Design and scale-up of a practical enantioselective route to 5-Phenylbicyclo[2.2.2]oct-5-en-2-one

Article abstract of DOI:10.1021/op200305y

A practical enantioselective route to chiral 5-phenylbicyclo[2.2.2]oct-5- en-2-one 1 has been designed and developed. The target compound has been obtained as colorless crystals in 22% yield from 2-cyclohexenone, with an enantiomeric ratio higher than 99.5:0.5 and notably high chemical purity (> 99%). Three intermediates out of nine chemical steps are isolated. It is noteworthy that this process is devoid of any chromatography or distillation although all but one intermediate are oils. Key to success was the optimization of an intramolecular aldol reaction of an in situ prepared ketone aldehyde leading to the solid intermediate (1R,4R,4S,6S)-6-hydroxybicyclo[2.2.2]octan-2- one 9a that is isolated in very high chemical and chiral purity. This is an example of an intramolecular crystallization-induced diastereomer transformation (CIDT). The dehydration of this secondary alcohol to 1 required an extensive screen of reaction conditions to secure an excellent purity, essential for crystallization of this low-melting compound. The final process is simple and concentrated as demonstrated by an expeditious synthesis of 1 kg of 1 in a 30-L reactor in 10 working days.

Full text of DOI:10.1021/op200305y

Article
pubs.acs.org/OPRD
Design and Scale-Up of a Practical Enantioselective Route to
5-Phenylbicyclo[2.2.2]oct-5-en-2-one
Stefan Abele,* Roman Inauen, Jacques-Alexis Funel, and Thomas Weller
Process Research Chemistry, Actelion Pharmaceuticals Ltd., Gewerbestrasse 16, CH-4123 Allschwil, Switzerland
S
* Supporting Information
ABSTRACT: A practical enantioselective route to chiral 5-phenylbicyclo[2.2.2]oct-5-en-2-one 1 has been designed and
developed. The target compound has been obtained as colorless crystals in 22% yield from 2-cyclohexenone, with an
enantiomeric ratio higher than 99.5:0.5 and notably high chemical purity (> 99%). Three intermediates out of nine chemical
steps are isolated. It is noteworthy that this process is devoid of any chromatography or distillation although all but one
intermediate are oils. Key to success was the optimization of an intramolecular aldol reaction of an in situ prepared ketone
aldehyde leading to the solid intermediate (1R,4R,4S,6S)-6-hydroxybicyclo[2.2.2]octan-2-one 9a that is isolated in very high
chemical and chiral purity. This is an example of an intramolecular crystallization-induced diastereomer transformation (CIDT).
The dehydration of this secondary alcohol to 1 required an extensive screen of reaction conditions to secure an excellent purity,
essential for crystallization of this low-melting compound. The final process is simple and concentrated as demonstrated by an
expeditious synthesis of 1 kg of 1 in a 30-L reactor in 10 working days.
enolate alkylation.7 Several adjustments of the oxidation stage
would lead to a Michael adduct 7, which is conceivably
synthesized from malonate addition to 1,4-dioxaspiro[4.5]dec-6-
en-8-one 8, a transformation which is lacking an enantioselective
INTRODUCTION
■
Enantiomerically pure (R,R)-phenylbicyclo[2.2.2]oct-5-en-2-one
11 was required as key intermediate for an active pharmaceutical
ingredient (API). Larger amounts (90 kg) were produced with
an initial Diels−Alder step2 and a final chromatographic
example in literature. The intramolecular aldol reaction of a
putative ketone aldehyde intermediate 10,8 path B, was
separation of the enantiomers (Scheme 1). Bicyclo[2.2.2]-
perceived as feasible yet highly risky endeavor due to the
octenes are important skeletons in various fields of organic
chemistry,3 including chiral dienes as a new privileged class of
many isomeric products which can evolve from such a seemingly
ligands as pioneered by the groups of Hayashi4 and Carreira.5
labile substrate. Ketone aldehyde 10 or a masked form thereof
was deemed accessible either via the nitrile 11 or the ester 12,
both being conceivably derived from asymmetric Michael
additions of phenyl acetonitrile or malonates, respectively,
onto cyclohexenone. There is ample precedence for the latter
even on kilogram scale,9 whereas the enantioselective Michael
addition of phenyl acetonitrile to this acceptor was unprece-
dented.10 These paths have been assessed in parallel, some first
as a racemic proof-of-concept, in order to quickly sort out the
best route for development of the enantioselective route.
Further alternative approaches to 1 were checked with
racemates to assess the feasibility: (i) asymmetric 1,4-addition
of organovinyl reagent onto cyclohexenone.11 The ensuing
epoxidation of the double bond failed, giving exclusively the
Baeyer−Villiger lactone product. (ii) On paper, one of the
shortest routes involves the Michael addition of (1-phenylvinyl)-
magnesium bromide12 onto cyclohexenone. However, in
practice, the preparation of the Grignard reagent, the costs,
and the lack of enantioselective examples caused us to leave
this route.
Although this route proved scalable and the separation was
highly efficient, it suffered from some impediments in view of
larger scales: (i) the diene 2 was prepared at low temperatures
and required distillation, (ii) the Diels−Alder reaction was
dependent on the quality of α-chloroacrylonitrile which might
have to be distilled prior use, (iii) the intrinsic loss of more
than half of the material in the final chiral separation without
the possibility for racemization and recycling, and (iv) the
moderate assay of the product (87% w/w) as obtained after
preparative chromatography. The product was obtained in 16%
overall yield as a waxy black solid. On the other hand, the
published route to 1 suffers from a yield of <0.2% over nine steps,
relying on an enantioselective host−guest complex formation,
starting from hydroquinone and maleic anhydride.1a An
enantioselective route was therefore required to secure the access
to larger scales.6 Besides the usual requirements like cost reduc-
tion including a low priced and early introduction of chirality,
high throughput and short cycle times, this specific target placed
a great demand on yields and selectivities, as purification of
the liquid intermediates should be restricted to extraction only.
Clearly, a solid intermediate as a cleanup stage was desired.
Intramolecular Ketone Enolate Alkylation (Path A).
Path A (Scheme 2) starts with 1,4-dioxaspiro[4.5]dec-6-en-8-
one, 8, which was best prepared from commercially available
1,4-dioxaspiro[4.5]decan-8-one.13 Alternative syntheses of 8
ALTERNATIVE ROUTES
■
The main retrosynthetic bond breaks are depicted in Figure 1.
The retrosynthetic path A leads to a known precursor, ketal
ketone 5,2a which was thought to arise from intramolecular
Received: October 24, 2011
Published: December 16, 2011
© 2011 American Chemical Society
129
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
Scheme 1. Original route used for the first GMP campaign2a
Reagents and conditions: (a)2a 66% (R = OAc), 84% (R = H), >88% (R = Cl); (b)2a 4−5 steps, 86% (R = OAc), 39% (R = H), 44% (R = Cl);
(c) preparative chiral chromatography, Chiralpak AS-V (20 μm, 250 mm × 80 mm), eluent: MeOH (0.1% Et3N), er > 98:2, 85% recovery of both
enantiomers; or Chiralpak IA (20 μm, 250 mm × 76 mm), eluent: heptane/EtOAc 75:25, er > 99:1, 82% recovery of both enantiomers.
Figure 1. Proposed retrosyntheses of 1.
Scheme 2. Intramolecular ketone enolate alkylation, racemic proof-of-concept (path A)
Reagents and conditions: (a) dimethyl malonate (1.5 equiv), Et3N, LiCl (1 equiv), toluene, rt, 89%; (b) DABCO (4 equiv), H2O (2 equiv), toluene,
110 °C, sealed vial, 73%; (c) 2 M LiAlH4 in THF (2 equiv), THF, rt, 65%; (d) pTosCl (1 equiv), Et3N (1 equiv), Bu2SnO (0.02 equiv), rt, 84%;
(e) SO3.pyridine (1.5 equiv), Et3N (3 equiv), DMSO (5 equiv), DCM, rt, 61%; (f) KOtBu (1.2 equiv), THF, 5 °C, 60%. (g) ref 2a.
gave inferior results.14 The crucial intermediate as probe for the
cyclization (rac-6) was synthesized by a sequence of Michael
addition of dimethyl malonate to 8,15 decarboxylation in the
presence of 1,4-diazabicyclo[2.2.2]octane (DABCO) to afford
rac-7,16 exhaustive reduction with LiAlH4 to the diol rac-16
as isomeric mixture, selective monosulfonylation of the
primary alcohol with p-toluenesulfonyl chloride (pTosCl) in
the presence of catalytic dibutyltin oxide to get rac-17,17 and a
Parikh−Doering oxidation to yield rac-6.18 Gratifyingly, the
cyclization step proceeded dose-controlled at 5 °C by addition
of KOtBu in THF. The quality of ketone ketal rac-5, a key
precursor of rac-1, was acceptable after comparison of the GC
and NMR spectra with a known sample from another route.2a
Due to expected high costs for the starting material 819 and the
need for a development of a new catalytic enantioselective
addition to 8, we focused on path B.
130
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
Intramolecular Aldol Reaction of Masked Ketone-
Aldehyde 10 to 9 (path B). The intramolecular aldol
cyclization was described for the des-phenyl derivative 18
by Mori,20 Bettolo21 and Mattay,22 see Scheme 3.23 The
into the ketal formation step with ethylene glycol to afford 23
as oil with an er > 99.5:0.5.16,26
Decarboxylation Step to 12. Initial runs with DABCO
and water in toluene at reflux over 16 h led to a black oil of low
purity requiring purification by distillation.16 In practice, a
mixture of malonate 23, 2 equiv of LiCl and 1 equiv of water in
2 vol of DMAc was heated to 135−140 °C and stirred at this
temperature for 4−5 h.27 A white precipitate28 was formed after
reaching 120 °C, indicating the start of the decarboxylation. In
contrast to reported procedures which rely on addition of water
to dissolve the solids, we took advantage of this observation and
removed the salts by a filtration of the reaction mixture at rt.
