this characteristic by passing the reaction mixture through
silica, and flushing it with the ethyl acetate:hexane (4:6) mobile phase. The collected eluate was
reacted with intermediate 3 without further isolation and purification.
292
293
Scheme 1: Synthesis of derivatives 4-10.
15

294
3.2 Antiproliferative Study
295
296
297
298
4-anilinoquinazoline derivatives were evaluated for in-vitro antiproliferative activity
against NSCLC cell lines with wild-type and mutated EGFR kinases: A549 (WT), H1650
(ex19del), and H1975 (LRTM). Gefitinib was used as the standard drug. The results are
summarized in Table 1.
299
Table 1: The IC50 values against A549, H1650, and H1975 for derivatives 4-10 and gefitinib
IC50 (Mean ± SD, µM)
A549
H1650
H1975
Derivatives:
4, R:
25.13 ± 1.01 15.19 ± 0.58 23.29 ± 1.95
24.60 ± 0.75 14.83 ± 0.54 21.72 ± 1.21
29.25 ± 1.32 17.56 ± 0.20 23.09 ± 0.39
31.19 ± 0.86 20.66 ± 0.55 25.81 ± 0.24
35.16 ± 1.09 18.69 ± 0.90 25.54 ± 0.99
5, R:
6, R:
7, R:
8, R:
32.57 ± 1.74 24.75 ± 0.89 28.86 ± 1.43
38.90 ± 1.89 28.50 ± 0.83 27.11 ± 0.81
9, R:
10, R:
16

Gefitinib
34.32 ± 1.30 19.17 ± 1.90 31.12 ± 0.38
300
301
302
303
304
305
Compared to the standard drug, five derivatives (4-7 and 9) recorded better IC50 values in
the A549 cell line, four derivatives (4-6 and 8) in the H1650 cell line, and seven derivatives (4-
10) in the H1975 cell line. Three derivatives, 4-6, consistently performed better than the standard
drug in all cell lines, and derivative 5 recorded the best activity in all cell lines (A549: 24.60 ±
0.75 µM, H1650: 14.83 ± 0.54 µM, H1975: 21.72 ± 1.21 µM).
306
307
308
309
310
311
312
313
314
The various functional groups next to the amide group of the derivatives showed some,
albeit not very significant, patterns in the inhibition study,. Derivatives 4 and 5 had non-polar
hydrocarbon chains, and they collectively recorded the best IC50 values. The rest of the
derivatives (6-10) incorporated several polar atom/s, and this modification seemed to contribute
to their relatively higher IC50 values as compared to derivatives 4 and 5. Derivative 10 had the
most polar atoms on its chain (1 nitrogen and 1 oxygen) and this might account for its highest
IC50 in the A549 (38.90 ± 1.89 µM) and H1650 (28.50 ± 0.83 µM) cell lines. However, in the
H1975 cell line, derivative 10 recorded a relatively better IC50 value (27.11 ± 0.81 µM), which
was comparable to gefitinib (31.12 ± 0.38 µM).
315
3.3 Molecular Docking Study
316
317
318
319
320
To gain insight into the probable binding mechanism of the derivatives onto the EGFR
kinase, we selected the derivative with the highest potency, derivative 5, and subjected it to
molecular docking into WT-EGFR (PDBID: 1M17), and LRTM-EGFR (PDBID: 5EDR)
kinases. Table 2 summarizes the interactions of derivative 5 in the kinases. The 2D and 3D
interactions diagrams of derivative 5 in the kinases are depicted in Figure 2.
17

