Synthesis and Evaluation of a 2,11-Cembranoid-Inspired Library

Article abstract of DOI:10.1002/chem.201505093

The 2,11-cembranoid family of natural products has been used as inspiration for the synthesis of a structurally simplified, functionally diverse library of octahydroisobenzofuran-based compounds designed to augment a typical medicinal chemistry library screen. Ring-closing metathesis, lactonisation and SmI₂-mediated methods were exemplified and applied to the installation of a third ring to mimic the nine-membered ring of the 2,11-cembranoids. The library was assessed for aqueous solubility and permeability, with a chemical-space analysis performed for comparison to the family of cembranoid natural products and a sample set of a screening library. Preliminary investigations in cancer cells showed that the simpler scaffolds could recapitulate the reported anti-migratory activity of the natural products. Building a bridge: A library of compounds based on the naturally occurring 2,11-cembranoid family has been synthesised by using different synthetic strategies. The library was found to cover the space between the natural products and a screening library sample set (see figure). The new compounds were shown to recapitulate reported activities of the natural products.

Full text of DOI:10.1002/chem.201505093

DOI: 10.1002/chem.201505093
Full Paper
&
Drug Design
Synthesis and Evaluation of a 2,11-Cembranoid-Inspired Library
Amanda J Welford,[a] John J. Caldwell,*[a] Manjuan Liu,[a] Meirion Richards,[a] Nathan Brown,[a]
Cara Lomas,[a] Graham J. Tizzard,[b] Mateusz B. Pitak,[b] Simon J. Coles,[b] Suzanne A. Eccles,[a]
Florence I. Raynaud,[a] and Ian Collins[a]
Abstract: The 2,11-cembranoid family of natural products
has been used as inspiration for the synthesis of a structurally
simplified, functionally diverse library of octahydroisobenzo-
furan-based compounds designed to augment a typical me-
dicinal chemistry library screen. Ring-closing metathesis, lac-
tonisation and SmI2-mediated methods were exemplified
and applied to the installation of a third ring to mimic the
nine-membered ring of the 2,11-cembranoids. The library
was assessed for aqueous solubility and permeability, with
a chemical-space analysis performed for comparison to the
family of cembranoid natural products and a sample set of
a screening library. Preliminary investigations in cancer cells
showed that the simpler scaffolds could recapitulate the re-
ported anti-migratory activity of the natural products.
Introduction
lenge being the introduction of the nine-membered cyclic
ether ring and the installation of the specific complex substitu-
tion patterns associated with each natural product. Methods
used to form such a ring include Nozaki–Hiyama–Kishi cou-
pling,[7,8,11,14] Claisen ring-expansion,[6] ring-closing metathesis
(RCM),[4,9,13,15] diazo ketone cyclisations[10,16] and oxidative ring-
expansion.[17] A number of these total synthesis routes have
been shown to afford access to more than one member of the
natural product family[5,7,8,10,11,13,17] allowing the preparation of
small libraries of the natural products. The broad range of bio-
logical activities of the cembranoid family has been reviewed,[1]
and has been found to encompass anticancer, anti-inflammato-
ry, antiviral and antibacterial properties, amongst others.
The 2,11-cembranoids belong to a structurally complex family
of natural products isolated from the Octocorallia species, con-
sisting of a polyoxygenated 2,11-cyclised diterpenoid core scaf-
fold.[1] The structures are further subcategorised into cladiellins
1, briarellins 2 and asbestinins 3 (Figure 1). Due to their struc-
Natural products have been a great source of inspiration in
drug discovery, being used either in their own right as drugs
or as the starting point for drug discovery projects.[18] Libraries
of natural products, natural-product-inspired or other structur-
ally complex scaffolds have been shown to be a rich source of
hits in phenotypic screens,[19] which probe phenotypes arising
from biological manipulation with small molecules in cells or
small organisms (e.g., zebra fish embryos[20]). As biological
assay screening tools they frequently occupy a complementary
chemical space to commercially available synthetic libraries,
providing novelty from both a structural and intellectual prop-
erty perspective.[21] Those natural products found to display in-
teresting biological activities can, however, be stereochemically
complex molecules, thus presenting significant synthetic chal-
lenges regarding analogue synthesis. Numerous approaches
have been taken to overcome such challenges, in particular di-
versity-oriented synthesis (DOS),[19,22] biology-oriented synthesis
(BIOS)[23] and lead-oriented synthesis.[24] A principal challenge is
to retain structural features of the natural products that differ-
entiate them from chemical space explored by traditional
small molecule screening libraries, and may be associated with
the bioactivities of the natural products, while simplifying the
Figure 1. Cladiellin, briarellin and asbestinin core scaffolds. Carbon atoms
bearing an asterisk are commonly oxygenated or unsaturated within the
class.
tural complexity and broad ranging biological activities, the
members of these structural class have been and continue to
be attractive targets for total synthesis.[2–17] By necessity, such
approaches require multi-step routes, with a particular chal-
[a] Dr. A. J. Welford, Dr. J. J. Caldwell, Dr. M. Liu, M. Richards, Dr. N. Brown,
Dr. C. Lomas, Prof. S. A. Eccles, Dr. F. I. Raynaud, Prof. I. Collins
Division of Cancer Therapeutics
The Institute of Cancer Research, London, SM2 5NG (UK)
E-mail: john.caldwell@icr.ac.uk
[b] Dr. G. J. Tizzard, Dr. M. B. Pitak, Dr. S. J. Coles
UK National Crystallography Service
University of Southampton, Southampton SO17 1BJ (UK)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201505093.
