Routes for the synthesis of (2S)-2-methyltetrahydropyran-4-one from simple optically pure building blocks

Article abstract of DOI:10.1021/op900163a

Routes to (2S)-2-methyltetrahydropyran-4-one of high optical purity starting from readily available chiral pool precursors and suitable for large-scale manufacture are described. In one approach, the key step is cyclisation of (S)-5-hydroxyhex-1-en-3-one, derived either from an alkyl (S)-3-hydroxybutyrate or (S)-propylene oxide. Formation of the tetrahydropyran ring directly via an intramolecular oxy-Michael reaction under acid-catalysed conditions resulted in loss of optical purity, whereas proceeding through the intermediate (2S)-2-methyl-2,3-dihydropyran-4-one, via an oxidative Pd-catalysed ring closure, followed by hydrogenation of the alkenyl bond, preserved the optical purity. An alternative approach to (2S)-2-methyl-2,3-dihydropyran-4-one is also reported, again starting from an alkyl (S)-3-hydroxybutyrate by elaboration to a carbonyl-protected (6S)-6-methyl-5,6-dihydropyran-2,4-dione derivative, followed by partial reduction and dehydration. Alternatively, the carbonyl group can be reduced out completely in one step to furnish (2S)-2-methyltetrahydropyran-4-one directly after deprotection.

Full text of DOI:10.1021/op900163a

Organic Process Research & Development 2010, 14, 58–71
Routes for the Synthesis of (2S)-2-Methyltetrahydropyran-4-one from Simple
Optically Pure Building Blocks
Kevin R. Anderson,† Ste´phanie L. G. Atkinson,† Takahiro Fujiwara,‡ Melvyn E. Giles,† Takaji Matsumoto,‡ Eric Merifield,*,†
John T. Singleton,† Takao Saito,‡ Tsukasa Sotoguchi,‡ James A. Tornos,†,§ and Edward L. Way†
AstraZeneca R&D Charnwood, Process Research and DeVelopment, Bakewell Road, Loughborough,
Leicestershire LE11 5RH, United Kingdom, and Takasago International Corporation, Corporate Research and DeVelopment
DiVision, 4-11, 1-Chome, Nishi-Yawata, Hiratsuka City, Kanagawa, Japan
Scheme 1. Retrosynthesis of AZD4407
Abstract:
Routes to (2S)-2-methyltetrahydropyran-4-one of high optical
purity starting from readily available chiral pool precursors and
suitable for large-scale manufacture are described. In one ap-
proach, the key step is cyclisation of (S)-5-hydroxyhex-1-en-3-one,
derived either from an alkyl (S)-3-hydroxybutyrate or (S)-propy-
lene oxide. Formation of the tetrahydropyran ring directly via an
intramolecular oxy-Michael reaction under acid-catalysed condi-
tions resulted in loss of optical purity, whereas proceeding through
the intermediate (2S)-2-methyl-2,3-dihydropyran-4-one, via an
oxidative Pd-catalysed ring closure, followed by hydrogenation of
the alkenyl bond, preserved the optical purity. An alternative
approach to (2S)-2-methyl-2,3-dihydropyran-4-one is also reported,
again starting from an alkyl (S)-3-hydroxybutyrate by elaboration
to a carbonyl-protected (6S)-6-methyl-5,6-dihydropyran-2,4-dione
derivative, followed by partial reduction and dehydration. Alter-
natively, the carbonyl group can be reduced out completely in one
step to furnish (2S)-2-methyltetrahydropyran-4-one directly after
deprotection.
needed an efficient, robust, low-cost synthesis of pyranone 4
of high optical purity (>99%). This paper describes the
identification, development, and scale-up of some routes to this
intermediate, starting from small chiral molecules.
The initial discovery synthesis of pyranone 4 has been
published;3 however, this was not envisaged as a potential route
for production of kilogram quantities due to its length, starting
material availability, and use of some undesirable reagents for
large-scale manufacture. Synthesis of pyranone 4 has also
previously been reported Via reductive resolution of the racemic
mixture by a horse liver alcohol dehydrogenase;4 however, the
yield was low and the enantiomeric excess only 85%. In view
of the requirement of a scaleable synthesis to supply pyranone
4 with an optical purity in excess of 99% as the (S)-enantiomer,
an alternative resolution approach was devised, Scheme 2.5 It
was found that a racemic mixture, containing predominantly
pyranols 6a and 6c, generated from a Prins reaction between
acetaldehyde and 3-buten-1-ol 5,6 could be resolved by selective
esterification of the unwanted (2R,4S)-isomer using Novozyme
435 and vinyl butyrate. Separation of compounds 6a and 8 by
an aqueous/organic partition was facilitated by the greater
aqueous solubility of the pyranol. The desired enantiomer 6a
was then oxidised under Jones conditions to provide pyranone
4. This route was used to manufacture around 100 kg of this
intermediate to supply the initial needs of the project, and whilst
this strategy was very attractive for a large-scale manufacturing
route, it had some drawbacks. The undesired trans-pyranol
isomers 6b and 6d formed in small amounts in the Prins
reaction, which are left unreacted in the enzymatic resolution
step, and would progress downstream with the major product
6a, thus approximately doubling in level. One of these [the
(2R,4R)-isomer] oxidised to (2R)-2-methyltetrahydropyran-4-
Introduction
Leukotrienes are a family of important biological molecules
which display powerful spasmogenic action in vascular and
bronchial tissue. 5-Lipoxygenase is the first enzyme in the
biosynthesis of these leukotrienes from arachidonic acid. Drugs
which block this biosynthetic pathway by inhibition of 5-li-
poxygenase may have therapeutic potential in a variety of
inflammatory conditions including inflammatory bowel disease,
asthma, chronic obstructive pulmonary disease (COPD), and
psoriasis. A compound which displays such activity is AZD4407
(1),1 which possesses three distinct moieties, an N-methylox-
indole bridged by a thioether linkage to a 2,4-disubstituted
thiophene and a chiral 2-methyltetrahydropyran-4-ol. Discon-
nection of 1 conveniently gives rise to the thienyl Grignard
reagents 2 or 3 and (2S)-2-methyltetrahydropyran-4-one 4
(Scheme 1).2 To support the development of AZD4407 we
* Author to whom correspondence may be sent. E-mail:
eric.merifield@astrazeneca.com.
† AstraZeneca.
(3) Crawley, G. C.; Briggs, M. T.; Dowell, R. I.; Edwards, P. N.; Hamilton,
P. M.; Kingston, J. F.; Oldham, K.; Waterson, D.; Whalley, D. P.
J. Med. Chem. 1993, 36, 295. Crawley, G. C.; Briggs, M. T. J. Org.
Chem 1995, 60, 4264.
‡ Takasago.
§ Current address: Pfizer Global R&D, Ramsgate Road, Sandwich, UK, CT13
9NJ.
(1) Bird, T. G. C.; Ple, P.; Crawley, G. C.; Large, M. S. Preparation of
4-aryl-4-hydroxy-tetrahydropyrans and 3-aryl-3-hydroxy-tetrahydro-
furans as 5-lipoxygenase inhibitors. EP 623614 A1, 1994.
(2) Alcaraz, M.-L.; et al. Org. Process Res. DeV. 2005, 9, 555.
(4) Haslegrave, J. A.; Jones, J. B. J. Am. Chem. Soc. 1982, 104, 4666.
(5) Holt, R. A.; Rigby, S. R.; Waterson, D. Production of optically active
2-substituted tetrahydropyran-4-ones. WO 9719185 A1, 1997.
(6) Hanschke, E. Chem. Ber. 1955, 88, 1053.
58
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development
10.1021/op900163a  2010 American Chemical Society
Published on Web 12/03/2009

Scheme 2. Synthesis of pyranone 4 Wia enzymatic resolution
Scheme 3. Hetero Diels-Alder approach to a pyranone
precursor
Scheme 4. Strategies for the synthesis of pyranone 4 from
small chiral molecules
one, with a consequent reduction of the optical purity of
pyranone 4. To overcome this and minimise another trouble-
some impurity,7 the crude pyranol mixture 6a-d was converted
to the crystalline p-nitrobenzoate ester 7, which underwent
several recrystallisations, after which the mixture of pyranols
6a and 6c was regenerated by hydrolysis with much higher
purity. As well as increasing the number of steps, this additional
processing regime led to an overall yield of less than 10% for
this approach. A Jones oxidation was also found to be the most
efficient oxidation method (at the time this work was carried
out), and this was deemed unacceptable on environmental
grounds for future production campaigns. These factors prompted
us to focus our efforts on finding an alternative, more efficient
synthesis, which would ultimately become a manufacturing
route for this intermediate, in preference to development and
optimisation of the chemistry shown in Scheme 2.
One alternative approach to a precursor 10, namely the hetero
Diels-Alder reaction between Danishefsky’s diene 9 and
acetaldehyde using a chiral titanium BINOL catalyst, has been
published (Scheme 3).8 This approach afforded dihydropyranone
10 in up to 80% yield and with an enantiomeric excess of up
to 93%. A more recent publication9 has reported the low-
temperature hetero Diels-Alder reaction between a 1-amino-
3-silyloxybutadiene 11 and acetaldehyde using a BAMOL
catalyst system to provide dihydropyranone 10 in 75% yield
and 97% enantiomeric excess, Scheme 3. Although the absolute
stereochemistry of the product was not determined, in principle
either enantiomer of the dihydropyranone can be obtained by
selection of the appropriate enantiomer of the catalyst. The
utility of the hetero Diels-Alder reaction in general for the
synthesis of the dihydropyranone ring has been amply demon-
strated over recent years by the optimisation of various catalyst
systems and application to the synthesis of many related
products.10
To satisfy ongoing and future demand for this project, there
was a clear need for a short route to pyranone 4 with good
stereocontrol so that material of high optical purity could be
obtained without recourse to time-consuming purification
protocols. Consequently, strategies involving resolution, enan-
tioselective synthesis, and synthesis from readily available chiral
pool building blocks were all considered during the course of
this research programme. This paper describes the results of
our endeavours starting from two types of readily available small
chiral molecules incorporating the desired stereocentre and
having suitable functionality for further elaboration to pyranone
4, namely esters of (S)-3-hydroxybutyric acid 12 and (S)-
propylene oxide 14, Scheme 4. These chiral pool building
blocks gave us a variety of options for building up pyranone 4
(7) The synthesis of pyranol 6 by an intramolecular Prins cyclisation
between 3-buten-1-ol and acetaldehyde is reversible and can go back
either to the same starting materials or to formaldehyde and 4-penten-
2-ol. The latter will undergo a Prins reaction with acetaldehyde to
generate 2,6-dimethylpyran-4-ol which remains essentially unreacted
in the resolution step.
(8) Mitsuda, M.; Hasegawa, J. Stereoselective preparation of 2-hydro-
carbyl-2,3-dihydro-4H-pyran-4-ones. GB 2304339 A1, 1997.
