Efficient synthesis of the A-ring phosphine oxide building block useful for 1α,25-dihydroxy vitamin D3 and analogues

Article abstract of DOI:10.1021/jo0161577

The 1α-hydroxy A-ring phosphine oxide 1, a useful building block for vitamin D analogues, was synthesized from (S)-carvone in nine synthetic operations and a single chromatographic purification in 25% overall yield. The synthesis features two novel efficient synthetic transformations: the Criegee rearrangement of α-methoxy hydroperoxyacetate 10 in methanol to obtain directly the desired secondary 3β-alcohol 11 and the highly chemo- and stereoselective isomerization of dieneoxide ester (E)-7 to the 1α-allylic alcohol with an exocyclic double bond (E)-8. Further insight into the selectivity control of the latter rearrangement was obtained from the reactions of (Z)epimeric substrates. The new synthetic approach leading to the 1α-hydroxy epimers complements our previously reported synthesis of the corresponding 1β-epimers, thus producing all stereoisomers of these versatile building blocks efficiently from carvone.

Full text of DOI:10.1021/jo0161577

1580
J . Org. Chem. 2002, 67, 1580-1587
Efficien t Syn th esis of th e A-Rin g P h osp h in e Oxid e Bu ild in g Block
Usefu l for 1r,25-Dih yd r oxy Vita m in D3 a n d An a logu es
Andrzej R. Daniewski, Lisa M. Garofalo, Stanley D. Hutchings, Marek M. Kabat, Wen Liu,
Masami Okabe, Roumen Radinov,* and George P. Yiannikouros
Chemical Synthesis - Process Research, Non-Clinical Development, Pre-Clinical Research and
Development, Hoffmann-La Roche, Inc., Nutley, New J ersey 07110
roumen.radinov@roche.com
Received September 28, 2001
The 1R-hydroxy A-ring phosphine oxide 1, a useful building block for vitamin D analogues, was
synthesized from (S)-carvone in nine synthetic operations and a single chromatographic purification
in 25% overall yield. The synthesis features two novel efficient synthetic transformations: the
Criegee rearrangement of R-methoxy hydroperoxyacetate 10 in methanol to obtain directly the
desired secondary 3â-alcohol 11 and the highly chemo- and stereoselective isomerization of
dieneoxide ester (E)-7 to the 1R-allylic alcohol with an exocyclic double bond (E)-8. Further insight
into the selectivity control of the latter rearrangement was obtained from the reactions of (Z)-
epimeric substrates. The new synthetic approach leading to the 1R-hydroxy epimers complements
our previously reported synthesis of the corresponding 1â-epimers, thus producing all stereoisomers
of these versatile building blocks efficiently from carvone.
In tr od u ction
CD-ring fragments, which are separately synthesized.3
Thus, a versatile and reliable method was pioneered by
Lythgoe6 that proceeded by the direct construction of the
conjugated triene system with complete stereoselectivity7
via Wittig-Horner coupling of the A-ring phosphine oxide
18 with a CD-ring ketone of type 2 (Scheme 1).9 Subse-
quently, this modular approach has been applied consis-
tently for the synthesis of a multitude of analogues with
modified CD-rings and side chains while retaining the
natural A-ring fragment intact by using the phosphine
oxide building block 1.10
The hormonally active metabolite of vitamin D3, 1R,-
25-dihydroxy vitamin D3 (3),1 has a broad spectrum of
potent biological activities that spreads across several
important therapeutic areas such as dermatology, meta-
bolic diseases, oncology, and autoimmune diseases.2
However, 3 is also a regulator of calcium homeostasis
causing dose-limiting hypercalcemia and related side
effects, which have thus far restricted its clinical use. The
search for structural analogues with attenuated calcemic
effect and improved target specificity has led to increased
synthetic activity in the field.3 Through these efforts a
number of practical synthetic routes have been developed,
mostly starting from steroid precursors or vitamin D2.4
More recently, however, extensive modifications of the
side chain and the CD-ring portion of the molecule have
led to new structures, which diverge significantly from
the natural motif and so cannot be easily derived from
natural precursors.5 Hence, in many cases the classic
approaches have become impractical and an efficient de
novo synthesis of such analogues is highly desirable.
The known total syntheses are inherently convergent,
on the basis of the final coupling of the A-ring and the
Unfortunately, however, the known preparations of 1
are long and tedious, discouraging the practical applica-
(5) (a) Guyton, K. Z.; Kensler, T. W.; Posner, G. H. Annu. Rev.
Pharmacol. Toxicol. 2001, 41, 421. (b) Gabrie¨ls, S.; Van Haver, D.;
Vandewalle, M.; De Clercq, P.; Verstuyf, A.; Bouillon, R. Chem. Eur.