321
322
Table 2: List of interactions for derivative 5 and gefitinib in WT-EGFR and LRTM-EGFR
kinases.
WT-EGFR/
Derivative 5
-8.0
WT-EGFR/
Gefitinib
-9.3
LRTM-EGFR/ LRTM-EGFR/
Derivative 5
Gefitinib
-7.9
Binding energy
(kcal/mol)
-8.3
Hydrogen bond
1MET793,
ASP855
1MET793
1MET793
1MET793
ARG841
-
Carbon
GLN791
GLN791,
ASP855
hydrogen bond
Halogen bond
Pi-Cation
LEU788
GLU762
-
-
-
LYS745
LEU788,
THR790,
LEU792,
GLY796,
THR854,
PHE856
Van der Waals
GLU762,
THR790,
LEU792,
GLY796,
THR854
PHE723,
ILE744,
GLY719,
MET766,
LEU788,
LEU792,
PRO794,
GLY796,
CYS797,
THR854
LEU718,
VAL726,
ALA743,
LYS745,
2MET790,
LEU844
GLU762,
MET766,
LEU792,
GLY796,
THR854
Hydrophobic
(Pi-Sigma,
Alkyl, Pi-
Alkyl, Pi-
Sulfur)
LEU844,
ALA743,
LYS745,
MET766,
LEU788
LEU718,
PHE723,
VAL726,
ALA743,
MET766,
LEU844
LEU718,
VAL726,
ALA743,
LYS745,
2MET790,
LEU844
1
323
the essential amino acid residues of the EGFR kinase for substrate binding at the active site. 2 the mutated amino acid.
18

324
325
Figure 2: (A) The 3D interaction diagram of derivative 5 in WT-EGFR. Gefitinib (orange) is
overlayed for comparison. LYS745, MET793, ASP855, THR790 and LEU858 are labeled. (B)
The 2D interaction diagram of derivative 5. (C) The 2D interaction diagram of gefitinib. (D) The
interaction indicator for 2D interaction diagrams (B) and (C).
326
327
328
329
330
19

331
332
Figure 3: (A) The 3D interaction diagram of derivative 5 in LRTM-EGFR. Gefitinib (orange) is
overlayed for comparison. LYS745, MET793, ASP855, MET790 and ARG858 are labeled. (B)
The 2D interaction diagram of derivative 5. (C) The 2D interaction diagram of gefitinib. (D) The
interaction indicator for 2D interaction diagrams (B) and (C).
333
334
335
336
337
Figure 2 and Figure 3 show that derivative 5 and the docked gefitinib bound to the
kinases in a similar orientation, where their quinazoline rings could be observed to be
20

338
339
340
341
overlapped. This orientation was also observed in other known anilinoquinazoline inhibitors in
their respective crystal structures [20, 21]. Hydrogen bonding with MET793, which was one of
the main interactions of ATP, erlotinib, gefitinib, lapatinib and other known inhibitors with
EGFR kinase, was observed for derivative 5 in both kinases (Figure 2B and 3B).
342
343
344
345
In WT-EGFR, the hydrogen bondings with THR790 and ASP855 were also observed
(Figure 2B). However, in LRTM-EGFR, apart from the previously mentioned MET793, there
was no additional hydrogen bonding observed for derivative 5 (Figure 3B). Derivative 5 was
observed to form a Pi-sulfur interaction with the mutated MET790 residue.
346
3.4 Molecular Dynamics Study
347
348
349
350
To further understanding of the binding interaction of derivative 5 in the EGFR kinases,
molecular dynamics simulations were performed on the derivative in-complex with WT-EGFR
and LRTM- EGFR kinases, where the best docked conformations were selected for the purpose.
The simulations were also performed on gefitinib in-complex with the kinases as comparison.
351
3.4.1 RMSD and RMSF analysis
352
353
354
355
356
357
358
359
The stability and the fluctuation of the systems were evaluated by calculating the root-
mean-square deviation (RMSD) between the backbone residues of the trajectories and the
reference structures over 50 ns. The respective energy-minimized structures of the complexes
were used as the references. Ten amino acid residues from both terminals of each kinase were
removed from the RMSD calculations due to their highly fluctuating nature, which might have
interfered with the interpretation of the results [49]. The backbone of WT-EGFR/derivative 5
experienced higher fluctuations during the period of 20 ns to 35 ns, but remained stable during
the rest of the simulation with a fluctuation under 0.25 nm (2.5 Å) (Figure 4). The backbone of
21

