Substituted 1-methyl-4-phenylpyrrolidin-2-ones – Fragment-based design of N-methylpyrrolidone-derived bromodomain inhibitors

Article abstract of DOI:10.1016/j.ejmech.2020.112120

N-Methylpyrrolidone is one of several chemotypes that have been described as a mimetic of acetyl-lysine in the development of bromodomain inhibitors. In this paper, we describe the synthesis of a 4-phenyl substituted analogue – 1-methyl-4-phenylpyrrolidin-2-one – and the use of aryl substitution reactions as a divergent route for derivatives. Ultimately, this has led to structurally complex, chiral compounds with progressively improved affinity as inhibitors of bromodomain-containing protein 4.

Full text of DOI:10.1016/j.ejmech.2020.112120

Journal Pre-proof
Substituted 1-methyl-4-phenylpyrrolidin-2-ones – Fragment-based design of N-
methylpyrrolidone-derived bromodomain inhibitors
J.P. Hilton-Proctor, O. Ilyichova, Z. Zheng, I.G. Jennings, R.W. Johnstone, J. Shortt,
S.J. Mountford, M.J. Scanlon, P.E. Thompson
PII:
S0223-5234(20)30087-8
DOI:
https://doi.org/10.1016/j.ejmech.2020.112120
EJMECH 112120
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 5 December 2019
Revised Date: 16 January 2020
Accepted Date: 3 February 2020
Please cite this article as: J.P. Hilton-Proctor, O. Ilyichova, Z. Zheng, I.G. Jennings, R.W. Johnstone, J.
Shortt, S.J. Mountford, M.J. Scanlon, P.E. Thompson, Substituted 1-methyl-4-phenylpyrrolidin-2-ones
–
Fragment-based design of N-methylpyrrolidone-derived bromodomain inhibitors, European Journal of
Medicinal Chemistry (2020), doi: https://doi.org/10.1016/j.ejmech.2020.112120.
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 Masson SAS.

Graphical abstract

Substituted 1-methyl-4-phenylpyrrolidin-2-ones – Fragment-
Based Design of N-Methylpyrrolidone-Derived Bromodomain
Inhibitors
Authors
1
1
1
1
1
2,3
Hilton-Proctor JP, Ilyichova O, Noor A, Zheng Z, Jennings IG, Johnstone RW, Shortt
4
,5
1
1
1
J, Mountford SJ, Scanlon MJ, Thompson PE
Addresses
1
Medicinal Chemistry, Monash Institute of Pharmaceutical Science, Monash University, 381
Royal Parade, Parkville, Victoria, Australia
2
Peter MacCallum Cancer Centre, 305 Grattan Street, Melbourne, Victoria, Australia
3
The Sir Peter MacCallum Department of Oncology, University of Melbourne, Parkville,
Victoria, Australia
4
Blood Cancer Therapeutics Laboratory, School of Clinical Sciences at Monash Health,
Monash University, 246 Clayton Road, Clayton, Victoria, Australia
5
Monash Haematology, Monash Health, 246 Clayton Road, Clayton, Victoria, Australia
Keywords
BRD4, Bromodomain, Epigenetics, K-ac, NMP
Abbreviations
BET, Bromodomain and Extra Terminal; BRD4 BD1, First Bromodomain of Bromdomain-
containing Protein 4; FBDD, Fragment-based Drug Design; HPLC, High Performance Liquid
Chromatography; K-ac, Acetyl-lysine; LLE, Lipophilic Ligand Efficiency; NHA, Non-
1

Hydrogen Atom; NMP, N-Methylpyrrolidone; RMS, Root Mean Square; SAR, Structure-
Activity Relationship; vdW, van der Waals.
Graphical abstract
2

