Diversification of ortho-Fused Cycloocta-2,5-dien-1-one Cores and Eight- to Six-Ring Conversion by σ Bond C?C Cleavage

Article abstract of DOI:10.1002/chem.201601970

Sequential treatment of 2-C₆H₄Br(CHO) with LiC≡CR¹(R¹=SiMe₃, tBu), nBuLi, CuBr?SMe₂and HC≡CCHClR²[R²=Ph, 4-CF₃Ph, 3-CNPh, 4-(MeO₂C)Ph] at ?50 °C leads to formation of an intermediate carbanion (Z)-1,2-C₆H₄{C_A(=O)C≡C_BR¹}{CH=CH(CH^?)R²} (4). Low temperatures (?50 °C) favour attack at C_Bleading to kinetic formation of 6,8-bicycles containing non-classical C-carbanion enolates (5). Higher temperatures (?10 °C to ambient) and electron-deficient R²favour retro σ-bond C?C cleavage regenerating 4, which subsequently closes on C_Aproviding 6,6-bicyclic alkoxides (6). Computational modelling (CBS-QB3) indicated that both pathways are viable and of similar energies. Reaction of 6 with H⁺gave 1,2-dihydronaphthalen-1-ols, or under dehydrating conditions, 2-aryl-1-alkynylnaphthlenes. Enolates 5 react in situ with: H₂O, D₂O, I₂, allylbromide, S₂Me₂, CO₂and lead to the expected C-E derivatives (E=H, D, I, allyl, SMe, CO₂H) in 49–64 % yield directly from intermediate 5. The parents (E=H; R¹=SiMe₃, tBu; R²=Ph) are versatile starting materials for NaBH₄and Grignard C=O additions, desilylation (when R¹=SiMe) and oxime formation. The latter allows formation of 6,9-bicyclics via Beckmann rearrangement. The 6,8-ring iodides are suitable Suzuki precursors for Pd-catalysed C?C coupling (81–87 %), whereas the carboxylic acids readily form amides under T3P conditions (71–95 %).

Full text of DOI:10.1002/chem.201601970

DOI: 10.1002/chem.201601970
Full Paper
&
Synthetic Methods
Diversification of ortho-Fused Cycloocta-2,5-dien-1-one Cores and
Eight to Six-Ring Conversion by s Bond CÀC Cleavage
Lee Eccleshare, Leticia Lozada-Rodrꢀguez, Phillippa Cooper, Laurence Burroughs,
John Ritchie, William Lewis, and Simon Woodward*[a]
Abstract: Sequential treatment of 2-C6H4Br(CHO) with LiCꢀ with H+ gave 1,2-dihydronaphthalen-1-ols, or under dehy-
CR1 (R1 =SiMe3, tBu), nBuLi, CuBr·SMe2 and HCꢀCCHClR2
[R2 =Ph, 4-CF3Ph, 3-CNPh, 4-(MeO2C)Ph] at À508C leads to
drating conditions, 2-aryl-1-alkynylnaphthlenes. Enolates 5
react in situ with: H2O, D2O, I2, allylbromide, S2Me2, CO2 and
formation of an intermediate carbanion (Z)-1,2-C6H4{CA(= lead to the expected C-E derivatives (E=H, D, I, allyl, SMe,
O)CꢀCBR1}{CH=CH(CHÀ)R2} (4). Low temperatures (À508C)
favour attack at CB leading to kinetic formation of 6,8-bicy-
cles containing non-classical C-carbanion enolates (5). Higher
temperatures (À108C to ambient) and electron-deficient R2
favour retro s-bond CÀC cleavage regenerating 4, which
subsequently closes on CA providing 6,6-bicyclic alkoxides
(6). Computational modelling (CBS-QB3) indicated that both
pathways are viable and of similar energies. Reaction of 6
CO2H) in 49–64% yield directly from intermediate 5. The pa-
rents (E=H; R1 =SiMe3, tBu; R2 =Ph) are versatile starting
materials for NaBH4 and Grignard C=O additions, desilylation
(when R1 =SiMe) and oxime formation. The latter allows for-
mation of 6,9-bicyclics via Beckmann rearrangement. The
6,8-ring iodides are suitable Suzuki precursors for Pd-cata-
lysed CÀC coupling (81–87%), whereas the carboxylic acids
readily form amides under T3Pꢁ conditions (71–95%).