J . 2001, 7, 520. (c) Uskokovic´, M. R.; Studzinski, G. P.; Gardner, J . P.;
Reddy, S. G.; Campbell, M. J .; Koeffler, H. P. Curr. Pharm. Des. 1997,
3, 99.
(6) Lythgoe, B.; Moran, T. A.; Nambudiry, M. E. N.; Tideswell, J .;
Write, P. W. J . Chem. Soc., Perkin Trans. 1 1978, 590.
(7) Kociensky, P. J .; Lythgoe, B.; Waterhouse, I. J . Chem. Soc.,
Perkin Trans. 1 1980, 1045.
(8) Baggiolini, E. G.; Iacobelli, J . A.; Hennessy, B. M.; Uskokovic´,
M. R. J . Am. Chem. Soc. 1982, 104, 2945.
(9) For recent syntheses of the CD-ring ketone fragment, see: (a)
Daniewski, A. R.; Liu, W. J . Org. Chem. 2001, 66, 626. (b) Van Gool,
M.; Vandewalle, M. Eur. J . Org. Chem. 2000, 3427. (c) Fall, Y.; Vitale,
C.; Mourin˜o, A. Tetrahedron Lett. 2000, 41, 7337. (d) Stork, G.; Ra, C.
S. Bull. Korean Chem. Soc. 1997, 18, 137. (e) J ankowski, P.; Marczak,
S.; Wicha, J . Tetrahedron 1998, 54, 12071 and references therein.
(10) For recent examples, see: (a) Calverley, M. J . Steroids 2001,
66, 249. (b) Norman, A. W.; Manchand, P. S.; Uskokovic´, M. R.;
Okamura, W. H.; Takeuchi, J . A.; Bishop, J . E.; Hisatake, J .-I.; Koeffler,
H. P.; Peleg, S. J . Med. Chem. 2000, 43, 2719. (c) Sestelo, J . P.;
Mourin˜o, A.; Sarandeses, L. A. J . Org. Chem. 2000, 65, 8290. (d)
Ferna´ndez-Gacio, A.; Vitale, C.; Mourin˜o, A. J . Org. Chem. 2000, 65,
6978. (e) Zhou, X.; Zhu, G.-D.; Van Haver, D.; Vandewalle, M.; De
Clercq, P. J .; Verstuyf, A.; Bouillon, R. J . Med. Chem. 1999, 42, 3539.
(f) Posner, G. H.; Wang, Q.; Han, G.; Lee, J . K.; Crawford, K.; Zand,
S.; Brem, H.; Peleg, S.; Dolan, P.; Kensler, T. W. J . Med. Chem. 1999,
42, 3425. (g) de los Angeles Rey, M.; Martı´nez-Pe´rez, J . A.; Ferna´ndez-
Gacio, A.; Halkes, K.; Fall, Y.; Granja, J .; Mourin˜o, A. J . Org. Chem.
1999, 64, 3196. (h) Grue-Sørensen, G.; Hansen, C. M. Bioorg. Med.
Chem. 1998, 6, 2029.
(1) The R and â designations and atom numbering used in the text
follow the conventional steroid nomenclature; see: Rose, I. A.; Hanson,
K. R.; Wilkinson, K. D.; Wimmer, M. J . Proc. Natl. Acad. Sci. U.S.A.
1980, 77, 2439.
(2) (a) 1st International Conference on Chemistry and Biology of
Vitamin D Analogues, Steroids-Special Issue 2001, 66, 127-472. (b)
Vitamin D Endocrine System: Structural, Biological, Genetic and
Clinical Aspects; Norman, A. W., Bouillon, R., Thomasset, M., Eds.;
University of California Riverside: Riverside, CA, 2000. (c) Vitamin
D, Chemistry, Biology and Clinical Applications of the Steroid Hor-
mone; Norman, A. W., Bouillon, R., Thomasset, M., Eds.; University
of California Riverside: Riverside, CA, 1997. (d) Bouillon, R.; Okamura,
W. H.; Norman, A. W. Endocrine Rev. 1995, 16, 200.
(3) For reviews, see: (a) Zhu, G.-D.; Okamura, W. H. Chem. Rev.
1995, 95, 1877. (b) Dai, H.; Posner, G. H. Synthesis 1994, 1383.
(4) Okabe, M. Chemistry of Vitamin D: A Challenging Field for
Process Research. In Process Chemistry in the Pharmaceutical Indus-
try; Gadamasetti, K. G., Ed.; Marcel Dekker: New York, Basel, 1999;
pp 73-89.
10.1021/jo0161577 CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/09/2002

Efficient Synthesis of the A-Ring Phosphine Oxide
J . Org. Chem., Vol. 67, No. 5, 2002 1581
Sch em e 1
Sch em e 2
bond. Instead, it requires several additional steps, mul-
tiple chromatographic purifications, and use of Martin’s
sulfurane for the dehydration.