Abstract
N-Methylpyrrolidone is one of several chemotypes that have been described as a mimetic of
acetyl-lysine in the development of bromodomain inhibitors. In this paper, we describe the
synthesis of a 4-phenyl substituted analogue – 1-methyl-4-phenylpyrrolidin-2-one – and the
use of aryl substitution reactions as a divergent route for derivatives. Ultimately, this has led
to structurally complex, chiral compounds with progressively improved affinity as inhibitors
of bromodomain-containing protein 4.
1
.
Introduction
Fragment-based drug design (FBDD) is a molecular design strategy that has gained
significant traction as a means to develop new drug candidates.[1, 2] It fundamentally enacts
a process of “growing” a small molecule with weak, but ligand efficient interactions with a
target macromolecule into a high-affinity drug-like molecule. That growth is typically
achieved by identifying a set of vectors and using functional group manipulations to extend
away from the fragment to gain new interactions and thus improve binding affinity.[3-5]
In practice, the nature of the FBDD process has shown a tendency towards extended
structures as the functional group transformations are based around building block
chemistry.[6, 7] Obtaining more compact structures demands accessing multiple functional
groups for elaboration on the same core fragment. There has also been a tendency to avoid
chirality in fragment sets as it can lead to more complex syntheses.[8] However, recent
efforts have embraced the discovery, synthesis and elaboration of racemic and chiral
fragments to generate libraries and inhibitors.[9-15]
In our work, we wanted to consider both of these features as they might contribute towards
compounds with increased novelty and characteristics such as structural rigidity that might
provide for improved target affinity. Here, we report on the elaboration by successive aryl
3

substitution of 1-methyl-4-phenylpyrrolidin-2-one, a chiral, phenyl-substituted version of the
solvent N-methylpyrrolidone (NMP) as a bromodomain and extra terminal (BET)
bromodomain inhibitor.
Despite the vast amount of effort that has gone into elaborating the many fragments that are
known to fit inside the acetyl-lysine (K-ac) binding site,[14, 16-22] NMP has been neglected
in these efforts. NMP’s historic use has been in industrial applications such as paint removal
and coatings, along with pharmaceutical purposes including drug solubilisation[23] and depot
injections, remarked for its low reactivity and toxicity. NMP has affinity for several
bromodomains[24] and crystallographic work has established that NMP fits into the
conserved areas of the K-ac binding pocket. NMP shows antimyeloma and
immunomodulatory activity in multiple myeloma models, which is attributed to pleiotropic
bromodomain inhibition.[25] At the moment, it is in a phase I clinical trial for the treatment
of multiple myeloma.[26] NMP also has a well-documented history with bone diseases like
rheumatoid arthritis and osteoporosis[27] by preventing bone resorption[28] and fat
accumulation in bone marrow.[29]
NMP has also served as the inspiration in the development of other novel mimetic warheads,
including the dimethylisoxazole[16] and dihydroquinazolinone.[30] Both of these mimetics
have been greatly utilised in the development of clinical inhibitors[31] and probes.[21, 32,
3
3] We recently described the elaboration of the NMP structure with a series of aliphatic
extensions, which include preparing a mimic of the acetamide-containing bromodomain
inhibitor, Olinone.[34] In this manuscript, we describe the efforts undertaken to develop aryl-
functionalised derivatives of NMP as inhibitors of the first bromodomain of bromodomain-
containing protein 4 (BRD4 BD1).
4

2
.
Results and Discussion
Based upon the available crystal structures of NMP (Supplementary Video S1), we decided
that the 4-position would be the most appropriate place to substitute a phenyl ring, which is a
suitable group to further functionalise through electrophilic aromatic substitution reactions.
Substituted analogues have the potential to interact with the ZA channel and WPF shelf
regions of the bromodomain binding pocket in a similar manner to other inhibitors reported
previously.[21, 31, 32]
2
.1.
Synthesis
The scheme for preparing 1-methyl-4-phenylpyrrolidin-2-one 5 is outlined in Scheme 1.
Commercially-available methyl trans-cinnamate 1 was reacted with nitromethane and
1
,1,3,3-tetramethylguanidine via Michael addition to afford the nitroester 2 in 82% yield as
previously reported[35]. The reduction of 2 to the amine 3 and then cyclisation to the lactam
was achieved by hydrogenation with acetic acid as the solvent. Reduction under these
4
conditions gave the amine 3 as an acetate salt. Refluxing 3 in a mixture of ethanol and
triethylamine drove cyclisation of 3 to the lactam 4 in good yields across the 2 steps. The
lactam was then methylated by methyl iodide and sodium hydride to give the desired phenyl
NMP derivative 5.[36] This procedure reliably gave 5 on gram scale with 50% overall yield.
5