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons At-
tribution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
Chem. Eur. J. 2016, 22, 1 – 9
1
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
structures to give easier synthetic access and control of the
physicochemical properties.
proach is outlined in Scheme 1. From (S)-5-(hydroxymethyl)fur-
an-2(5H)-one (9), a Diels–Alder reaction with a range of dienes
was envisaged to install the [6,5] core of compound 7, which
can be further manipulated to produce bicyclic analogues of
the general structure 5, or further cyclised to tricyclics such as
compound 4. Following the generation of the library, aqueous
solubility, permeability and structural diversity in relation to in-
house screening libraries were assessed. In addition, the ability
to recapitulate some of the biological activities of the more
complex natural products with simpler structures from the li-
brary was demonstrated.
Within the context of drug discovery, numerous metrics
have been suggested as design aids in attempts to predict the
likelihood of drug-likeness,[25] one of the most prevalent being
Lipinski’s rule-of-five.[26] The central tenet of many of these
metrics is to predict the ability of a compound to dissolve in
aqueous media and to pass through a lipid membrane, both
properties that may also be readily measured. To be useful for
cell- or organism-based biology, compounds must be suffi-
ciently cell permeable and water soluble. It is therefore desira-
ble to design new libraries around scaffolds where these prop-
erties are intrinsic.
Here, we present the synthesis of a 2,11-cembranoid-inspired
small library of compounds, designed to retain the biological
activities of the natural product class within structurally simpli-
fied library analogues, while displaying desirable lead-like
properties. The cembranoid class of natural products has dem-
onstrated multiple biological activities in cells and parasites,[1]
making the core scaffold an ideal starting point to make libra-
ries of cell-permeable compounds. It was envisaged that the
octahydroisobenzofuran core, with and without the third fused
medium ring from the 2,11-cembranoid motif, could be used
as a privileged structure.[28] It was anticipated that the library
of compounds produced from such a privileged core would be
able to recapitulate some of the biological activities observed
in the natural product class, while reducing the synthetic com-
plexity. The main synthetic challenge, therefore, was to find
quick and concise routes to introduce substituent variations
around the chosen core and methods to install a third ring
system to mimic the nine-membered ring of the 2,11-cembra-
noids.
Results and Discussion
Protection of the commercially available alcohol 9 with tert-bu-
tyl(chloro)diphenylsilane (TBDPSCl) gave the lactone 8 in 87%
yield; chiral HPLC confirmed the presence of a single enantio-
mer (Scheme 2). Reaction of the silyl ether 8 with butadiene in
From the many successful synthetic strategies exemplified
for the 2,11-cembranoid natural products, we chose to explore
a Diels–Alder reaction to construct the [6,5]-bicyclic core of the
library[6,7] We anticipated a rapid and flexible entry to the octa-
hydroisobenzofuran scaffold through a Diels–Alder cyclisation,
with the potential to vary the substituent patterns through
choice of starting materials or manipulation of intermediates.
Such an approach would provide a platform to investigate
both the functionalization of the bicycle as well as the installa-
tion of a medium or larger ring equivalent to the nine-mem-
bered ring of the natural products. The retrosynthetic ap-
Scheme 2. Reagents and conditions: a) TBDPSCl, imidazole, DMF, 08C–RT,
2.5 h; b) butadiene, TfN(AlMeCl)2, CH2Cl2, 608C; c) DIBAL-H, CH2Cl2, À788C;
d) allylTMS (TMS=trimethylsilyl), BF3·OEt2, À788C; e) tetrabutylammonium
fluoride (TBAF), 08C; f) acid chloride, pyridine or Et3N; g) Grubbs II, 0.0005m
CH2Cl2, RT; h) Pd/C (5 wt%), H2, EtOAc (Ac=acetyl), RT.
the presence of AlCl3 in dichloromethane at 558C in a sealed
tube for one week under literature conditions[29] gave the bicy-
clic compound 10 in 40% yield. Although forming the desired
product, yields were found to be highly variable. Screening of
[30]
a number of Lewis acids found TfN(AlMeCl)2 (Tf=triflate) to
give a consistently high yield of around 70–80% after a two
day reaction. Furthermore, it was found possible to use a solu-
tion of 30% butadiene in toluene as the diene source, avoiding
the inconvenience and hazards associated with condensing
gaseous butadiene. Reduction of the lactone 10 with diisobu-
tylaluminum hydride (DIBAL-H) gave the lactol 11 in high yield,
Scheme 1. Library design and retrosynthetic analysis.
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
2
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
which in turn was reacted with allyltrimethylsilane[31,32] to give
compound 12 as a single enantiomer. The exclusive cis selec-
tivity from reaction of the bicyclic lactol 11 contrasts with liter-
ature examples of monocyclic lactols, where the trans isomer
predominates from an intermolecular allyltrimethylsilane addi-
tion.[30] The observed difference can be rationalised by looking
at a proposed conformation of the oxonium intermediate 13,
where the least hindered approach of the nucleophile leads to
the cis product 12.