As previously reported,14 we have devised a very
efficient isomerization of the epoxide moiety of the
respective 1â-dieneoxide (E)-ester to the corresponding
dienol using palladium catalysis. Thus, we anticipated
that if an efficient synthesis of the requisite 1R-dieneox-
ide (E)-7 could be achieved, its palladium-catalyzed
isomerization would directly give the desired dienol (E)-
8.
Herein, we report a new, very efficient process for the
preparation of the key building block 1 that minimizes
intermediate isolation and purification steps.
Resu lts a n d Discu ssion
tion of this useful approach. Thus, while many ingenious
syntheses of 1 have been reported,11 only very recently
has a synthesis of less than 10 steps been realized.12
Furthermore, these preparations involve multiple isola-
tions and chromatographic purifications, rendering them
unacceptable for larger kilo-scale production.
The seminal Baggiolini-Uskokovic´ synthesis (Scheme
2),13 while affording 1 in an overall yield of 21% from (S)-
carvone, requires 14 steps and at least eight chromato-
graphic purifications. In this approach, the conversion
of the 1R-epoxide (E)-4, obtained in two steps from (S)-
carvone, to the silylated diol (E)-5 required eight steps.
Whereas oxidative degradation of the isopropenyl sub-
stituent was accomplished in three steps, four steps were
needed for the formal isomerization of the dieneoxide
moiety in 4 to the allylic alcohol, with an exocyclic double
bond, in compound 5.
The 1R-dieneoxide 7 was expediently obtained from 1R-
carvone oxide 9 without intermediate purification, as
outlined in Scheme 3.
The stereoselective epoxidation of (S)-carvone has been
previously described, leading to 9 as the major diastere-
omer, in 89% yield.15 Nevertheless, following the known
procedure, we had to use a large excess of 3 equiv of 30%
hydrogen peroxide and 0.5 equiv of sodium hydroxide in
methanol at temperatures from -15 °C to 0 °C to effect
complete reaction. The low reaction temperature is
required as the product slowly decomposes under these
strongly basic reaction conditions. To reduce the amounts
of oxidant and base, tert-butyl hydroperoxide was exam-
ined as a substitute for hydrogen peroxide. It was found
that only 1.22 equiv of 70% aqueous tert-butyl hydrop-
eroxide and 0.1 equiv of 25% sodium methoxide in
methanol as the base source were needed to complete the
reaction and obtain cleanly the ca. 15-20:1 mixture of
diastereomers in 94% yield. The desired product, 9, was
then isolated in 80% yield by low-temperature (below -30
°C) crystallization from hexane,16 resulting in essentially
complete removal of the minor 1â-isomer.
Indeed, the apparent shortcoming of this process is the
inability to convert the epoxide moiety in (E)-4 directly
into the desired allylic alcohol with an exocyclic double
(11) For recent syntheses of 1, see: (a) Hatakeyama, S.; Okano, T.;
Maeyama, J .; Esumi, T.; Hiyamizu, H.; Iwabuchi, Y.; Nakagawa, K.;
Ozono, K.; Kawase, A.; Kubodera, N. Bioorg. Med. Chem. 2001, 9, 403.
(b) Miyata, O.; Nakajima, E.; Naito, T. Chem. Pharm. Bull. 2001, 49,
213. (c) Koiwa, M.; Hareau, G. P. J .; Sato, F. Tetrahedron Lett. 2000,
41, 2389. (d) Anne´, S.; Wu, Y.; Vandewalle, M. Synlett 1999, 1435 and
references therein. For earlier syntheses, see ref 3.
(14) Kabat, M. M.; Garofalo, L. M.; Daniewski, A. R.; Hutchings, S.
D.; Liu, W.; Okabe, M.; Radinov, R.; Zhou, Y. J . Org. Chem. 2001, 66,
6141.
(15) (a) Klein, E.; Ohloff, G. Tetrahedron 1963, 19, 1091. (b)
Okamura, W. H.; Aurrecoechea, J . M.; Gibbs, R. A.; Norman, A. W. J .
Org. Chem. 1989, 54, 4072.
(12) Hiyamitzu, H.; Ooi, H.; Inomoto, Y.; Esumi, T.; Iwabuchi, Y.;
Hatakeyama, S. Org. Lett. 2001, 3, 473.
(13) Baggiolini, E. G.; Iacobelli, J . A.; Hennessy, B. M.; Batcho, A.
D.; Sereno, J . F.; Uskokovic´, M. R. J . Org. Chem. 1986, 51, 3098.
(16) At ambient temperature 9 is a yellow oil, as its mp is below 0
°C.

1582 J . Org. Chem., Vol. 67, No. 5, 2002
Daniewski et al.