H N
3
O N
2
O
O
O
(
a)
(b)
O
O
O
82%
-
CH COO
3
1
2
3
8
3%
(
c)
(
2 steps)
O
O
(
d)
N
NH
88%
5
4
Scheme 1. Synthesis of 5. Reagents and conditions: (a): CH NO , 1,1,3,3-
3
2
tetramethylguanidine, 42 h, rt; (b): H , Pd/C, CH COOH, 4 d, rt; (c): Et N, EtOH, o/n, reflux;
2
3
3
(
d): N , NaH, MeI, THF, o/n, 0 °C ‰ rt.
2
With the core scaffold 5 in hand, the next step was to functionalise the phenyl group which
was achieved with chlorosulfonation followed by amine substitution of the resultant sulfonyl
chloride. Treatment of 5 with an excess of chlorosulfonic acid gave the intermediate 6. The
para-substituted sulfonyl chloride 6 was treated with a range of primary and secondary
amines to generate sulfonamides 7a-7n in yields ranging from 21-69% (Scheme 2).
Scheme 2. General formation of sulfonamides 7. (a): ClSO H, CH Cl , o/n, rt; (b): NR R ,
3
2
2
1 2
Et N, CH Cl , o/n, rt.
3
2
2
6

Alternatively, bromination using sodium bromide and sulfuric acid[37] gave a mixture of
both ortho- (8) and para-substituted (9) products (Scheme 3) which were separated by
column chromatography. These aryl bromides underwent Suzuki coupling with
phenylboronic acid, giving the biphenyl NMP derivatives 10 and 11 in respectable yields.[38]
Scheme 3. Formation of bromides 8 and 9 with subsequent Suzuki coupling. (a): NaBr,
H SO , 2 h, 80 °C. (b): N , PhB(OH) , Pd(PPh ) , 1,2-dimethoxyethane, 2 M Na CO , 4 h,
2
4
2
2
3 4
2
3
reflux.
The alkyl sulfonamides showed promising affinity (discussed later), so the diethyl
sulfonamide 7a was used to further functionalise the aryl group via bromination to give tri-
substituted benzene products. As outlined in Scheme 4, 7a was treated with a mixture of N-
bromosuccinimide and sulfuric acid to give 12 in good yield.[39] The aryl bromide 12 was
then subject to cross-coupling reactions. Heck coupling gave a series of alkenes (13) which
were in turn reduced to the corresponding alkanes (14). Suzuki coupling using the method
described above gave a series of biaryl compounds (15).[40] The coupling of a thianaphthene
group (15d) resulted in atropisomerism due to the hindered rotation between the
7

thianaphthene and NMP moieties. This was shown by the broadening of the NMP signals in
1
the H NMR, while the signals for the other compounds in this series were well-defined.
Scheme 4. Formation of sulfonamide 7a with bromination and cross-coupling of 12. (a):
ClSO H, CH Cl , o/n, rt; (b): Et NH, Et N, CH Cl , o/n, rt. (c): NBS, H SO , o/n, rt. (d): N ,
3
2
2
2
3
2
2
2
4
2
R-alkene, P(o-tol) , Pd(OAc) , DIPEA, DMF, o/n, 90 °C; (e): H , Pd/C, EtOAc, o/n, rt; (f):
3
2
2
N , R-B(OH) , Pd(PPh ) , 1,2-dimethoxyethane, 2 M Na CO , 4 h, reflux.
2
2
3 4
2
3
8

The chiral nature of all these products led us to examine the generation of single enantiomers
of 5, as well as the biaryl derivative 15c by diastereomeric resolution. As shown in Scheme 5,
lactam 4 was treated with lithium diisopropylamide and (S)-Naproxen chloride 16[41] to give
diastereomeric imides 17, which were separated carefully by column chromatography to
1
acquire the (R,S)-17 and (S,S)-17. In agreement with the literature, the H NMR signals of the
first eluted product corresponded with the (R,S)-diastereomer, and the proton signals of the
second product matched the (S,S)-diastereomer. The imides were hydrolysed with aqueous
potassium hydroxide to give (R)-4 and (S)-4. The specific rotation of each enantiomer was
consistent with literature values. They were then methylated as above to give the respective
isomers (R)-5 and (S)-5, which were carried forward to (R)-15c and (S)-15c using the same
conditions described as in Scheme 4. The enantiomers resolved well by chiral
chromatography and matched the corresponding peaks in the racemic mixture, with the first
eluting product being the S-enantiomer.
9