Deprotection of silyl ether 12 using TBAF gave the alcohol
14 in 82% yield, which was esterified to form esters of various
chain lengths, that is, compounds 15, 16, 17 and 18. Following
considerable optimisation, ring-closing metathesis (RCM) cycli-
sation using the Grubbs II catalyst at high dilution in dichloro-
methane gave the tricyclic compounds 20, 21 and 22. The
nine-membered lactone 19 could not be formed, whereas the
ten-membered lactone 20 was produced in low yield. The
alkene conformation of compound 20 was presumed to be cis
due to the strain of a trans double bond in such a ten-mem-
bered lactone. The larger, less strained eleven- and twelve-
membered lactones were more readily constructed. The alkene
conformation of compound 21 was undetermined due to over-
lapping NMR signals, whereas compound 22 was assigned to
be trans by analogy to the crystal-structure conformations of
compounds 40 and 42 (see below). By using a similar reaction
sequence, but reducing the alkene moiety of compound 10
earlier in the route, the cyclohexane analogues 26, 27 and 28
were also formed. The structure of compound 26 was con-
firmed by X-ray crystallography (see the Supporting Informa-
tion).
Scheme 4. Reagents and conditions: a) OsEnCat (0.5 mol%), N-methylmor-
pholine N-oxide (NMO); b) 2,2-dimethoxypropane, pyridinium tosylate
(PPTS); c) DIBAL-H, À788C; d) allylTMS, BF3·OEt2, À788C; e) TBAF; f) hex-5-
enoyl chloride, pyridine (py), CH2Cl2; g) Grubbs II, 0.0005m CH2Cl2; h) HCl,
THF.
31 was reduced to lactol 32, reacted with allyltrimethylsilane
in the presence of BF3·OEt2 and deprotected to give alcohol
33. Esterification, ring-closing metathesis and acetal deprotec-
tion with acid gave the diol 36.
Analogue synthesis by changing the diene component was
also demonstrated (Scheme 5). The use of 2-methylbutadiene
As lactone 19 could not be formed, speculated to be due to
the transoid nature of the ester moiety in compound 15
adding to the strain in forming a nine-membered ring, synthe-
sis of the ether analogue 30 was attempted (Scheme 3). Alkyla-
Scheme 5. Reagents and conditions: a) TfN(AlMeCl)2, CH2Cl2, 608C, 100%;
b) BH3·THF, then H2O2, H2O 55%.
Scheme 3. Reagents and conditions: a) NaH, allylbromide; b) Grubbs II,
0.0005m, CH2Cl2, RT. 0%.
gave an 85:15 inseparable mix of methyl regioisomers (major
isomer 37 shown). The major 5-methyl component was isolat-
ed following oxidative hydroborylation to give the alcohol 38.
As above, the alcohol 38 was in turn converted to the lactone
39. As further exemplification of diene variations in the Diels–
Alder reaction, cyclohexa-1,3-diene and cyclopentadiene were
used to make the tetracyclic compounds 40, 41 and 42. Confir-
mation of the structure of compounds 40 and 42 was achieved
by X-ray crystallography (see the Supporting Information).
During the synthesis of compound 42, it was found necessary
to reduce the alkene derived from the diene due to undesired
ring-opening occurring in the subsequent RCM reaction, pre-
sumably as a consequence of release of inherent ring strain in
the embedded cyclopentene ring that was not present in the
cyclohexene analogue.
tion of alcohol 14 with allyl bromide gave the cyclisation pre-
cursor 29. However, no RCM product could be observed under
the above-described optimised conditions, highlighting the dif-
ficulty in forming the strained nine-membered ring. Previous
RCM approaches to nine-membered ring formation in the con-
text of 2,11-cembranoid synthesis have met with mixed suc-
cess,[3,4,15,33] and appear highly substrate dependent.
By using the alkene functionality of lactone 10 as a point of
derivatisation, dihydroxylation of the alkene from the least
sterically crowded face using catalytic osmium tetroxide (OsEn-
Cat) followed by dimethylacetal protection gave the tricyclic
compound 31 in 52% yield over two steps (Scheme 4). By
using a similar sequence to that shown in Scheme 2, lactone
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
3
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
The Mukaiyama aldol reaction on lactol 11 provided
a means of introducing further substitution and a handle for
elaboration. To this end, adducts of the general structure 43
were formed as single enantiomers by reacting the hemiacetal
11 with a range of silyl enol ethers. As in the example of allyl-
trimethylsilane addition described above (Scheme 2), the cis
products were observed exclusively or as the major compo-
nent (85:15 in the formation of compound 48). The silyl ethers
44 and 45 were obtained, as well as alcohols 46, 47 and 48
following silyl deprotection (Scheme 6).
An alternative lactonisation approach to forming an oxygen-
containing nine-membered ring system and thus providing li-
brary members with closer correspondence to the [6,5,9]-tricy-
clic core of the natural product family was attempted from al-
dehyde 44. A Horner–Wadsworth–Emmons reaction gave the
a,b-unsaturated ester 61 (Scheme 8). Reduction of both al-
Scheme 8. Reagents and Conditions: a) triethylphosphonoacetate, NaH, THF;
b) Pd/C, EtOAc, H2; c) TBAF, THF; d) LiCl, H2O, THF; e) 2-methyl-6-nitrobenzo-
ic anhydride, DMAP, CH2Cl2.