electrophile (typically, simple H+) the phenylene derivatives
1 were attained in moderate yields (25 examples at 26–72%,
14 above 50% yield). This ready access prompted us to consid-
er the applicability of the chemistry to provide libraries of di-
versified cores of potential future use in medicinal chemistry.[2]
Specifically, we raised the questions: 1) What further diversity
can be introduced into 1 via simple derivatisation; and 2) How
far can the diversity of units R1, R2 and E+ units be extended.
In fact, the group R2 became of immediate initial importance,
because one of our preliminary studies showed that this group
could profoundly affect the reaction’s chemoselectivity
(Scheme 1, 1a vs. formation of 2b instead).
Introduction
We recently described a new method for the synthesis of
eight-membered rings[1] that allows the preparation of gram
quantities of cycloocta-2,5-dien-1-one cores (Scheme 1). When
2-bromobenzaldehydes were treated sequentially with acety-
lides, nBuLi, CuBr·SMe2, propargylic chlorides and finally an
Results and Discussion
Factors in eight- versus six-ring formation
Model reactions by using 2-bromobenzaldehyde, TMS-acety-
lene and various propargylic electrophiles were used to gain
understanding of the effect of group R2 on the reaction
(Table 1). Under our standard conditions,[1] highly chemoselec-
tive eight-ring formation was observed for 1a, and this was
isolated in good yield (70%). This chemoselectivity was not af-
fected by the leaving group (Cl, O2CCF3, OAc), although the
last of these gave lower yields of 1a due to incomplete con-
version. Decreasing the electron density within the aryl group
was confirmed to promote specific formation of six-ring 2b. X-
ray studies (see the Supporting Information) confirmed a syn
arrangement of the alcohol and aryl (R2) group in the major
diastereomer of 2b. In this syn isomer, the chemical shift and
Scheme 1. Methodology for the preparation of eight-rings (1a) and initially
unexplained competing six-ring (2b) when R2 is 4-(MeO2C)Ph.
[a] L. Eccleshare, L. Lozada-Rodrꢀguez, P. Cooper, Dr. L. Burroughs, J. Ritchie,
Dr. W. Lewis, Prof. Dr. S. Woodward
School of Chemistry, University of Nottingham
University Park, Nottingham NG7 2RD (UK)
E-mail: simon.woodward@nottingham.ac.uk
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201601970.
Chem. Eur. J. 2016, 22, 1 – 7
1
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Table 1. Substituent and condition effects on the chemoselectivity of eight- versus six-ring formation.[a]
R2
X
Conditions prior to quench
(1)/(2)[b] d.r. (2)[d] Isolated product [%][c]
Ph
Ph
Ph
Ph
Ph
Cl
warmed to À108C over 0.5h and hold for 1h (standard)
100:0
100:0
100:0
1:4
–
–
–
1a (70)
1a (58)
1a (19)
3a[e,f] (48)
–
O2CCF3 standard
OAc
Cl
standard
standard, then warm to rt over 0.5h and hold for 1h
2:1
2:1
5:1
5:1
2:1
–
2:1
2:1
3:1
1:1
–
Cl
standard, then warm to rt over 0.5h and hold for 1 h, then re-cool to À108C for 1h 1:4
4-(MeO2C)Ph Cl
4-(MeO2C)Ph Cl
3-CNPh
3-CNPh
4-(CF3)Ph
4-(CF3)Ph
4-(MeO2C)Ph Cl
4-(CF3)Ph Cl
4-(MeO2C)Ph Cl
4-(CF3)Ph Cl
standard
0:100
0:100
1:9
2:1
1:11
1:1
0:100
1:5
100:0
1:3
2b (65)
–
kept at À508C for 16h
Cl
Cl
Cl
Cl
standard
2c (48)
1c (42)
3d[e,f] (47)
1d (35)
2e[e] (52)
3e[e,f] (49)
1e[e] (27)
3 f[e,f] (42)
kept at À508C for 16h