360
361
WT-EGFR/gefitinib fluctuated from the beginning of the simulation, and gradually stabilized
from 35 ns onwards.
362
363
364
365
366
The RMSD analysis showed that the backbone of LRTM-EGFR/gefitinib gained stability
from the beginning of the simulation until 20 ns, where there was an increase in fluctuation, but
this was followed by stabilization from 30 ns until the end of the simulation (Figure 5).
Meanwhile, LRTM-EGFR/derivative 5 showed a gradual increase in fluctuation from the
beginning of the simulation and started to stabilize from 25 ns onwards.
367
368
369
370
371
372
373
374
375
376
377
To further investigate the effect of the ligands binding to their respective EGFR kinases,
the amino acid residues within 0.03 nm of the ligands were extracted and their RMSF was
analyzed [50]. The backbone RMSF values for these residues are plotted in Figure 6. The
difference in the RMSF values of WT-EGFR/derivative 5 and WT-EGFR/gefitinib was about
0.02 nm. In comparison to the backbone RMSF of WT-EGFR, LRTM-EGFR was found to have
a similar pattern, where the difference in RMSF values when in-complex with different ligands
was about 0.02 nm. The only residue in LRTM-EGFR that showed a difference in fluctuation of
more than 0.03 nm when in-complex with different ligands was PHE723 residue. In WT-EGFR,
there was a significant difference between the fluctuation of PRO794 for derivative 5 (0.083 nm)
and gefitinib (0.061 nm). The difference may be attributed to the methoxy group on the
quinazoline ring of gefitinib that helped stabilization of PRO794.
22

Backbone RMSD
0.275
0.225
0.175
0.125
0.075
0.00 10.00 20.00 30.00 40.00 50.00
Time (ns)
WT-EGFR/Derivative 5
WT-EGFR/Gefitinib
378
379
Figure 4: RMSD of the backbone residues of WT-EGFR/derivative 5 and WT-EGFR/gefitinib
Backbone RMSD
0.50
0.40
0.30
0.20
0.10
0.00 10.00 20.00 30.00 40.00 50.00
Time (ns)
LRTM-EGFR/Derivative 5
380
381
Figure 5: RMSD of the backbone residues of the LRTM-EGFR/derivative 5 and LRTM-EGFR/gefitinib
Backbone RMSF
0.15
0.10
0.05
0.00
catalytic
loop
DFG,
activation loop
Gly-rich loop
α-helix
Residue number
WT-EGFR/Derivative 5 WT-EGFR/Gefitinib LRTM-EGFR/Derivative 5 LRTM-EGFR/Gefitinib
382
383
Figure 6: Backbone RMSF of WT-EGFR/derivative 5, WT-EGFR/gefitinib, LRTM-EGFR/derivative 5 and LRTM-EGFR/gefitinib
23

Sidechain RMSF
0.30
0.20
0.10
0.00
Gly-rich loop
α-helix
catalytic
loop
DFG,
activation loop
Residue number
WT-EGFR/Derivative 5 WT-EGFR/Gefitinib LRTM-EGFR/Derivative 5 LRTM-EGFR/Gefitinib
384
385
386
Figure 7: Side chain RMSF of WT-EGFR/derivative 5, WT-EGFR/gefitinib, LRTM-EGFR/derivative 5 and LRTM-EGFR/gefitinib
The side chains of both EGFR kinases showed more significant differences in
387
388
389
390
391
392
393
394
395
396
397
398
fluctuation compared to the backbone when they were in-complex with different ligands. For
WT-EGFR, the CYS797 residue of WT-EGFR/gefitinib had a higher fluctuation (0.126 nm)
compared to WT-EGFR/derivative 5 (0.071 nm) (Figure 7). However, this difference was not
observed in LRTM-EGFR (Figure 7). The explanation for this observation could be that the
flexibility of the morpholine tail on gefitinib caused it to move away(as observed in the
optimized position in WT-EGFR) from the initial comformation (Figure 8B), thus, affecting the
fluctuation of the nearest residue, CYS797. For LRTM-EGFR, the most significant difference in
fluctuation was observed on ARG841, where the RMSF for the residue in LRTM-
EGFR/derivative 5 (0.336 nm) was much higher than that of LRTM-EGFR/gefitinib (0.147 nm).
Through visualization of the molecular dynamic trajectory, it was found that the side chain of
ARG-841 in LRTM-EGFR/derivative 5 was flipped in the opposite direction (Figure 8C)
compared to its position in LRTM-EGFR/gefitinib (Figure 8D).
24