Sch em e 3
The subsequent Criegee rearrangement17 of peracetate
10 was key to the successful oxidative cleavage of the
isopropenyl group in 9 to unmask the requisite 3â-
alcohol. Ozonolysis of 9 at -70 °C in the presence of
methanol and in situ acetylation of the hydroperoxide
intermediate 15 gave crude peracetate 10, which could
be used directly for the Criegee rearrangement.
ature and extracted with hexane. The extracts were
washed to remove acidic and basic byproducts, which are
otherwise detrimental in the subsequent rearrangement
step.
Crude peracetate 10 thus obtained could then be used
advantageously for the Criegee rearrangement, after
solvent exchange, without further purification. Indeed,
the stereospecific rearrangement of peracetate 10 pro-
ceeded cleanly in methanol at 37 °C to give directly the
desired secondary alcohol 11. Remarkably, alcohol 11 was
the single product isolated, even at a lower temperature
of 20 °C, and its acetate was never observed when the
reaction was run in methanol. In contrast, with nonpolar
aprotic solvents such as dichloromethane or chloroform,
the reaction was sluggish and produced a complex
mixture of products.
The rearrangement of related R-methoxy hydroperoxy-
acetates has been previously carried out in aprotic
solvents such as dichloromethane,18 where the corre-
sponding acetate of the alcohol product was obtained
predominantly upon prolonged reflux. The alcohol, usu-
ally the minor product under those conditions, was
produced from hydrolysis of the acetate on aqueous
workup. Thus, it has been postulated that a dioxenium
ion intermediate such as 17 should suffer demethylation
by nucleophilic attack of its acetate counterion to gener-
ate the acetate of the alcohol, and methyl acetate as a
byproduct.
As it is not desirable to isolate these unstable inter-
mediates, the one-pot process required considerable
optimization. Excess methanol is necessary during the
ozonolysis to trap the short-lived carbonyl oxide inter-
mediate 14, which is formed via retro 1,3-dipolar cycload-
dition from ozonide 13. In the absence of methanol, 14
undergoes recombination with formaldehyde to give the
thermally more stable ozonide 16. However, since excess
methanol would interfere with the subsequent acylation
of 15 to 10, it is beneficial to minimize the amount of
methanol in the process. It was found that clean forma-
tion of 15 could be achieved with only 4 equiv of methanol
in dichloromethane. When methanol was further reduced
to 3 equiv the reaction was not as clean presumably due
to competitive formation of ozonide 16, leading to forma-
tion of the corresponding methyl ketone as a major
byproduct.
The hydroperoxide 15 thus obtained was acetylated in
situ at -5 °C with 7 equiv each of acetic anhydride and
triethylamine in the presence of a catalytic amount of
DMAP. Acylation at a higher temperature (e.g., at room
temperature) gave a product of a much darker color. The
reaction was quenched with methanol at a low temper-
In methanol, however, the transient intermediate 17
should be swiftly trapped to produce acetic acid and
orthoacetate 18 instead, which would then rapidly un-
(17) (a) Criegee, R.; Kaspar, R. J . Liebigs Ann. Chem. 1948, 560,
127. (b) Criegee, R. Angew. Chem., Int. Ed. Engl. 1975, 14, 745.
(18) Schreiber, S. L.; Liew, W.-F. Tetrahedron Lett. 1983, 24, 2363.

Efficient Synthesis of the A-Ring Phosphine Oxide
J . Org. Chem., Vol. 67, No. 5, 2002 1583
Ta ble I.
dergo acetic acid-catalyzed methanolysis to yield cleanly
alcohol 11 along with trimethyl orthoacetate as the
byproduct.19 A substantial rate acceleration in methanol
should normally result17a from the stabilization of the
highly polarized transition state leading to the dioxenium
ion 17 in this polar solvent.
1
Indeed, H NMR analysis of a reaction run in deute-
riomethanol (CD3OD) revealed the clean formation of 11,
without accumulation of intermediates, while producing
three additional singlet peaks. These peaks were as-
signed to CH3OD (δ 3.35 ppm, originating from the
methoxy group in 10), CH3CO2D (δ 1.99 ppm, originating
from the acetoxy group in 10), and CH3C(OCD3)3 (δ 1.40
ppm). As a reference, trimethyl orthoacetate (δ 3.24 and
1.40 ppm) in CD3OD slowly produced two peaks at 3.35
ppm (CH3OD) and 1.40 ppm. Methanol and acetic acid
in CD3OD have peaks at 3.35 and 1.99 ppm, respectively.
The direct formation of alcohol 11 from the rearrange-
ment of 10, when carried out in methanol, is synthetically
significant because it has been reported20 that 11 could
not be produced from its acetate by either chemical or
enzymatic hydrolysis due to the facile elimination of the
acetoxy group. In fact, attempts to produce the acetate
for reference purpose, by acylation of 11, failed completely
due to elimination.