Scheme 6. Reagents and conditions: a) silyl enol ether, BF3·OEt2 (TMSOTf for
compound 46), CH2Cl2, À788C 36–97%.
kenes by hydrogenation over palladium gave the ester 62. De-
protection of the silyl group with TBAF followed by ester hy-
drolysis gave the acid 64 in 71% over two steps. Lactonisation
was initially attempted by using the protocol of Yamaguchi
et al.,[35] through the intermediate 2,4,6-trichlorobenzoyl anhy-
dride. None of the desired lactone was observed in a number
of attempts, with a dimeric species being the only isolated
product seen. A more successful approach used conditions de-
veloped by Shiina et al.,[36] whereby a solution of the acid 64
was added slowly by using a syringe pump over a time period
of 15 h to a solution of 2-methyl-6-nitrobenzoic anhydride and
4-dimethylaminopyridine (DMAP) in dichloromethane. By using
these conditions, lactone 65 was isolated in 30% yield. Inter-
estingly, an IR absorption for the carbonyl group was observed
at n˜ =1730 cmÀ1, suggesting the carbonyl group to be signifi-
cantly twisted out of plane from the lowest energy lactone
conformation (compare with the crystal structure of compound
69, Figure 2).
With the aldehyde 44 in hand, derivatisation was exempli-
fied either by reduction to the diol 49, or through reductive
amination with the amines listed in Scheme 7. Deprotection of
amines 50–54 with TBAF provided alcohols 55–59, respective-
ly, in good yield.
Modification of the terminal olefin in compound 24 was also
exemplified. Oxyarylation[34] by using a gold(I) catalyst in the
presence of Selectfluor with phenyltrimethylsilane gave the
ether 60 as an inseparable 4:1 mix of diastereomers
(Scheme 7).
A SmI2-mediated Reformatsky reaction approach was also
found successful in the synthesis of a nine-membered lactone.
The Reformatsky reaction has been shown previously to give
high yields and a controlled stereochemistry in the synthesis of
a variety of natural product ring systems,[37] and more perti-
nently in medium ring synthesis.[38] Here, the alcohol 23 was
subjected to the Mukaiyama aldol/TBAF deprotection se-
quence described above to give the unstable alcohol 67
(Scheme 9). Bromoacetyl bromide acylation followed by SmI2-
induced cyclisation gave the alcohol 69 in 45% yield. In addi-
tion, the alcohol 69 could be further derivatised if required,
such as to the acetate 70. The structure and absolute stereo-
chemistry of compound 69 was confirmed by X-ray crystallog-
raphy analysis (Figure 2). As with lactone 65, an unusual IR car-
bonyl stretch at n˜ =1723 cmÀ1 was observed for compound 69
Scheme 7. Reagents and Conditions: a) NaBH4, THF, MeOH 82%; b) TBAF,
THF, 41–93%; c) amine, NaBH(OAc)3, 1,2-dichloroethane (DCE), 52–64%;
d) PhSiMe3, [AuCl(PPh3)], selectfluor, MeOH, MeCN, then TBAF 29%.
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
4
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
in 97% yield. A NOESY correlation
between both the [6,5] bridgehead
hydrogen atoms and the siloxy-
substituted CÀH hydrogen atom in
compound 75 as highlighted in
Figure 3 was used to confirm the
sterochemistry of the Grignard ad-
dition. Deprotection with TBAF
gave the alcohol 76 in 55% yield.
The 2,11-cembranoid based li-
Figure 3. 1H–1H NOESY corre-
lations in the tricyclic com-
pound 75.
brary of 44 compounds synthesised by using the routes and
methods outlined above is shown in full in Figure 4. As the
compounds were intended for the use in drug discovery
screening programmes, an assessment of the library in terms
of structural diversity, physicochemical properties and potential
biological activity was undertaken. An analysis of the molecular
properties of the compounds given in Figure 4 is compiled in
Table 1. For comparison, the properties of 183 cembranoid nat-
Figure 2. X-ray crystal structure of compounds a) 42, b) 40, and c) 69. d) Par-
tial view of compound 69 along the OÀC lactone single bond.
Table 1. Arithmetic mean and standard deviation of the molecular prop-
erties of the cembranoid-inspired library, the 2,11-cyclised cembranoid
natural products and a random sampling set from the Institute of Cancer
Research screening collection.
Cembranoid- 2,11-Cyclised ICR screening
inspired
library
cembranoids collection
sample
no. of compound
44
183
358
MW
AlogP
275 (28)
2.6 (0.8)
441 (77)
3.1 (1.1)
6.4 (2.0)
1.0 (0.9)
3.4 (0.6)
0 (0)
6.4 (2.0)
39 (5.9)
8.9 (1.5)
34 (5.4)
0 (0)
332 (59)
2.6 (1.1)
4.1 (1.4)
1.1 (0.8)
2.9 (0.9)
2.0 (1.0)
2.4 (1.4)
18 (5.3)
0.04 (0.33)
25 (4.7)
11 (5.2)
4.5 (2.0)
Scheme 9. Reagents and conditions: a) BF3·OEt2, trimethyl((2-methylprop-1-
en-1-yl)oxy)silane, CH2Cl2, À788C–RT; b) TBAF, THF; c) bromoacetyl bromide,
Et3N, CH2Cl2, 34% over three steps; d) SmI2, THF; e) Ac2O, DMAP, RT.
no. of hydrogen acceptors 3.2 (0.7)
no. of hydrogen donors
no. of rings
no. of aromatic rings
no. of oxygen atoms
no. of hydrogen atoms
no. of stereoatoms
no. of bonds
0.7 (0.8)
2.6 (0.6)
0.07 (0.26)
3.1 (0.7)
25 (2.9)
4.7 (0.9)
21 (2.2)
0.43 (1.6)
4.8 (3.2)
as a consequence of a 338 twist out of conjugation with the
ester oxygen atom.