(9) Unni, A. K.; Takenaka, N.; Yamamoto, H.; Rawal, V. H. J. Am. Chem.
Soc. 2005, 127, 1336.
(10) See for example: Chaladaj, W.; Kwiatkowski, P.; Jurczak, J. Tetra-
hedron Lett. 2008, 49, 6810. Du, H.; Zhang, X.; Wang, Z.; Bao, H.;
You, T; Ding, K. Eur. J. Org. Chem. 2008, 2248. Wang, Y.; Wolf, J.;
Zavalij, P.; Doyle, M. P. Angew. Chem., Int. Ed. 2008, 47, 1439. Seki,
K.; Ueno, M.; Kobayashi, S. Org. Biomol. Chem. 2007, 5, 1347.
Berkessel, A.; Vogl, N. Eur. J. Org. Chem. 2006, 5029. Du, H.; Zhang,
X.; Wang, Z.; Ding, K. Tetrahedron 2005, 61, 9465. Anada, M.;
Washio, T.; Shimada, N.; Kitagaki, S.; Nakajima, M.; Shiro, M.;
Hashimoto, S. Angew. Chem., Int. Ed. 2004, 43, 2665.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
59

Scheme 5. Small-scale synthesis of lactone 17
Scheme 6. Options for reduction of lactone 17
by incorporation of a C2 or C3 unit by simple functional group
manipulation.
Results and Discussion
Lactone Reduction Route. Esters of (S)-hydroxybutyric
acid are readily available materials which can be prepared by
hydrogenation of the corresponding acetoacetates in good yield
and high optical purity using transition metal catalysts with
chiral ligands.11 One strategy utilising such a chiral molecule
as the starting point for synthesis of pyranone 4 was to install
the required additional two-carbon unit by incorporation of an
acetate equivalent, Scheme 5. The resulting ꢀ-keto ester 16
would then undergo cyclisation together with selective protec-
tion of the more electrophilic ketonic carbonyl group to form
the protected pyran-2,4-dione 17, which would then be subject
to reduction and deprotection steps to generate the desired
pyranone 4. Whilst methods of preparing the unprotected lactone
13 (Scheme 4) are known in the literature, either by enanti-
oselective hydrogenation12 or Via biotransformations13 we chose
to evaluate the chemistry given in Scheme 5 due to requirements
for high optical purity and high throughput. On a small lab
scale,14 a short route to the reduction substrate was demonstrated
Via a low-temperature Claisen condensation between tert-butyl
acetate and ethyl (S)-3-hydroxybutyrate 15 according to the
reported conditions15,16 to furnish ꢀ-keto ester 16 in 86% yield.
Cyclisation to the lactone16 with concomitant ketalisation was
achieved in good yield, utilising ethylene glycol with TsOH as
catalyst to generate lactone 17. Despite the requirement for
cryogenic reaction conditions, this provided the basis of an
efficient route to the key reduction substrate.
The various options for reduction of lactone 17 are given in
Scheme 6. Use of lithium aluminium hydride unfortunately led
to over-reduction yielding diol 18 in 79% yield. Whilst this
could be converted to pyran 20 in 59% overall yield, by a
selective monomesylation followed by treatment with base, the
optical purity of the product dropped slightly to 98%. This was
attributed to some mesylation at the secondary hydroxyl centre
with resultant inversion of stereochemistry upon SN2 ring
closure. Similarly, reduction using sodium borohydride in THF
under reflux gave mainly diol 18.
Clearly, reduction of the lactone carbonyl group without ring-
opening was highly desirable to maintain the chiral integrity.
Reduction using one equivalent of DIBALH at low temperature
afforded lactol 21 as a mixture of epimers at the newly formed
hydroxyl centre. Without purification, this underwent acid-
catalysed deprotection to regenerate the carbonyl group, with
concomitant dehydration to furnish dihydropyranone 10, having
an identical optical purity to that of the starting material 15.
Conversion of this intermediate through to pyranone 4 by
hydrogenation was demonstrated on the same substrate prepared
Via an alternative route and with no loss in optical purity (Vide
infra).
In order to shorten the route, direct reduction of lactone 17
to pyran 20 was sought. Many reagent systems have been
identified for achieving such transformations in pyran and furan
synthesis and we opted for sodium borohydride in the presence
of acid. In the presence of sulphuric acid, deketalisation occurred
resulting in low selectivity for the lactone carbonyl group,
whereas boron trifluoride etherate allowed much better selectiv-
ity, giving no deprotection of the ketal, although the yield of
isolated purified pyran 20 was low (20% following distillation
or 35% if purified by chromatography on a 120 g scale).
Deketalisation of 20 was then carried out under mild conditions
due to the sensitivity of pyranone 4 to acid (Vide infra), for
which PPTS proved most amenable. In the presence of acetone
a slow conversion at elevated temperature was observed;
however, replacing acetone by formaldehyde allowed much
faster reaction, albeit still at elevated temperature. Pyranone 4
(11) A Ru-BINAP catalyst can be used for the preparation of methyl (S)-
hydroxybutyrate 36 from methyl acetoacetate; see: Kitamura, M.;
Tokunaga, M.; Ohkuma, T.; Noyori, R. Organic Syntheses; Wiley &
Sons: New York, 1998; Collect. Vol. 9, p 589. SEGPHOS, however,
was used as ligand in place of BINAP since it gives better results for
reduction of ꢀ-keto esters: Saito, T.; Yokozawa, T.; Ishizaki, T.; Moroi,
T.; Sayo, N.; Miura, T.; Kumobayashi, H. AdV. Synth. Catal. 2001,
343, 264. Shimizu, H.; Nagasaki, I.; Saito, T. Tetrahedron 2005, 61,
5405.
(12) Huck, W.-R.; Burgi, T.; Mallat, T.; Baiker, A. J. Catal. 2001, 200,
171. Fehr, M. J.; Consiglio, G.; Scalone, M.; Schmid, R. J. Org. Chem.
1999, 64, 5768.
(13) Sato, M.; Sakaki, J.; Sugita, Y.; Nakano, T.; Kaneko, C. Tetrahedron
Lett. 1990, 31, 7463. Tamarez, M. M.; Franck, R. W.; Geer, A.
Tetrahedron 2003, 59, 4249.
(14) On a kilogram scale, lactone 17 was prepared from methyl-(S)-
hydroxybutyrate by conversion to methyl-(S)-3-oxo-5-(tetrahydropy-
ran-2-yloxy)hexanoate using methods developed from a published
synthesis. Liu, L.; Tanke, R. S.; Miller, M. J. J. Org. Chem. 1986,
51, 5332. followed by acid-catalysed cyclisation to (S)-6-methyl-5,6-
dihydropyran-2,4-dione and subsequent carbonyl protection.
(15) Deschenaux, P.-F.; Kallimpoulos, T.; Stoeckli-Evans, H.; Jacot-
Guillarmod, A. HelV. Chim. Acta 1989, 72, 731.
(16) Drochner, D.; Mu¨ller, M. Eur. J. Org. Chem. 2001, 211.
60
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Scheme 7. Hydrazone route to pyranone 4
Scheme 8. Synthesis of pyranones by an oxy-Michael
reaction
constituted a short synthesis of pyranone 4 of high optical purity
and once again demonstrated the utility of dihydropyranone 10
as a stable intermediate, which could be converted to the target
pyranone 4 under mild conditions without loss of optical purity.
Intramolecular Oxy-Michael Approach. It is known19 that
R,ꢀ-unsaturated ketone 26 can be cyclised to 2-methyltetrahy-
dropyran-4-one 27 by treatment with Amberlyst 15 resin in
chloroform. In a similar manner, we expected that the R,ꢀ-
unsaturated ketone 28, incorporating the required stereocentre,
would provide a route to the optically pure product 4, Scheme
8. A detailed study20 of reagents capable of effecting such a
transformation on related, more substituted substrates has
appeared after this work was carried out. Furthermore, a paper21
has also recently been published describing an antibody-
mediated kinetic resolution of hydroxyenones and cyclisation
of the enantioenriched products to the tetrahydropyranones in
the presence of TMSOTf and Hu¨nig’s base with no loss in
optical purity.
It was envisaged that the enone functionality would be
constructed by addition of a vinyl Grignard reagent into a
suitable amide derived from an ester of (S)-hydroxybutyric acid,
necessitating hydroxyl protection. With appropriate choice of
protecting group, deprotection could then either be accomplished
in situ under the anticipated acidic cyclisation conditions or as
a separate step if necessary.The chemistry used for conversion
of ethyl (S)-hydroxybutyrate 15 through to oxy-Michael precur-
sor 29 is shown in Scheme 9.22 Initially, we chose to carry out
TBS protection of starting material, which was achieved in
quantitative yield under standard conditions, and followed this
by Weinreb amide23 formation, to ensure control during the
enone formation. Preliminary studies demonstrated that Weinreb
amide formation on TBDMS-protected ethyl (S)-hydroxybutyate
using carbonyl diimidazole (CDI) as a coupling agent was
inefficient, whilst addition of vinyl magnesium bromide directly
to ester 15 was also unsuccessful. Therefore, we turned to a
published protocol for direct conversion of esters to Weinreb
amides using isopropylmagnesium chloride24 rather than using
was thus obtained in 78% yield and with an optical purity of
99.5% on a 76-g scale. The difficulties encountered with this
step (and indeed a consideration with all approaches to this
molecule) were due to the significant volatility and water
solubility of the product 4, requiring multiple extractions during
work-up and care during concentration and distillation. Nev-
ertheless, this approach was used for manufacture of 250 g of
pyranone 4, providing material with 99.5% optical purity and
98% chemical purity, a reflection of the mild conditions used
for the final step, under which the product was shown to be
stable. Overall, this route appeared very attractive, offering a
four-step synthesis of pyranone 4 with high optical purity from
readily available materials. Nevertheless, this approach would
have required development and refinement to improve yields,
particularly for the reduction step, and to move away from the
use of chromatography and cryogenic reaction conditions
associated with the efficient route to the reduction substrate 17.
Cyclisation onto Aldehyde Equivalents. It has recently
been reported that 4-substituted 2-oxetanones (ꢀ-lactones) can
be used as precursors for the preparation of 2,3-dihydropyra-
nones.17 Rather than look at this approach starting from
ꢀ-butyrolactone, we chose to focus on a variation by attempting
this chemistry on acyclic substrates derived from the much more
readily available ethyl 3-(S)-hydroxybutyrate, Scheme 7. Hy-
drazone 24 derived from acetaldehyde and 1-aminopiperidine18
was lithiated and then reacted with the protected ethyl (S)-
hydroxybutyrate 22 to furnish ꢀ-keto hydrazone 25. Subsequent
acid mediated hydrolysis gave the desired dihydropyranone 10,
though in only 25% yield. It was determined by isolation of
the intermediate ꢀ-keto hydrazone 25 that the low yield was
due to the poor reaction of ester 22 with the lithiated hydrazone.