After solvent exchange with acetonitrile, alcohol 11 was
then silylated without isolation to give ketone 12 (Scheme
3). The relatively volatile silyl byproducts were removed
at 45 °C under high vacuum, and crude 12, thus obtained
in 82% yield from epoxide 9, was directly subjected to
the Wittig-Horner reaction yielding the desired dieneox-
ide 7. The olefination was carried out using 2.2 equiv of
triethyl phosphonoacetate and 1.8 equiv of lithium hy-
dride in a concentrated THF solution at the lower
temperature of 11 °C to minimize elimination of the
silyloxy group.21
E:Z of
E:Z of
product
entry X (substrate) substrate
dieneol:enonea
1
2
3
4
5
6
7
8
9
CO2Et (7)
CO2Et (7)
CO2Et (7)
CO2Et (7)
CO2Et (7)
4:1
7:1
9:1
E only
Z only
E only: (E)-8
E only: (E)-8
E only: (E)-8
E only: (E)-8 >99:1 (8:19)
E only: (E)-8 14:86 (8:19)
E only: (E)-21 >99:1 (21:22)
E only: (E)-21 8:92 (21:22)
80:20 (8:19)
88:12 (8:19)
90:10 (8:19)
CO2t-Bu (20) E only
CO2t-Bu (20) Z only
CN (23)
CN (23)
E only
Z only
E only: (E)-24 >99:1 (24:25)
Z on ly: (Z)-24 >99:1 (24:25)
a
HPLC analysis was performed on a Nucleosil 5 µm, 4.6 × 250
mm column with 2% 2-propanol in hexane at 0.5 mL/min.
Percentages given are the area percentages of the corresponding
peaks detected at 220 nm.
epimers was subjected to the same reaction conditions,
an 88:12 mixture of (E)-dienol 8 and enone 19 was
obtained, while the corresponding (Z)-dienol product was
not observed. This unexpected result warranted further
examination.
Since the E:Z-ratio of the starting materials used
seemed to correspond exactly to the ratio of these two
products (see Table 1, entries 1-3), the pure epimers
(E)-7 and (Z)-7, as well as the related tert-butyl esters
(E)-20 and (Z)-20, and nitriles (E)-23 and (Z)-23 were
separately prepared in an attempt to clarify the issue.
Indeed as expected, the (E)-ester substrates, ethyl ester
(E)-7 and tert-butyl ester (E)-20, gave the (E)-dienols,
(E)-8 and (E)-21, respectively (Table 1), with high selec-
tivities (>99:1) as previously noted for the corresponding
1â-epimers.14 However, the (Z)-esters, (Z)-7 and (Z)-20,
gave the corresponding enones 19 and 22 as the major
products. Moreover, the minor products from the rear-
rangement of the (Z)-esters were the (E)-dienols, (E)-8
and (E)-21, respectively, while the epimeric (Z)-dienols
were undetectable. In contrast, however, the epimeric
nitriles 23 rearranged in a stereospecific fashion, as (E)-
23 gave only dienol (E)-24; (Z)-23 gave dienol (Z)-24
exclusively, and the corresponding enone 25 was never
observed.
These results suggested that the inverse selectivity for
the palladium-catalyzed isomerization of the (Z)-esters
may be due to a different coordination of palladium,
which is available for the (Z)-esters, but not for the (E)-
esters and the nitriles. Hence, for the (Z)-esters, unlike
the (E)-esters, internal coordination of the ester carbonyl
to the allylic palladium such as in intermediate 26 would
promote the sterically favored equilibration of 26 to the
palladium enolate 27 with concomitant loss of stereo-
chemistry at the exocyclic double bond. Subsequent
Under these conditions, 7 was obtained in 76% yield
as a mixture of E/ Z epimers in a ratio of 8.5:1. However,
a slight variation in yield in the range of 70-85% was
noted with reaction time, which also affected the E:Z
ratio of products. Thus, the E:Z ratio varied accordingly
from 9:1 to 7:1, with higher yields correlating with lower
selectivity, suggesting that selective decomposition of the
minor (Z)-epimer was occurring under the reaction condi-
tions. The concentration of the reaction mixture also
affected the E:Z ratio, as a 2-fold dilution resulted in a
ca. 2-fold decrease in selectivity for the (E)-epimer. Other
phoshonoacetates, i.e., (EtO)2P(O)CH2CO2Me, (EtO)2P-
(O)CH2CO2t-Bu, (MeO)2P(O)CH2CO2Me, and (i-PrO)2P-
(O)CH2CO2Et, as well as (EtO)2P(O)CH2CN were exam-
ined but gave lower yields and selectivities.