With different synthetic routes to nine-membered lactones
in hand, our focus returned to accessing nine-membered
ethers. Starting from alcohol 25, synthesised as shown in
Scheme 2, oxidation to the aldehyde 71 and subsequent reac-
tion with but-3-en-1-ylmagnesium bromide gave alcohols 72
and 73 as a 5:3 separable mixture of isomers (Scheme 10). Fol-
lowing TBS protection, only the major isomer 72 was found to
cyclise in the RCM reaction, giving the tricyclic compound 75
no. of aromatic bonds
no. of rotatable bonds
5.9 (3.1)
ural products taken from reviews[1,39] and a random sampling
set of 358 compounds taken from a substantially larger 75000
screening library at the Institute of Cancer Research (ICR) are
also listed. The ability of a random sampling set of a large
screening library to represent the entirety has been demon-
strated by Feher and Schmidt,[21] and has been applied here.
The ICR screening library itself was chosen by typical criteria
for lead-like/drug-like libraries, including limits on predicted/
calculated physicochemical parameters and choice of scaffolds,
with a high frequency of heterocyclic aromatic and aliphatic
scaffolds.[26,27] As a means of visually assessing the diversity be-
tween the groups, a principal component analysis (PCA)[41] was
performed. PCA has been used widely in chemoinformatics[42]
and particularly in the description of chemical space.[22,43] The
analysis, carried out by using SIMCA-P+[44] and visualised by
using Spotfire,[45] showed (Figure 5) that the ICR screening
compounds (red in Figure 5) cover a wide range of chemical
Scheme 10. Reagents and conditions: a) Dess–Martin reagent, CH2Cl2;
b) but-3-en-1-ylmagnesium bromide, THF; c) trimethylsilyl trifluoromethane-
sulfonate (TBSOTf), diisopropylethylamine (DIPEA), CH2Cl2; d) Grubbs II
(10 mol%), CH2Cl2, heating to reflux; e) HCl, EtOH, Et2O.
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
5
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Figure 4. Cembranoid-like library.
branoid-inspired library as synthesised here (green in Figure 5)
was found to occupy an area of chemical space between the
natural products and the ICR screening collection, showing
that the cembranoid-inspired library begins to effectively
bridge the chemical space between classical drug discovery
screening compounds and the natural products.
The library prepared was assessed for both solubility and
permeability, with representative results shown in Table 2. Due
Table 2. Solubility and permeability of selected representative com-
pounds.
Compound
28
15
65
76
solubility [mm]
permeability (Pe) [ꢁ10À6 cmÀ1
550
52
500
25
550
>150
500
]
not detected[a]
Solubility measured by nephelometry. [a] Undetectable due to low ionisa-
tion.
Figure 5. Principal component analysis (PCA) of the cembranoid-inspired li-
brary (green, 44 structures), the 2,11-cyclised cembranoid natural products
(blue, 183 structures) and the ICR screening sample (red, 358 structures).
Performed by using SIMCA-P+[44] and displayed by using Spotfire.[45]
to a lack of UV chromophore and/or poor mass spectrometry
ionisation, nephelometry[46] was used to measure the solubility.
PAMPA (parallel artificial membrane permeability assay)[47] was
used to assess the permeability. Generally, the compounds
were found to have high aqueous solubility and permeability.
Of the 44 compounds given in Figure 4, all were measured to
have greater than 100 mm aqueous solubility, with 34 greater
than 500 mm and 21 greater than 1000 mm. In regard to the 28
compounds that ionised sufficiently well to be detected by the
space as would be expected from a well-designed screening
deck;[40] however, little overlap exists with the cembranoid nat-
ural products (blue in Figure 5), which occupied a discrete and
well-defined region defined by the PCA. By contrast, the cem-
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
6
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
PAMPA, 25 were found to have a high permeability, greater
than 25ꢁ10À6 cmsÀ1. For the specific example of compound
28 in addition to a good aqueous solubility and high PAMPA,
the compound showed moderate permeability across a mono-
layer of Caco-2 human intestinal cancer cells,[48] (16ꢁ
10À6 cmsÀ1) with no evidence of transporter efflux.
liferation in PC3 cells at 50 mm. Therefore, within this context,
compound 28 displays typical cytotoxicity for the natural prod-
uct family,[39] with reduced structural and synthetic complexity.
Whereas cembranoid natural products may not be inherent-
ly cytotoxic, an appreciable number do display an ability to in-
hibit the migratory and invasive properties of cancer cells.[50]
The effect of compound 28 on a PC3-based migratory assay
was therefore determined. The tricyclic compound 28 at 3-
and 5-fold GI50 concentrations demonstrated a reduction to
near basal levels of migration (Figure 6). Although less potent
than a number of the 2,11-cembranoid natural products (most
active being sclerophytin A, with 85% inhibition of migration
at 50 mm),[50] it does demonstrate that the anti-migratory activi-
ty can be retained in a greatly simplified structural motif.