Therefore, we envisaged the possibility of using Weinreb amide
23 as the substrate. The first step proceeded cleanly to give
ꢀ-keto hydrazone 25 as the major product, which was treated
with Amberlyst 15 resin in THF to give dihydropyranone 10
in 59% yield after purification by chromatography and with
98.5% optical purity, unchanged from the starting material.
Hydrogenation over a Pd/C catalyst afforded the desired
pyranone 4, again with an unchanged optical purity of 98.5%.
Although this route was not particularly atom efficient, it
(19) Prandi, C.; Venturello, P. J. Org. Chem. 1994, 59, 3494.
(20) Reiter, M.; Turner, H.; Gouverneur, V. Chem.sEur. J. 2006, 12, 7190.
(21) Baker-Glenn, C.; Hodnett, N.; Reitner, M.; Ropp, S.; Ancliff, R.;
Gouverneur, V. J. Am. Chem. Soc. 2005, 127, 1481.
(22) Atkinson, S.; Tornos, J. Preparation of Optically Active Pyranone
Derivatives. WO 2003051862 A1, 2003.
(23) Sibi, M. P. Org. Prep. Proced. Int. 1993, 25, 15.
(24) Williams, J. M.; Jobson, R. B.; Yasuda, N.; Marchesini, G.; Dolling,
U.-H.; Grabowski, E. J. J. Tetrahedron Lett. 1995, 36, 5461. Colle,
S.; Taillefumier, C.; Chapleur, Y; Liebl, R.; Schmidt, A. Bioorg. Med.
Chem. 1999, 7, 1049.
(17) Zipp, G. G.; Hilfiker, M. A.; Nelson, S. G. Org. Lett. 2002, 4, 1823.
(18) Marques-Lopex, E.; Herrera, R. P.; Fernandez, R.; Lassaletta, J. M.
Eur. J. Org. Chem. 2008, 3457.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
61

Scheme 9. Oxy-Michael route to pyranone 4
Scheme 10. Pathways for loss of optical purity
cyclisation conditions and observing complete consumption and
formation of racemic pyranone 27. Cyclisation of TBS ether
29 took up to 24 h for complete reaction, and therefore the
pyranone formed was exposed to the acidic conditions for a
significant period of time, causing the loss in optical purity. In
an attempt to speed up the cyclisation reaction and thus
minimise racemisation, use of the more labile triethylsilyl (TES)
protecting group was evaluated. Ethyl (S)-3-hydroxybutyrate
15 was converted to TES ether 30 then through to vinyl ketone
32 in an analogous manner to the TBDMS series in 50% overall
yield, Scheme 9. This substrate was then subjected to the in
situ silyl deprotection/ring-closure reaction under the same
conditions, which reached completion much more rapidly within
2 h. Isolated pyranone 4 was found to have an optical purity of
94% as the (S)-enantiomer, a significant improvement over
cyclisation of TBS-protected substrate 29, but still well below
our requirement. An investigation into alternative solvents for
this reaction showed that chloroform could be replaced by
toluene, although product 4 was obtained with slightly lower
optical purity, whereas use of alcoholic solvents led to com-
plications due to Michael addition of the solvent into the enone.
Elevating the reaction temperature to 40 °C in order to increase
the rate of reaction led to a more dramatic lowering of the
product optical purity, whereas cyclisation at a lower temper-
ature (0 °C) in chloroform gave a product with 97% optical
purity over an 8 h reaction time.
trimethylaluminium.25 Thus, TBS ether 22 was treated with
N,O-dimethylhydroxylamine hydrochloride in the presence of
isopropylmagnesium chloride to afford amide 23 in 76% yield
and with good purity. Interestingly, a recent paper has reported
conversion of unprotected methyl hydroxybutyrate to the
Weinreb amide in good yield under similar conditions at low
temperature;26 however, subsequent reaction with vinyl mag-
nesium bromide proceeded better if the hydroxyl group was
protected. Reaction of 23 with vinyl magnesium bromide then
afforded enone 29 in 42% yield27 which was subjected to an in
situ silyl ether cleavage and ring-closure process by treatment
with Amberlyst 15 resin in chloroform. This provided pyranone
4 in 40% yield, which disappointingly had an optical purity of
only 82%. Interestingly, deprotected enone 28 was not detected
by GC during the course of the reaction, suggesting a facile
ring closure. In a control experiment to assess product stability,
pyranone 4 was subjected to the acidic cyclisation conditions
and after 24 h was recovered with unchanged chemical purity
1
as judged by H NMR spectroscopy and analysis by GC.
Attempts at acid-mediated oxy-Michael ring closure on an
isolated, purified sample of hydroxyenone 28 were also
undertaken. This substrate was prepared by deprotection of 29
using methanesulphonic acid in 85% yield and was purified by
chromatography rather than by distillation to avoid thermal
decomposition (Vide infra). However, under conditions where
good conversion to pyranone 4 was achieved (MsOH in a range
of solvents at low temperature), a reduction in optical purity of
the product resulted. We then turned to Lewis acid mediated
oxy-Michael ring-closure, whereby catalysts such as FeCl3,
Pd(OAc)2, Fe(acac)3 and Cu(OTf)2 gave poor conversion or
significant loss in optical purity. However, a breakthrough came
during screening of some palladium catalysts where certain
species were found to promote the desired cyclisation reaction.
PdCl2(MeCN)2 in THF was found to be most active, providing
pyranone 4 with very good optical purity (99.2%) although the
reactions were slow (13 h at 40 °C and 24 h at rt). Moreover,
these transformations required a high catalyst loading (10 wt
%) to achieve a good conversion, which would have had a
significant impact on cost. On the other hand, attempts to cyclise
either hydroxyenone 28 under basic conditions (KOtBu) or
However, the optical purity had decreased from 99% to 59%,
suggesting that a ring-opening/ring-closure reaction was taking
place, Via primary alcohol 26 (Scheme 10), despite no acyclic
intermediates being observed by GC.
Attempts to facilitate the ring closure within alternative acidic
media (TsOH, camphor sulphonic acid, PPTS, sulphuric acid
in THF or dichloromethane) led to either poor conversion,
unwanted reactions, or in the case of PTSA in dichloromethane,
pyranone formation but with complete racemisation. Isolation
of dienone 33 from this reaction provided evidence for an
alternative racemisation pathway (Scheme 10), which was
confirmed by submitting a sample of this compound to the
(25) Hodgetts, K. J. Tetrahedron Lett. 2001, 42, 3763.
(26) Gebauer, J.; Rost, D.; Blechert, S. Heterocycles 2006, 68, 2129.
(27) Prior studies on the racemic version of 23 demonstrated that the yield
was compromised by formation of the product of aza-Michael addition
of N-methoxymethylamine to the newly formed enone system 29.
Investigation into the formation of this impurity demonstrated that it
was formed during work-up, and by careful control of temperature
and addition rate of the reaction quench this impurity could be
minimised to less than 5%. Use of more acidic quench conditions
was expected to suppress the formation of the impurity but led to
some deprotection.
62
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Scheme 11. Impurities derived from pyranone 4
Scheme 12. Pd-catalysed oxidative cyclisation route
of a few palladium complexes and either CuCl2 under air or
with benzoquinone as stoichiometric co-oxidant was carried out,
Table 1.
From this screening, PdCl2(MeCN)2 or PdCl2(PhCN)2 were
almost equally effective in allowing low catalyst loading and
giving the cleanest reaction profile with an excess of benzo-
quinone as the stoichiometric oxidant, of which the former was
selected for further development. An extensive solvent screen
was not carried out; however, THF as solvent gave a far better
conversion and product purity profile compared with ethyl
acetate at 1 mol % catalyst loading. Under such conditions,
with 1.2 equiv of benzoquinone in THF, dihydropyranone 10
was formed cleanly together with only 1% of the reduced
tetrahydropyranone 4 within 1 h at elevated temperature. The
direct cyclisation Via a Pd-assisted oxy-Michael reaction to form
tetrahydropyranone 4 required controlling since it was shown
subsequently that, when formed in this manner, this product
could have reduced optical purity and would therefore contribute
partially to an overall depletion of optical purity on progression
into the reduction step.
Because of the observed instability of hydroxyenone 28 in
the deprotection step and the requirement for purification by
chromatography due to decomposition encountered during
distillation, avoiding its isolation through a one-pot deprotection/
oxidative ring-closure was sought. Aqueous hydrochloric acid
was found suitable for the deprotection reaction and aqueous
acetone an ideal solvent for both steps. However, it was found
that the effect of water was opposed in that the rate of
deprotection was better with a low concentration of water,
whereas the oxidative cyclisation reaction proceeded best with
a higher concentration of water present. Thus, the process was
established with initially one volume of water in acetone for
deprotection and then additional water added part way through
to facilitate ring-closure. Formation of hydroxyenone 28 was
almost complete within 1 h under these conditions and cycli-
sation proceeded over 4.5 h total reaction time. The appreciable
solubility of dihydropyranone 10 in water was a feature of the
work-up, where removal of the silyl protecting group residues
was achieved by washing an aqueous solution of the product
with heptane. Following extraction into toluene, the mixture
was separated by distillation to afford the pure dihydropyranone
10 in 53% yield on an 800 g (input scale), with a chemical
purity of 94.6 area % and an optical purity of 99.5%, very close
to that of the starting material. Although the isolated yield from
this process was typically only 50%, a small amount of material
did remain in the aqueous layer, and it is believed that
polymerisation may account for some of the mass balance.
Stability studies conducted on the reaction solution containing
dihydropyranone 10 after pH adjustment demonstrated no loss
in optical purity at pH 5-8 on heating to 80 °C for 9 h. This
compound was also stable towards racemisation at pH 2 at rt,
whereas after 9 h at 80 °C, a drop in optical purity of 5% was
observed, indicating that high temperatures had to be avoided
during this step.
TBDMS ether 29 by treatment with tetrabutylammonium
fluoride were unsuccessful. It was also later demonstrated that
pyranone 4 was susceptible to Lewis acid and base mediated
aldol reactions to generate impurities 34 and 35 (Scheme 11).
Given the susceptibility of the product towards racemisation in
the presence of acid (hexadiene 33 was observed in most of
the acid catalysed processes) and the high catalyst loading
required for the only process which provided a product with
an acceptable optical purity, we shifted focus to study formation
of dihydropyranone 10 in the ring-closure step, which would
allow conversion to pyranone 4 under mild, neutral conditions
to avoid degradation in the final step.
Oxidative Pd-Catalysed Ring Closure. Formation of the
dihydropyranone skeleton from hydroxyenone 28 via an oxida-
tive cyclisation (intramolecular Wacker-type reaction), Scheme
12, was appealing as an alternative means of synthesising the
previously prepared dihydropyranone 10. Following earlier
studies (Schemes 6 and 7), it was expected that dihydropyranone
10 would be relatively stable towards loss of optical purity since
it is believed that a higher pKa of the C3 methylene protons
and reduced electrophilicity would infer greater stability com-
pared with tetrahydropyranone 4. It was noted during studies
into the Lewis-acid mediated oxy-Michael cyclisations of this
substrate described above that a small amount of dihydropy-
ranone 10 generally accompanied formation of the major
product 4 when Pd catalysis was used. This was rationalised
by a Pd-mediated cyclisation of the hydroxyl group onto the
enone system28 followed by ꢀ-hydride elimination to generate
dihydropyranone 10 together with a reduced Pd (0) species.