standard
kept at À508C for 16h
by using 2-bromo-5-methoxybenzaldehyde and warm to 08C (for 1h)
by using 2-bromo-5-flourobenzaldehyde warm to 08C and hold (for 1h)
by using 3-bromothiophene-2-carbaldehyde and warm to 08C (for 1h)
by using 2-bromo-5-methoxybenzaldehyde and warm to 08C (for 1h)
1:2
[a] Reagents: LiCCTMS (3.4 mL of 0.5m THF/hexane solution) treated with 2-bromobenzaldehyde (1.62 mmol), followed by nBuLi (1.06 mL of 1.6m hexane
solution), CuBr·SMe2 (0.81 mmol), propargylic electrophile (0.79 mmol) added at À508C. [b] From crude reaction mixture, determined by 1H NMR spectros-
copy. [c] Isolated. [d] syn (OH to R2)/anti ratio; stereochemical assignment is based on NMR correlation to 2b (X-ray). [e] Structures of products 1e–3 f are
shown below. [f] Isolated as naphthalene 3; dehydration from 2 to 3 achieved with Amberlystꢁ 15 and 3 ꢃ molecular sieves.
coupling constants of the CHR2 proton (dH 4.08, J=
2.9, 2.6 Hz) are distinct from the same proton in the
anti diastereomer (dH 4.01, J=4.6, 1.6 Hz); the other
derivatives had their relative stereochemistries as-
signed on this basis. For propargylic chlorides based
on 3-CNPh substituents, the six- vs. eight-ring che-
moselectivity is controlled by reaction temperature;
reactions at À50 and À108C, prior to quenching,
maximise the yields of 1c and 2c, respectively. The
bis-trifluoromethyls 1d/2d behaved analogously.
When the parent (R1 =SiMe3, R2 =Ph) reaction mix-
ture was warmed to room temperature (over 30 min),
3a (48%) was isolated upon protonation and dehy-
dration. Conversely, the chemoselectivity in reactions using the
4-(CO2Me)Ph substituted propargylic chloride could not be af-
fected by changing the reaction temperature. Even at À508C
(prior to quenching), no eight-ring formation was observed.
The use of lower temperatures (to favour eight-ring formation)
was considered, but deemed impractical, because the reactions
become very slow below À508C. Use of both electron-rich and
deficient substituted (MeO vs. F) 2-bromobenzaldehydes also
resulted in product distributions showing increased six-ring
formation at 08C (Table 1, 2e and 3e–f), when using electron-
withdrawing R2 groups in the propargylic chloride. However,
use of 3-bromothiophene-2-carbaldehyde even with an elec-
tron-withdrawing R2 group led to exclusive formation of only
eight-ring 1e, even under the highest temperature usable that
did not induce competing extensive decomposition (08C).
The R2 substituent and condition-dependant formation of
mixtures of 1 and 2 is most in accord with interconversion of
a common intermediate (4) with both a non-classical C-enolate
(5) and alkoxide (6) prior to final protonation (Scheme 2). A
carbanion formulation for 5 was supported by our previous
computational studies (formation of an eight-ring allenoate en-
genders too much ring strain).[1] The antiperiplanar arrange-
ment in 5 favours CÀC s-bond cleavage regenerating 4
(Scheme 2). Higher temperatures (À108C to ambient) resulted
in thermodynamic capture of 4 forming alkoxide 6. Formation
of 6 is not reversible, because re-cooling solutions enriched
with this anion did not result in the isolation of eight-ring
products (see behaviour of 1a/2a in Table 1). Formation of 6 is
favoured by groups that stabilise the resultant benzyl anion.