This significantly reduced the volumes of both water and
toluene for the workup with this dipolar aprotic solvent.29 The
filter cake was washed with 0.5 vol toluene, and the product 12
was obtained as low-viscosity oil in 88% yield30 after a final
water extraction step and solvent removal. In the absence of
water, degradation occurred. Best conversion was achieved with
2 equiv of LiCl; the use of 1 equiv led to 20% conversion and
degradation. Besides being operationally simpler, this process is
faster than in published protocols.31 Viable solvents are DMAc,
NMP, DMF, and DMSO. DMSO was not judged as the ideal
solvent due to its instability at elevated temperatures. The
evolving chloromethane was trapped in a scrubber filled with a
1:1 mixture of ethanolamine and water.
α-Arylation to 24. The α-arylation of esters is an
established method for larger scales as well32 since the seminal
work by the Hartwig and Buchwald groups.33 We used lithium
diisopropylamide (LDA) as base. P(tBu)3·HBF4 was chosen as
a hindered phosphine ligand due to its stability and ease of
handling. Pd(OAc)2 and Pd2(dba)3 gave similar results.
Replacement of bromobenzene by chlorobenzene was only
feasible with Pd2(dba)3 and gave similar yields.34 Excess
bromobenzene was difficult to remove from the oil; therefore,
only 1 equiv was used.35 The reaction displayed a slight delayed
exotherm which did not vary much upon scale-up (internal
temperature rose to 25−29 °C with a constant jacket
temperature at 20 °C on 40 g-, 400 g-, and 4-kg scale); on
the other hand, dosage to the enolate of bromobenzene alone
or as a mixture with 12 at 40 °C led to byproducts. Dosage of
a mixture of Pd2(dba)3 and P(tBu)3·HBF4 in toluene did not
seem advisable as this mixture gave a precipitation when
Scheme 3. Literature precedence for the racemic des-phenyl
derivative 1821a
cyclization afforded an endo/exo mixture of ∼85:15, and rac-
19a was obtained in 61% yield after flash chromatography.21a
To probe the unprecedented cyclization to 9 we synthesized
the required α-phenyl derivative rac-21 as depicted in Scheme 4.
Aldehyde rac-21 was prepared according to published
procedures. A Michael addition of phenylacetonitrile 14 to
cyclohexenone 13 afforded rac-20,24 which was protected as
ketal rac-11 and finally reduced to aldehyde 21. Gratifyingly, the
pivotal intramolecular aldol reaction proceeded with 5 M
H3PO4 in quantitative (uncorrected) yield using crude rac-21.22a
The dehydration was achieved by heating the intermediate
mesylate rac-22 in the presence of Li2CO3 and LiBr in DMF at
150 °C. It is noteworthy that all intermediates (oils) have
been processed as crude products into the next stages without
chromatography, the only purification being short-path
distillation of rac-1.
ENANTIOSELECTIVE ROUTE
■
With the racemic proof-of concept for the crucial stereospecific
C−C bond formation at hand and the availability of an efficient
synthesis of the chiral starting material 23,16,25 we set out to
develop a robust route building on the Shibasaki reaction as
enantioselective step (Scheme 5). As a testimony to the
practicality of this robust transformation, the enantiomerically
pure malonate 23 was synthesized on multi kg scale, starting
with a Michael addition of dimethyl malonate to cyclohexenone
13 catalyzed by (R)-ALB (Al−Li-bis(binaphthoxide complex);
without crystallization the crude Michael adduct was telescoped
Scheme 4. Phenylacetonitrile route, racemic proof-of-concept (paths B, B′)
Reagents and conditions: (a)24 1:1 mixture of diastereoisomers; (b) ethylene glycol (10 equiv), pTsOH·H2O (0.05 equiv), toluene, 110 °C, 1.5 h,
99%; (c) 1 M diisobutyl aluminium hydride in heptane (1.7 equiv), THF, 25 °C, 59%; (d) 5 M H3PO4 (6.5 equiv), THF, 80 °C 5 h, 101%;
(e) Methanesulfonyl chloride (MsCl, 1.9 equiv), Et3N (2.0 equiv), DCM, 10−20 °C, 1 h, 89%; (f) i) Li2CO3 (1.0 equiv), LiBr (1.0 equiv), DMF,
150 °C, 1 h, 91% crude, ii) distillation, 120 °C, 0.001 mbar, 37%.
131
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
Scheme 5. Preferred route for scale up (paths B, B″)
Reagents and conditions. Yields are corrected for NMR assay of starting material and product: (a) LiCl (2.0 equiv), H2O (1.0 equiv),
dimethylacetamide (DMAc), 140 °C, 5 h, 88%; (b) i) 33% HexLi in hexane (1.1 equiv), diisopropylamine (DIPA, 1.2 equiv), toluene, 0−10 °C,
30 min, (ii) 12, 5−10 °C, 45 min, iii) tris(dibenzylideneacetone)dipalladium (Pd2(dba)3, 0.01 equiv), P(tBu)3·HBF4 (0.01 equiv), bromobenzene
(1.0 equiv), 10−29 °C, 3 h, toluene-solution of 24 telescoped into next step; (c) 2.4 M LiAlH4 in THF (0.55 equiv), toluene, 5−15 °C, 30 min, 84%
(2 steps); (d) NaOCl solution (12% w/w, 1.2 equiv), aqueous sat. NaHCO3 solution, KBr (0.1 equiv), 2,2,6,6-tetramethyl-piperidine-1-oxyl, (TEMPO,
0.01 equiv), 5−10 °C, 1 h, EtOAc-solution of 21 (2 vol) telescoped into next step; (e) 32% HCl (0.3 equiv), 50 °C, 2 h, 45% (2 steps); (f) MsCl
(1.3 equiv), Et3N (1.5 equiv), toluene, 10−20 °C, 10 min, telescoped into next step; (g) 2,4,6-collidine, 140−145 °C, 1.5 h, two crops, 79% (2 steps).
Scheme 6. Small-scale synthesis of unprotected keto-aldehyde rac-10 for initial test of cyclization
prepared separately. Once the catalytic cycle with the
halobenzene was runningas judged by product formation
according to HPLCtemperatures up to 50 °C were not
detrimental. Excess of LDA (1.25 equiv) did not improve
conversion. A concern was the removal of the color as no
purification of the downstream intermediates by crystallization
seemed possible at this stage. The black precipitation (Pd black
and residues of the ligand) was triggered by lower pH. The best
compromise between color removal and ease of phase split was
accomplished by a citric acid quench, followed by a phase split,
water extraction, and a final charcoal treatment of the organic
layer.36 Scavengers (Silicycle, Si-Diamine, 1.52 mmol/g)37 were
not superior for decoloration. On kilogram scale, neat 12 was
added to freshly prepared LDA in hexane and toluene at
0−10 °C. After 10 min, Pd2(dba)3 and P(tBu)3·HBF4 were
added, followed by the dosage of bromobenzene. Following
aqueous workup, the toluene solution was concentrated,
thereby removing water by azeotropic distillation and adjusting
the desired concentration of 24 for the next step.38 Ester 24
consisted of a mixture of like and unlike epimers 24a and 24b in
a ratio of 60:40 to 70:30 by GC, both epimers being productive
substrates for downstream chemistry. For analytical purposes
we isolated both racemic epimers 24a and 24b allowing for the
determination of the relative configuration by X-ray structure
analysis (rac-24a, Supporting Information [SI]).
Reduction to 25. The reduction of the ester 24 to the
alcohol 25 was straightforward. We concentrated on stream-
lining the process and minimizing the equivalents of LiAlH4
and solvent volumes. On 5-kg scale, 0.55 equiv of LiAlH4 was
dosed as 2.4 M solution in THF to the toluene stream of the
preceding step, the exothermic reaction being dose controlled
at 5−15 °C. The classical three-stage quench employing the
least amount of water was employed, thus producing a nicely
granulated suspension for easy filtration.39 Simple removal of
solvent afforded the product 25 as low-viscosity oil in 84% yield
over two steps. Remarkably, this reduction is very concentrated,
i.e. 4.2 vol solvent was required for reaction, including quench.
Oxidation/Cyclization Sequence to 9a. Oxidation to 21.
The masked precursor for the cyclization 21 was accessed from 25
by the TEMPO-bleach oxidation protocol, see Scheme 5.40 It was
anticipated that cyclization would ensue directly after
deketalization in a one-pot fashion.
We prepared the putative precursor for the intramolecular
aldol reaction, the labile free ketone rac-1041 via keto-alcohol
rac-26 to get more insight into this intramolecular aldol reaction
(Scheme 6). In situ prepared rac-10 cyclized spontaneously to
132
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
Scheme 7. Isomeric distribution of the products of the intramolecular aldol reaction
rac-9 upon standing as neat oil at rt, one diastereoisomer being
enriched. This made us confident to find mild conditions for the
cyclization of 21. It is noteworthy that basic conditions, too, were
beneficial for the cyclization of rac-10 to rac-9, for example KOH
in aqueous MeOH, or KOtBu in THF.