1
0

Scheme 5. Stereoisomer resolution of 4 with corresponding synthesis of phenyl NMP
isomers. (a): LDA, THF, 3 h, –78 °C. (b): Column chromatography, 10% EtOAc in
petroleum benzine. (c): 1 M KOH, THF, o/n, rt. (d): NaH, MeI, THF, N , o/n, 0 °C ‰ rt. (e):
2
ClSO H, CH Cl , o/n, rt; (f): Et NH, Et N, CH Cl , o/n, rt. (g): NBS, H SO , o/n, rt. (h): N ,
3
2
2
2
3
2
2
2
4
2
(
CH O) -C H -B(OH) , Pd(PPh ) , 1,2-dimethoxyethane, 2 M Na CO , 4 h, reflux.
3 2 6 3 2 3 4 2 3
2
.2. BRD4 BD1 FRET Assay Results
The phenyl-substituted NMP compounds were all assessed for inhibition of a biotinylated
5
8
12 16
tetra-acetylated histone peptide (H4K K K K (ac )) binding to BRD4 BD1 in a FRET-
4
based assay. The results of the first series of assays are shown in Table 1.
Firstly, the addition of a phenyl group to NMP (5) increased the binding affinity for BRD4
BD1 by approximately 2-fold. This confirmed the suitability of the aryl substituent for
functionalisation of NMP.
Introduction of the sulfonamide group gave some significant improvements with alkyl
sulfonamides 7a, 7b, 7h and 7i having IC values between 120 – 150 µM, nearly a 10-fold
5
0
increase from 5. However, other alterations to the sulfonamide – arylpiperidines,
terahydroquinolines, benzylamines and nipecotic acids – resulted in a loss of activity. An
interesting observation was that the activity of 7n was similar to compounds 7a, 7b, 7h and
7
i, despite possessing a bulky adamantyl group. These observations, paired with the
generated X-ray crystal structures to be discussed later, suggest that certain aliphatic
substituents are best tolerated.
The introduction of a bromine group resulted in improved affinity compared to 5; the ortho-
bromide 8 (321 µM) had marginally affinity than the para-bromide 9 (417 µM). The
substitution to a phenyl group was also tolerated, but was preferable in the para position 11
(434 µM) compared to the ortho position 10 (575 µM).
1
1

Table 1. SAR data on initial phenyl NMP derivatives.
Compound
R=
IC (µM
± SEM)
d
e
5
0
a
cLogP LLE
Number
H
2660 ± 480
-0.40
1.16
2.97
1.78
5
1140 ± 10
7
a
123 ± 29
144 ± 18
276 ± 12
473 ± 13
1.08
1.55
0.30
2.96
2.83
2.29
3.26
0.37
7b
7
c
7d
7
e
450 ± 33
2.48
1.71
0.87
1.65
c
7
f
443
b
7
g
250 ± 3
0.14
0.67
3.46
3.16
b
7h
147 ± 5
b
7
i
134 ± 7
0.45
0.55
3.42
3.10
b
7
j
223 ± 23
7k
1680 ± 240
0.76
2.01
1
2

Table 1 (continued).
7
l
1960 ± 90
1390 ± 20
151 ± 43
0.76
0.19
2.60
1.95
2.67
1.22
7m
7n
8
321 ± 48
417 ± 25
2.02
2.02
1.47
1.36
9
1
0
575 ± 177
2.75
3.05
0.49
0.31
11
434 ± 116
a
b
c
d
Unless stated otherwise, all results of means are n ≥ 2. SD reported. n = 1. cLogP values
e
were calculated using ChemDraw 18.1. Lipophilic ligand efficiency = –log(IC ) – cLogP.
5
0
Control compound used: I-BET726 (IC : 0.01 µM).
5
0
The synthetic elaboration of 7a yielded di-substituted aryl NMP analogues that produced a
range of activities, which are shown in Table 2. The addition of the bromide to the selected
sulfonamide (12) resulted in a 5-fold poorer inhibitor compared to 7a. The products from the
Heck/hydrogenation steps (14a – 14b) were found to be inactive. The change to a phenyl
group (15a) was detrimental to activity, but compound 15c showed much improved inhibition
with an IC of 52 µM.
5
0
1
3