The behaviour of 4 has been computationally modelled by
using simplified structure A (n=0) at the CBS-QB3 level of
theory. To decrease computational cost in the model system
(A), the atom count has been reduced by using R1 =H, R2 =
phenyl and CuC6H4(2-CHOLiCꢀCH) has been simplified to
&
&
Chem. Eur. J. 2016, 22, 1 – 7
www.chemeurj.org
2
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Scheme 2. Mechanistic proposal for s (CÀC) cleavage processes in eight- versus six-ring formation.
Diversification by non-classical enolate
CuPh. Viable transition states for both eight- and six-rings (B
and C, respectively, see the Supporting Information) could be
identified with very similar energies. These simple (metal-free)
models suggest that six-ring formation should be kinetically
favoured, and that eight-ring products would be attained as
the thermodynamic product—the opposite of what was ac-
tually observed. This failure of the computational model is ap-
parently caused by the absence of metal ligation, in particular:
1) the absence of a strong LiÀO bond (112 kcalmolÀ1)[3] in the
reaction coordinate approaching (6); and 2) the lack of Cu···al-
kyne coordination in events leading to (5) (see the Supporting
Information). It is clear from Table 1 that the six- versus eight-
ring reaction pathways are rather close in energy, such that
they may be biased by experimental conditions in a useful
manner. Support for Cu···alkyne coordination in events leading
to 5 were provided by reaction of Li[Cu{C6H4(2-CHOLiCꢀCPh)}2]
(formed from enantioenriched (R)-1,2-C6H4Br(CHOHCꢀCPh),
88% ee) with (Æ)-PhCHClCꢀCH (see the Supporting Informa-
tion). This resulted in the formation of (À)-1j with 20% ee.
Thus, both chiral cuprate coordination to prochiral 4 and par-
tial chirality transfer appear to be possible. The alternative pos-
sibility of kinetic resolution of (Æ)-PhCHClCꢀCH cannot be
completely discounted; however, even if this was the case, the
reaction manifold would still have to pass through 4 and
a “memory of chirality” effect would be required to explain the
observed enantiomeric excess (ee).[4] This seems to be less
likely than alkyne coordination, which was also favoured in
preliminary computational studies by using A (n=1, see the
Supporting Information). Conversion of 4 to 6 can be affected
In many cases, the C-enolates (5) are stable at À108C and thus
offer an excellent point for rapid core diversification. We
sought to extend the preliminary observations we had made
in this area.[1] It was found that a range of electrophiles can be
added directly into the one-pot cascade sequence to give
a wider variety of eight-ring derivatives or to install reactive
handles (Scheme 3). Trapping enolate 5 with reactive deuteri-
um oxide, allyl bromide, and dimethyl disulfide allowed prepa-
ration of functionalised ring systems 7–9 (52–64%). These pro-
cedures are not appropriate for lower reactivity electrophiles,
for example, DEAD and tributyltin chloride resulted in the for-
mation of the parent 1a after aqueous work-up. Addition of el-
emental iodine led to the formation of 10, in good yield,
which lends itself well to further diversification by Suzuki–
Miyaura cross-coupling reactions giving 11 a–c (79–87%) as
representative examples. Amides 13a–c and pyrazolone 14
can conveniently be synthesised from our previously de-
scribed[1] carboxylic acid 12 in 71–95% yield. The rapid synthe-
sis of pyrazolone 14 (47% overall yield, 2 steps from 2-bromo-
benzaldehyde), amides 13a–c (35-40% overall yield, 2 steps
from 2-bromobenzaldehyde) and furyl 11 c (46% overall yield,
2 steps from 2-bromobenzaldehyde) showed how this method-
ology can swiftly incorporate heteroatoms and drug-like motifs
into the core scaffold.