Some process parameters of the oxidation required
optimization. Decomposition was observed with the less
expensive 4-hydroxy-TEMPO.42 Both DCM and EtOAc
Figure 2. X-ray structures of rac-9a, rac-9b, and rac-9c.
proved successful for this oxidation. Toluene in alcohol 25
slightly decreased both yield and assay purity. In practice,
a freshly prepared bleach solution of pH 8−9.5, was dosed at
5−10 °C to a mixture of 25, KBr (dissolved in water), and
TEMPO in EtOAc. The concentrated (8.5 vol) exothermic
oxidation was dose-controlled. After quench of excess bleach
with sodium thiosulfate and an aqueous wash of the organic
phase, the concentration for the next step (2 vol EtOAc) was
adjusted by distillation under reduced pressure.43 The ketal
aldehyde 21 proved to be sufficiently stable to allow for
an aqueous workup and a telescoping into the next stage,
vide infra.44
Cyclization to 9. The projected cyclization was associated
with the formation of many conceivable products and isomeric
mixtures thereof. In addition to the four possible isomeric
alcohols 9a−d (Scheme 7) which could be produced by intra-
molecular attack of the ketone enolate onto the aldehyde, the
following byproducts are conceivable: oligomers derived from
intermolecular aldol reactions, ketals of all products, cyclo-
butanols45 arising from cyclization to the opposite methylene
of the keto group in 10, and hemiacetals45b derived from an
attack of the aldehyde enolate to the ketone. Original
literature conditions for the racemic des-phenyl derivative
18 (Scheme 3) or for rac-21 (Scheme 4) were quite harsh,
employing either phosphoric acid22 or 2 N HCl in refluxing
THF21a for 4 and 24 h, respectively; the product was isolated
after evaporation to dryness, aqueous workup, and chroma-
tography. Employing these conditions our phenyl derivative
21 afforded semisolid material 9 as isomeric mixture that
was used as such in the last two steps. For the first racemic
batches on 100-g scale we applied a telescoped process:
dosage of the DCM-solution containing the aldehyde rac-21
to a solution of 32% HCl in 2-methyl-THF at 65−75 °C,
affording 9 as a yellow, greasy solid in 80% assay-corrected
yield. Residual 2-methyl-THF and ethylene glycol proved
difficult to be removed and led to decreased purity in the
ensuing stages.
observation that the isomer 9a (and rac-9a) was isolated with
>99% diastereoisomeric purity after simple filtration of the
reaction mixture (sum of isomers 9a and 9b was 70−85% a/a
by LC−MS, with a best ratio of 9a and 9b of 85:15), when we
turned to solvents like esters or toluene. The diastereomeric
ratio of starting ketal aldehyde 21 was between 60:40 and 70:30
(by GC−MS) in favor of the isomer with like topicity (R,R).48
However, in-process controls (IPC) by HPLC of the mixture
indicated 9a with (S)-configuration of the benzylic position as
major isomer. We think that the equilibrium between these
isomers is driven by the higher melting pointhence lower
solubilityof 9a, thus depleting the mixture in 9a. Epimeriza-
tion should occur via the reversible aldehyde enolate
reprotonation (Figure 3).49
To elucidate the utility of each single isomer for downstream
chemistry we submitted these three diastereomerically pure
racemic isomers separately to the elimination conditions of the
next step (Scheme 5): whereas rac-9a and rac-9b displayed
comparable kinetics and IPC purity, rac-9c turned out to be
unproductive resulting in decomposition.
With these data at hand, a first specific goal for the optimiza-
tion of the cyclization was to reach the highest possible isomeric
ratio of diastereoisomer 9a in the reaction mixture. The isomeric
ratio was determined by chiral HPLC separating all six
isomers.50 Consistent purity indication was obtained by thin
layer chromatography (TLC), whereas LC or GC often faked
too high a purity which was not corroborated by NMR assay or
TLC.51 After having reached the highest possible ratio, the yield
was to be optimized.
A set of experiments performed with 21 is presented in
Table 1. Running the reaction more concentrated led to better
results (entries 1 and 2). The kinetics were elucidated (entries
3, 4): an IPC at 20 min showed a multitude of products in
addition to the three major isomers which coalesced to two
products, 9a and 9b, after 100 min. Raising the temperature
from 50 to 70 °C did not significantly improve the ratio of 9a
(entry 3, 5). 50 °C was found to be the optimum temperature
considering the overall purity by HPLC which decreased with
higher temperatures (not shown). In addition, important conclu-
sions have been obtained from screening reaction conditions
with racemic 2152 (1 equiv 32% HCl, 2 h, 20 °C): (i) the
reaction requires temperatures higher than rt; the isomeric ratio
A chromatographic separation of the three main (racemic)
products allowed for the determination of the relative
configuration by X-ray crystal structure analysis (Figure 2).46
In addition, the melting points of these bicyclic isomers led
to a better understanding of the isomeric ratio in the reaction
mixture, which is best rationalized as crystallization-induced
diastereomer transformation (CIDT).47 Key to success was the
133
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
Figure 3. Tentative mechanistic explanation of the observed isomeric ratio by CIDT.
a
Table 1. Optimization of cyclization conditions with
enantiopure 21
(dr > 99:1) and enantiomerically pure (dr > 99:1, er > 99:1)
isomers have been used and gave a similar trend. Gratifyingly,
the diastereoisomer rac-9a was stable (entry 1), a third of
rac-9b isomerized to the desired rac-9a (entry 2), and rac-9c
led to a 1:1 mixture of rac-9a and rac-9b (entry 3). The
enantiomer 9a isomerized to a small extent (entry 4) and 9b
led to a similar ratio as its racemate (entry 5).
Following the optimized conditions (Table 1, entry 1), 32%
HCl was added to the EtOAc-solution containing 3 kg of crude
21 from the oxidation. After stirring at 50 °C for 2 h
(conversion to the sum of all three diastereoisomers of 9 was
97% by LC−MS) the suspension was aged first at 10 °C for
16 h, then at 0 °C for 1 h. The main diastereoisomer 9a
was isolated by filtration in 46% yield over two steps with
er > 99.5:0.5 (HPLC).54 We found that ∼10% toluene in 21
slightly decreased the yield: a second run on 1 kg scale with a
toluene content of 2% w/w in 21 gave 51% corrected yield
from 25 to 9a. This telescoped oxidation/cyclization protocol is
characterized by a highly concentrated CIDT protocol and a
remarkably high diastereomeric purity (>99.5%).55
b
b
b
entry
solvent
temp. °C
time
2 h
9a
%
9b
%
9c
%
1
2
3
4
5
2 vol EtOAc
5 vol EtOAc
1 vol EtOAc
1 vol EtOAc
1 vol EtOAc
50
50
50
50
70
85
68
78
86
80
15
25
13
14
15
0
7
9
0
5
2 h
20 min
100 min
20 min
a
b
5−45 g scale (21), 32% HCl (0.3 equiv). Ratio of each isomer
normalized to the sum of 9a, 9b, and 9c in the reaction mixture, by %
a/a (chiral HPLC).
for rac-9a/rac-9b/rac-9c at 20 °C in EtOAc was 76:16:8; (iii)
reactions in the following solvents gave a lower content of rac-9a
and 5−20% unproductive rac-9c as compared to the results in
EtOAc: acetone, toluene, isopropanol, THF, tert-butyl methyl
ether (TBME). Clearly, EtOAc was the first choice to get both a
high content of the desired isomer and a high-yielding reactive
crystallization.53
We subjected the isomerically pure products 9a,b,c to the
best reaction conditions to shed light on their interconversion
potential, to define the aging time prior to filtration, and to
determine the limits of this process (Table 2). Both racemic
Elimination to 1. Originally (Scheme 4), we engaged
crude 9 consisting of the three isomers (as greasy solid) with a
moderate purity in the mesylate-elimination sequence. The
poor quality of crude 1 hampered an efficient purification by
crystallization or distillation. With 9a of very high purity at
hand, we now embarked on the stepwise optimization. The
obvious one-step dehydration from 9a to 1 was studied first
(Scheme 8). Either no reaction (HOAc, with/without NaOAc;
a
Table 2. Isomerization studies with diastereomerically
pure isomers of 9
starting
material
rac-
b
rac-
b
rac-
b
entry
solvent
9a
9b
9c
1
2
3
rac-9a
rac-9b
rac-9c
2 vol EtOAc
2 vol EtOAc
2 vol EtOAc
100
30
0
67
41
0
2
0
Scheme 8. Dehydration of 9a to 1
59
c
c
c
9a
9b
9c
4
5
9a
9b
2 vol EtOAc
2 vol EtOAc
93
49
7
0
0
51
a
b
Conditions: 0.5 g scale, 32% HCl (0.3 equiv), 50 °C, 18 h. Ratio of
each isomer normalized to the sum of rac-9a, rac-9b, and rac-9c in the
reaction mixture, by % a/a (chiral HPLC). Ratio of each isomer
c
BF3·Et2O in DCM; SiO2), decomposition (thionyl chloride
in toluene; phosphoryl chloride; Ac2O/HOAc; pTsOH in
toluene; polyphosphoric acid; dicyclohexylcarbodiimide with
normalized to the sum of 9a, 9b, and 9c in the reaction mixture, by %
a/a (chiral HPLC).
134
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
catalytic CuCl56), or undesired product formation (o-formyl
product with HCO2H or trimethylorthoformiate; two un-
desired products with Burgess reagent57) was observed. Best
IPC purity (55% a/a by LC−MS) was achieved by heating
in neat thionyl chloride for 3 h, albeit with many byproducts.
The ketone group was suspected of intervening in the skeletal
rearrangements.58,59 However, we could not protect 9a as
ketal without acid-catalyzed ring-opening and formation of a
multitude of products. On the other hand, treatment of 9a with
NaH and benzyl bromide to protect the secondary alcohol led
to isomerization yielding a 2:1 mixture of 9a and 9b.
respectively, should enable an easy control of the process
temperature (Table 3). Pyridine gave a spot-to-spot reaction
a
Table 3. Eliminations of 22a in pyridine bases
b
temp.