399
400
Figure 8: 3D interaction diagrams of the lowest energy trajectories of (A) WT-
EGFR/derivative 5, (B) WT-EGFR/gefitinib, (C) LRTM-EGFR/derivative 5 and (D)
LRTM-EGFR/gefitinib
401
402
403
3.4.2 Hydrogen bond occupancy
404
405
406
407
408
409
The data for hydrogen bond occupancy are summarized in Table 3. For the hydrogen
bonding with MET793 residue, derivative 5 recorded a slightly higher occupancy in both
kinases, at 94.16% in WT-EGFR as compared to gefitinib at 91.80%; and 93.68% in LRTM-
EGFR as compared to gefitinib at 91.48%. In LRTM-EGFR, additional hydrogen bonding pairs
were observed. The amide group of derivative 5 was found to interact with two oxygen atoms of
ASP855 residue (namely, OD1 and OD2), with an occupancy of 47.74% and 12.87%,
25

410
411
412
413
414
respectively. From the visualization of the 3D interaction diagram of LRTM-EGFR/derivative 5
obtained from the simulation (Figure 8C), the presence of bulky mutant arginine residue
(ARG858) pushed the ASP855 residue closer towards derivative 5, allowing the amide group of
derivative 5 to form a hydrogen bond with ASP855, whereas this interaction was absent in WT-
EGFR/derivative 5 (Figure 8A).
415
416
Table 3: The hydrogen bond occupancy of derivative 5 and gefitinib in WT-EGFR and
LRTM-EGFR
Kinase
Ligand
Derivative 5
Gefitinib
Acceptor
Donor
Occupancy (%)
Derivative 5-N
MET793-NH
94.16
WT-EGFR
Gefitinib-N
Derivative 5-N
ASP855-OD1
ASP855-OD2
Gefitinib-N
MET793-NH
MET793-NH
91.80
93.68
12.87
47.74
91.48
Derivative 5
Gefitinib
Derivative 5-N21
Derivative 5-N21
MET793-NH
LRTM-EGFR
417
418
3.4.3 Molecular mechanics Poisson-Boltzmann surface area (MM-PBSA)
419
420
421
422
423
424
425
426
The MM-PBSA results are summarized in Table 4. In both kinases, derivative 5 showed
better binding affinities compared to gefitinib. In WT-EGFR, derivative 5 recorded a binding
energy of -35.287 kcal/mol as opposed to gefitinib (-26.071 kcal/mol). This result is in
agreement with their antiproliferative activity in the A549 cell line, where derivative 5 recorded
a better IC50 value (24.60 ± 0.75 µM) than gefitinib (34.32 ± 1.30 µM). Similarly, in LRTM-
EGFR, derivative 5 recorded slightly better binding energy (-42.867 kcal/mol) as compared to
gefitinib (-41.778 kcal/mol), which is also in agreement with their antiproliferative activity in the
H1975 cell line (IC50, derivative 5: 21.72 ± 1.21 µM, gefitinib: 31.12 ± 0.38 µM).
26

427
Table 4: MM-PBSA values of derivative 5 and gefitinib in WT-EGFR and LRTM-EGFR
MM-PBSA (kcal/mol)
Total
binding
energy
van der
Waals
Polar
solvation
Kinase
Ligand
Electrostatic
SASA
Derivative
-48.286
-53.312
-52.369
-55.340
-19.255
-19.544
-20.059
-15.725
37.109
52.415
34.590
34.725
-4.855
-5.630
-5.029
-5.439
-35.287
-26.071
-42.867
-41.778
5
WT-EGFR
Gefitinib
Derivative
5
LRTM-EGFR
Gefitinib
428
429
3.4.4 Per-residue binding energy decomposition
430
431
432
433
434
435
436
437
438
439
440
441
For WT-EGFR/gefitinib, the negative value binding energy (favorable interaction) was
contributed by LEU718, PHE723, VAL726, LEU792, MET793, and LEU844 residues, while the
positive value binding energy (unfavorable interaction) was contributed by LYS745, GLU762,
ASP800, and ASP855 (Figure 9). Meanwhile, for WT-EGFR/derivative 5, the negative value
binding energy (favorable interaction) was contributed by LEU718, VAL726, LEU792,
MET793, and LEU844 residues, while the positive value binding energy (unfavorable
interaction) was contributed by GLU762 and ARG841 (Figure 9). The binding energy of
GLU762 in WT-EGFR/gefitinib (4.101 kcal/mol) was more positive than that of WT-
EGFR/derivative 5 (1.651 kcal/mol), suggesting that the fluoro group on the gefitinib (the closest
to GLU762) interacted unfavorably with the residue, whereas the fluoro group was not a part of
derivative 5. To compare the binding energy influenced by derivative 5 and gefitinib, ASP800
contributed a much higher positive value when in-complex with gefitinib (1.309 kcal/mol) as
27