Diversification of eight-ring core structures
1
by polar co-solvents. The H NMR spectrum of the crude reac-
We have also explored the potential for representative eight-
rings 1a and f to provide additional derivatives through simple
modification (Scheme 4). Ring expansions to nine-membered
cyclic amides were achieved via the formation of oximes 15a–
b. These are attained in high yield in the presence of hydroxyl-
amine hydrochloride and sodium acetate providing 15a (83%,
1.7:1 E:Z) and 15b (98%, 1.5:1 E:Z). Proton, 13C and 2D NMR
analysis of 15b are in accord with the major isomer to be the
E-oxime. Based on the X-ray structure of the parent carbonyl
1a and f, we believe that proximity of the OH to the a-vinylic
CH in the E-oxime causes characteristic shifts for this signal in
1H and 13C NMR spectrums, as has been observed before.[6] In
tion mixture containing 2a (prior to conversion to 3a) showed
that addition of DMF (10 equiv compared to substrate) before
warming to room temperature can increase the diastereoselec-
tivity from 2:1 to 4:1. The facility of six-ring formation is more
affected by geometrical factors than the electronic effects of
the aldehyde. Both electron-rich and -poor substitution of the
1,2-C6H4 unit provides six-ring major products (2e and 3e and
f). However, replacement of the phenylene unit by a five ring
1,2-thiophenylene negates six-ring formation completely (1e).
Calculations (B3LYP/6-31G+(d), vacuum, see Supporting Infor-
mation) on the ground states of the model anions (Z)-1,2-
Aryl{CA(=O)CꢀCBH}{CH=CH(CCHÀ)C6H4CO2Me} indicated that for
Aryl=C6H4, CC is significantly closer to CA (CA·CC 3.34 ꢃ, CB·CC
3.67 ꢃ). For Aryl=C4S (thienyl), the opposite is true (CA·CC
3.63 ꢃ, CB·CC 3.29 ꢃ), in line with experimentally observed for-
mation of 1e. Finally, we note that cross-over from eight- to
six-ring formation has been noted once before, but with alkox-
ide nucleophiles not carbanions.[5]
1
practice, the H NMR E-oxime a-CH signal of 15b resonates at
d=7.35 ppm versus Z-oxime at approximately 6.75 ppm,
whereas in the 13C NMR spectra, the E-oxime a-CH showed dC
ꢁ118 vs. dC ꢁ124 for the Z-oxime. The Beckmann rearrange-
ment of oximes 15a–b in the presence of toluenesulfonyl chlo-
ride gave nine-membered cyclic amides 16a–b in moderate
yields. Such rearrangements generating medium-sized rings
Chem. Eur. J. 2016, 22, 1 – 7
www.chemeurj.org
3
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Scheme 3. Diversification via non-classical C-enolate (5, R1 =SiMe3, R2 =Ph). Reagents and conditions: (i) D2O (10 equiv), À108C; (ii) allyl bromide (5 equiv),
À108C to ambient; (iii) dimethyl disulfide (5 equiv), À108C to ambient; (iv) I2 (5.0 equiv), À108C to ambient; (v) arylboronic acid (1.05–1.50 equiv),
[PdCl2(dppf)] (3–5 mol%), sodium bicarbonate (5 equiv) in THF; (vi) solid CO2 (ca. 20-fold excess), À108C to ambient; (vii) amine (2 equiv), NEt(iPr)2 (3 equiv),
T3Pꢁ (2 equiv) in EtOAc; (viii) HOBt (1.2 equiv), EDC·HCl (1.2 equiv), hydrazine (2 equiv) in MeCN/THF.
Scheme 4. Diversification of eight-ring cores. Reagents and conditions: (i) hydroxylamine hydrochloride (20 equiv) and sodium acetate (20 equiv) in ethanol
upon heating at reflux; (ii) sodium hydride (3 equiv) and toluenesulfonyl chloride (1.1 equiv) in THF; (iii) neat trifluoroacetic acid (5.5 equiv) from 1a only; (iv)
tetrabutylammonium fluoride (2 equiv) in THF from 1a only; (v) boron trifluoride diethyl etherate (0.3 equiv) in toluene (mw, 1508C); (vi) lithium aluminium hy-
dride (1 equiv) in THF; (vii) methylmagnesium bromide (1.7 equiv) in THF; (viii) allylmagnesium bromide (1.4–1.7 equiv) in THF.
