Design and synthesis of a fragment set based on twisted bicyclic lactams

Article abstract of DOI:10.1016/j.bmc.2018.02.027

Current fragment sets tend to be dominated by flatter molecules, and their shape diversity does not reflect that of the fragments that are theoretically possible. The design and synthesis of a set of bridged fragments containing a bridgehead nitrogen is d

Full text of DOI:10.1016/j.bmc.2018.02.027

Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Design and synthesis of a fragment set based on twisted bicyclic lactams
a,b,
Haitham Hassan a, Stephen P. Marsden a, , Adam Nelson
⇑
⇑
a School of Chemistry, University of Leeds, Leeds LS2 9JT, UK
b Astbury Centre for Structural Molecular Biology, University of Leeds, Leeds LS2 9JT, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Current fragment sets tend to be dominated by flatter molecules, and their shape diversity does not
reflect that of the fragments that are theoretically possible. The design and synthesis of a set of bridged
fragments containing a bridgehead nitrogen is described. Many of these fragments contain twisted lac-
tams whose modulated electronic properties may present unusual opportunities for interaction with tar-
get proteins. The demonstrated novelty, three-dimensionality and molecular properties of the set of 22
fragments may provide valuable, and highly distinctive, starting points for fragment-based drug
discovery.
Received 17 January 2018
Revised 15 February 2018
Accepted 16 February 2018
Available online xxxx
Keywords:
Fragments
Drug discovery
Twisted amides
Shape diversity
Ó 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://
creativecommons.org/licenses/by/4.0/).
1. Introduction
extent of deformation.9,10 In extremely twisted amides, for exam-
ple, the electronic properties of the functional group are more rem-
Over the last 15–20 years, fragment-based discovery has
become a mainstream approach in medicinal chemistry.1 Conse-
quently, guidelines have been formulated to facilitate the assembly
of fragment sets that target diverse relevant chemical space.2 How-
ever, current fragment sets tend3 to be dominated by flatter (gen-
erally heteroaromatic) molecules whose shape diversity is not
representative of the fragments that are theoretically possible.4
As a result, significant effort has been invested in the design of
fragment sets with higher shape diversity.3a,d More three
dimensional (3D) fragments have inherently higher molecular
complexity which has been argued5 to result in lower hit rates in
fragment screens.6 Yet, such fragments are likely to offer distinc-
tive opportunities for subsequent growth along specific vectors.
Here, we describe the design and synthesis of a fragment set
that is based on a number of bicyclic ring systems containing a
bridgehead nitrogen atom. Such ring systems are substructures
within the frameworks of a diverse range of alkaloid natural prod-
ucts (see Fig. 1 for examples7) which can serve as an inspiration for
drug discovery.8 The geometric constraints imposed by these ring
systems can perturb functional group properties and characteris-
tics by restricting or preventing the overlap of the bridgehead
iniscent of those of unconjugated amino ketones (for examples, see
Fig. 2): this is reflected in the electrophilic reactivity of the car-
bonyl group10,11 and in the full12 or partial13 N-protonation of
the amide in contrast to the O-protonation observed with non-
twisted amides. Despite the unusual and distinctive hydrogen-
bonding opportunities offered by such motifs, bicyclic lactams
have barely been explored in a medicinal chemistry context, and
may therefore provide new opportunities in drug discovery.14
2. Results and discussion
Our synthetic approach to bridged bicyclic fragments is shown
in Scheme 1. Thus, 3-(x-carboxylate)-substituted piperidines 1
would be lactamised to yield a range of bridged lactams 2 (with
variable n and R). It was envisaged that, with appropriate choice
of substituent, addition of ring(s) (? 3) or functionalisation of
the twisted amide (? 4) might be possible to yield related scaf-
folds. Finally, decoration would yield corresponding fragments
for addition to a screening set.