°C
time
h
IPC %
a/a
entry
base
comments
1
2
3
4
5
5 vol pyridine
115
143
150
143
150
24
3
60
90
66% conversion
byproduct
2 vol 2,6-lutidine
2 vol 2,4,6-collidine
1 vol 2,4,6-collidine
0.5 vol 2,4,6-collidine
3
98
1.5
2
99
100
in 0.5 vol NMP
We then turned to a two-stage process by first synthesizing
the mesylate 22a which was used as substrate for the sub-
sequent screens. Optimization targets were the IPC purity
of 1 by HPLC, followed by the color and the isolated yield. At
first, the amidine bases DBU (1,8-diazabicyclo[5.4.0]undec-7-ene),
DBN (1,5-diazabicyclo[4.3.0]non-5-ene), MTBD (7-methyl-1,5,7-
triazabicyclo[4,4,0]dec-5-ene), TMPDA (N,N,N′,N′-tetramethyl-
propane-1,3-diamine), and DABCO (1,4-diazabicyclo[2.2.2]-
octane) were examined for the eliminations in neat conditions
or in THF, DMF, toluene, diglyme, n-butyronitrile, or acetonitrile
at 85−150 °C (SI). These reactions either proved to be slow at
lower temperatures or produced many byproducts, some arising
from ring-opened amidine bases. The best result was obtained
with 2 equiv of DBU in toluene at 140 °C (external temp.) for
1 h, with a purity of ca. 90% a/a of an IPC sample.
a
0.5 g scale (22a), full conversion unless otherwise stated, IPC a/a of
b
desired product 1 (LC method 1). External temperature.
on TLC; however, the temperature was obviously too low to
achieve full conversion (entry 1). With 2,6-lutidine the purity
was pushed up to 90% which guided us to use 2,4,6-collidine,
the solvent of choice (entries 2, 3). Higher concentrations are
favorable for conversion and purity (entries 3, 4). By LC−MS,
the addition of NMP seemed favorable, but the TLC revealed a
baseline spot and a trace of alcohol 9 as compared to a single
spot for the run in neat collidine (entries 4, 5).
To assess the safety of these reactions at high temperature,
differential scanning calorimetry (DSC) was performed, see
Table 4. On the basis of these results, NMP and collidine were
The risk of ring-opening of DBU and DBN at the required
high temperatures directed us to screen inorganic bases such as
Li2CO3, K2CO3 in o-xylene, DMF, NMP, DMSO, DBU, or
toluene, with and without additives (NaI or LiBr). The elimina-
tion in the presence of Li2CO3 in DBU gave a high purity;
however, degradation started already after 25 min. Alkoxides
(NaOMe, NaOtBu, and KOtBu) in diglyme or THF led to a
low purity of 1. We were surprised to observe a decent purity
(89% a/a) without any base, when running the reaction in the
presence of silica gel. This finding drove the design of the next
experiments in the absence of base.
Amongst the solvents tested (toluene, o-xylene, chloroben-
zene, DMPU (1,3-dimethyl-3,4,5,6-tetrahydro-2(1H)-pyrimidi-
none), DMF, NMP, DMSO, sulfolane), NMP and sulfolane
gave the highest purities (92−99% a/a 1) the latter ending up
in a black solution. We envisaged a melt process heating neat
22a (melting point of 87 °C for mesylate 22a), but this led to
many byproducts.
The finding that the elimination of the mesylate 22a worked
well without base allowed for a hypothesis about the mechanism.
With the E2 anti elimination being impeded by the bicyclic
structure and the E1cb being ruled out by the observation above,
the E2 syn mechanism seems to be prevalent with release of
methanesulfonic acid and explains the observations that only the
two isomers 22a and 22b with the required cis-stereochemistry
of the OMs group and the β-H finally produced 1.60
A one-pot elimination protocol was tempting; after
formation of the mesylate 22a in NMP with MsCl and Et3N,
salts were filtered off, and the filtrate was heated directly to
135 °C for 80 min. The product 1 was obtained as black oil in
91% yield, however, with a low assay of 51% w/w which was
upgraded to only 66% and 84% w/w by short-path distillation61
or chromatography, respectively.
Table 4. DSC data for the elimination of 22a
entry
sample
left limit °C peak °C exotherm kJ/kg
a
1
2
3
4
neat 22a
124
149
100
176
142
177
146
249
−549
−158
−169
−739
22a in 2,4,6-collidine
22a in NMP
rac-22a in DMSO
a
Mp at 87 °C (peak).
considered as acceptable solvents. The DSC charts of neat
solvents and of 1 are disclosed in the SI.
Crystallization of 1 proved challenging due to the low melting
point (66−68 °C) and its lipophilic and unpolar nature.
Residual impurities and solvents combined with tar formation
thwarted crystallization. Phenylketone 1 of inferior quality (i.e.,
< 90% a/a or w/w) was impossible to crystallize without oil
formation. A breakthrough was obtained with heptane for the
crystallization that not only removed most of the impurities
but enabled as well an efficient removal of tarry byproduct
accompanying all runs.62 Alternative crystallization solvents
were heptane/TBME 4:1 v/v or neat TBME, both very con-
centrated (1−3 vol).
Following the preferred protocol on 1.6-kg scale, mesylate 22a
was accessed in 30 min with MsCl and Et3N in toluene from
alcohol 9a. After water extraction, toluene was removed under
reduced pressure.63 The mesylate 22a was dissolved in 1 vol of
2,4,6-collidine and heated at 140−145 °C for 1.5 h. Heptane was
added followed by an acidic aqueous workup. Concentration of
the heptane solution to the desired dilution, seeding at 40 °C,
aging, and filtration gave a first crop of 1 in 63% yield. A second
crop was obtained from the mother liquor in 17% yield, both as
an off-white crystalline solid with an er > 99.5:0.5.64
Even though the number of chemical steps of this new route
increased from six (including chiral separation) to nine this
processstarting from 2-cyclohexenoneis more cost-efficient
as compared to the first Diels−Alder route (Scheme 1).65
We speculated that the less basic pyridines could play the
role of both base and solvent. The pKa and stability seemed
ideal for the elimination and the boiling points of pyridine,
2,6-lutidine, and 2,4,6-collidine at 115, 144, and 171 °C,
135
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
a
Scheme 9. Final process showing the three isolated intermediates; common solvents for telescoping are highlighted
a
Reagents and conditions, refer to Scheme 5 for more details. Yields are corrected for NMR assay of starting material and product: (a) i) LiCl, H2O,
DMAc, 140 °C, 5 h, toluene for workup, ii) HexLi, DIPA, hexane, toluene, Pd2(dba)3, P(tBu)3·HBF4, bromobenzene, 10−20 °C, 3 h, toluene for
workup, iii) LiAlH4 in THF, toluene, 5−15 °C, 0.5 h, 74%; (b) i) NaOCl, NaHCO3, KBr, TEMPO, water, EtOAc, 5−10 °C, 1 h, ii) 32% HCl,
EtOAc, 50 °C, 2 h, 45%; (c) i) MsCl, Et3N, toluene, 10−20 °C, 10 min, ii) 2,4,6-collidine, 140−145 °C, 1.5 h, 79%.
As compared to this preceding route, no time-consuming and
expensive chiral separation is required, cyanides are omitted,
and cryogenic conditions are no longer used. The overall yield
increased from 16% to 22% (Scheme 9).
(85% w/w) gave 4.2 kg of 12 (84% w/w, 12% w/w toluene):
88% yield; Purity (GC−MS): 97% a/a, Rt 2.37 min, [M + 1]+ =
1
215; H NMR (CDCl3): δ 3.95 (s, 4H), 3.68 (s, 3H), 2.26 (d,
J = 6.9 Hz, 2H), 2.03−2.20 (m, 1H), 1.70−1.87 (m, 4H),
1.38−1.66 (m, 2H), 1.26 (t, J = 12.5 Hz, 1H), 0.86−1.06 (m, 1H).
Methyl 2-Phenyl-2-((R)-1,4-dioxaspiro[4.5]decan-7-
yl)-acetate (24, Mixture of Diastereoisomers). All liquids
were degassed by three vacuum (500 mbar) − nitrogen purge
(1 atm) cycles prior charging into the reactor. HexLi in hexane
(33%, 8.3 L, 1.1 equiv) was added to a dry 30-L steel-enameled
reactor, followed by toluene (16 L). Diisopropylamine (3.2 L,
1.2 equiv) was added at 0−10 °C over 30 min (exothermic
reaction), followed by a toluene (0.5 L) rinse. Ketal 12 (4.0 kg,
86% w/w, 18.7 mol) was added neat at 5−10 °C over 45 min
(exothermic reaction), followed by a toluene (1.5 L) rinse. The
milky, light-yellow mixture was stirred for 10 min at 5−10 °C.
Pd2(dba)3 (172 g, 0.01 equiv) and P(tBu)3·HBF4 (55 g,
0.01 equiv) were added, followed by three cycles of evacuation
to 500 mbar and nitrogen purge to 1000 mbar. Bromobenzene
(2.94 kg, 1.0 equiv) was added at 10−15 °C over 15 min. The
black mixture was warmed up to 20 °C by adjusting the jacket
temperature to 20 °C and was stirred for 2 h 45 min (an
exotherm occurred with a maximal internal temp. of 28.6 °C, see
text, for larger scales reaction calorimetry is recommended). An
IPC (GC) showed 99% conversion. A solution of citric acid
monohydrate (2.4 kg, 0.6 equiv) in water (9.6 L) was added at
20−30 °C (exothermic reaction). After separation of the dark
layers the organic phase was washed with water (2 × 12 L).