We have previously described14 the palladium-cata-
lyzed isomerization of the corresponding 1â-dieneoxide
(E)-esters, which produced dienols exclusively over enones
when using a fluoro alcohol cocatalyst. However, the
effects of the relative configuration of the double bond
and C-1 of the substrates on selectivity had not been
clarified. When 1R-dieneoxide 7 as a 7:1 mixture of E/Z
(19) The ionic decomposition of 2-methoxy-2-propyl per-p-nitroben-
zoate in methanol has been reported to produce quantitatively methyl
acetate and p-nitrobenzoic acid: Hedaya, E.; Winstein, S. Tetrahedron
Lett. 1962, 563.
(20) De Brabander, J .; Kulkarni, B. A.; Garcia-Lopez, R.; Vande-
walle, M. Bull. Soc. Chim. Belg. 1997, 106, 665.
(21) When sodium ethoxide was used as the base, significant
amounts of tert-butyldimethylsilanol were produced via elimination.

1584 J . Org. Chem., Vol. 67, No. 5, 2002
Daniewski et al.
Sch em e 4
elimination from 27 would occur via enolate 28 to form
enone 19 preferentially,22 along with a lesser amount of
the corresponding dienol (E)-8. As this internal equilibra-
tion is not possible for the nitrile (Z)-23, stereospecific
isomerization leads to dienol (Z)-24 exclusively by â-hy-
dride elimination from the methyl group as previously
described.14
to be equally effective as 9-fluorenone but easily removed
from the product by filtration through silica gel. Thus,
after photolysis with 10 mol % of 9-fluorenone-4-carboxy-
lic acid followed by filtration, crystallization from aceto-
nitrile afforded the known13 alcohol (Z)-6 in greater than
99% purity.
The isomeric 9-fluorenone-2-carboxylic acid and 9-fluo-
renone-1-carboxylic acid were less effective as sensitizers
in this reaction, giving 98.2 and 96.9% conversion,
respectively. 9-Anthracenecarboxylic acid was virtually
inactive. Alternatively, the previously described photoi-
somerization of ester (E)-513 was examined but was found
to be inefficient. For instance, with 9-fluorenone and
9-acetylanthracene the highest conversions achieved
were 88.8 and 11.7%, respectively.
Alcohol (Z)-6 was converted to the desired phosphine
oxide 1 as previously described.23 Crude phosphine oxide
1 was crystallized from a mixture of hexane and cyclo-
hexane as a 2:1 solvate with cyclohexane, mp 54-55 °C.
However, because recovery from this crystallization (ca.
60%) was modest, its chromatographic purification was
preferred instead, increasing the yield of 1 to 84% over
these two steps.
In summary, we have developed a new practical
preparation of the vitamin D building block 1, which is
useful for the synthesis of many vitamin D analogues
bearing this structural motif via the dependable Lythgoe
method. The process uses inexpensive (S)-carvone as the
starting material, is scaleable, and requires a minimal
number of chromatographic purification steps. Using this
process, we obtained the key precursor (E)-7 in five
synthetic operations and 42% overall yield from (S)-
carvone, which compares very favorably to previous
syntheses.11 Ultimately, phosphine oxide 1 is obtained
in nine synthetic operations and 25% overall yield from
(S)-carvone, requiring only one chromatographic purifica-
tion, that of the final product 1. While carvone oxide 9
and alcohol (Z)-6 were crystallized, intermediate purifica-
tion, where necessary (as of 7, (E)-5, (E)- and (Z)-6), could
be achieved by simple filtration through silica to remove
polar impurities, obviating extensive chromatographic
separation.
While high selectivity was achieved with the pure (E)-
epimer 7, the separation of the E/Z epimers was not
practical. Thus, in practice, the mixture was subjected
to the rearrangement conditions, and after solvent ex-
change with DMF, the resulting mixture of dienol (E)-8
and enone 19 was subjected to silylation (Scheme 4).
Since dienol (E)-8 was converted to the nonpolar product
(E)-5, while the polar enone 19 remained unchanged,
pure (E)-5 was then isolated, by simple silica gel filtra-
tion, in 64% yield over the three steps from epoxyketone
12.
DIBALH reduction of ester (E)-5 in hexane unevent-
fully produced allylic alcohol (E)-6, which was photoi-
somerized to (Z)-6 in tert-butyl methyl ether, in the
presence of a sensitizer. With 10 mol % of 9-fluorenone,
greater than 98.7% conversion was achieved. However,
this sensitizer was difficult to remove from the product
without chromatography. Hence, other sensitizers were
evaluated, and 9-fluorenone-4-carboxylic acid was found
The present synthesis, affording the 1R-hydroxy A-ring
alcohols 6 with the E- and Z-configuration of the substi-
(22) Coordination of the cationic palladium in ester enolate 27 to
the 1R-OH group is likely to favor the â-hydride migration yielding
enolate 28 (cf. ref 14). For an elimination of a palladium O-enolate to
enone in the context of an intramolecular redox reaction, see: Ho¨ge-
nauer, K.; Mulzer, J . Org. Lett. 2001, 3, 1495.