&
&
Chem. Eur. J. 2016, 22, 1 – 7
www.chemeurj.org
4
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
are known to be challenging.[7] However, it should be noted
that the consolidated yields of nine-rings 16a–b over the
three-step sequence are competitive with current literature ap-
proaches to such systems.[8] The regiochemistry of the rear-
rangements was confirmed by X-ray analysis of 16a. Subject-
ing 1a to neat trifluoroacetic anhydride cleaved the TMS
group from the eight-membered ring and triggered alkene iso-
merisation to synthesise enone 17 in excellent yield. Converse-
ly, removal of the TMS group with tetrabutylammonium fluo-
ride triggered a different isomerisation that gave conjugated
diene 18 in 82% yield. The reasons for differing isomerisation
mechanisms are not completely understood, but 18 and 17
are not inter-convertible in the presence of trifluoroacetic acid.
Fused cyclobutene 19 was synthesised in 79% yield from the
intramolecular [2+2] cycloaddition of diene 18. Microwave
heating (1508C) of the diene in the presence of boron trifluor-
ide etherate gives the Woodward–Hoffmann thermally allowed
trans-19, the stereochemistry of which was confirmed by NOE
NMR experiments (see the Supporting Information).
Scheme 5. Conformational bias of eight-membered rings controls Grignard
1,2-addition diastereoselectivity and X-ray structural observation of H···C=C
contact in anti-(21a) R=Me.
proton is indicative of the diastereoisomer formed, a higher
frequency signal (dH ꢁ4.6 ppm) is indicative of anti diastereo-
isomer formation (vs. dH ꢁ4.4 ppm for the equivalent syn
signal). On this basis, the diastereoselectivity of allyl Grignard
additions were also assigned. Relative signal integrations were
used to determine the syn/anti ratio in 21 and 22. Blocking of
the Bꢄrgi–Dunitz[11] (p*) approach leads to excellent syn diaste-
reoselectivity (>40:1) in the formation of 21b and 22b. The
anti selectivity of 21a and 22a is not as good (7:1 and 3:1, re-
spectively), because both diastereofaces of the carbonyl in 1a
are more open. There are hints that the distal cis-CÀC p bond
might also act as a directing group (Mg···p bond in the
Grignard addition), because product anti-(21a) showed an un-
usual, but potentially related, H···p-bond hydrogen bond
(2.2 ꢃ, structure in the centre of Scheme 5). However, this sug-
gestion should be regarded as tentative. It seems that similar
syn/anti control issues account for the stereochemical reversal
seen in hydride addition to form 20a and b. Due to the pro-
tection offered to the top diastereotopic face in conformation
(Y), it is tentatively assumed that hydride addition into carbon-
yl 1 f will also result in the formation of predominantly syn-
(20b).