Charcoal (0.4 kg) was added to the organic phase. After stirring
at 20−30 °C for 30 min the organic phase was filtered over a
pad of Celite (750 g) followed by a final rinse of the filter cake
with toluene (2 L). Solvent (28.6 L) was removed from the
organic phase by distillation at 60 °C jacket temperature and
210−58 mbar, then toluene (2.5 L) was added. This solution
(8.85 kg, 0.9 vol per kg 24) was used in the next step. An aliquot
was withdrawn and evaporated to dryness (to a red-brown oil)
for yield and purity determination. Yield: 4.97 kg content of 24
(exact yield was determined after the next step). NMR assay of
aliquot: 82% w/w; purity (GC−MS): 96% a/a, Rt 3.34, 3.38 min
(pair of diastereoisomers, 1:2), [M + 1]+ = 291.
CONCLUSION
■
The enantioselective process as depicted in Scheme 9 yielded
1 in 22% overall yield starting from 2-cyclohexenone, as
compared to < 1% yield following published protocols.66,67
Essentially, this new route afforded 1 in > 99% chemical purity
and er > 99.5:0.5 as crystalline solid. It is noteworthy that both
enantiomers of 1 are accessible via this route by using either
(R)- or (S)-BINOL in the powerful Shibasaki reaction. The
intramolecular aldol reaction of ketal aldehyde 21 to secondary
alcohol 9a is a rare example of a practical intramolecular CIDT
under mild conditions. Three major isomers of 9 have been
isolated, and the relative configuration was unambiguously
determined by single-crystal X-ray crystallography. A very
concentrated elimination protocol from mesylate 22a to 1 was
designed, optimizing purity and color to enable an isolation
of 1 by crystallization. The nine steps with three isolated inter-
mediates enable a smooth and rapid scale-up to 1 kg of 1 with a
high throughput.
EXPERIMENTAL SECTION
■
General remarks: 1 vol or 1 wt means 1 L of solvent or 1 kg of
reagent, respectively, with respect to the reference starting
material; er (enantiomeric ratio), dr (diastereomeric ratio); er
and dr reported in this paper have not been validated by
calibration.68 Compounds are characterized by 1H NMR
(400 MHz, Bruker) or 13C NMR (100 MHz, Bruker); internal
standard for quantitative NMR was 1,4-dimethoxybenzene.
Details for the GC−MS, LC−MS, and DSC methods are listed
in the SI. All temperatures are internal temperatures, and yields
are presented as is, unless otherwise stated.
Methyl 2-((R)-1,4-Dioxaspiro[4.5]decan-7-yl)-acetate
(12). A 30-L steel-enameled reactor was charged with
malonate 23 (5.99 kg, 85% w/w, 22 mol) and DMAc (15 L).
Water (0.4 L, 1 equiv) and LiCl (1.872 kg, 2 equiv) were
added. The solution was heated at 120−136 °C during 5 h.
After reaching 120 °C, a precipitate formed, and gas evolution
started without foaming. IPC by GC−MS showed more than
99% conversion. The mixture was cooled to 30 °C and filtered
over a 10-L nutsche, filled with Celite (1.2 kg), and rinsed with
toluene (3.5 L). The filtrate was washed with water (3 × 12 L).
The organic phase was concentrated at 100 °C jacket
temperature and 500−110 mbar to afford 12 as a low-viscosity
oil. Yield: 4.2 kg (89%). Assay-corrected yield: 5.99 kg of 23
Analytical reference samples of rac-24a and rac-24b were
obtained from a racemic run by chromatography on silica gel
with toluene/EtOAc (95:5) as eluent. The more polar rac-24a
(TLC) was further crystallized from TBME.
(R*)-Methyl 2-phenyl-2-((R*)-1,4-dioxaspiro[4.5]-
decan-7-yl)-acetate (rac-24a). Colorless crystalline solid, like
configuration as proven by single-crystal X-ray structure analysis;
mp 87 °C (peak by DSC); TLC: Rf 0.30 (toluene/EtOAc 9:1);
136
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
(0.5 N solution)/starch (1% aqueous solution) was negative.
Water (65 mL) was added and the mixture filtered over Celite
(30 g). The organic phase was extracted with 1/2-sat. NaCl
solution (2 × 60 mL). Total mass of the light-red EtOAc
solution was 134 g. An aliquot was withdrawn and evaporated
to dryness at 40 °C under reduced pressure during 15 min for
purity and yield determination. Estimated yield of 21: 27.4 g
(85%, corr. for NMR assay). Analytical data for aliquot (yellow
oil): NMR assay: 79% w/w; purity (GC−MS): 96% a/a, Rt
3.25, 3.29 min (40:60), [M + 1]+ = 261. 1H NMR (CDCl3,): δ
9.71 (d, J = 3.2 Hz, 0.6H), 9.70 (d, J = 3.3 Hz, 0.4H), 7.29−
7.42 (m, 3H), 7.18−7.24 (m, 2H), 3.77−4.04 (m, 4H),
3.30−3.36 (m, 1H), 2.42−2.57 (m, 1 H), 1.23−0.90 (m, 8H).
(1R,4R,5S,6S)-6-Hydroxy-5-phenylbicyclo[2.2.2]octan-
2-one (9a). Preparation of a bleach solution with pH 8.5−9.5:
Commercial bleach (12−15% aqueous NaOCl solution) was
titrated with the KI/Na2SO3 couple to determine its hypo-
chlorite content: 1.92 N, 12% w/w. This bleach solution (7.2 L,
1.2 equiv) was diluted with aqueous sat. NaHCO3 solution
(2.9 L) to achieve pH 9.3. Alcohol 25 (3 kg, 81% w/w,
11.4 mol) was dissolved in EtOAc (15 L). A solution of KBr
(136.1 g, 0.1 equiv) in water (300 mL) was added at 20 °C,
followed by TEMPO (17.9 g, 0.01 equiv). The freshly prepared
bleach solution was added at 4−8 °C over 45 min (exothermic
reaction). IPC (GC−MS and LC−MS method 1) showed >98%
conversion. Aqueous sat. Na2S2O3 solution (50 mL, 0.02 equiv)
was added at 10−15 °C until the test against KI (0.5 N
solution)/starch (1% aqueous solution) was negative. The
organic phase was washed with water (9 L) and 1/2-sat. NaCl
solution (2 × 4 L). Solvent (9 L) was removed by distillation at
50 °C jacket temperature at 250−150 mbar. The target con-
centration (2 vol with respect to 21) was adjusted by the
addition of EtOAc (2 L) to get a total vol of 9 L. An aliquot
(138.67 g) was withdrawn and evaporated to dryness (29.33 g,
yellow oil) for purity and yield determination. Estimated crude
purity (GC−MS): 97% a/a, Rt 3.43 min, [M + 1]+ = 291; purity
(LC−MS method 1): 100% a/a, Rt 1.74 min, [M + 1]+ = 291; 1H
NMR (CDCl3): δ 7.23−7.41 (m, 5H), 3.93−4.08 (m, 4H), 3.66
(s, 3H), 3.30 (d, J = 10.5 Hz, 1H), 2.33−2.46 (m, 1H), 1.84−1.93
(m, 1H), 1.61−1.80 (m, 2H), 1.26−1.56 (m, 4H), 0.70−0.85 (m,
1H); 13C NMR (CDCl3): δ 173.7, 137.6, 128.6, 128.5, 127.3,
108.9, 64.3, 64.3, 58.1, 51.8, 40.0, 39.0, 34.8, 29.0, 22.7.
(R*)-Methyl 2-phenyl-2-((S*)-1,4-dioxaspiro[4.5]-
decan-7-yl)-acetate (rac-24b). Yellow oil, unlike configu-
ration by deduction with single-crystal X-ray structure analysis
of rac-24a; TLC: Rf 0.33 (toluene/EtOAc 9:1); purity (GC−
MS): 96% a/a, Rt 3.41 min, [M + 1]+ = 291; purity (LC−MS
1
method 1): 95% a/a, Rt 1.67 min, [M + 1]+ = 291. H NMR
(CDCl3): δ 7.23−7.39 (m, 5H), 3.72−3.95 (m, 4H), 3.66 (s,
3H), 3.31 (d, J = 10.5 Hz, 1H), 2.34−2.48 (m, 1H), 1.33−1.86
(m, 6H), 0.97−1.15 (m, 2H); 13C NMR (CDCl3): δ 128.6,
127.3, 64.0, 64.2, 58.2, 51.8, 38.80, 38.6, 34.8, 30.5, 22.8.
2-Phenyl-2-((R)-1,4-dioxaspiro[4.5]decan-7-yl)-etha-
nol (25, Mixture of Diastereoisomers). The reactor was
charged with toluene (5.9 L) and a solution of 2.4 M LiAlH4 in
THF (3.9 L, 0.55 equiv). The feed tank was charged with the
toluene solution (8.85 kg) of 24 (4.914 kg, 82% w/w of aliquot
stripped to dryness, 17.1 mol) and additional toluene (3.9 L) to
adjust to the target concentration of 1.6 vol per kg 24. This
solution was added to the LiAlH4 solution at 5−15 °C over 1 h
(very exothermic reaction). The reaction was stirred at 10−20 °C
for 30 min. IPC (GC−MS) indicated > 99% conversion. A
mixture of water (350 mL, 0.07 vol) and THF (990 mL) was
added at 13−22 °C over 40 min (exothermic reaction with
evolution of H2). 15% NaOH-solution (350 mL, 0.07 vol)
was added at 10−20 °C over 20 min (exothermic reaction).
Water (1.1 L, 0.22 vol) was added at 10−20 °C over 5 min
(exothermic reaction). Charcoal (0.5 kg, granulated, 0.1 wt.) was
added, and the mixture was stirred at 20 °C for 2 h prior to
filtration over Celite (0.6 kg, 0.2 wt.). The filter cake was
washed with toluene (2 L). The filtrate was concentrated at
50−55 °C batch temperature and 500−25 mbar to afford 25 as
a low-viscosity dark oil. Yield: 4.4 kg (98%). Assay-corrected
yield over two steps: 4 kg of 12 (86% w/w) gave 4.4 kg of 25
(81% w/w, 11% w/w toluene): 84% yield; purity (GC−MS):
96% a/a, Rt 3.40, 3.49 min (30:70), [M − 18 + 1]+ = 245.