Table 2. SAR data on halide-coupled NMP derivatives from 7a.
Compound
Number
IC (µM
± SEM)
c
d
5
0
R=
a
cLogP
LLE
7
1
a
2
H
Br
123 ± 29
652 ± 116
1.08
1.95
2.83
1.24
1
4a
4500 ± 40
>5000
3.48
4.01
2.67
2.90
-1.13
-1.70
-0.18
0.77
1
4b
1
5a
3200 ± 80
217 ± 19
1
5b
1
5c
51.6 ± 2.5
2.35
3.72
1.75
1
5d
2670 ± 220
-1.14
a
c
Unless stated otherwise, all results of means are n ≥ 2. cLogP values were calculated using
d
ChemDraw 18.1. Lipophilic ligand efficiency = –log(IC ) – cLogP. Control compound
5
0
used: I-BET726 (IC : 0.01 µM).
5
0
As described above, we also had the single enantiomers of compounds 5 and 15c. There was
a modest level of chiral recognition shown with the (R)-enantiomer being the eutomer in both
cases (Table 3). (R)-15c proved to be the highest affinity compound of this study with an IC50
1
4

of 14.1 µM, making it 190-fold more potent than NMP itself (dose response curves of 15c,
R)-15c and (S)-15c are available in Supplementary Figure S3).
(
Table 3. Assay data on resolved isomers.
Compound
R=
IC (µM
± SEM)
c
d
5
0
a
cLogP
LLE
Number
5
1140 ± 6
1.16
1.78
2.04
1.69
b
(
R)-5
629 ± 26
1.16
(
S)-5
1430 ± 170
1.16
1
5c
51.6 ± 2.5
14.1 ± 0.3
2.35
1.75
2.50
2.10
(
R)-15c
2.35
2.35
(
S)-15c
35.2 ± 1.3
a
b
c
Unless stated otherwise, all results of means are n ≥ 2. SD reported. cLogP values were
d
calculated using ChemDraw 18.1. Lipophilic ligand efficiency = –log(IC ) – cLogP.
5
0
Control compound used: I-BET726 (IC : 0.01 µM).
5
0
With the elaboration of the phenyl NMP structure, we were interested to see how the
structural modifications impacted the efficiency metrics. While 5 has an initial LLE of 1.78,
1
5

the increases in affinity, while associated with a growing molecular weight, also results in a
rise in LLE due to the inclusion of the sulfonamide group (reducing cLogP), as shown in
Table 1. Further substitution via Suzuki reactions gave further improved affinity, but at the
expense of LLE (Table 2). Overall, the development of (R)-15c led to an improvement in
affinity by over 2 orders of magnitude with an LLE of 2.50. From the perspective of fragment
based elaboration this compares well to the development of BRD4 inhibitor I-BET726 (IC50
=
0.003 µM, LLE = 1.93)[42] if considered as deriving from a racemic tetrahydroquinaldine
fragment (IC = 158 µM, LLE = 1.51).[43, 44]
5
0
2
.3. X-ray Crystallography
We were able to gather a number of X-ray structures of our compound series to support the
SAR studies of these compounds.
Firstly, a crystal structure of 5 bound to BRD4 BD1 was solved (Figure 2, Supplementary
Video S2). The NMP segment of 5 retains the K-ac mimetic function, replicating the
interactions between Asn140 and Tyr97. The phenyl group moiety of 5 emerges from the
pocket and rests on the WPF shelf and appears to form weak interactions of its own with the
WPF shelf residues Trp81 and Pro82 and the gatekeeper residue Ile146. While 5 was
submitted as a racemic mixture, only the (R)-enantiomer was observed in the X-ray crystal
structure of BRD4 BD1. This observation is consistent with the binding activity acquired
from both (R)-5 and (S)-5. The (S)-isomer in this pose would have more than likely produced
a steric clash with the wall created by the ZA loop.
1
6

Tyr97
Tyr97
A
B
Asn140
Asn140
Ile146
Ile146
Pro82
Pro82
Trp81
Figure 2. Comparison of binding substrates in BRD4 BD1. [A] K-ac (purple blue) (PDB:
Trp81
3
UVW). [B] (R)-5 (purple, resolution: 1.5 Å). Water molecules shown as red spheres. Pro86,
Val87 and Asp88 were hidden for better visualisation. Distances in Å, highlighted by red
dashes.
Crystal structures of the sulfonamides 7b and 7h were also solved. As shown in Figure 3,
only the (R)-isomer was evident in these co-crystal structures and the orientation of the NMP
ring in the binding pocket matches closely with compound 5. Despite the NMP backbone
atoms overlaying closely (RMS = 0.20 and 0.37, respectively), the phenyl group appears to
have shifted slightly (RMS = 1.25 and 1.63, respectively). This change has positioned the
phenyl group closer towards the side chain of Leu92, but further away from Trp81, Pro82 and
Ile146. Both the piperidyl group of 7b and n-propyl moiety of 7h rest in a pocket adjacent to
the WPF shelf and form weak vdW interactions with Trp81, Asp145 and Met149 (Figure 4,
Supplementary Video S3).
1
7