2.1. Synthesis of bridged bicyclic lactams
nitrogen lone pair with an adjacent
p-system, and in the case of
twisted amides, parameters have been developed to describe the
Initially, a range of 3-(x-carboxylate)-substituted piperidine
substrates 7 was prepared (Scheme 2 and Table 1). The reductive
aminations between the piperidin-3-one 5 and the amino esters
6a–d gave the corresponding 3-amino piperidines 7a–d. Alterna-
tively, condensation of the piperidin-4-one 9 with pyrrolidine gave
an enamine that was reacted directly with ethyl acrylate to give
⇑
Corresponding authors at: School of Chemistry, University of Leeds, Leeds LS2
9JT, UK (A. Nelson).
E-mail addresses: s.p.marsden@leeds.ac.uk (S.P. Marsden), a.s.nelson@leeds.ac.
uk (A. Nelson).
https://doi.org/10.1016/j.bmc.2018.02.027
0968-0896/Ó 2018 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Please cite this article in press as: Hassan H., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.02.027

2
H. Hassan et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
H
O
H
N
OH
Me
N
O
N
H
H
Tubifoline
E
Lycoposerramine
O
HO
N
H
HO
N
N
N
H
A
Eucophylline
Kopsihainanine
Fig. 1. Ring systems with bridgehead nitrogens embedded in alkaloids.
O
Fig. 3. X-ray crystal structures of 8a, 8c and 16.
N
N
O
12
13
bicyclo[4.3.1]decane bicyclic ring systems in moderate yield.
Unfortunately, the approach did not enable the synthesis of the
alternate bicyclo[3.2.2]nonane, bicyclo[4.2.2]decane and bicyclo
[4.1.1]octane ring systems, presumably owing to increased ring
strain in these species. Furthermore, deprotection of benzyl-pro-
tected 8a and 8c and Cbz-protected 8d was not successful under
a range of hydrogenation conditions (see Fig 3).
Fig. 2. Unstrained amide 12 and Kirby’s most twisted amide 13.
R1
X
n
Add ring
N
O
The distortion10 of the bridged bicyclic lactams 8a–d and 11,
and their derivatives (see below), was assessed. 13C NMR spec-
troscopy is a valuable technique for investigating amide distortion,
with the chemical shift of the carbonyl carbon serving as a sensi-
tive probe of strain (see Fig. 2; unstrained d-lactam 12, dC: 165
ppm; Kirby’s most twisted amide 13, dC: 200 ppm).11,16 The lac-
tams based on bicyclo[3.3.1]nonane ring systems (e.g. 8a, 8b and
11; amide carbonyl dC: 181–185 ppm) were markedly more
strained than those (e.g. 8c and 8d; amide carbonyl dC: 174–177
ppm) based on bicyclo[4.3.1]decane ring systems.
R1
3
R1
X
X
n
cyclise
n
CO2H
N
N
R1
X
Boc
O
2
1
n
Functionalise
twisted amide
N
R2
4
Scheme 1. Envisaged synthetic approach to bridged bicyclic fragments.
Strain was also assessed by analysis of the X-ray crystal struc-
tures of 8a and 8c and a derivative of 11 (lactam 16, see below),
and determination of standard amide distortion parameters (Sup-
porting Information). For example the sum of the bond angles
around the amide nitrogen (b) deviated further from 360° in the
bicyclo[3.3.1]nonanes (8a: b = 337°; 16: b = 341°) than in the bicy-
clo[4.3.1]decane ring system 8c (b = 349°).
2.2. Synthesis of hetaryl-annulated scaffolds
The potential to access alternative scaffolds by annulation of
heteroaromatic rings by exploiting the ketone in bicyclic lactam
11 was then explored (Scheme 3). Thus, reaction of the ketone
11 and propargylamine, catalyzed by 2.5 mol% NaAuCl4,17 gave
the related pyrido-fused lactam 14 in 43% yield. In a similar vein,
reaction of 11 with 2-iodoaniline, catalyzed by 20 mol% Pd(OAc)2,
gave the indolo-fused lactam 15 in 53% yield.