By inclusion of a stoichiometric oxidant in the reaction, a more
complete conversion was anticipated by regeneration of the Pd
(II) catalyst, in agreement with recent publications reporting
similar oxidative Pd-catalysed cyclisations for construction of
tetrahydropyran ring systems.20,28 A subsequent selective hy-
drogenation of the alkene bond would transform the product
into the target tetrahydropyranone 4 under mild, neutral condi-
tions, Scheme 12.
Initial studies into the oxidative ring-closure were carried
out on isolated, purified hydroxyenone 28, and a simple screen
To achieve the final transformation to tetrahydropyranone
4, a selective reduction of dihydropyranone 10 was required.
To ensure a high yield and to minimise over-reduction to
pyranols 6a/6b, extensive screening of catalysts and conditions
(28) Reiter, M.; Ropp, S.; Gouverneur, V. Org. Lett. 2004, 6, 91. Reiter,
M.; Turner, H.; Mills-Webb, R.; Gouverneur, V. J. Org. Chem. 2005,
70, 8478.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
63

Table 1. Conditions screened for oxidative cyclisation of 28
entrya
reagent system
solvent
t (h)
conversion (%)
10 (area %)
4 (area %)
1
2
3
4
5
PdCl2 (6 mol %), air
H2O
24
13
10
95
99
63
99
3
5
PdCl2 (10 mol %), CuCl2 (10 mol %), air
PdCl2 (MeCN)2 (2 mol %), BQb (1 equiv)
PdCl2 (MeCN)2 (1 mol %), BQb (1.5 equiv)
5% Pd/C (10% w/w), PdCl2 (MeCN)2 (2 mol %)c BQb
(1.2 equiv)
H2O
3
99
THF
EtOAc
EtOAc
1
1
99
75
>99
1
12
1
3 + 1
6
7
PdCl2 (MeCN)2 (1 mol %), BQb (1.2 equiv)
PdCl2 (PhCN)2 (1 mol %), BQb (1.2 equiv)
THF
THF
1
0.5
>99
>99
99
99
1
1
a All reactions run at 60 °C except entry 1 (rt). b BQ ) benzoquinone. c 3 h with Pd-C catalyst (<1% conversion), then 1 h with PdCl2(MeCN)2.
Table 2. Screening of catalyst for reduction of dihydropyranone 10
entry
catalysta
solvent
H2 (Mpa)
T (°C)
t (h)
10b
4b
6a/6bb
1
3% Pd-Al2O3
PhMe
PhMe
PhMe
PhMe
AcOMe
THF
Me2CO
PhMe
PhMe
PhMe
PhMe
PhMe
3
3
3
3
3
3
3
1
3
1
1
1
50
50
50
50
50
50
50
50
50
27
60
60
3
3
-
-
89.8
93.3
93.1
89.7
81.3
86.5
85.9
87.5
92.1
69.6
45.8
82.9
5.4
2.1
0.9
7.2
14.2
9.0
9.5
3.5
1.9
2.6
2.0
1.7
2
5% Pd-1.5% S-Al2O3
3% Pd-1.5% S-MgO-Al2O3
3% Pd-Al2O3
3
12
6
-
4
0.3
-
5
3% Pd-Al2O3
5
6
3% Pd-Al2O3
3
0.11
-
7
3% Pd-Al2O3
3
8
3% Pd-Al2O3
2
-
9
3% Pd-Al2O3
2
-
10
11
12c
3% Pd-Al2O3
3
22.0
3% Pd-Al2O3
2
45.8
11.4
3% Pd-Al2O3
17
a 5 wt % except entries 11 (3 wt %) and 12 (1 wt %). b GC area % c Prior reaction at 27 °C for 6 h.
Scheme 13. Large-scale synthesis of pyranone 4
was required, and some results are presented in Table 2. Of
the catalysts investigated, 3% Pd on alumina was selected,
affording a 94:6 ratio of 4:6a/6b based on analysis by GC (entry
1). Although not the most selective catalyst under the conditions
screened,29 it was readily available in quantities required for
the planned large-scale production. Choice of solvent had only
a slight effect on the level of pyranol produced (entries 4-7),
and so toluene was selected; more significant effects were seen
with catalyst loading, hydrogen pressure, and reaction temper-
ature (entries 8-12). From this series of experiments, the
conditions given for entry 9 were clearly the best, and a scale-
up reaction (180 g input) under these conditions afforded
pyranone 4 in 73% yield with 99.8% chemical purity and 99.4%
optical purity after isolation by distillation.
Large-Scale Synthesis of Pyranone 4. During the course
of studies into the oxy-Michael ring closure and Pd-mediated
oxidative cyclisation, the synthesis of enone 29 was evolving.
Together with the cyclisation and hydrogenation steps, these
were being developed into a route suitable for manufacture of
initially a 10 kg batch of optically pure pyranone 4 and
subsequently a 100 kg batch (Scheme 13). The starting material
used was methyl (S)-hydroxybutyrate 36, a commercially
available compound which can be readily prepared in good yield
and very high optical purity by asymmetric hydrogenation of
methyl acetoacetate.11 The dimethylamide was the preferred
electrophilic moiety for introduction of the vinyl group30 rather
than the Weinreb amide on cost grounds and was put in place
prior to the protection step. Thus, methyl (S)-hydroxybutyrate
36 (optical purity 99.7%) was converted to dimethylamide 37
by reaction with dimethylamine hydrochloride and sodium
methoxide in methanol in 64% yield on a 300 kg scale. The
sodium methoxide stoichiometry was set so as to achieve
maximum conversion to product while minimising byproduct
formation from dehydration (methyl crotonate and its amide).
The high water solubility of amide 37 limited the range of
efficient extraction solvents, but n-butanol was found to be ideal
for this purpose, from which the pure product was isolated by
distillation. Direct vinylation of amide 37 proceeded in only
30% yield; therefore, TBDMS protection was retained and was
accomplished under standard conditions in 89% yield. This was
(29) Use of the alternative catalyst 3% Pd-0.3% S on MgO/Al2O3 was
found to give slightly better selectivity for pyranone 4 over pyranols
6a/6b, but the rate of reaction was slower.
(30) Olah, G. A.; Prakash, G. K. S.; Arvanaghi, M. Synthesis 1984, 228.
64
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Scheme 14. Route to allene alcohol 42
slightly lower than expected due to some decomposition of the
product during work-up whilst destroying excess TBDMS-Cl
by hydrolysis. For conversion to enone 29, amide 38 was slowly
added to a solution of excess vinyl magnesium chloride in THF,
allowing good control of the exotherm. The aqueous acidic
work-up was carefully controlled to avoid loss of the silyl
protecting group whilst breaking down the hemiaminal inter-
mediate to the desired product, which was achieved by
quenching with dilute sulphuric acid then extracting into toluene.
Pure enone 29 was isolated by distillation, carried out in the
presence of a small amount of hydroquinone to suppress
degradation. Smaller-scale reactions proceeded well with yields
of up to 85% at end of reaction and 75% after isolation.
Unfortunately, during scale-up to 320 kg, the yield at end of
reaction was 76%; however, this dropped to 60% after
concentration to remove THF and further still to 48% after
distillation of the product 29. The reasons for this were thought
to be a consequence of longer exposure to the acidic work-up
on scale leading to decomposition or polymerisation of the
enone, coupled with some degree of decomposition during
distillation, although it was deemed preferable in the short term
to have a purification sequence in place at this stage at the
expense of yield. Indeed, stability studies carried out on batches
of neat enone 29 indicated rapid decomposition at 50 °C (75%
loss over 5 h) to polymeric materials. Subsequent one-pot
deprotection and cyclisation of 29 were accomplished under
the conditions identified (described above) and afforded dihy-
dropyranone 10 in 31% yield on a 100 kg (input) scale, after
work-up and distillation. Further material was obtained from
re-extraction of the aqueous layer, bringing the total up to 40%
yield, which was still lower than anticipated, but with an optical
purity of 99.7%. The final hydrogenation step, using 3% Pd on
alumina in toluene at 45-55 °C, was only scaled up to 10-20
L due to some variability encountered in the rate of reaction.
However, complete conversion within 4.5 h was achievable on
a 2.5 kg (input) scale under these conditions, with a selectivity
of 98% for pyranone 4 over pyranols 6a/6b and an averaged
yield of 94% for formation of the crude product. Combination
of multiple batches of crude pyranone 4 and purification by
distillation provided a 10 kg batch of product having very high
optical and chemical purities by GC (99.7 area % and 99.6 area
%, respectively) in an average 67% yield from dihydropyranone
10.
Scheme 15. Ring-closure pathways for allene 42
Routes from (S)-Propylene Oxide. An alternative short
route to pyranone 4 Via dihydropyranone 10 starting from (S)-
propylene oxide was envisaged in which a suitably functiona-
lised C3 nucleophile would be incorporated to build up the
pyranone ring. It has been reported that methoxyallene can be
lithiated to generate 3-lithio-3-methoxy allene and that this reacts
readily with a variety of electrophiles.32,33 Thus, incorporating
the successful cyclisation methodology already developed, we
postulated that reaction of 3-lithio-3-methoxyallene 41 with (S)-
propylene oxide 14 followed by ring closure would construct
the 6-carbon framework of the target pyranone with the required
stereocentre installed, Scheme 14. Essentially, allene ether 42
is a latent enone and thus offered many possibilities for
cyclisation, Scheme 15. (S)-Propylene oxide 14 can be produced
with very high optical purity (typically >99.5%) and is available
in large-scale quantities, thus making it a convenient chiral
building block. Methoxyallene 40 was not commercially avail-
able at the time this work was carried out but is reportedly
generated in a straightforward manner in high yield from methyl
propargyl ether 39 by heating under reflux with catalytic
potassium tert-butoxide followed by isolation of the product
by distillation (bp 53 °C).32,34 However, it was immediately
apparent that this synthesis would require modification for scale-
up for reasons of safety. Acetylenes (especially metalloacetyl-
Whilst this chemistry did provide the first multikilogram
batch of pyranone 4 using a new route, some problems were
encountered during scale-up and these would have been
addressed had the project continued. Additionally, it would have
been highly desirable to shorten the route by dispensing with
the protection step during the enone synthesis or using an
alternative, cheaper protecting group, e.g., TMS or THP.
Dispensing with protection altogether, by addition of a vinyl
Grignard reagent into dimethylamide 37 as demonstrated
recently for Weinreb amides,31 would be particularly appealing
due to the water-solubility of hydroxyenone 28, potentially
allowing its progression without isolation and purification into
the oxidative cyclisation.