2-Phenyl-2-((R)-1,4-dioxaspiro[4.5]decan-7-yl)-acetal-
dehyde (21). Preparation of a bleach solution with pH
8.5−9.5: commercial bleach (12−15% aqueous NaOCl
solution) was titrated with the KI/Na2SO3 couple to determine
its hypochlorite content: 1.9 N, 12% w/w. This bleach solution
(65 mL, 1.1 equiv) was diluted with aqueous sat. NaHCO3
solution (26.4 mL) to achieve pH 8.7. Alcohol 25 (28.9 g, 89%
w/w, 0.11 mol) was dissolved in EtOAc (110 mL). A solution
of KBr (0.993 g, 0.08 equiv) in water (2.2 mL) was added at
0 °C, followed by TEMPO (130 mg, 0.008 equiv). The freshly
prepared bleach solution was added at 0−10 °C over 20 min
(exothermic reaction). IPC (GC and LC−MS method 1)
showed >98% conversion. Aqueous sat. Na2S2O3 solution
(0.3 mL) was added at 10−15 °C until the test against KI
1
mass was 3.095 kg of aldehyde 21 with an H NMR assay of
61% w/w (14% w/w EtOAc) corresponding to 1.86 kg 21 (78%
yield, corr. for NMR assay).
An aqueous solution of HCl (32%, 333 mL, 0.3 equiv) was
added at 25 °C to the EtOAc solution of 21 and the mixture
was stirred at 50 °C for 2 h. IPC (LC−MS method 1) showed
> 97% conversion. The mixture was cooled to 10 °C, stirred at
this temp. for 16 h, then further cooled to 0 °C and stirred at
this temp. for 1 h. The suspension was filtered and washed with
EtOAc (2 × 1.5 L). The filter cake was dried by applying
vacuum for 2 h to afford 9a as colorless crystalline solid. Yield:
0.921 kg. Assay-corrected yield: 3 kg of 25 (81% w/w) gave
0.921 kg 9a (98% w/w): 45% yield. Mp 191 °C; relative con-
figuration as proven by single-crystal X-ray structure analysis
of rac-9a; chiral HPLC method: er > 99.5:0.5, diastereomeric
purity: >99.5%; purity (LC−MS method 1): 100% a/a, Rt 1.23
1
min, [M − 18 + 1]+ = 199; H NMR (CDCl3): δ 7.34−7.42
(m, 4 H), 7.27−7.32 (m, 1 H), 4.48 (t, J = 3.7 Hz, 1 H),
2.93−2.97 (m, 1 H), 2.58 (q, J = 3.1 Hz, 1 H), 2.49−2.56 (m,
1 H), 2.35−2.44 (m, 2 H), 1.87−1.95 (m, 3 H), 1.72−1.83 (m,
1 H), 1.42−1.53 (m, 1 H); 13C NMR (CDCl3): δ 215.4, 142.2,
128.6, 127.6, 126.6, 74.4, 52.8, 51.5, 45.6, 34.4, 20.2, 18.2.
(1R,2S,3S,4R)-6-Oxo-3-phenylbicyclo[2.2.2]octan-2-yl
Methanesulfonate (22a). Bicyclic alcohol 9a (25 g, 0.12
mol) was dissolved in DCM (125 mL) followed by Et3N
(24 mL, 1.5 equiv). The suspension was cooled to 0 °C and
MsCl (11.6 mL, 1.3 equiv) was added at 10−20 °C. After 1.5 h,
the mixture was filtered and the filtrate washed with water
137
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
(125 mL) and 1/2-sat. NaCl solution (2 × 125 mL). The
organic phase was dried over Na2SO4 and concentrated to
dryness under reduced pressure to afford 22a as a yellow oil,
which solidified at rt. Yield: 32.5 g (96%). Twenty grams
thereof was dissolved in heptane (350 mL) and EtOAc (350 mL)
at 50 °C and filtered over silica gel (15 g). The filtrate was cooled
to 0 °C and filtered, and the filter cake was washed with heptane
(100 mL) to afford a first crop of 22a as a colorless solid. Yield
first crop: 8.33 g (42% recovery). Additional crystals were filtered
off from the mother liquor to afford a second crop of 22a as a
colorless solid. Yield of second crop: 2.75 g (14% recovery). Mp
87 °C (peak by DSC); purity (LC−MS method 1): 100% a/a, Rt
1.4 min, [M − 96 + 1]+ = 199; 1H NMR (CDCl3): δ 7.38−7.49
(m, 2H), 7.30−7.38 (m, 3H), 5.45 (t, J = 3.8 Hz, 1H), 3.22−3.30
(m, 1H), 2.88−3.00 (m, 4H), 2.54−2.63 (m, 1H), 2.44−2.53
(m, 1H), 2.35−2.42 (m, 1H), 1.96−2.08 (m, 2H), 1.71−1.88 (m,
1H), 1.43−1.60 (m, 1H); 13C NMR (CDCl3): δ 211.0, 140.0,
129.0, 127.3, 82.6, 50.6, 48.6, 45.5, 39.5, 35.4, 20.2, 18.0.
known chirality. X-ray crystal structure analysis of this
intermediate allowed for the determination of the absolute
configuration.
ASSOCIATED CONTENT
■
S
* Supporting Information
Details for the GC−MS, LC−MS, and DSC methods.
Experimental data for the reactions as depicted in Scheme 4.
Tables with screening results of the cyclization and elimination
1
step. H- and 13C NMR data (12, rac-24, 25, 21, 9a, 22a, 1),
GC−MS (12, 24, rac-24a, 25, 21), LC−MS (9a, 22a, 1), chiral
HPLC (9a, mixture of rac-9a, rac-9b, and rac-9c, 1, mixture of 1
and ent-1), DSC (22a), X-ray crystal structure analysis for rac-
24a, rac-9a, rac-9b, rac-9c, and 1. This material is available free
of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
(1R,4R)-5-Phenylbicyclo[2.2.2]oct-5-en-2-one (1). Teles-
coped Process. MsCl (0.56 L, 1.3 equiv) was added at
10−20 °C to a suspension of 9a (1.2 kg, 5.55 mol) in toluene
(6 L) and Et3N (1.15 L, 1.5 equiv) (exothermic reaction). IPC
(LC−MS method 1) showed >99% conversion after 10 min.
The mixture was washed with water (2 × 3 L) and concentrated
to dryness under reduced pressure to afford compound 22a
as a yellow oil, which solidified at rt. Yield of crude 22a: 1.6 kg
(99%). NMR assay: 95% w/w; purity (LC−MS method 1): 98%
a/a, Rt 1.44 min, [M − 96 + 1]+ = 199. 1H NMR data in CDCl3
correspond to the structure. Mesylate 22a (1.6 kg, 5.50 mol)
was dissolved in 2,4,6-collidine (1.2 L) and stirred at 140−
145 °C for 1.5 h. IPC (LC−MS method 1) showed > 99%
conversion. HCl (1 N, 3 L) and heptane (19 L) were added,
and the layers were separated. The organic phase was washed
with 1 N HCl (3 L), then with water (2 × 3 L) and filtered over
Na2SO4 (1.7 kg).69 The cake was washed with heptane (3 L).
Solvent (11.5 L) was removed from the filtrate at 110 °C jacket
temp. under reduced pressure. Seed crystals of 1 (300 mg) were
added at 40 °C, the mixture was stirred at 39−40 °C for 1 h,
cooled to 0 °C within 0.5 h, and stirred at 0 °C for 10 min. The
suspension was filtered and the filter cake washed with heptane
(1 L) to afford 1 as a beige crystalline solid. Yield first crop:
Corresponding Author
*Corresponding author. E-Mail: stefan.abele@actelion.com.
Telephone: +41 61 565 67 59.
ACKNOWLEDGMENTS
■
We thank Luke Harris and Step
́
hanie Combes for their skillful
experimental work, Rene
́
Vogelsanger for the expeditious
development of the chiral HPLC method, Christine Metzger
for the DSC measurements, and Dr. Daniel Bur for the X-ray
structures. We are indebted to Dr. Kurt Hilpert for fruitful
discussions.
REFERENCES
■
(1) (a) From host−guest complexation of rac-bicyclo[2.2.2]oct-7-
ene-2,5-dione and (S)-(−)-10,10′-dihydroxy-9,9′-biphenanthryl, see:
Kinoshita, T; Haga, K; Ikai, K; Takeuchi, K. Tetrahedron Lett. 1990,
31, 4057−4060. (b) Analytical data for racemic 1, see: Ikai, K;
Takeuchi, K.; Kinoshita, T; Haga, K.; Komatsu, K; Okamoto, K. J. Org.
Chem. 1991, 56, 1052−1058. (c) Hilpert, K; Hubler, F; Renneberg, D.
PCT Int. Appl. WO/2008/132679, 2008.
(2) (a) Funel, J.-A.; Schmidt, G.; Abele, S. Org. Process Res. Dev. 2011,
15, 1420−1427. (b) Abele, S.; Stoessel, F. et al. Safety Assessment of
Diels−Alder Reactions with Highly Reactive Reagents. Org. Process Res.
Dev, 2012. Manuscript in preparation.
0.69 kg (63%). Mp 66.3−67.5 °C; [α]26 = +547° (c = 1;
D
CDCl3) ([α]19D = +447° (c = 1; CDCl3));1a NMR assay: 100%
w/w; chiral HPLC method: er > 99.5:0.5, diastereomeric purity:
>99.5%; purity (LC−MS method 1): 100% a/a, Rt 1.60 min,
(3) Holub, N.; Jiang, H.; Paixao, M. W.; Tiberi, C.; Jørgensen, K. A.