A selection of fifteen of the cembranoid-inspired com-
pounds was screened against a panel of 60 cancer cell lines at
the National Cancer Institute screening panel (see the Support-
ing Information).[49] This cell-based assay aims to discover com-
pounds with a strong cytostatic or cytotoxic growth inhibition
over 48 h. Unfortunately, only low levels (up to 20%) of cell
growth inhibition were seen when screened at a single con-
centration of 10 mm. Of these, compound 28 gave 17% inhibi-
tion of proliferation in the PC3 human prostate cancer cell line.
A full dose response curve obtained for compound 28
(Figure 6) found a GI50 value of 63 mm. It is important to note
Conclusion
The 2,11-cembranoid class of natural products has served as
a rich source of inspiration from both a synthetic and biologi-
cal perspective. Here, we have developed a number of synthet-
ic approaches to produce a library of structurally simplified
cembranoid-inspired derivatives. An assessment of the molecu-
lar properties of the octahydroisobenzofuran-based library
found them to be complementary to both the cembranoid
natural products as well as a sample set of an in-house medici-
nal chemistry screening collection, providing potential access
to novel chemical space. A screen of the physicochemical
properties found the library to have desirable properties,
namely aqueous solubility and cell membrane permeability, es-
sential to be useful in both biochemical and cell-based assays.
The illustrative synthetic routes developed could readily be ex-
panded to more densely populate the chemical space occu-
pied by the small library and the natural products, or indeed
to follow up any hits from future screening. A preliminary
screen of a sample set of the library found compound 28 to
weakly inhibit the growth of PC3 cells, as well as having an in-
hibitory effect when screened in a cell migration assay. These
data suggest that some of the bioactivities observed in cancer
cells for the structurally complex 2,11-cembranoid natural
products can be recapitulated with much simpler scaffolds de-
rived from the embedded octahydroisobenzofuran core.
Experimental Section
All synthetic methods are included in the Supporting Information.
CCDC 1010197 (26), 1011320 (40), 1435146 (42) and 1435148 (69)
contain the supplementary crystallographic data. These data can
be obtained free of charge from The Cambridge Crystallographic
Data Centre.
Figure 6. a) Concentration–response curve for the antiproliferative effect of
compound 28 on PC3 cells in vitro 96 h assay. b) Compound 28 inhibiting
the migration properties of PC3 TEM cancer cells. FCS=fetal calf serum. The
y axis shows the fold-changes in the cell migration normalised to the ab-
sence of FCS.
that the cembranoid family itself is not overtly cytotoxic, with Acknowledgements
most examples screened to date ranging from single figure mi-
cromolar IC50 values to inactive in the cytotoxic assays used.[1]
One obvious exception to this is sclerophytin A, with an in
vitro IC50 value of 3 nm in L1210 mouse lymphocytic leukemia
cells,[51] although the natural product showed no effect on pro-
This work was supported by CRUK grant C19524A8027 (PhD
studentship to A.J.W.), the CRUK grant C309A11566, and The
Institute of Cancer Research, London. We also acknowledge
NHS funding to the NIHR Biomedical Research Centre. We
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
7
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[22] a) G. L. Thomas, E. E. Wyatt, D. R. Spring, Curr. Opin. Drug Discovery
Devel. 2006, 9, 700–712; b) W. R. J. D. Galloway, A. Bender, M. Welch,
D. R. Spring, Chem. Commun. 2009, 2446–2462.
thank Amin Mirza and Sarah Langdon for helpful discussions.
G.J.T., M.B.P. and S.J.C. thank the EPSRC for funding the UK Na-
tional Crystallography Service. The authors declare no compet-
ing financial interest.
[23] a) R. Breinbauer, I. R. Vetter, H. Waldmann, Angew. Chem. Int. Ed. 2002,
41, 2878–2890; Angew. Chem. 2002, 114, 3002–3015; b) D. Brohm, S.
Metzger, A. Bhargava, O. Mꢃller, F. Lieb, H. Waldmann, Angew. Chem. Int.
Ed. 2002, 41, 307–311; Angew. Chem. 2002, 114, 319–323.
[24] a) R. G. Doveston, P. Tosatti, M. Dow, D. J. Foley, H. Y. Li, A. J. Campbell,
D. House, I. Churcher, S. P. Marsden, A. Nelson, Org. Biomol. Chem. 2015,
13, 859–865; b) R. Doveston, S. Marsden, A. Nelson, Drug Discovery
Today 2014, 19, 813–819.
[25] E. H. Kerns, L. Di in Drug-Like Properties: Concepts, Structure Design and
Methods, Vol. 1, Elsevier, Amsterdam, 2008.
[26] C. A. Lipinski, F. Lombardo, B. W. Dominy, P. J. Feeney, Adv. Drug Delivery
Rev. 1997, 23, 3–25.
Keywords: medicinal chemistry · medium-ring compounds ·
molecular diversity · oxygen heterocycles · synthesis design
[1] A. J. Welford, I. Collins, J. Nat. Prod. 2011, 74, 2318–2328.
[2] a) J. M. Ellis, M. T. Crimmins, Chem. Rev. 2008, 108, 5278–5298; b) M. E.
Jung, J. Pontillo, J. Org. Chem. 2002, 67, 6848–6851.
[3] M. E. Jung, J. Pontillo, Tetrahedron 2003, 59, 2729–2736.