(32) Hoff, S.; Brandsma, L.; Arens, J. F. Recl. TraV. Chim. Pays-Bas 1968,
87, 916.
(33) Hoff, S.; Brandsma, L.; Arens, J. F. Recl. TraV. Chim. Pays-Bas 1968,
87, 1179. Tius, M. A. 1-Methoxyallenyllithium. In e-EROS: Ency-
clopedia of Reagents for Organic Synthesis; Wiley: New York, 2001.
(34) See for example: Weiberth, F. J.; Hall, S. S. J. Org. Chem. 1985, 50,
5308. Perez, M.; Canoa, P.; Gomez, G.; Teijeira, M.; Fall, Y. Synthesis
2005, 411. Chengebroyen, J.; Linke, M.; Robitzer, M.; Sirlin, C;
Pfeffer, M. J. Organomet. Chem. 2003, 687, 313.
(31) Cohen, F.; Overman, L. E. J. Am. Chem. Soc. 2006, 128, 2594.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
65

enes) can be thermally unstable and with low-molecular weight
structures, decomposition is inevitably highly energetic and
potentially explosive (the heat of decomposition of methyl
propargyl ether 39 was determined by DSC to be ∼1600 J/g).
These problems were confounded in the literature synthesis by
the absence of solvent (to provide dilution and a heat sink) and
the “all-in” nature of the reaction in which there is no control
of the heat of the primary (desired) chemical reaction by reagent
addition. These initial concerns were duly confirmed by running
the reaction in a PHI-TEC II adiabatic calorimeter. In this test,
self-heating was detected immediately on mixing the potassium
tert-butoxide and methyl propargyl ether 39 (at ambient
temperature) that led to an extremely violent reaction resulting
in the complete destruction of the internals of the calorimeter
less than eight minutes later. Thus, a modification of the process
was developed to overcome the all-in nature of many of the
reported syntheses,32,34 by adding a solution of methyl propargyl
ether in THF in a controlled manner over 1 h to a solution of
potassium tert-butoxide in THF maintained at 50 °C. Under
these conditions, the reaction is addition-rate controlled, and
less than 6% methyl propargyl ether remained immediately upon
completion of the addition. Methoxyallene 40 was then isolated
in 95% yield as a solution in THF, ready for use directly in the
downstream chemistry as a telescope, following distillation of
the solution under vacuum (bp 15 °C at 70 mmHg). Chemically,
such solutions showed no significant degradation after storage
at rt over a period of 6 weeks. The thermal stability of the THF
solution of methoxyallene 40 was assessed by DSC and by
running an adiabatic PHI-TEC test which showed that the
reaction could probably be operated safely at scale with the
provision of suitable control measures. Having a significantly
higher onset temperature meant that this modified process was
deemed acceptable to run at up to 2 L scale within our
laboratories. Whilst efficient conversion of methyl propargyl
ether 39 to methoxyallene 40 could be achieved using the
alternative system of potassium hydroxide in DMSO at ambient
temperature,35 this method was not adopted due to the difficulty
of isolating the product from such a reaction mixture.
Lithiation of methoxyallene was carried out in THF solution
at 0 °C, and the resulting anion 41 was reacted with (S)-
propylene oxide 14 at the same temperature. A regioselective
reaction followed, and under these conditions, the yield of allene
alcohol 42 was typically 41% after isolation by distillation.
Forming the lithio species at -30 °C and reacting with
benzaldehyde gave a quantitative yield of the addition product,
indicating that lithiation was complete under these conditions
and that the problem was the reactivity of the epoxide. The
reaction profiles of lithio anions formed and quenched with (S)-
propylene oxide at -30 and 0 °C were comparable, demon-
strating cryogenic temperatures were not critical to the success
of the reaction. The propensity of allene alcohol 42 towards
degradation (over a few days at 4 °C, faster at rt) was also a
key consideration and meant that a fast low-temperature reaction
was imperative. Several Lewis acids were evaluated in this
process and presence of lithium bromide in a reaction carried
out at 0 °C to rt was found to improve the yield significantly to
52% after isolation, and most importantly, analysis by GC
demonstrated no loss in optical purity. Of the other Lewis acids
investigated it is noteworthy that BF3OEt2 caused violent
polymerisation of methoxyallene. Due to termination of the
project prior to running all but small-scale reactions, no in-
depth safety evaluation of this stage was undertaken. It could
be expected, however, that allene alcohol 42 may also present
safety issues for scale-up.
The versatile allene alcohol 42 offered various options for
conditions under which to attempt ring-closure; basic, acidic,
Lewis acidic, or transition-metal-mediated, Scheme 15. We
found that acid-mediated cyclisation of allene alcohol 42 by
protic or Lewis acids to generate pyranone 4, as expected,
proceeded through the intermediate enone 28. However, as
before, under the most successful conditions, loss of optical
purity was evident accompanied by formation of a variety of
side products. A base-mediated cyclisation was also investi-
gated,36 using a catalytic amount of potassium tert-butoxide,
anticipating formation of methyl enol ether 43, and this did
provide the desired product together with another unidentified
compound. Subsequent acid-mediated hydrolysis generated
pyranone 4 as the major component of a mixture; however,
this was not developed further due to the high level of byproduct
formation and drop in optical purity of pyranone 4 to 90%.
Following the successful use of Wacker-type conditions for
cyclisation of enone 28 to generate dihydropyranone 10
(Scheme 13), we envisaged subjecting allene alcohol 42 to these
conditions to generate this intermediate (Scheme 15). Initially,
we found that 10 mol % palladium chloride and copper(II)
chloride in the presence of benzoquinone gave the best
conversion in a mixture of acetone and water. Typically, enone
28 was observed almost immediately after addition of the
palladium and copper species, indicating that the first transfor-
mation was hydrolysis of the enol ether moiety. However,
formation of tars indicated polymerisation was a significant side
reaction. Whilst reducing the palladium to 5 mol % still gave
product 10, reducing the palladium amount below this led to
formation of pyranone 4 directly, presumably from competing
oxy-Michael cyclisation promoted by HCl generated in situ
during the reaction. It was essential that direct pyranone
formation in this step was minimised due to its propensity for
racemisation and potential for decomposition (Vide supra). The
conditions found to be successful earlier were then applied to
allene alcohol 42 in an attempt to improve selectivity. Thus,
PdCl2(MeCN)2 (0.5 mol %) and benzoquinone (1 equiv) were
added to the allene alcohol in aqueous acetone, generating
dihydropyranone 10 cleanly in 48% yield, again following rapid
formation of the intermediate enone 28. The drawback of using
these conditions was separation of the product from large
amounts of benzoquinone and hydroquinone during work-up
(36) Previous work in this area has focused on the base-mediated
cyclisations of the products from lithio-methoxyallene additions to
carbonyls. These have been shown to cyclise to generate furan species;
see: Reissig, H.-U.; Hormuth, S.; Schade, W.; Amombo, M. O.;
Toshiko, W.; Pulz, R.; Hausherr, A.; Zimmer, R. J. Heterocycl. Chem.
2000, 37, 597, and references therein. It was our suspicion that the
same type of transformation to form pyran ring systems may occur
from allene alcohols derived from epoxide ring opening. This was in
contrast, however, to a literature publication which reported formation
of furans: Hoff, S.; Brandsma, L.; Arens, J. F. Recl. TraV. Chim. Pays-
Bas 1969, 88, 609.
(35) Eroshchenko, S. V.; Sinegovskaya, L. M.; Tarasova, O. A.; Frolov,
Y. L.; Trofimov, B. A.; Ignatyev, I. S. Spectrochim. Acta, Part A 1990,
46A, 1505.
66
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

for which use of toluene as an extraction solvent was found to
minimise progression of these materials downstream. Dihydro-
pyranone 10 prepared in this way was then hydrogenated to
afford pyranone 4, which was shown have an optical purity
greater than 99% as the (S)-enantiomer. Typically in this
cyclisation, conversion of allene 42 to enone 28 is observed
almost immediately on adding the palladium and copper species
to the reaction mixture. A comparison study in which allene
alcohol 42 and enone 28 were subjected to the same conditions
demonstrated no difference in reactivity. This approach has
demonstrated on a small lab scale the utility of methoxyallene
for generation of a C3 anion containing latent enone functionality
and coupling with (S)-propylene oxide as an alternative ap-
proach to the cyclisation precursor 28. In addition, subsequent
in situ hydrolysis releases the enone moiety ready for the
oxidative Pd-catalysed ring closure under mild conditions.
Although defining a safe mode of operation with such energeti-
cally unstable intermediates on scale would need addressing,
perhaps through continuous manufacture, this approach dem-
onstrates in principle an alternative highly efficient four-step
synthesis of pyranone 4 from simple starting materials.
starting materials and dry solvents were used as received without
further purification. Small-scale chromatography was carried
out on Biotage KP-Sil silica gel cartridges (40-63 µm), and
product-containing fractions were concentrated to dryness under
vacuum. 1H NMR spectra were recorded at a probe temperature
of 25 °C, and chemical shifts are reported in ppm downfield
relative to TMS as an internal standard in CDCl3 unless
otherwise stated. Assays were performed against purified,
characterised reference materials. MS data were obtained on
an HP6890 series GC and HP5973 Mass Selective Detector
(MSD) using electron impact ionisation at 70 eV. HRMS were
performed on an LC-MS-IT-TOF spectrometer (Shimadzu
corporation) using positive electrospray ionisation (ESI) at 4.5
keV. GC method used for determination of the chemical purity
of 4: DB-FFAP column, 30 m × 0.25 mm id, 0.25 µm, oven
temp 60 °C/10 min then to 200 at 5 °C/min, He carrier gas at
0.5 mL/min, injector 180 °C, detection by FID at 230 °C. GC
method used for determination of the optical purity of 4 and
10: Chiraldex G-TA column, 50 m × 0.25 mm id, 0.125 µm,
oven temp 100 °C, He carrier gas at 2.2 mL/min, injector 220
°C, detection by FID at 220 °C. HPLC method used for
determination of the optical purity of 10: Chiralcel OB column,
4.6 mm × 250 mm, at 40 °C, detector 220 nm, mobile phase
90:10 n-hexane/iPrOH, flow rate 1.0 mL/min. GC method used
for determination of the chemical purity of 29, 37, and 38:
Neutrabond 1 column, 30 m × 0.25 mm id, 0.25 µm, oven
temp 120 °C/5 min then to 230 at 5 °C/min, He carrier gas at
0.5 mL/min, injector 180 °C, detector FID at 230 °C. For other
intermediates, the following GC purity method can be used:
J&W DB-5 column, 30 m × 0.32 mm id × 1 µm, oven temp
70 °C/5 min then to 220 at 15 °C/min, carrier gas hydrogen at
60 kPa, injector 250 °C, detection by FID at 300 °C.