Scheme 2. Synthesis of bridged bicylic lactams (see Table 1).
the 3-substituted piperidin-4-one 10. Boc deprotection of 7a–d
and 10, and ester hydrolysis, was followed by Bu2SnO-mediated
cyclisation to give the corresponding lactams 8a–d and 11.15 The
approach enabled the synthesis of both bicyclo[3.3.1]nonane and
Table 1
Synthesis of bridged bicyclic lactams (see Scheme 2).
Intermediate synthesis
Lactam synthesis
Yieldb%
dC
c
Substrates
Product
Yield,%
Product
5, 6a (n = 1; R1 = Bn; R2 = Et)
5, 6b (n = 1; R1 = Me; R2 = Me)
5, 6c (n = 2; R1 = Bn; R2 = Et)
5, 6d (n = 2; R1 = H; R2 = Et)
9, ethyl acrylate
7a (n = 1; R1 = Bn; R2 = Et)
7b (n = 1; R1 = Me; R2 = Me)
7c (n = 2; R1 = Bn; R2 = Et)
7d (n = 2; R1 = Cbz; R2 = Et)
10
68
57
72
74a
60
8a (n = 1; R1 = Bn)
8b (n = 1; R1 = Me)
8c (n = 2; R1 = Bn)
8d (n = 2; R1 = Cbz)
11
46
45
59
52
57
182.9
182.6
176.2
176.7
182.4
a
b
c
Isolated after Cbz protection: CbzCl (1.1 eq.), NaHCO3 (6.0 eq.) in CH2Cl2.
Yield over 3 steps.
Lactam carbonyl.
Please cite this article in press as: Hassan H., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.02.027

H. Hassan et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
3
treatment of 8c and 8d with POCl3, and direct reaction with aniline,
gave the amidines 20a and 20b in good yield (for another example,
see Supporting Information). The amidines 20a–b were shown to
be stable, being unchanged after heating at 65 °C in DMSO for 3
days.
Third, we prepared fragments in which the amino group within
the bicyclic amino lactam had been decorated (Scheme 6). Because
N-benzyl and N-Cbz deprotection had proved difficult, we investi-
gated the cyclisation of precursors with an unprotected secondary
amine. NaBH(OAc)3-mediated reductive amination of N-Boc piper-
idin-3-one with the hydrochloride salts of glycine ethyl ester 6e
Scheme 3. Synthesis of hetaryl-annulated ring systems.
2.3. Synthesis of a diverse fragment set
A range of approaches was exploited to enable decoration at dif-
ferent positions within the bicyclic scaffolds. It was decided to
focus on the synthesis of a diverse set of fragments with 17 or
fewer heavy atoms and with clogP < 2.5.
First, we investigated O-functionalisation of the alcohol 16
derived from the bicyclic keto lactam 11 (Scheme 4). Treatment
of 11 with NaBH4 resulted in attack onto the less hindered face
of the ketone to yield the secondary alcohol 16 with >95:<5
diastereoselectivity. Treatment of 16 with NaH, followed by reac-
tion with 2-bromothiazole, 2-fluoropyridine or methyl iodide gave
the ethers 17a, 17b and 17c respectively.
Scheme 5. Twisted lactam functionalization. Panel A: Chloroenamine formation
and Suzuki reaction. Panel B: Synthesis of bicyclic amidines.