[4] B. Wang, A. P. Ramirez, J. J. Slade, J. P. Morken, J. Am. Chem. Soc. 2010,
132, 16380–16382.
[5] H. Kim, H. Lee, J. Kim, S. Kim, D. Kim, J. Am. Chem. Soc. 2006, 128,
15851–15855.
[6] a) L. A. Paquette, O. M. Moradei, P. Bernardelli, T. Lange, Org. Lett. 2000,
2, 1875–1878; b) P. Bernardelli, O. M. Moradei, D. Friedrich, J. Yang, F.
Gallou, B. P. Dyck, R. W. Doskotch, T. Lange, L. A. Paquette, J. Am. Chem.
Soc. 2001, 123, 9021–9032.
[7] F. Gallou, D. W. C. MacMillan, L. E. Overman, L. A. Paquette, L. D. Pen-
nington, J. Yang, Org. Lett. 2001, 3, 135–137.
[8] a) D. W. C. MacMillan, L. E. Overman, L. D. Pennington, J. Am. Chem. Soc.
2001, 123, 9033–9044.
[9] a) M. T. Crimmins, B. H. Brown, J. Am. Chem. Soc. 2004, 126, 10264–
10266; b) M. T. Crimmins, J. M. Ellis, J. Am. Chem. Soc. 2005, 127, 17200–
17201; c) M. T. Crimmins, M. C. Mans, A. D. Rodrꢂguez, Org. Lett. 2010,
12, 5028–5031; d) M. J. Campbell, J. S. Johnson, J. Am. Chem. Soc. 2009,
131, 10370–10371.
[10] J. S. Clark, R. Berger, S. T. Hayes, L. H. Thomas, A. J. Morrison, L. Gobbi,
Angew. Chem. Int. Ed. 2010, 49, 9867–9870; Angew. Chem. 2010, 122,
10063–10066.
[11] O. Corminboeuf, L. E. Overman, L. D. Pennington, J. Org. Chem. 2009,
74, 5458–5470.
[12] D. W. C. MacMillan, L. E. Overman, J. Am. Chem. Soc. 1995, 117, 10391–
10392.
[13] a) M. T. Crimmins, B. H. Brown, H. R. Plake, J. Am. Chem. Soc. 2006, 128,
1371–1378; b) M. T. Crimmins, J. M. Ellis, J. Org. Chem. 2008, 73, 1649–
1660.
[14] L. E. Overman, L. D. Pennington, Org. Lett. 2000, 2, 2683–2686.
[15] a) J. Becker, K. Bergander, R. Frçhlich, D. Hoppe, Angew. Chem. Int. Ed.
2008, 47, 1654–1657; Angew. Chem. 2008, 120, 1678–1681; b) G. Yue,
Y. Zhang, L. Fang, C.-C. Li, T. Luo, Z. Yang, Angew. Chem. Int. Ed. 2015,
54, 6531–6535; Angew. Chem. 2015, 127, 6631–6635.
[16] a) J. S. Clark, S. T. Hayes, C. Wilson, L. Gobbi, Angew. Chem. Int. Ed. 2007,
46, 437–440; Angew. Chem. 2007, 119, 441–444; b) J. S. Clark, L. Delion,
L. J. Farrugia, Org Lett 2014, 16, 4300–4303; c) J. S. Clark, L. Delion, L. J.
Farrugia, Chem. Eur. J. 2015, 21, 4772–4780.
[17] a) G. A. Molander, D. J. St. Jean, J. Haas, J. Am. Chem. Soc. 2004, 126,
1642–1643; b) J. S. Clark, R. Berger, S. T. Hayes, H. Martin Senn, L. J. Far-
rugia, L. H. Thomas, A. J. Morrison, L. Gobbi, J Org Chem 2013, 78, 673–
696.
[18] a) D. J. Newman, G. M. Cragg, J. Nat. Prod. 2007, 70, 461–477; b) A. L.
Harvey, Drug Discovery Today 2008, 13, 894–901; c) A. L. Harvey, R.
Edrada-Ebel, R. J. Quinn, Nat. Rev. Drug Discovery 2015, 14, 111–129;
d) D. J. Newman, G. M. Cragg, J. Nat. Prod. 2012, 75, 311–335.
[19] I. Collins, A. M. Jones, Molecules 2014, 19, 17221–17255.
[20] a) B. Dasari, S. Jogula, R. Borhade, S. Balasubramanian, G. Chandrase-
khar, S. S. Kitambi, P. Arya, Org. Lett. 2013, 15, 432–435; b) M. Aeluri, C.
Pramanik, L. Chetia, N. V. Mallurwar, S. Balasubramanian, G. Chandrase-
khar, S. S. Kitambi, P. Arya, Org. Lett. 2013, 15, 436–439; c) M. Aeluri, J.
Gaddam, D. V. K. S. Tinath, G. Chandrasekhar, S. S. Kitambi, P. Arya, Eur. J.
Org. Chem. 2013, 3955–3958; d) S. Chamakuri, S. K. R. Guduru, S. Pamu,
G. Chandrasekhar, S. S. Kitambi, P. Arya, Eur. J. Org. Chem. 2013, 3959–
3964.
[27] D. J. Huggins, A. R. Venkitaraman, D. R. Spring, ACS Chem. Biol. 2011, 6,
208–217.