Conclusions
(S)-2-Methyltetrahydropyran-4-one has proved a challenging
molecule to synthesise due to its sensitivity to acidic, basic,
and Lewis acidic reaction media giving rise to decomposition,
condensation products, and loss of optical purity. Despite this
inherent instability, the basis for short, efficient routes to this
target of high optical purity from readily available small chiral
building blocks has been established and demonstrated suc-
cessfully on multigram to multikilogram scale. Synthesis of an
appropriately functionalised δ-lactone followed by reduction
and deprotection steps afforded the target compound in four
steps in 21% overall yield, although some of the reaction
conditions used in their present form are undesirable for large-
scale manufacture and would have required further develop-
ment. Identification of dihydropyranone 10 as a robust inter-
mediate allowed conversion to the target pyranone 4 under mild,
neutral conditions with preservation of optical purity. A
straightforward synthesis of this intermediate Via an intramo-
lecular Pd-mediated oxidative cyclisation of a hydroxyl group
onto an enone provided an efficient five-step synthesis of the
target pyranone in 19% overall yield on 100s-g scale, although
the overall yield was lower for multikilogram manufacture,
reflecting difficulties of handling and isolation of the hydroxy-
enone intermediates. Furthermore, if the highly energetic
intermediates in the methoxyallene route can be safely handled
at scale, the number of steps is reduced, affording an alternative
highly efficient four-step synthesis that was demonstrated in
an overall yield of 19% on a small laboratory scale. Clearly,
much of this chemistry would have required further develop-
ment to improve yields and gain a better understanding of key
parameters to improve performance on multikilogram scale;
however, termination of the project left many of these aspects
incomplete.
(S)-9-Methyl-1,4,8-trioxaspiro[4.5]decan-7-one (17). To a
stirred solution of 1615 (3.5 g, 17.1 mmol) in dichloromethane
(15 mL) at rt was added ethylene glycol (3.55 mL, 63.7 mmol)
followed by TsOH ·H2O (0.66 g, 3.5 mmol). The solution was
heated under reflux with stirring for 6 h, then allowed to cool
before saturated aqueous sodium bicarbonate (50 mL) was
added. The layers were separated, the aqueous phase was
extracted with dichloromethane (30 mL), the combined organic
phases were washed with saturated aqueous sodium chloride
(50 mL), dried (Na2SO4) and then concentrated under reduced
pressure to give 17 as a yellow oil (2.6 g, 88%) which was
used without further purification. Purity by GC 75.6 area %.
Spectral data were obtained on a sample purified by chroma-
tography, eluting with 30% ethyl acetate in isohexane. 1H NMR
(400 MHz) δ 4.58 (dqd, J ) 11.9, 6.1, 2.9 Hz, 1H), 4.03-3.95
(m, 4H), 2.78 (dd, J ) 17.2, 1.8 Hz, 1H), 2.70 (d, J ) 17.4
Hz, 1H), 2.05 (dt, J ) 13.9, 2.2 Hz, 1H), 1.83 (dd, J ) 13.8,
11.8 Hz, 1H), 1.42 (d, J ) 6.2 Hz, 3H). 13C NMR (100 MHz)
δ 169.5, 105.6, 73.4, 64.8, 64.7, 40.9, 40.7, 21. [R]25 -19.5
D
(c 1.2, CHCl3).
(S)-7-Methyl-1,4,8-trioxaspiro[4.5]decane (20). A solution
17 (118.1 g, 84.7% w/w, contained weight 100 g, 581 mmol)
in THF (100 mL) was added dropwise over 1 h to a cooled
(<0 °C) suspension of sodium borohydride (22.0 g 582 mmol)
in THF (100 mL) with stirring (hydrogen evolution observed).
BF3 ·OEt2 (147.2 mL, 164.9 g, 1.16 mol) was then added
Experimental Section
General. All experiments were carried out under a nitrogen
atmosphere unless otherwise stated. Commercially available
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
67

dropwise over 30 min, maintaining the reaction temperature
below 0 °C (exothermic and continued gas evolution). On
completion of the addition, the viscous reaction mixture was
maintained at between 20 and 30 °C for 2 h with stirring
(continued exotherm). The reaction was quenched by the
dropwise addition of a mixture of ethyl acetate (500 mL) and
aqueous NaHCO3 (10% w/w, 500 mL). The organic layer was
separated, washed with water (500 mL), and then evaporated
to dryness under reduced pressure. The residue was purified
by chromatography, eluting with a 2:1 mixture of hexane and
ethyl acetate to afford 20 (32.3 g, 35.1%). 1H NMR (400 MHz)
δ 4.06-3.94 (m, 5H), 3.68-3.60 (m, 2H), 1.78 (dt, J ) 12.9,
5.4 Hz, 1H), 1.69 (m, 1H), 1.61 (m, 1H), 1.50 (t, J ) 16 Hz,
1H), 1.21 (d, J ) 8 Hz, 1H). HRMS. Calcd for C8H15O3 (MH+)
159.1016. Found 159.1010.
purity by GC 98.6% as the (S)-enantiomer. [R]25 +199.5 (c
D
0.46, CHCl3). 1H NMR (400 MHz) δ 7.35 (d, J ) 5.9 Hz, 1H),
5.41 (d, J ) 6.2 Hz, 1H), 4.60-4.52 (m, 1H), 2.56-2.42 (m,
2H), 1.47 (d, J ) 6.4 Hz, 3H). 13C NMR (100 MHz) δ 192.6,
163.2, 106.9, 76.0, 43.5, 20.3. MS (m/z) 112 [M+].
(S)-2-Methyl-2,3-dihydropyran-4-one (10) via Hydrazone
25. To a solution of diisopropylamine (1.57 mL, 11.2 mmol)
in tetrahydrofuran (45 mL) cooled to 0 °C was added dropwise
a solution of n-butyllithium (2.5 M in hexanes, 4.2 mL, 10.5
mmol) over a period of 5 min. The resulting pale-yellow
solution was stirred for 15 min at this temperature, and then
hydrazone 2418 (1.32 g, 10.5 mmol) was added dropwise. The
reaction mixture was stirred at 0 °C for a further 1 h during
which time a precipitate formed. The suspension was cooled
to -78 °C, and then a solution of 23 (1.83 g, 7 mmol) in
tetrahydrofuran (5 mL) was added dropwise. After stirring at
-78 °C for 3 h the mixture was allowed to warm to rt and
stirred overnight. Water (3.5 mL) and diethyl ether (140 mL)
were added, the suspended solids were removed by filtration,
the filtrate was dried (MgSO4), and the solvent was evaporated
under reduced pressure to afford intermediate 25 as a yellow
oil (3.5 g) which was used without further purification. This
material was dissolved in tetrahydrofuran (50 mL), and Am-
berlyst 15 ion-exchange resin (3.5 g) was added and the
suspension heated under reflux for 2 h. After cooling to rt, the
resin was removed by filtration, the filtrate was concentrated
under reduced pressure, and the residue was purified by
chromatography, eluting with a 1:9 mixture of ethyl acetate and
isohexane to afford 10 as a yellow oil (0.46 g, 59%). Optical
purity by GC 98.8% as the (S)-enantiomer. Purity by GC 97.8
area %.
(S)-2-Methyltetrahydropyran-4-one (4) via Deprotection
of 20. A mixture of crude 20 (121.8 g, 76.46 g contained wt,
483 mmol), aqueous formaldehyde solution (37% w/w, 156.9
g, 1.93 mol), and PPTS (3.06 g, 12 mmol) was heated at 80 °C
with stirring for 5 h. After cooling to rt, the reaction mixture
was extracted with ethyl acetate (1 × 360 mL, 4 × 180 mL).
The organic layers were combined and washed with aqueous
NaHCO3 (5% w/w, 60 mL) and water (60 mL). The solvent
was removed under reduced pressure and the residue purified
by distillation under reduced pressure to give 4 (43.12 g, 78%).
Bp 57-58 °C/19 mmHg. Purity by GC 91.9 area %. Optical
1
purity by GC 99.6%. H NMR (400 MHz) δ 4.28 (ddd, J )
11.5, 7.4, 1.3 Hz, 1H), 3.78-3.70 (m, 1H), 3.68 (ddd, J ) 12.3,
11.5, 2.8 Hz, 1H), 2.58 (dddd, J ) 14.6, 12.3, 7.4, 1.0 Hz, 1H),
2.40 (dt, J ) 14.6, 2.6 Hz, 1H), 2.35-2.25 (br m, 2H), 1.28
(d, J ) 6.2 Hz, 3H). 13C NMR (100 MHz) δ 206.9, 74.3, 66.4,
49.9, 41.9, 22.0. MS (m/z) 114 [M]+.
Ethyl (S)-3-(tert-Butyldimethylsilyloxy)butyrate (22). Eth-
yl (S)-3-hydroxybutyrate 15 (5 g, 37.8 mmol) was dissolved in
dichloromethane (10 mL). After cooling to 10 °C, tert-
butylchlorodimethylsilane (6.84 g, 45.3 mmol) was added and
when most of the solid had dissolved this was followed by
imidazole (5.15 g, 75.6 mmol) in one portion (exotherm to 30
°C). After stirring the thick suspension at rt for 16 h, water (25
mL) was added, the layers were separated and the aqueous layer
was extracted with dichloromethane (3 × 25 mL). The
combined organic layers were washed with water (25 mL), dried
(MgSO4) and the solvent evaporated under reduced pressure
to leave crude 22 as a colourless oil (10 g, .100%). Purity by
GC 95.0 area %. A sample was purified by chromatography,
eluting with 5% ethyl acetate in isohexane, to afford a pure
sample of the title compound as a colourless oil for character-
(S)-2-Methyl-2,3-dihydropyran-4-one (10) via lactol 21.
A stirred solution of 17 (2.5 g, 14.5 mmol) in toluene (40 mL)
was cooled to -70 °C, and a solution of diisobutylaluminium
hydride in toluene (1.0 M, 14.5 mL, 14.5 mmol) was added
dropwise over 30 min. The solution was stirred at -70 °C for
a further 2 h, and then saturated aqueous ammonium chloride
(5 mL) was added dropwise. The solution was then allowed to
warm to rt and stirred for 30 min. Sodium sulfate (5 g) was
added followed by diethyl ether (200 mL) and stirring continued
for 30 min. The solids were removed by filtration through Celite
and washed with diethyl ether (100 mL), and the combined
filtrates were concentrated under reduced pressure to give the
intermediate lactol 21 as a yellow oil comprising a mixture of
epimers at C-7 (2.1 g, 83%) which was used without purification
and characterisation.