Tyr97
A
C
B
Leu92
Asn140
Tyr97
Leu92
Asn140
Ile146
Pro82
Pro82
Trp81
Trp81
Tyr97
D
Asn140
Leu92
Tyr97
Leu92
Ile146
Asn140
Pro82
Pro82
Trp81
Trp81
1
8

E
Tyr97
Ile146
F
Leu92
Asn140
Tyr97
Leu92
Asn140
Pro82
Pro82
Trp81
Trp81
Figure 3. Comparison of X-ray crystal structures of NMP derivatives. [A, C, E] Binding site.
Pro86, Val87 and Asp88 were hidden for better visualisation. [B, D, F] Emphasis on phenyl
group. [A, B] 5 (purple, resolution: 1.5 Å). [C, D] (R)-7b (raspberry, resolution: 1.55 Å). [E,
F] (R)-7h (orange, resolution: 1.59 Å). Water molecules shown as red spheres. Distances in
Å, highlighted by red dashes.
A
B
Trp81
Trp81
Met149
Met149
Asp145
Asp145
1
9

Figure 4. X-ray crystal structures illustrating the interactions between sulfonamide alkyl
moieties and BRD4 BD1. [A] (R)-7b (raspberry, resolution: 1.55 Å). [B] (R)-7h (orange,
resolution: 1.59 Å). Water molecules shown as red spheres. Distances in Å, highlighted by
red dashes.
Finally, we were able to obtain a co-crystal structure of our highest affinity compound, 15c
(Figure 5 and Supplementary Video S4). To our surprise, the inclusion of the
dimethoxyphenyl moiety had completely changed the orientation of the molecule. The
dimethoxyphenyl group fills the acetamide binding pocket, using the two oxygen atoms to
form a pair of hydrogen bonds to Asn140. The sulfonamide was re-oriented into the ZA
channel, with the sulfonyl forming a hydrogen bond with the backbone nitrogen of Asp88
and the diethyl group lining up against Gln85. One of the ethyl moieties is situated in a small
pocket that is in close proximity to three backbone carbonyls: Pro82, Gln85 and Pro86.
Finally, the NMP segment rested in the pocket adjacent to Trp81, close to the sulfur atom of
Met149. Again, submitted as a racemate, the (S)-enantiomer is observed in the co-crystal
(Figure 6), although (S)-15c was found to have lower affinity than (R)-15c albeit only 2-fold.
This is potentially due to the differences in crystallisation and assay conditions (electron
density maps can be found in Supplementary Figure S4).
2
0

A
C
Tyr97
B
Asn140
Trp81
Ile146
Met149
D
Asp88
Leu92
Pro86
Gln85
Met149
Pro82
Trp81
Figure 5. X-ray crystal structure of (S)-15c (cyan, resolution: 1.48 Å) bound to BRD4 BD1,
illustrating interactions with the bromodomain. [A] Binding site. Pro86, Val87 and Asp88
were hidden for better visualisation. [B] NMP. [C] Phenyl ‘hub’. [D] Sulfonamide. Water
molecules shown as red spheres. Distances in Å, highlighted by red dashes.
2
1

A
B
C
D
Figure 6. Comparison of X-ray crystal structures of bound NMP isomers in BRD4 BD1. [A]
(
R)-5 (purple, resolution: 1.5 Å). [B] (R)-7b (raspberry, resolution: 1.55 Å). [C] (R)-7h
orange, resolution: 1.59 Å). [D] (S)-15c (cyan, resolution: 1.48 Å).
(
The presence of catechol ethers as acetamide mimics in BET inhibitors is precedented. The
kinase inhibitor GW612286X was assessed among other inhibitors by Ember et al.[45] for
BET inhibition. In their crystallographic studies, the trimethoxy group of GW612286X
served as the K-ac mimetic in the binding pocket of BRD4 BD1. Chen et al.[46] evaluated
the trimethoxy group as a mimetic for novel BRD4 inhibitors. In one of their compounds,
DC-BD-29, the 3- and 4-methoxy groups also formed the dual interaction to Asn140 seen in
1
5c. While the catechol ether of each compound fits in the binding pocket of BRD4 BD1,
2
2

they all have subtle differences with how they interact with the conserved tyrosine and
asparagine residues (Figure 7).
A
B
Tyr97
Asn140
C
D
Tyr97
Asn140
2
3