[28] K. C. Nicolaou, J. A. Pfefferkorn, A. J. Roecker, G.-Q. Cao, S. Barluenga,
H. J. Mitchell, J Am Chem. Soc. 2000, 122, 9939–9953.
[29] a) J. Mann, A. Thomas, Chem. Commun. 1985, 737–738; b) M. G. B.
Drew, J. Mann, A. Thomas, J. Chem. Soc. Perkin Trans. 1 1986, 2279–
2285.
[30] A. Saito, H. Yanai, T. Taguchi, Tetrahedron Lett. 2004, 45, 9439–9442.
[31] R. A. Pilli, V. B. Riatto, Tetrahedron: Asymmetry 2000, 11, 3675–3686.
[32] J. P. Wolfe, M. B. Hay, Tetrahedron 2007, 63, 261–290.
[33] D. Joe, L. E. Overman, Tetrahedron Lett. 1997, 38, 8635–8638.
[34] L. T. Ball, M. Green, G. C. Lloyd-Jones, C. A. Russell, Org. Lett. 2010, 12,
4724–4727.
[35] a) J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, M. Yamaguchi, Bull. Chem.
Soc. Jpn. 1979, 52, 1989–1993; b) P. Basabe, O. Bodero, I. S. Marcos, D.
Diez, A. Blanco, M. de Roman, J. G. Urones, J. Org. Chem. 2009, 74,
7750–7754; c) M. Hikota, H. Tone, K. Horita, O. Yonemitsu, J. Org. Chem.
1990, 55, 7–9.
[36] a) I. Shiina, M. Kubota, H. Oshiumi, M. Hashizume, J. Org. Chem. 2004,
69, 1822–1830; b) I. Shiina, M. Kubota, R. Ibuka, Tetrahedron Lett. 2002,
43, 7535–7539; c) I. Shiina, H. Oshiumi, M. Hashizume, Y. Yamai, R.
Ibuka, Tetrahedron Lett. 2004, 45, 543–547.
[37] D. J. Edmonds, D. Johnston, D. J. Procter, Chem. Rev. 2004, 104, 3371–
3404.
[38] G. A. Molander, J. B. Etter, L. S. Harring, P. J. Thorel, J. Am. Chem. Soc.
1991, 113, 8036–8045.
[39] P. Bernardelli, L. A. Paquette, Heterocycles 1998, 49, 531–556.
[40] A. A. Shelat, R. K. Guy, Nat. Chem. Biol. 2007, 3, 442–446.
[41] a) S. Wold, K. Esbensen, P. Geladi, Chemom. Intell. Lab. Syst. 1987, 2, 37–
52; b) J. E. Jackson in A User’s Guide to Principal Components, Wiley,
New York, 1991.
[42] K. Varmuza, P. Filzmoser in Multivariate Statistical Analysis in Chemomet-
rics, CRC Press, London, 2009.
[43] a) C. Lipinski, A. Hopkins, Nature 2004, 432, 855–861; b) C. M. Dobson,
Nature 2004, 432, 824–828; c) N. J. Rosꢄn, J. Gottfries, S. Muresan, A.
Backlund, T. I. Oprea, J. Med. Chem. 2009, 52, 1953–1962.
[44] Simca-P+ 12.0, Umertrics, Umeꢅ, Sweden, http://www.umetrics.com/.
[45] Spotfile Decision Site, Tibco, Palo Alto, CA, USA, http://spotfire.tibco.
com/.
[46] C. D. Bevan, R. S. Lloyd, Anal. Chem. 2000, 72, 1781–1787.
[47] M. Kansy, F. Senner, K. Gubernator, J. Med. Chem. 1998, 41, 1007–1010.
[48] a) P. Artursson, K. Palm, K. Luthman, Adv. Drug Delivery Rev. 2001, 46,
27–43; b) E. H. Kerns, L. Di, S. Petusky, M. Farris, R. Ley, P. Jupp, J.
Pharm. Sci. 2004, 93, 1440–1453.
[49] a) R. H. Shoemaker, Nat. Rev. Cancer 2006, 6, 813–823; b) http://dtp.nci.
nih.gov.
[50] H. M. Hassan, M. A. Khanfar, A. Y. Elnagar, R. Mohammed, L. A. Shaala,
D. T. A. Youssef, M. S. Hifnawy, K. A. El Sayed, J. Nat. Prod. 2010, 73, 848–
853.
[51] P. Sharma, M. Alam, J. Chem. Soc. Perkin Trans. 1 1988, 2537–2540.
Received: December 18, 2015
[21] M. Feher, J. M. Schmidt, J. Chem. Inf. Comput. Sci. 2002, 42, 218–227.
Published online on && &&, 0000
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
8
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
Building a bridge: A library of com-
pounds based on the naturally occur-
ring 2,11-cembranoid family has been
synthesised by using different synthetic
strategies. The library was found to
cover the space between the natural
products and a screening library sample
set (see figure). The new compounds
were shown to recapitulate reported ac-
tivities of the natural products.
&
Drug Design
A. J. Welford, J. J. Caldwell,* M. Liu,
M. Richards, N. Brown, C. Lomas,
G. J. Tizzard, M. B. Pitak, S. J. Coles,
S. A. Eccles, F. I. Raynaud, I. Collins
&& – &&
Synthesis and Evaluation of a 2,11-
Cembranoid-Inspired Library
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
9
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