Lactol 21 (2.0 g, 11.5 mmol) was dissolved in a mixture of
acetone (20 mL) and water (20 mL) at rt, and TsOH ·H2O (2.2
g, 11.5 mmol) was added. The solution was heated to 50 °C
for 2 h and then allowed to cool to rt before a saturated aqueous
solution of sodium bicarbonate (20 mL) was added. The solution
was partially concentrated and extracted with ethyl acetate (3
× 40 mL); the combined organic layers were dried (Na2SO4),
and the solvent was evaporated under reduced pressure to give
a yellow oil which was purified by chromatography, eluting
with 20% ethyl acetate in isohexane, to afford 10 as a pale-
yellow oil (0.75 g, 58%). Purity by GC 97.3 area %. Optical
ization purposes. Purity by GC 100 area %. [R]25 +21.7 (c
D
1.05, CHCl3).37 1H- and 13C NMR data in agreement with
published data.37 MS (m/z) 245 [M - H]+.
(S)-3-(tert-Butyldimethylsilyloxy)-N-methoxy-N-methyl-
butyramide (23). Crude 22 (5 g, 20.3 mmol) was dissolved in
tetrahydrofuran (50 mL) and N,O-dimethylhydroxylamine
hydrochloride salt (3.07 g, 31.5 mmol) added. After cooling
the slurry to -20 °C, a solution of isopropylmagnesium chloride
in THF (2.0 M, 30 mL, 60 mmol) was added dropwise over
30 min, maintaining the reaction temperature at below -15 °C.
After keeping at -20 °C for a further 2 h, the reaction was
(37) Wattanasereekul, S.; Maier, M. E. AdV. Synth. Catal. 2004, 346, 855.
68
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

quenched with a saturated aqueous solution of ammonium
chloride (100 mL) and diluted with diethyl ether (40 mL). The
layers were separated, the aqueous layer was extracted with
diethyl ether (100 mL), the combined organic layers were
washed with water (100 mL) and dried (Na2SO4), and the
solvent was evaporated under reduced pressure to leave crude
23 as a pale-yellow oil (4.8 g, 90%). Purity by GC 88.7 area %
(main impurity TBDMS-OH, 7.8 area %). A portion of the
residue (1.3 g) was purified by chromatography, eluting with a
1:9 mixture of ethyl acetate and isohexane, to afford the title
compound as a colourless oil (1.09 g, 84% recovery). Purity
pressure to leave crude 30 as a colourless oil (9.53 g, >100%).
Purity by GC 97.9 area %. Spectral data were obtained on a
sample purified by chromatography, eluting with 1% ethyl
acetate in isohexane. Purity by GC 100 area %. [R]25D +18.2
1
(c 0.99, CHCl3). H NMR (400 MHz) δ 4.33-4.28 (m, 1H),
4.07-4.18 (m, 2H), 2.50 (dd, J ) 14.6, 7.2 Hz, 1H), 2.36 (dd,
J ) 14.6, 5.6 Hz, 1H), 1.26 (t, J ) 7.1 Hz, 3H), 1.21 (d, J )
6 Hz, 3H), 0.95 (t, J ) 8 Hz, 9H), 0.56 (q, J ) 8 Hz, 6H). 13C
NMR (100 MHz) δ 171.6, 65.6, 60.2, 44.9, 24.0,14.2, 6.7, 4.8.
MS (m/z) 245 [M - H]+.
(S)-3-(Triethylsilyloxy)-N-methoxy-N-methylbutyra-
mide (31). This was prepared in a similar manner to that
described above for compound 23, starting from 13.3 g of crude
30), with appropriate scaling of quantities, affording 31 as a
colourless oil (12.6 g, 89%). [R]25D +6.4 (c 1.00, CHCl3). 1H
NMR (400 MHz) δ 4.39 4.32 (m, 1H), 3.69 (s, 3H), 3.18 (s,
3H), 2.77 (dd, J ) 14.6, 7.2 Hz, 1H), 2.40 (dd, J ) 14.6, 5.9
Hz, 1H), 1.22 (d, J ) 6 Hz, 3H), 0.95 (t, J ) 7.9 Hz, 9H), 0.56
(q, J ) 7.9 Hz, 6H). 13C NMR (100 MHz) δ 170.0, 65.7, 61.3,
41.8, 32.0, 24.2, 6.7, 4.7. MS (m/z) 260 [M - H]+.
(S)-5-(Triethylsilyloxy)hex-1-en-3-one (32). This com-
pound was prepared in a similar manner to that described above
for compound 29, starting from 10 g (38.2 mmol) of crude 31,
with appropriate scaling of quantities. After purification by
chromatography, 32 was obtained as a colourless oil (4.9 g,
56%). Purity by GC 80.4 area %. [R]25D +20.7 (c 1.20, CHCl3).
1H NMR (400 MHz): δ 6.36 (dd, J ) 17.7, 10.5 Hz, 1H), 6.22
(dd, J ) 17.7, 0.9 Hz, 1H), 5.85 (dd, J ) 10.5, 1.0 Hz, 1H),
4.38-4.31 (m, 1H), 2.86 (dd, J ) 15.1, 6.7 Hz, 1H), 2.58 (dd,
J ) 15.1, 5.9 Hz, 1H), 1.21 (d, J ) 6.2 Hz, 3H), 0.95 (t, J )
7.9 Hz, 9H), 0.54 (q, J ) 7.9 Hz, 6H). 13C NMR (100 MHz)
δ 199.6, 137.4, 128.5, 65.4, 49.1, 24.2, 6.8, 4.7. MS (m/z) 227
[M - H]+.
(S)-2-Methyltetrahydropyran-4-one (4) via Intramolecu-
lar Oxy-Michael Addition on 32. Crude 32 (0.2 g, 0.88 mmol)
was treated with Amberlyst 15 ion-exchange resin (0.1 g) in
toluene (10 mL) at 0 °C for 8 h. Work-up and purification as
described above afforded 4 as a pale-yellow oil (0.05 g, 50%).
Purity by GC 97.6 area %. Optical purity by GC 94.6% as the
(S)-enantiomer.
(S)-3-Hydroxy-N,N-dimethylbutyramide (37). Sodium
methoxide in methanol (28% w/w, 930 kg, 4.82 kmol) was
added over a period of 95 min to a stirred, cooled (4 °C) solution
of dimethylamine hydrochloride (250 kg, 3.07 kmol) in
methanol (640 kg). During the course of the addition, the
reaction temperature increased to 12.8 °C and was brought back
to 3.1 °C during a subsequent 2 h stir. Methyl (S)-3-hydroxy-
butyrate 36 (316 kg, 2.68 kmol) was added to the reaction
mixture over a period of 85 min, and stirring continued for 18 h
at between 2 and 4 °C. After cooling to 0 °C, aqueous sulphuric
acid (10% w/w, 316 kg) was added over 57 min, which brought
the pH to 8-9; then further aqueous sulphuric acid (20% w/w,
42.7 kg) was charged over 40 min, bringing the pH to 7. The
reaction temperature was allowed to rise to a maximum of 14
°C during these exothermic additions. The inorganic byproduct
was removed by centrifugation and the filtrate concentrated
under vacuum (100 mmHg) at below 51 °C to remove methanol
(1770 kg). The residue was extracted with n-butanol (3 × 316
by GC 100 area %. [R]25 +18.1 (c 1.37, CHCl3).38 1H- and
D
13C NMR data in agreement with published data.38 MS (m/z)
260 [M - H]+.
(S)-5-(tert-Butyldimethylsilyloxy)hex-1-en-3-one (29) from
Amide 23. Crude 23 (3.5 g, 13.4 mmol) was dissolved in
tetrahydrofuran (70 mL) at rt then cooled to 0 °C. A solution
of vinylmagnesium bromide in THF (1.0 M, 20 mL, 20 mmol)
was added dropwise over 30 min, maintaining the reaction
temperature below 5 °C. After keeping for a further 1 h at
between 0 and 5 °C, the solution was quenched into a stirred
mixture of saturated aqueous ammonium chloride (70 mL) and
diethyl ether (175 mL). The layers were separated, the aqueous
layer was extracted with diethyl ether (100 mL), the combined
organic layers were washed with water (100 mL) and dried
(Na2SO4), and the solvent was evaporated under reduced
pressure. The residue was purified by chromatography, eluting
with a 1:20 mixture of ethyl acetate and isohexane, to provide
29 as a colourless oil (1.27 g, 42%). Purity by GC 96.9 area
%. [R]25D +30.9 (c 1.08, CHCl3). 1H NMR (400 MHz) δ 6.36
(dd, J ) 17.7, 10.5 Hz, 1H), 6.22 (dd, J ) 17.7, 1.3 Hz, 1H),
5.85 (dd, J ) 10.5, 1.3 Hz, 1H), 4.37-4.33 (m, 1H), 2.85 (dd,
J ) 14.9, 7.2 Hz, 1H), 2.54 (dd, J ) 14.9, 5.4 Hz, 1H), 1.20
(d, J ) 6.2 Hz, 3H), 0.86 (s, 9H), 0.06 (s, 6H), 0.02 (s, 6H).
13C NMR (100 MHz) δ 199.8, 137.4, 128.5, 65.8, 49.1, 25.8,
25.7, 24.2, 18.0, -4.6, -5.0. MS (m/z) 227 [M - H]+.
(S)-2-Methyltetrahydropyran-4-one (4) via Intramolecu-
lar Oxy-Michael Addition on 29. 29 (0.4 g, 1.75 mmol) was
dissolved in chloroform (30 mL) at rt. Amberlyst 15 ion-
exchange resin (0.15 g) was added in one portion, and then the
reaction mixture was held at rt with stirring for 24 h. After
removal of the resin by filtration, solvent was removed by
evaporation under reduced pressure and the residue purified by
chromatography, eluting with 20% ethyl acetate in petroleum
ether, to afford the title compound as a pale-yellow oil (80 mg,
40%). Optical purity by GC 81.6% as the (S)-enantiomer.
Ethyl (S)-3-(Triethylsilyloxy)butyrate (30). Ethyl (S)-3-
hydroxybutyrate 15 (5 g, 37.8 mmol) was dissolved in dichlo-
romethane (20 mL), and then triethylamine (6.4 mL, 46 mmol)
was added followed by chlorotriethylsilane (7.6 mL, 45 mmol)
dropwise, resulting in precipitation of a white solid (exotherm
to 35 °C over 30 min). After allowing to cool back to rt and
stirring for 5 h, water (30 mL) was added, the layers were
separated, and the aqueous layer was extracted with dichlo-
romethane (3 × 30 mL). The combined organic layers were
washed with water (50 mL) followed by brine (50 mL), and
dried (Na2SO4), and the solvent was evaporated under reduced
(38) Denmark, S. E.; Fujimori, S. Org. Lett. 2002, 4, 3477.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
69

kg), and the combined extracts were concentrated by distillation
(50 mmHg/88 °C) to remove a mixture of n-butanol and water
(1895 kg). Additional inorganic salts were removed by cen-
trifugation and washed with n-butanol (316 kg). The combined
filtrates were separated by distillation to afford 37 (223 kg,
63.5%). Bp 88-90 °C at 0.3 mmHg. Purity by GC 97.5 area
% with 1.56 area % 36. 1H NMR (300 MHz) δ 4.27-4.12 (m,
1H), 4.07 (br s, 1H), 2.97 (s, 6H), 2.49 (ABX system, J ) 15,
2.4 Hz, 1H), 2.30 (ABX system, J ) 15, 9 Hz), 1.22 (d, J )
6.3 Hz, 3H). HRMS. Calcd for C6H13NO2Na ([M + Na]+)
154.0839. Found 154.0834.
this temperature for 65 min. Further water (400 kg) was added,
and the reaction mixture was held at 35-40 °C for another
255 min when analysis showed 97 area % conversion. The
reaction mixture was brought to pH 5-6 by the addition of
solid sodium bicarbonate in portions (total 8.9 kg, 106 mol
added) and then concentrated under vacuum (350 mmHg, 67.5
°C) until 249 kg of solvent had been removed. After cooling,
heptane (68.4 kg) was added to the residue, and the layers were
allowed to separate. The organic phase was extracted with water
(2 × 150 kg), and then the combined aqueous layers were
extracted with toluene (4 × 340 kg). The combined toluene
phases were concentrated under vacuum (70 mmHg/81.3 °C)
until 1298 kg of solvent had been removed. PEG#400 (9.8 kg)
was added to the residue which was then purified by distillation
to afford 10 (17.28 kg, 15.22 kg corrected for assay, 31.0%).