(23) For improved preparation of A-ring phosphine oxides from
allylic alcohols, see: Daniewski, A. R.; Garofalo, L. M.; Kabat, M. M.
Synth. Commun. 2002, in press.

Efficient Synthesis of the A-Ring Phosphine Oxide
J . Org. Chem., Vol. 67, No. 5, 2002 1585
9 (20.0 g, 120 mmol) and methanol (20 mL, 494 mmol) in
dichloromethane (200 mL) was cooled with a dry ice/acetone
bath; the nitrogen inlet tube was replaced with a gas disper-
sion tube, and the gas outlet tube was connected to a trap
through a wide 4 mm i.d. tube immersed in a 1 M solution of
potassium iodide (2 L). Then, ozonized air (4 LPM) was
continuously passed through the reaction mixture at -68 ( 3
°C. The reaction turned pale blue after 65 min, indicating
complete reaction. Then, excess ozone was removed by purging
with nitrogen (4 LPM) for 30 min, and the mixture was allowed
to warm to 14 °C over 40 min to ensure complete conversion
of the initially formed ozonide 14 to the desired hydroperoxide
15. Then, the mixture was cooled to -25 °C, and triethylamine
(117 mL, 839 mmol) was added over 5 min while maintaining
the temperature of the mixture below -25 °C. Then, DMAP
(2.0 g, 16.4 mmol) was added in one portion, followed by the
slow addition of acetic anhydride (79.6 mL, 843 mmol) over
10 min while maintaining the reaction temperature between
-25 and -38 °C. The mixture was allowed to warm to -8 °C
over 30 min and stirred at -7 ( 1 °C for 1.5 h. TLC analysis
indicated complete reaction. The reaction was quenched by the
slow addition (over 7 min) of methanol (33 mL) while main-
taining the temperature of the mixture below 10 °C. After
stirring at 5 °C for 5 min, the mixture was diluted with hexane
(220 mL), washed consecutively with 10% aqueous citric acid
(2 × 150 mL) and saturated aqueous KHCO3 (2 × 80 mL),
dried over Na2SO4, and concentrated to dryness at 30 °C under
reduced pressure to give 38.2 g (overweight) of crude 10 as a
yellow oil. This material is relatively unstable and was
immediately used in the next step without further purification.
TLC (2:1 hexanes-EtOAc; PMA stain): Rf 9 ) 0.8 and Rf 15
) 0.45. TLC (40:2:1 dichloromethane-ethyl acetate-methanol;
PMA stain): Rf 15 ) 0.4 and Rf 10 ) 0.8.
(1S,4S,6S)-4-[[(1,1-Dim eth yleth yl)d im eth ylsilyl]oxy]-1-
m eth yl-7-oxa bicyclo[4.1.0]h ep ta n -2-on e (12). Crude 10 as
obtained above (38.2 g, 120 mmol in theory) and NaOAc (2 g,
24.4 mmol) in methanol (245 mL) were stirred at 37 °C
overnight. TLC analysis indicated complete reaction. Then, the
mixture was concentrated to dryness at 39 °C, and the residue
(29 g) was dissolved in acetonitrile (40 mL). The resulting
solution was concentrated to dryness at 35 °C under reduced
pressure, and additional acetonitrile (40 mL) was added. The
resulting solution was again concentrated to dryness at 35 °C
under reduced pressure, and then acetonitrile (35 mL) and
imidazole (29.5 g, 433 mmol) were added. After the mixture
was cooled with an ice-water bath, tert-butylchlorodimethyl-
silane (32.6 g, 217 mmol) was added. The cold bath was
removed, and the mixture was stirred at room temperature
for 4 h. TLC analysis indicated the presence of only a trace
amount of 11. The reaction was quenched by the addition of
methanol (10 mL). A mild exotherm ensued, which raised the
temperature of the mixture by 2 °C. After the mixture was
stirred for 5 min, ice-water (55 mL) was added and the
mixture was extracted with hexanes (2 × 50 mL). The
combined organic extracts were washed with a 2:3 v/v mixture
of methanol and water (50 mL), dried over Na2SO4, and
concentrated to dryness at 40 °C under reduced pressure.
Further drying of the residue at 46 °C and 0.4 mmHg for 1 h
gave 25.2 g (81.7% from 9) of crude 12 as a pale yellow oil.
This material, which according to NMR analysis contained
silanol (ca. 2%; δ 0.09 ppm) and siloxane (ca. 12%; δ 0.00 ppm)
byproducts, was used directly in the next step without further
purification. TLC (40:2:1 dichloromethane-ethyl acetate-
methanol; PMA stain): Rf 10 ) 0.8, Rf 11 ) 0.4, and Rf 12 )
0.95.
tuted exocyclic double bond, complements our previously
reported synthesis of the corresponding 1â-epimers,14
thus making all stereoisomers of the versatile A-ring
phosphine oxide building blocks readily available23 from
the respective enantiomers of carvone.