̃
Chem.Eur. J. 2010, 16, 4337−4346, and references cited therein.
(4) (a) Hayashi, T.; Ueyama, K.; Tokunaga, N.; Yoshida, K. J. Am.
Chem. Soc. 2003, 125, 11508−11509. (b) For bod* ligands, see:
Tokunaga, N.; Otomaru, Y.; Okamoto, K.; Ueyama, K.; Shintani, R.;
Hayashi, T. J. Am. Chem. Soc. 2004, 126, 13584−13585. (c) Otomaru,
Y.; Okamato, K.; Shintani, R.; Hayashi, T. J. Org. Chem. 2005, 70,
2503−2508.
1
[M + 1]+ = 199. H NMR (CDCl3): δ 7.44−7.49 (m, 2H),
7.36−7.43 (m, 2H), 7.29−7.36 (m, 1H), 6.46 (dd, J1 = 6.7 Hz,
J2 = 2.2 Hz, 1H), 3.53−3.58 (m, 1H), 3.30−3.35 (m, 1H),
2.19−2.23 (m, 2H), 1.97−2.06 (m, 1H), 1.83−1.96 (m, 1H),
1.64−1.80 (m, 2H); 1H NMR (CD3OD): δ 7.48−7.55 (m, 2H),
7.34−7.43 (m, 2H), 7.22−7.33 (m, 1H), 6.49 (d, J = 6.4 Hz,
1H), 3.55−3.62 (m, 1H), 3.22−3.30 (m, 1H), 2.09−2.30 (m,
2H), 1.88−2.04 (m, 2H), 1.59−1.83 (m, 2H). 13C NMR
(CDCl3): δ 212.4, 148.0, 137.6, 128.7, 127.8, 124.9, 122.1, 49.2,
40.4, 35.4, 24.6, 23.2.
(5) Fischer, C.; Defieber, C.; Suzuki, T.; Carreira, E. M. J. Am. Chem.
Soc. 2004, 126, 1628−1629.
(6) Abele, S.; Funel, J.-A. Patent application filed, 20 October 2010.
Preparation of bicyclo[2.2.2]octane-2-one compounds.
(7) Reported for the precursor lacking the ketal moiety, see: La Bella,
A.; Popolla, I.; Felici, M.; Leonelli, F.; Ceccacci, F.; Filocamo, L.;
Migneco, L. M.; Bettolo, R. M. ARKIVOC 2008, iii, 80−90.
(8) Published after the first 100 g batches of 1 had been produced by
our present method according to Scheme 9: Bella, M.; Schietroma,
D. M. S.; Cusella, P. P.; Gasperi, T. Chem. Commun. 2009, 597−599.
(9) Xu, Y.; Ohori, K.; Ohshima, T.; Shibasaki, M. Tetrahedron 2002,
58, 2585−2588.
The mother liquor was evaporated to dryness to get a red,
oily residue (310 g). This was dissolved in heptane (1 L) at
50 °C, cooled down to 20 °C, and filtered to get a second crop
of 1 as a light-brown crystalline solid. Yield: 0.186 kg (17%).
NMR-assay: 94% w/w. Corrected yield: 16%. Chiral HPLC
method: er > 99.5:0.5; purity (LC−MS method 1): 100% a/a.
The structure was proven by single-crystal X-ray analysis. 1
was processed to a crystalline intermediate with a residue of
(10) Enantioselective Michael addition of 2-phenylcyanoacetate to
vinyl ketones, see: Jautze, S.; Peters, R. Angew. Chem., Int. Ed. 2008,
47, 1−6.
138
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
(11) Duursma, A.; Boiteau, J. G.; Lefort, L.; Boogers, J. A. F.;
de Vries, A. H. M.; de Vries, J. G.; Minnaard, A. J.; Feringa., B. L.
J. Org. Chem. 2004, 69, 8045−8052.
(12) (a) Krasovsky, A.; Straub, B. F.; Knochel, P. Angew. Chem., Int.
Ed. 2006, 45, 159−162. (b) Nakao, Y.; Chen, J.; Imamaka, H.;
Hiyama, T.; Ichikawa, Y.; Duan, W.-L.; Shintani, R.; Hayashi, T. J. Am.
Chem. Soc. 2007, 129, 9137−9143.
(13) Obtained by β-elimination of benzenesulfenic acid from the
intermediate α-phenylsulfoxide, see: Resek, J. E.; Meyers, A. I.
Tetrahedron Lett. 1995, 36, 7051−7054.
(28) We recovered ∼90% of the expected LiHCO3 salt on the filter.
(29) Delhaye, L.; Ceccato, A.; Jacobs, P.; Kottgen, C.; Merschaert, A.
̈
Org. Process Res. Dev. 2007, 11, 160−164.
(30) (a) This step can be telescoped into the next step as toluene is
the solvent for both steps. At this scale 12 was isolated in the reactor as
oil to get an intermediate yield. (b) In other runs, the yield was
increased by 5% by a back-extraction of the aqueous phase with
toluene, followed by two water washes.
(31) LiCl, water, DMSO, 140 °C, 17 h, followed by aqueous workup
in EtOAc and flash chromatography, see ref 25.
(14) (a) Allylic oxidation of 1,4-dioxaspiro[4.5]dec-6-ene with tert-
butylhydroperoxide and catalytic Pd(OH)2, see: Yu, J.-Q.; Corey, E. J.
J. Am. Chem. Soc. 2003, 125, 3232−3233. (b) Large-scale example
with RuCl3, see: Harre, M.; Haufe, R.; Nickisch, K.; Weinig, P.;
Weinmann, H.; Kinney, W. A.; Zhang, X. Org. Process Res. Dev. 1998,
2, 100−104. (b) Allylic oxidation catalyzed by Mn(OAc)3, see: Shing,
T. K. M.; Yeung, Y.-Y.; Su, P. L. Org. Lett. 2006, 8, 3149−3151.
(c) Saegusa oxidation of the TMS-enol ether derived from cyclo-
hexanedione monoethylene ketal with Pd(OAc)2, see: Kerr, W. J.;
McLaughlin, M.; Morrison, A. J.; Pauson, P. L. Org. Lett. 2001, 3,
2945−2948.
(32) (a) Busacca, C. A.; Cerreta, M.; Dong, Y.; Eriksson, M. C.;
Farina, V.; Feng, X.; Kim, J.-Y.; Lorenz, J. C.; Sarvestani, M.; Simpson,
R.; Varsolona, R.; Vitous, J.; Campbell, S. J.; Davis, M. S.; Jones, P.-J.;
Norwood, D.; Qiu, F.; Beaulieu, P. L.; Duceppe, J.-S.; Hache, B.;
́
Brong, J.; Chiu, F.-T.; Curtis, T.; Kelley, J.; Lo, Y. S.; Powner, T. H.
Org. Process Res. Dev. 2008, 12, 603−613. (b) Jiang, X.; Gong, B.;
̂
Prasad, K.; Repic, O. Org. Process Res. Dev. 2008, 12, 1164−1169.
(33) (a) Review: Johansson, C. C. C.; Colacot, T. J. Angew. Chem.,
Int. Ed. 2010, 49, 676−707. (b) Review: Culkin, D. A.; Hartwig, J. F.
Acc. Chem. Res. 2003, 36, 234−245. (c) Lee, S.; Beare, N. A.; Hartwig,
J. F. J. Am. Chem. Soc. 2001, 123, 8410−8411. (d) Jørgensen, M.; Lee,
A.; Liu, X.; Wolkowski, J. P.; Hartwig, J. F. J. Am. Chem. Soc. 2002, 124,
12557−12565. (e) Moradi, W. A.; Buchwald, S. L. J. Am. Chem. Soc.
2001, 123, 7996−8002.
(15) Cyclohexenone as Michael acceptor, see: Widegren, M.; Dietz,
M.; Friberg, A.; Frejd, T.; Hahn-Hagerdahl, B.; Gorwa-Grauslund,
̈
M. F.; Katz, M. Synthesis 2006, 3527−3530.
(16) Jiricek, J.; Blechert, S. J. Am. Chem. Soc. 2004, 126, 3534−3538.
(17) Martinelli, M.; Moher, E. D. PCT Int. Appl. WO/1998/09942,
1998. Due to the toxicity of tin organyls, this is not appropriate for
larger scales.
(34) Hama, T.; Hartwig, J. F. Org. Lett. 2008, 10, 1549−1552.
(35) Remaining traces of bromobenzene were removed in the
subsequent step where residual Pd catalyzed the debromination to
benzene with LiAlH4.
(36) A second citric acid wash (pH 3) led to further black precipitate
with coating of the glass. The remaining traces of Pd were removed at
the next stage.
(18) (a) Parikh, J. R.; Doering, W. v. E. J. Am. Chem. Soc. 1967, 89,
5505−5507. (b) Ripin, D. H. B.; Abele, S.; Cai, W.; Blumenkopf, T.;
Casavant, J. M.; Doty, J. L.; Flanagan, M.; Koecher, C.; Laue, K. W.;
McCarthy, K.; Meltz, C.; Munchhoff, M.; Pouwer, K.; Shah, B.; Sun, J.;
Teixeira, J.; Vries, T.; Whipple, D. A.; Wilcox, G. Org. Process Res. Dev.
2003, 7, 115−120.
́
(37) Perez-Balado, C.; Willemsens, A.; Ormerod, D.; Aelterman, W.;
Mertens, N. Org. Process Res. Dev. 2007, 11, 237−240.