Reduction of the ketones 1a and f with lithium aluminium
hydride gave alcohols 20a and b in 71 and 85% yield, respec-
tively. Reduction of the carbonyl removes the Michael accept-
ing enone moiety, potentially increasing its appeal for medici-
nal chemistry screening. Michael acceptor units are frequently
excluded from initial screening protocols to avoid toxicologi-
cal/suicide substrate issues.[9] Interestingly, our own attempts
at 1,4-conjugate additions into 1a showed a complete resil-
ience to nucleophilic attack at this position. Simple 1,2-addi-
tion of methylmagnesium bromide and allylmagnesium bro-
mide into ketones 1a and f proceeded with high yield of 82–
97%. When analysing the stereo outcome of reactions, which
involved direct 1,2-addition of Grignard reagents to the car-
bonyl, it became clear that opposite diastereoselectivity was
attained from the two parent substrates 1a (R1 =SiMe3) and 1 f
(R1 =tBu); R2 =Ph in both cases. Conformational bias in the
tert-butyl (1 f) versus TMS derivative (1a) starting materials is
the source of this behaviour, based on analysis of their X-ray
crystal structures (see the Supporting Information). For com-
pound 1a, conformation (X) is seen in the solid state
(Scheme 5). The phenyl group is pseudo equatorial placing the
CHPh proton endocyclic, presumably to minimise transannular
Finally, as the majority of our studies had involved the use
strain. In the X-ray structure of 1 f, the shorter =CÀCMe3 versus of simple aryl groups in the R2 position of structure 1, we were
=CÀSiMe3 bond (1.53 compared to 1.90 ꢃ) leads to significant
increase in allylic A1,2 strain,[10] destabilising conformer (X) and
leading to the observation of (Y) in the solid state for 1 f (see
the Supporting Information). The subsequent conformational
change results in endocyclic placement of the phenyl group di-
rectly obscuring one of the diastereotopic faces of the carbon-
yl. The hypothesis that these solid-state observations also op-
erate in solution is corroborated by the X-ray structures of the
major diastereoisomers resulting from MeMgBr addition to 1a
and f. For tert-butyl (1 f; conformer Y), addition to the bottom
face is seen (see the Supporting Information) leading to syn-
21b (diastereomeric assignment relates to the orientation of
the OH in respect to the phenyl group in all cases, as shown in
Scheme 5). Addition to the TMS derivative 1a led to anti-21a
as the major product. The 1H NMR chemical shift of CHPh
very keen to extend the range of propargylic chlorides to in-
clude electron-rich (hetero)aromatics, electron-poor (hetero)ar-
omatics and alkyl functionality. Although electron-rich aromatic
propargylic chlorides proved too reactive to be used in the re-
action (e.g., 2-furanylCHClCꢀCH spontaneously decomposed at
room temperature in our hands), the less reactive acetate al-
lowed the one-pot synthesis of electron-rich 1g. Conversely,
electron-poor 1h containing the pyrazolyl function could be
prepared from the corresponding propargylic chloride
(Scheme 6). By using 3-chlorooct-1-yne and a slightly modified
procedure (see the Supporting information), we were also able
to synthesise alkyl-chain-containing 1i.
Chem. Eur. J. 2016, 22, 1 – 7
www.chemeurj.org
5
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
EPSRC’s National Service for Computational Chemis-
try Software (NSCCS) for computational facilities and
for support of X-ray facilities at Nottingham (award
EP/K038869/1). L.L.-R. is grateful for financial support
(Grant 266665) by Consejo Nacional de Ciencia y Tec-
nologꢀa (CONACYT, Mexico) towards supporting her
fellowship at Nottingham.
Scheme 6. Representative variation of substituent R2 in compounds (1).
Keywords: annulation · CÀC activation · medium-
ring compounds · ring expansion · synthetic methods
Conclusion
Due to the simplicity of their synthesis,[1] bicyclic 6,8-ring sys-
tems containing the cycloocta-2,5-dien-1-one core are attrac-
tive starting materials for diversity chemistry potentially aimed
at the development of biologically active libraries. Addition of
LiCꢀCR1 to 2-bromobenzaldehyde, followed by subsequent
cuprate formation and reaction with R2CHClCꢀCH gave a non-
classical C-enolate as the kinetic product at lower temperatures
that can be trapped with a wide range of electrophiles to rap-
idly provide diverse libraries with little synthetic effort (1-3
steps total, consolidated yields>25%). These products them-
selves are attractive starting points for ring expansion and dia-
stereoselective addition reactions. If allowed to warm, the C-
enolate rearranges via an unusual retro s CÀC bond cleavage
to provide some 6,6-ring systems. In support of the assertion
that the strategies presented herein are useful in biological ac-
tivity discovery libraries, several of the compounds described
herein have already revealed significant activity in open-inno-
vation drug-discovery programmes.[12]
[1] L. Burroughs, L. Eccleshare, J. Ritchie, O. Kulkarni, B. Lygo, S. Woodward,
W. Lewis, Angew. Chem. Int. Ed. 2015, 54, 10648–10651; Angew. Chem.
2015, 127, 10794–10797.