Tyr97
E
F
Asn140
Figure 7. Comparison of catechol ether groups bound to BRD4 BD1. [A, B] (S)-15c (cyan,
resolution: 1.48 Å). [C, D] GW612286X (pink) (PDB: 4O78). [E, F] DC-BD-29 (sand)
(PDB: 5H21). [A, C, E] Chemical structure. [B, D, F] Binding site. Water molecules shown
as red spheres. Pro86, Val87 and Asp88 were hidden for better visualisation. Distances in Å,
highlighted by red dashes.
3
.
Conclusion
BRD4 is a well-characterised target for FBDD approaches and provides a useful exemplar of
bromodomain inhibitor discovery in general.[21, 47, 48] NMP itself has been co-crystallised
with several bromodomains and could be a starting point for campaigns against each one.
Here, we have tested these principles with a 4-phenyl derivative of NMP.
We devised a synthetic strategy built around electrophilic aromatic substitution as a
chemically tractable way to introduce functionality into the “naked” ring of 5 which is
available in multigram quantities. We were able to go from a mono-substituted precursor to a
tri-substituted analogues series, and in doing so, improve affinity 190-fold with just 30
compounds prepared. In terms of efficiency metrics, it is notable that the increase affinity has
come with a well preserved LLE with the sulfonamide, a positive influence in keeping the
2
4

logP of analogues down. As a proof-of-concept, this work supports further investigation of
direct aromatic substitution reactions to drive FBDD programs forward. Here, we have
utilised the ortho- and para- directing effects of the NMP group to create the vectors, but
other means of ring activation could generate pathways to meta substitution, relative to the
NMP group. The chiral nature of the precursor compound 5 might typically be considered a
disincentive to progression in an FBDD campaign, but here, we show that ready access to
both enantiomers gave rise to pairs of analogues that could be assessed either as racemates or,
where warranted, in enantiopure form.
Crystallography was also an important support for the development of the series with the
elaboration of NMP to 5 and subsequently to 7e showing the expected poses in response to
structure elaboration. In an unexpected turn, a crystal structure of the lead compound 15c
adopted a new binding orientation in the BRD4 bromodomain pocket. This change provides a
clear explanation of the SAR surrounding the tri-substituted compounds 14 and 15 and shows
that the tri-substituted aryl ring is perhaps not the best for preserving the NMP group in the
acetamide pocket. On the other hand, 15c itself may prove an excellent new starting point for
analogue design projecting the sulfonamide and NMP “vectors” in potentially productive
directions.
2
5

4
.
Experimental
4
.1.
Chemistry
1
13
H and C Nuclear Magnetic Resonance spectra were conducted on a Bruker Advance III
Nanobay 400 MHz spectrometer coupled to the BACS 60 automatic sample changer and
obtained at 400.1 MHz and 100.6 MHz. All spectra were processed using MestReNova 6.0
1
software. The chemical shifts of all H were measured relative to the expected solvent peaks
of the respective NMR solvents; CDCl , 7.26; MeOD, 3.31; DMSO, 2.50. The chemical
3
1
3
shifts of all C were measured relative to the expected solvent peaks of the respective NMR
solvents; CDCl , 77.2; MeOD, 49.0; DMSO, 39.5. The data for all spectra are reported in the
3
following format: chemical shift (integration, multiplicity, coupling constant, assignment).
Multiplicity is defined as; s = singlet, d= doublet, t = triplet, q = quartet, quint = quintet, dd =
doublet of doublets, dt = doublet of triplets, td = triplet of doublets, tt = triplet of triplets, qd =
quartet of doublets, ddd = doublet of doublet of doublets, m = multiplet. Coupling constants
1
13
are applied as J in Hertz (Hz). For H and C spectra, refer to Supplementary Figure S5.
All HRMS analyses were done on an Agilent 6224 TOF LC/MS Mass Spectrometer coupled
to an Agilent 1290 Infinity (Agilent, Palo Alto, CA). All data were acquired and reference
mass corrected via a dual‐spray electrospray ionisation (ESI) source. Each scan or data point
on the Total Ion Chromatogram (TIC) is an average of 13,700 transients, producing a
spectrum every second. Mass spectra were created by averaging the scans across each peak
and background subtracted against the first 10 seconds of the TIC. Acquisition was
performed using the Agilent Mass Hunter Data Acquisition software version B.05.00 Build
5
.0.5042.2 and analysis was performed using Mass Hunter Qualitative Analysis version
B.05.00 Build 5.0.519.13.
2
6