(38) Alternatively, 24 was distilled by short-path distillation (VTA,
VKL 70-4) at jacket temp. 175 °C, condenser temp. 55 °C, dropping
funnel temp. 55 °C, pressure ∼0.01 mbar with concomitant
improvement of the assay from 75 to 90% w/w and color removal.
(39) (a) Chemetall brochure, Organolithium compounds , http://
www.chemetalllithium.com/fileadmin/files_chemetall/Downloads/
DL_Organolithium.pdf (accessed Feb 17, 2011). (b) The relative
volume of the filter cake was 1.2 vol.
(19) Not available in the usual chemical catalogues.
(20) Phosphoric acid, 100 °C, chromatography, 35%, see: Mori, K.;
Nakahara, Y.; Matsui, M. Tetrahedron 1972, 28, 3217−3226.
(21) (a) Racemate: 2 N HCl, THF, reflux, 24 h, chromatography,
61% rac-19a, see: De Santis, B.; Iamiceli, A. L.; Bettolo, R. M.;
Migneco, L. M.; Scarpelli, R.; Cercichelli, G.; Fabrizi, G.; Lamba, D.
Helv. Chim. Acta 1998, 81, 2375−2387. (b) Discussion of the endo-
exo equilibrium of 19a and 19b, see: Di Stefano, S.; Leonello, F.;
Garofalo, B.; Mandolini, L.; Bettolo, R. M.; Migneco, L. M. Org. Lett.
2002, 4, 2783−2785.
(22) (a) Racemate: phosphoric acid, THF, reflux, 4 h, chromatography,
49% rac-19a, see: Schmoldt, P.; Mattay, J. Synthesis 2003, 1071−1078.
(b) Chirally pure compound: phosphoric acid, THF, reflux, 4 h,
chromatography, 38% 19a, see: Tzvetkov, N.; Schmoldt, P.; Neumann,
B.; Stammler, H.-G.; Mattay, J. Tetrahedron: Asymmetry 2006, 17, 993−
998.
(23) Likewise, as part of terpenoids to form polycyclic compounds:
(a) NaOH, MeOH, 70%, see: Bravetti, D.; Bettolo, R. M.; Lupi, A.
Helv. Chim. Acta 1982, 65, 371−376. (b) KOH, MeOH, 91%, see:
Nagata, W.; Narisada, M.; Sugasawa, T.; Wakabayashi, T. Chem.
Pharm. Bull. 1968, 16, 885−896.
(24) Strzalko, T.; Seyden-Penne, J.; Wartski, L.; Corset, J.; Castella-
Ventura, M.; Froment, F. J. Org. Chem. 1998, 63, 3295−3301.
(25) Ohshima, T.; Xu, Y.; Takita, R.; Shibasaki, M. Tetrahedron 2004,
60, 9569−9588.
(40) Anelli, P. L.; Biffi, C.; Montanari, F.; Quici, S. J. Org. Chem.
1987, 52, 2559−2562.
(41) Obtained from cyclohexenone and dimethyl malonate according
to ref 15, followed by ketalization with ethylene glycol and elaboration
as indicated in Scheme 5 herein.
(42) (a) Ciriminna, R.; Paliaro, M. Org. Process Res. Dev. 2010, 14,
245−251. (b) Fritz-Langhals, E. Org. Process Res. Dev. 2005, 9,
577−582.
(43) On scales ranging from 1 g to 3 kg, the yield determined with an
aliquot of this solution, was between 73 and 85%, corrected for NMR
assay of starting material and product.
(44) Stress test with a sample of 93% w/w assay and 100% a/a by
LC−MS showed that ketal aldehyde 21 was stable at rt neat or in THF
solution for 12 days, the stability clearly depending on the purity.
(45) (a) Reference 23a. (b) Kelly, R. B.; Harley, M. L.; Alward, S. J.;
Rej, R. N.; Gowda, G.; Mukhopadhyay, A.; Manchand, P. S. Can. J.
Chem. 1983, 61, 269−275.
(46) Details in SI for (a) rac-9a, (b) rac-9b, and (c) rac-9c.
(47) (a) Brands, K. M. J.; Davies, A. J. Chem. Rev. 2006, 106, 2711−
2733. (b) Anderson, N. G. Org. Process Res. Dev. 2005, 9, 800−813.
(c) This reaction was termed an intramolecular reactive crystallization
as well; see: Tung, H.-H.; Paul, E. L.; Midler, P. M.; McCauley, J. A.
Crystallization of Organic Compounds: An Industrial Perspective; John
Wiley & Sons, Inc.: Hoboken, NJ, 2009.
(48) This means that the reduction−oxidation sequence from 24 to
21 did not alter the epimeric ratio as proven for 24 by an X-ray
structure analysis.
(26) In one batch in our 30-L steel-enameled reactor we produced
7 kg of 23 within less than three 1-shift days in 85% assay-corrected yield
over two chemical steps, purity (GC−MS): 96.3% a/a, er > 99.5:0.5. The
1
assay was 94% w/w, containing 9% w/w residual toluene, by H NMR
with 1,4-dimethoxybenzene as internal standard. No recrystallization was
required.
(27) (a) Residual toluene (∼9% w/w) stemming from malonate 23
was distilled off. (b) Krapcho decarboxylation in DMSO, see: Krapcho,
A. P.; Weimaster, J. F.; Eldridge, J. M.; Jahngen, E. G. E. Jr.; Lovey,
A. J.; Stephens, W. P. J. Org. Chem. 1978, 43, 138−147.
139
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140

Organic Process Research & Development
Article
(49) Via the open ketone-aldehyde 10, see des-phenyl example 18:
(a) Reference 21. (b) Coons, S.; Javanmard, S.; Collard, D. M.; Kuhar,
M. J.; Schweri, M. M.; Deutsch, H. M. Med. Chem. Res. 2002, 11,
24−39.
(50) As representative sampling of the suspension was difficult,
samples were taken with strong stirring, stripped to dryness, and
dissolved in the solvent for chiral HPLC. Consistent results were
obtained with multiple sampling or by dissolving the total reaction
volume in the HPLC solvent.
monoketals of bicyclo[2.2.2]oct-7-ene-2,5-dione, see: Hill, R. K.;
Morton, G. H.; Peterson, J. R.; Walsh, J. A.; Paquette, L. A. J. Org.
Chem. 1985, 50, 5528−5533.
(67) The chiral intermediate bicyclo[2.2.2]octane-2,5-dione (2.3%
yield) was obtained by fractional crystallization of the (R)-5-(1-
phenylethyl)semioxamazide, see: Otomura, Y.; Okamoto, K.; Hayashi, T.
J. Org. Chem. 2005, 70, 2503−2508.
(68) Wernerova, M.; Hudlicky, T. Synlett 2010, 2701−2707.
(69) Na2SO4 can be replaced by Celite. This polish filtration removes
some residual black particles.
(51) Material with 100% a/a purity by GC or LC−MS could still
show a baseline spot on TLC.
(52) As the selectivity seemed to be driven partially by the solubility
differences, the results obtained by racemic compounds were not
representative for the enantiomerically pure compounds, although we
expected a similar distribution due to the same trend of melting points,
see Scheme 7.
(53) We did not quantify the amount of HOAc which might be
formed by hydrolysis of EtOAc at 50 °C.
(54) A rework of the mother liquor (9a/9b 30:70, 60% a/a chiral
HPLC) by extraction with NaHCO3-solution and crystallization from
either TBME or EtOAc gave an additional 3% yield.
(55) The assay-corrected yield for the cyclization was 60%.
(56) Miller, M. J. J. Org. Chem. 1980, 45, 3131−3132.
(57) Methyl-N-(triethylammoniumsulfonyl)carbamate, see: Atkins,
G. M.; Burgess, E. M. J. Am. Chem. Soc. 1968, 90, 4744−4745.
(58) A hyperconjugative interaction was postulated between the
n(CO), the σ(C1)−(C2), and the forming empty p orital, see:
Carrupt, P.-A.; Vogel, P. Helv. Chim. Acta 1989, 72, 1008−1028.
(59) Skeletal rearrangements of bicyclo[2.2.2]octyl cation, see:
Walborsky, H. M.; Baum, M. E.; Youssef, A. A. J. Am. Chem. Soc.
1961, 83, 988−993.
(60) Mesylates 22a, 22b, and 22c showing their suitability for a syn-
elimination.
(61) VKL 70-4 SKR (VTA), at jacket temp. 137 °C, condenser
25 °C, dropping funnel temp. 60 °C, pressure ∼0.002 mbar.
(62) Solubilities of 1: toluene: 66−100 g/L, heptane: 40−50 g/L,
EtOAc and iPrOAc: > 200 g/L.
(63) The mesylate 22a was obtained as oil which solifidied at rt. The
two steps could be telescoped for larger scales. Toluene must be
removed completely as a run with 30% v/v toluene made removal of
black particles more difficult.
(64) In view of a nearly quantitative yield (for the sum of the three
isomers of 9) of the cyclization, in a separate experiment starting from
ketal aldehyde 21, the crude product 9 (a mixture of oil and waxy
solid) was processed as is to 1: the assay-corrected yield for the three
steps with crude 9 and crystallized 9a was 37% and 51%, respectively.
In addition, the material from the crude run was colored and could not
be purified with one crystallization. Hence, the isolation of a single
isomer 9a prior to the last steps is favored.
(65) Pd and ligand were the major cost drivers. The same yield was
obtained with 0.2 mol % Pd(OAc)2 and Pd2(dba)3 on 5-g scale.
(66) (a) Reference 1a. (b) Alternatively, chiral bicyclo[2.2.2]octane-
2,5-dione, an intermediate of Takeuchi’s synthesis was obtained in
∼0.45% yield by chromatographic separation of tartrate-derived
140
dx.doi.org/10.1021/op200305y | Org. ProcessRes. Dev. 2012, 16, 129−140