442
443
444
445
446
447
448
449
450
451
compared to derivative 5 (-0.098 kcal/mol). This might have been caused by the morpholine tail
of gefitinib that provided an unfavorable interaction with ASP800 (Figure 8B). On the other
hand, ARG841 contributed a much higher positive value when in-complex with derivative 5 as
compared to gefitinib. This could be explained by the n-butyl chain that extended from the amide
group of derivative 5 that interacted unfavorably with ARG841 (Figure 8A). LYS745 and
ASP855, which were in close proximity to the aniline ring of both ligands, showed better binding
energies when in-complex with derivative 5 (0.947 kcal/mol, 0.372 kcal/mol, respectively) as
compared to gefitinib (2.173 kcal/mol, 1.209 kcal/mol, respectively). This suggests that the
introduction of the amide group on derivative 5 exerted good effects on those residues as it
lowered the unfavorable interactions with those residues.
452
453
454
455
456
457
458
459
460
461
462
463
464
For LRTM-EGFR, the binding energies contributed by the residues when in-complex
with both ligands were almost similar to WT-EGFR. Similar to WT-EGFR, the binding energy
of ASP855 residue in LRTM-EGFR/derivative 5 was lower (more favorable) than that of
LRTM-EGFR/gefitinib (Figure 10). The ARG841 residue of LRTM-EGFR/derivative 5
recorded a lower energy (more favorable) value of 0.427 kcal/mol, as compared to the same
residue of WT-EGFR/derivative 5 (1.578 kcal/mol). Based on the RMSF analysis (Figure 7)
which was previously discussed, the side chain of the ARG841 residue of LRTM-EGFR was
flipped in the opposite direction compared to its conformation observed in WT-EGFR. This
movement avoided the unfavorable interaction with the n-butyl chain next to the amide group,
thus returning the lower binding energy of the ARG841 residue in LRTM-EGFR as compared to
the residue of WT-EGFR. LYS745 in LRTM-EGFR/derivative 5 had a much higher positive
value binding energy (1.267 kcal/mol) than in LRTM-EGFR/gefitinib (-0.732 kcal/mol), in
contrast to its observation in WT-EGFR. As observed in Figure 8C, the change in the orientation
28

465
466
467
468
469
470
of ARG841 in LRTM-EGFR/derivative 5 also affected the orientation of ARG858, where it was
pointed towards GLU762, instead of ASP837 as in LRTM-EGFR/gefitinib (Figure 8D). By
pulling GLU762 away, ARG858 indirectly destabilized LYS745 in LRTM-EGFR/derivative 5 as
compared to LRTM-EGFR/gefitinib. This observation could also explain why the binding energy
of GLU762 was much lower in LRTM-EGFR (-0.303 kcal/mol) than that of WT-EGFR (1.651
kcal/mol), as it was stabilized by ARG858 in LRTM-EGFR.
Per Residue Energy Decomposition
GLU762
5.000
4.000
3.000
LYS745
ARG841
ASP855
ASP800
2.000
1.000
0.000
-1.000
-2.000
-3.000
PHE723
LEU718
VAL726
MET793
LEU792
LEU844
847
697
747
797
897
947
997
Residue Number
WT-EGFR/Derivative 5
WT-EGFR/Gefitinib
471
472
Figure 9: Per residue energy decomposition of WT-EGFR/derivative 5 and WT-EGFR/gefitinib
Per Residue Energy Decomposition
3.000
GLU762
LYS745
ASP800
2.000
1.000
0.000
-1.000
-2.000
-3.000
ASP855
ARG841
CYS797
THR854
LEU844
LEU718
VAL726
LEU792
MET793
MET790
697
747
797
847
897
947
997
Residue Number
LRTM-EGFR/Derivative 5
LRTM-EGFR/Gefitinib
473
29