Bp 68-70 °C at 10 mmHg. Optical purity by HPLC 99.8
area %.
The aqueous layer remaining after toluene extraction was
re-extracted with the recovered toluene (1298 kg) and the
organic phase was concentrated under vacuum to leave 79 kg
of residue to which PEG#400 (0.5 kg) was added. Following
distillation, further dihydropyranone 10 was obtained (5.44 kg,
4.57 kg corrected for assay, 9.3%). Optical purity by HPLC
99.7%. Total yield 19.79 kg (40.3%).
(S)-2-Methyltetrahydropyran-4-one (4) from Dihydro-
pyranone 10. A mixture of 10 (5.48 kg, weight corrected for
assay 4.83 kg, 43.1 mol) and 3% Pd on alumina (0.28 kg) in
toluene (8.66 kg) was hydrogenated at 45-55 °C/30 bar
pressure until uptake of hydrogen had ceased (18.5 h) when
analysis by GC demonstrated 100% conversion and 99.6 area
% selectivity for pyranone 4. After cooling to rt, the catalyst
was removed by filtration and toluene was removed by
distillation (temperature 44 °C/45 mmHg pressure) to leave a
concentrated solution of pyranone 4 in toluene (8.16 kg. This
was combined with the product concentrates from several other
reactions carried out under similar conditions, starting from a
total weight of 11.57 kg of 10 (10.19 kg corrected for assay,
103.2 mol) and the mixture was purified by distillation under
reduced pressure to afford 4 as a colourless oil (10.2 kg, 67%).
Bp 52-54 °C/20 mmHg. Chemical purity by GC 99.6 area %.
Optical purity by GC 99.7% as the (S)-enantiomer.
(S)-3-(tert-Butyldimethylsiloxy)-N,N-dimethylbutyra-
mide (38). Imidazole (138.3 kg, 2.03 kmol) was added to a
solution of 37 (222 kg, 1.69 kmol) in ethyl acetate (603 kg)
with stirring. Once dissolution was complete, a solution of
TBDMS chloride in ethyl acetate (50% w/w, 561 kg, 280.5 kg
contained weight, 1.86 kmol) was added over a period of 125
min, during which the reaction temperature rose from 12.6 °C
to a maximum of 17.2 °C. After stirring at 15-20 °C for 16 h,
imidazole hydrochloride was removed by centrifugation, and
the solids were washed with ethyl acetate (100 kg). The
combined filtrates were concentrated under reduced pressure
(80 mmHg/76 °C) until 909 kg solvent had been removed.
Toluene (385 kg) was added to the residue, and the mixture
was washed with water (3 × 222 kg) and then concentrated
under vacuum (40 mmHg/86 °C) until 377 kg of solvent had
been removed. The residue was separated by distillation to
provide 38 (371 kg, 89.3%). Bp 80 °C at 0.3 mmHg. Purity by
GC 96.8 area %. 1H NMR (300 MHz) δ 4.40-4.28 (m, 1H),
3.05 (s, 3H), 2.94 (s, 3H), 2.65 (ABX system, J ) 14, 7.5 Hz,
1H), 2.28 (ABX system, J ) 14, 5 Hz), 1.21 (d, J ) 6 Hz,
3H), 0.86 (s, 9H), 0.06 (s, 3H), 0.02 (2, 3H). HRMS found
268.1705 calcd for C12H27NO2SiNa ([M + Na]+) 268.1703.
(S)-5-(tert-Butyldimethylsiloxy)hex-1-en-3-one (29) from
Dimethylamide 38. 38 (320 kg, 1.30 kmol) was added to a
cooled solution of vinyl magnesium chloride in THF (0.85
M/kg, 2783 kg, 2.37 kmol) over a period of 75 min (initial
temperature 4.9 °C, final temperature 11.6 °C). On completion
of the addition, the mixture was stirred for a further 110 min
whilst it was cooled back to 6.8 °C. The reaction mixture was
then added to a cooled solution of aqueous sulphuric acid (5%
w/w, 3500 L) over a period of 105 min (initial temperature 1.1
°C, final temperature 6.9 °C, maximum temperature 8.1 °C)
and stirred for a further 30 min. Toluene (510 kg) was added,
the layers were separated, and the organic layer was washed
sequentially with water (370 kg), aqueous sodium carbonate
(5% w/w, 368 kg), and water (370 kg). Hydroquinone (370 g,
3.36 mol) was added to the toluene phase which was then
concentrated under vacuum (70 mmHg/106 °C) until 2117 kg
of solvent had been removed. The residue was purified by
distillation under vacuum to afford 29 (142 kg, 47.7%). Bp 64
°C at 0.6 mmHg. Purity by GC 95.3 area %.
Methoxyallene (40). A solution of methyl propargyl ether
(90.6 g, 1.29 mol) in THF (109 mL) was added in 5 mL portions
over a period of 1 h to a solution of potassium tert-butoxide
(14.5 g, 0.129 mol) in THF (436 mL) at 55 °C. Once the
addition was complete, the solution was allowed to cool to rt,
and methoxyallene was isolated as a solution in THF by
1
distillation under vacuum (486 g). Assay by H NMR 17.8%
w/w, therefore contained weight of methoxyallene 86.3 g (95%).
Bp 15-20 °C at 70 mmHg. 1H NMR (400 MHz) δ 6.77 (t, J
) 6.1 Hz, 1H), 5.48 (d, J ) 6.1, 2H), 3.41 (s, 3H). 13C NMR
(75 MHz) δ 201.1, 122.8, 91.2, 55.8. MS: 70 (M+, 100%), 55
(M-15, 25%).
(S)-4-Methoxyhexa-4,5-dien-2-ol (42). A solution of meth-
oxyallene (40) in THF solution (150 g, 17.8% w/w, containing
methoxyallene 26.6 g, 0.379 mol) was cooled to -4 °C and
heptane (27.8 mL) added as an internal GC standard. A solution
of n-butyllithium in hexanes (2.5 M, 167 mL, 0.418 mol) was
(S)-2-Methyl-2,3-dihydropyran-4-one (10) via Oxidative
Cyclisation Route. Benzoquinone (56.1 kg, 519 mol), aqueous
hydrochloric acid (37% w/w, 2.16 kg) and PdCl2(CH3CN)2 (0.57
kg, 2.20 mol) were added to a solution of 29 (100 kg, 438 mol)
in a mixture of acetone (237 kg) and water (100 kg) at 15 °C.
The mixture was then heated to 35-40 °C and maintained at
70
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

then added over a period of 1 h, maintaining the reaction
temperature between -2 °C and +2 °C. Once the addition was
complete, the mixture was stirred at 0 °C for 0.5 h (complete
lithiation was confirmed by 1H NMR spectroscopy on a sample
quenched into d4-MeOH). The lithio-methoxyallene solution
was then added to a mixture of LiBr (32.9 g, 0.379 mol) and
(S)-propylene oxide (26.6 mL, 0.380 mol) in THF (100 mL) at
0 °C over a period of 30 min, maintaining the reaction
temperature between 0 °C and +5 °C. The reaction mixture
was then allowed to warm to rt (exotherm to 35 °C occurred,
mixture allowed to cool back to rt) and stirred for 1 h. Water
(350 mL) was slowly added (exothermic) and the product
extracted into diethyl ether (2 × 250 mL). The combined
organic layers were dried (Na2SO4), filtered, and concentrated
under reduced pressure (bath temperature 25 °C) to leave an
orange oil, 35.2 g (72.5%). This was purified by distillation to
afford 42 (28.1 g). Bp 45-50 °C at 1 mbar. Purity by GC 88.1
area %. Assay by GC 90% w/w., therefore yield corrected for
g, 0.13 mol) was then added portion-wise, washed in with
acetone (10 mL). Bis(acetonitrile)dichloropalladium(II) (0.173
g, 0.67 mmol) was then added, again washed in with acetone
(10 mL) before the mixture was heated to 55 °C and held at
this temperature for 2 h. The mixture was allowed to cool to rt
then diluted with water (100 mL) and extracted with toluene
(4 × 100 mL). The organic layers were combined, dried
(MgSO4), and evaporated (bath temperature 35 °C, vacuum 60
mmHg) to give the crude product as an oil (14.2 g). This was
purified by distillation to afford 10 as a colourless oil (7.2 g,
48%). Bp 59-61 °C at 6 mbar. Assay by GC 85% w/w. HRMS
calcd for C8H12NO2 (MH+ + CH3CN) 154.0868, found
154.0864. Optical purity by GC 100% as the (S)-enantiomer.
Acknowledgment
We thank the following from AstraZeneca for their contribu-
tion to this project: Paul Gadd for undertaking the process safety
studies, Annabel Germon, Jason Knott and Lee Mallard for
analytical support and also Anil Mistry and Liz Poulton for
project management. We thank Professor James Anderson of
University College London for helpful discussions on the
oxidative cyclisations.
1
assay 51.8%. H NMR (400 MHz) δ 5.45 (s, 2H), 4.05 (m,
1H), 3.42 (s, 3H), 2.35 (m, 2H), 2.30 (br s, 1H), 1.25 (d, J )
6.4 Hz, 2H). 13C NMR (75 MHz) 199.0, 132.3, 90.3, 66.2, 56.
2, 41.8, 22.6. MS 128 (M+, 50%), 113 (M+-15, 60%). Chiral
purity by GC: 100% as the (S)-enantiomer.
(S)-2-Methyl-2,3-dihydropyran-4-one 10 (via Methoxy-
allene Route). Allene 42 (17.1 g, 0.13 mol) was dissolved in
acetone (80 mL) and water (33 mL) added. Benzoquinone (14.4
Received for review June 30, 2009.
OP900163A
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
71