Second, we showed that the twisted amide of 11 could also be
functionalised (Scheme 5, Panel A). Thus, treatment of 11 with
POCl3 gave the stable and isolable chloroenamine 18. Subsequent
Suzuki coupling of the chloroenamine 18 with p-tolyl boronic acid
gave the enamine 19.18 Notably, reaction of the less strained bicyc-
lic lactams 8c and 8d with POCl3 did not yield isolable chloroe-
namine products; presumably, the twisted nature of 11 increases
its reactivity by disrupting conjugation of the nitrogen lone pair
and the alkene. Instead, the bicyclic lactams 8c and 8d were con-
verted into related bicyclic amidines (Scheme 5, Panel B). Thus,
N
O
N
A
A
N
O
a
CO2
7e
Et
STAB
70%
N
8e,
68%
Boc
O
O
O
H
N
N
Cl
N
CO2Et
NaHCO3
90%
N
N
O
CO2Et
Boc
N
Boc
a
,
7f
8f,
32%
21a
6e
74%
(from
)
O
O
O
OMe
O
N
Cl
N
N
A
OMe
CO2Et
N
O
NaHCO3
95%
Boc
63%a
Me
N
7g
8g,
O
O
O
N
A
Cl
N
CO2Et
N
Boc
NaHCO3
Quant%
O
a
7h
8h
35%
N
Ar
O2S
N
N
O2S
O2
S
H
N
N
A
Cl
Ar
NaHCO3
Quant%
N
CO2Et
N
CO2Et
N
Boc
Boc
O
,
a
21b 77%
6f
7i
(from
)
8i,
43%
CN
CN
Br
N
CN
N
A
K2CO3
77%
N
N
CO2Et
O
Boc
13%a
7j
8j,
Scheme 6. Synthesis of bicyclic lactams functionalised at a remote nitrogen atom.
Method A: (1) NaOH (1.1 equiv), 1:1 H2O-MeOH, 70 °C. (2) 6 N HCl. (3) Bu2SnO,
PhMe, 120 °C. aYield over 3 steps.
Scheme 4. Reduction of the bicyclic keto lactam 11, and O-functionalization of the
resulting alcohol.
Please cite this article in press as: Hassan H., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.02.027

4
H. Hassan et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
nonane and bicyclo[4.3.1]decane ring systems. In addition to hav-
ing appropriate molecular properties, the fragment set may offer
distinctive opportunities for fragment-based drug discovery. The
fragments were shown to be highly three-dimensional, and thus
are complementary to existing fragment sets. Furthermore, the
strained nature of the bicyclic frameworks will perturb the elec-
tronic properties of the embedded functional groups, presenting
unusual opportunities for interaction with target proteins. The
fragment set will be exploited in screens against a range of protein
targets, and the results of these investigations will be reported in
due course.
Acknowledgments
We acknowledge support from EPSRC (EP/N025652/1) and from
the Innovative Medicine Initiative Joint Undertaking under grant
number 115489, resources of which are composed on financial
contributions from the European Union’s Seventh Framework Pro-
gramme (F97/2008-2013) and EFPIA companies’ in kind contribu-
tion. We thank Dr Shiao Chow for assistance with figure
preparation.
A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.bmc.2018.02.027.
References
1. (a) Murray CW, Rees DC. Nat Chem. 2009;1:187–192;
(b) Zartler ER. ACS Med Chem Lett. 2014;5:952–953;
(c) Erlanson DA, Fesik SW, Hubbard RE, Jahnke W, Jhoti J. Nat Rev Drug Discov.
2016;15:605–619.
Fig. 4. Properties and diversity of the 22 synthesised fragments. Panel A: Molecular
properties. Panel B: Shape diversity. D, disk; R, rod; S, sphere.
2. Jhoti H, Williams G, Rees DC, Murray CW. Nat Rev Drug Discov.
2013;12:644–645.
3. (a) Morley AD, Pugliese A, Birchall K, et al. Drug Discov Today.
2013;18:1221–1227;
(b) Murray CW, Rees DC. Angew Chem Int Ed. 2016;55:488–492;
(c) Over B, Wetzel S, Grütter C, et al. Nat Chem. 2013;5:21–28;
(d) Twigg DG, Kondo N, Mitchell SL, et al. Angew Chem Int Ed.
2016;55:12479–12483.
4. Ruddigkeit L, Deursen R, van R, Blum LC, Reymond JL. J Chem Inf Model.
2012;52:2864–2875.
5. Leach AR, Hann MM. Curr Opin Chem Biol. 2011;15:489–496.
6. Fry DC, Wartchow C, Graves B, et al. ACS Med Chem Lett. 2013;4:660–665.
7. (a) Takayama H, Katakawa K, Kitajima M, Yamaguchi K, Aimi N. Tetrahedron
Lett. 2002;43:8307–8311;
(b) Kump WG, Patel MB, Rowson JM, Schmid H. Helv Chim Acta.