[2] For recent overviews of diversity synthesis as discovery tools in medici-
nal chemistry and the importance of medium rings, see: a) W. Galloway,
A. Isidro-Llobet, D. R. Spring, Nat. Commun. 2010, 1, 80; b) M. Dow, M.
Fisher, T. James, F. Marichetti, A. Nelson, Org. Biomol. Chem. 2012, 10,
17–28; c) W. Galloway, A. Bender, M. Welch, D. R. Springer, Chem.
Commun. 2009, 2446–2462; d) A. Hussain, S. K. Yousuf, D. Mukherjee,
RSC Adv. 2014, 4, 43241–43257.
[3] S. Woodward, Tetrahedron 2002, 58, 1017–1050.
[4] For a review of ‘memory of chirality effects’ see: H. Zhao, D. C. Hsu, P. R.
Carlier, Synthesis 2005, 1–16.
[5] S. L. Schreiber, S. E. Kelly, Tetrahedron Lett. 1984, 25, 1757–1760.
[6] K. Pandiarajan, R. T. Sabapathy Mohan, Magn. Reson. Chem. 1986, 24,
312–316.
[7] a) V. P. Srivastava, R. Patel, D. Garima, L. D. S. Yadav, Chem. Commun.
2010, 46, 5808–5810; b) Y. Furuya, K. Ishihara, H. Yamamoto, J. Am.
Chem. Soc. 2005, 127, 11240–11241.
[8] Representative examples of nine-membered lactam synthesis (number
of steps and overall yield are given in parenthesis); see: a) J. K. Jung,
N. S. Choi, Y. G. Suh, Arch. Pharmacal Res. 2004, 27, 985–989 (7, 24%);
b) L. Chacun-Lefꢅvre, V. Bꢆnꢆteau, B. Joseph, J. Y. Mꢆrour, Tetrahedron
2002, 58, 10181–10188 (6, 43%).
[9] a) W. P. Walters, M. A. Murcko, Adv. Drug Delivery Rev. 2002, 54, 255–
271; b) J. B. Baell, G. A. Holloway, J. Med. Chem. 2010, 53, 2719–2740.
[10] F. Johnson, S. K. Malhotra, J. Am. Chem. Soc. 1965, 87, 5492–5493.
[11] H. B. Bꢄrgi, J. D. Dunitz, J. M. Lehn, G. Wipff, Tetrahedron 1974, 30,
1563–1572.
Acknowledgements
Early parts of this study were facilitated by the European
Union’s Seventh Programme for research, technological devel-
opment and demonstration under grant agreement No
308768. L.E. thanks the University of Nottingham for provision
of a studentship. J.R. is grateful to European Thermodynamics
Ltd. and the Engineering Physical Sciences Research Council
(EPSRC) for provision of a studentship. We acknowledge
[12] M. Alvim-Gaston, T. Grese, A. Mahoui, A. D. Palkowitz, M. Pineiro-Nunez,
I. Watson, Curr. Top. Med. Chem. 2014, 14, 294–303.
Received: April 27, 2016
Published online on && &&, 0000
&
&
Chem. Eur. J. 2016, 22, 1 – 7
www.chemeurj.org
6
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
Medium-ring compounds: Simple 1–3
step processes provide access to 6,n-by-
cyclic ring systems (n=6,8,9) and other
rapidly prepared derivatives (see
&
Synthetic Methods
L. Eccleshare, L. Lozada-Rodrꢀguez,
P. Cooper, L. Burroughs, J. Ritchie,
W. Lewis, S. Woodward*
scheme). Several of the compounds de-
scribed herein have already revealed
significant activity in open-innovation
drug-discovery programmes.
&& – &&
Diversification of ortho-Fused
Cycloocta-2,5-dien-1-one Cores and
Eight to Six-Ring Conversion by
s Bond CÀC Cleavage
Chem. Eur. J. 2016, 22, 1 – 7
www.chemeurj.org
7
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