All LCMS analyses were carried out on an Agilent 6100 Series Single Quad LC/MS coupled
with an Agilent 1200 Series HPLC, 1260 Infinity G1312B Binary pump, 1260 Infinity
G1367E 1260 HiP ALS autosampler and 1290 Infinity G4212A 1290 DAD detector. The
liquid chromatography conditions were: reverse phase HPLC analysis fitted with a Luna
C8(2) 5 µL 50 X 4.6 mm 100Å at a temperature of 30 °C. The sample injection volume was
5
µL, which was run in 0.1% formic acid in acetonitrile at a gradient of 5-100% over 10
minutes. Detection methods were either 254 nm or 214 nm. The mass spectrum conditions
were: Quadrupole ion source with Multimode-ES. The drying gas temperature was 300 °C
and the vaporizer temperature was 200 °C. The capillary voltage in positive mode was
2
1
000V, while in negative mode, the capillary voltage was 4000V. The scan range was 100-
000 m/z with a step size of 0.1 second over 10 minutes.
TLCs were carried on Merck TLC Silica gel 60 F254 plates using the appropriate mobile
phase. Purification by column chromatography was conducted with Davisil Chromatographic
Silica LC60A (40-63 micron) using the specified mobile phases.
Purification on reverse-phase HPLC was done on a Waters Delta Prep 2000 Prep HPLC
System that was fitted with a Waters Delta Prep 2000 Pump and Controller. Samples were
injected into a Waters Prep Rack with Manual Injector, which were run through a Luna C8(2)
1
0 µL 50 X 21.20 mm 100Å and Waters 486 Tunable Absorbance Detector. The conditions
were: Solvent A (0.1% TFA in H O) and Solvent B (0.1% TFA in CH CN), with a gradient
2
3
of 0-80% Solvent B over a 20-minute period.
Compound purity was determined using an Agilent 1260 Infinity Analytical HPLC (1260
Infinity G1322A Degasser, 1260 Infinity G1312B Binary pump, G1367E HiP ALS
autosampler, 1260 Infinity G1316A Thermostatted Column Compartment, and 1260 Infinity
G4212B DAD detector. The liquid chromatography conditions were: reverse phase HPLC
2
7

analysis fitted with a Zorbax Eclipse Plus C18 Rapid Resolution 4.6 X 100 mm 3.5‐Micron.
The sample injection volume was 1 µL, which was run in Solvent A (0.1% TFA in H O) and
2
Solvent B (0.1% TFA in CH CN), with a gradient of 5-100% Solvent B over a 10-minute
3
period. All compounds submitted for assays and X-ray crystallography studies were assessed
for purity of 95% or greater on 214 nm and 254 nm.
Separation on chiral HPLC was done on an Agilent 1260 Infinity Prep HPLC (1260 Infinity
G1361A Prep Pump, 1260 Infinity G2260A Prep ALS, 1260 Infinity G1328C Man Inj, 1260
Infinity G1315D DAD VL, and 1260 Infinity G1364B FC‐PS. The liquid chromatography
conditions were: normal phase HPLC analysis fitted with a Phenomenex Lux 5 µm Cellulose-
1
4.6 x 150 mm. The sample injection volume was 10 µL at a concentration of 1 mg/mL,
which was run in Solvent A (40% petroleum ether) and Solvent B (60% ethanol), over a 10-
minute period.
Specific rotations were measured using a Jasco P-2000 polarimeter. The light source was a
sodium lamp with a wavelength of 589 nm. Compounds were loaded into a 3.5 mm x 100
mm cylindrical glass cell. Digital integration time was 5 seconds per cycle for 10 cycles.
4
.1.1. Methyl 4-nitro-3-phenylbutanoate (2)
Methyl trans-cinnamate 1 (5.01 g, 30.9 mmol) was weighed into a 250 mL round-bottom
flask with a magnetic stir-bar and dissolved in nitromethane (17 mL, 0.308 mol), to which
stirring commenced. 1,1,3,3-tetramethylguanidine (775 µL, 6.18 mmol) was added, and the
mixture was left to stir for 42 hours at room temperature, checking for completion with TLC.
After completion, the mixture was quenched with 5% HCl (75 mL) and washed with diethyl
ether (3 x 75 mL). The organic layers were combined, dried with MgSO , filtered and
4
concentrated under reduced pressure to afford an orange-coloured oil. The oil was then
purified using column chromatography (Mobile phase: 10% EtOAc in petroleum benzine,
2
8