1964;47:1497–1503;
(c) Chen J, Chen J-J, Yao X, Gao K. Org Biomol Chem. 2011;9:5334–5336;
(d) Deguchi J, Shoji T, Nugroho AE, et al. J Nat Prod. 2010;73:1727–1729.
8. Rodrigues T, Reker D, Schneider P, Schneider G. Nat Chem. 2016;8:531–541.
9. Winkler FK, Dunitz JD. J Mol Biol. 1971;59:169–182.
10. Szostak M, Aubé J. Chem Rev. 2013;113:5701–5765.
11. Kirby AJ, Komarov IV, Wothers PD, Feeder N. Angew Chem Int Ed.
1998;37:785–786.
and b-alanine ethyl ester 6f gave the corresponding secondary
amines 21a and 21b in 74% and 77% yield respectively. Unfortu-
nately, after Boc deprotection and ester hydrolysis, Bu2SnO-medi-
ated cyclisation was not successful in either case; perhaps the
presence of a free secondary amine enabled competing polymeri-
sation of the substrate. Therefore, the secondary amines 21a and
21b were decorated by alkylation, acylation or sulfonylation to
yield the intermediates 7e–j; Boc deprotection, ester hydrolysis,
and Bu2SnO-mediated cyclisation15 then gave the corresponding
decorated lactams 8a–f.
2.4. Analysis of shape diversity and molecular properties
The molecular properties and shape diversity of the fragment
set was analysed using LLAMA (Lead-Likeness And Molecular Anal-
ysis; llama.leeds.ac.uk), an open-access computational tool that we
have developed (Figure 4).19 The 22 synthesised fragments had
molecular properties that were appropriate for fragment-based
drug discovery. Furthermore, the fragments were highly distinctive
from existing fragment sets. None of the Murcko frameworks20 of
the fragments were found as a substructure in a random 2% sample
of the ZINC database21 of commercially-available compounds. Fur-
thermore, the bridged scaffolds were shown to result in a fragment
set that was highly three-dimensional.
12. Tani K, Stoltz BM. Nature. 2006;441:731–734.
13. Morgan J, Greenberg AJ. Chem Thermodyn. 2014;73:206–212.
14. A SciFinder structure search of 1-azabicyclo[3.3.1]nonan-2-ones, 1-azabicyclo
[4.3.1]decan-2-ones and all possible heterocyclic variants returned fewer than
20 appearances in medicinally-related patents, and a sole publication with
annotated biological activity: Sturm PA, Cory M, Henry DW, McCall JW, Ziegler
JB. J Med Chem. 1977;20:1327–1333.
15. (a) Steliou K, Poupart MA. J Am Chem Soc. 1983;105:7130–7138;
(b) McCabe PH, Milne NJ, Sim GA. J Chem Soc Perkin Trans. 1989;2:1459–1462.
16. (a) Kirby AJ, Komarov IV, Feeder N. J Chem Soc Perkin Trans. 2001;2:522–529;
(b) Buchanan GL. J Chem Soc Chem Commun. 1981;814–815.
17. Abbiati G, Arcadi A, Bianchi G, Giuseppe SD, Marinelli F, Rossi E. J Org Chem.
2003;68:6959–6966.
3. Conclusion
18. Hassan H, Mohammed S, Robert F, Landais Y. Org Lett. 2015;17:4518–4521.
19. Colomer I, Empson CJ, Craven P, et al. Chem Commun. 2016;52:7209–7212.
20. Bemis GW, Murcko MA. J Med Chem. 1996;39:2887–2893.
21. Irwin JJ, Sterling T, Mysinger MM, Bolstad ES, Coleman RG. J Chem Inf Model.
2012;52:1757–1768.
A set of 22 fragments was prepared that was based on some
bridged scaffolds containing a bridgehead nitrogen atom. The
key step was Bu2SnO-mediated amide formation, which enabled
the synthesis of alternative heteroatom-substituted bicyclo[3.3.1]
Please cite this article in press as: Hassan H., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.02.027