Exp er im en ta l Section
Gen er a l Ma ter ia ls a n d P r oced u r es. All reactions were
performed in dried glassware under a positive pressure of
nitrogen. Reaction extracts and chromatography fractions were
concentrated using a rotary evaporator at approximately 10
Torr using a diaphragm pump and then at high vacuum to
approximately 0.01 Torr using an oil pump. TLC analysis was
performed using silica gel 60 F254 precoated glass plates (EM
Science) and detected by UV254 or phosphomolybdic acid (PMA)
stain. 1H NMR spectra were recorded at 300 MHz. CDCl3 was
treated with basic alumina prior to use. Melting points were
obtained on a capillary melting point apparatus and are
uncorrected. A Polymetrics Laboratory Ozonator model T-816
instrument was used to generate ozonized air (shell pressure
) 6 PSIG; flow rate ) 4 LPM; 110 V). Commercial grade
reagents and solvents were used without purification, except
as indicated. Silica gel 60, particle size 0.040-0.063 mm (230-
400 mesh), was used. The preparation of diphenylphosphine
oxide has been previously described.14
(1S,4S,6S)-1-Meth yl-4-(1-m eth yleth en yl)-7-oxa bicyclo-
[4.1.0]h ep ta n -2-on e (9).15 A 70% solution of tert-butyl hy-
droperoxide in water (105 g, 816 mmol) was cooled to 5 °C,
and a 25% solution of sodium methoxide in methanol (15 mL,
65.3 mmol) was added in one portion. A mild exotherm ensued
and raised the temperature of the mixture to 7 °C. After the
mixture was cooled to 5 °C, a mixture of (S)-carvone (100 g,
666 mmol) and methanol (40 mL) was added dropwise over 2
h while maintaining the reaction temperature between 3 and
6 °C. The mixture was stirred at 3 °C for 3 h. At this point,
TLC analysis indicated ca. 80% conversion. After the mixture
was stored in an ice-water bath in a refrigerator overnight,
TLC analysis indicated the presence of only a trace amount
of starting material. Then, 5% aqueous sodium bicarbonate
(200 mL) was added over 10 min while maintaining the
temperature of the mixture below 10 °C. The mixture was
cooled to 3 °C, and an aqueous sodium dithionite solution,
prepared by dissolving 85% sodium dithionite (12 g, Stench!)
in water and dilution to 100 mL, was added over 7 min while
maintaining the temperature of the mixture below 12 °C. After
10 min, the cooling bath was removed. After stirring for 30
min, the mixture was extracted with tert-butyl methyl ether
(2 × 100 mL). The combined organic extracts were washed
with a mixture of saturated aqueous NaHCO3 (100 mL) and
saturated aqueous NaCl (100 mL), dried over Na2SO4, and
concentrated to dryness at 30 °C under reduced pressure. The
residue was dissolved in hexane (100 mL), and the resulting
solution was concentrated to dryness at 30 °C under reduced
pressure. The resulting yellow oil was dissolved in hexane (500
mL), and the stirred solution was cooled with a dry ice-
acetone bath. At -30 °C, crystallization commenced to give a
thick suspension. The mixture was then diluted further with
hexane (100 mL). After the mixture was stirred at -70 °C for
15 min, the solid was collected by filtration using a sintered
glass filter (pre-cooled with dry ice) and quickly washed with
cold (pre-cooled to -70 °C) hexane (200 mL). Then, the solid
was allowed to melt at ambient temperature. The resulting
oil was filtered through the glass filter, which was rinsed with
hexane (50 mL). The combined filtrate and washes were
concentrated at 30 °C under reduced pressure. The resulting
residue was dried at room temperature for 1.5 h under high
vacuum to give 88.0 g (79.5%) of 9 as a pale yellow oil. NMR
analysis of this material indicated the presence of a trace
amount of the 1â-diastereomer, while the mother liquor of the
crystallization was found to contain a 2:1 mixture of 9 and its
1â-diastereomer. TLC (9:1 hexanes-ethyl acetate; short-wave
UV detection): Rf (S)-carvone ) 0.4 and Rf 9 ) 0.35.
E t h yl 2-[(1S,4S,6S)-1-Met h yl-7-oxa -4-[[(1,1-d im et h yl-
eth yl)d im eth ylsilyl]oxy]-bicyclo[4.1.0]h ep t-2-ylid en e]a c-
eta te (7). Caution! When handling lithium hydride powder,
use a protective mask to prevent accidental inhalation. A
mixture of lithium hydride (1.41 g, 177 mmol) and triethyl
phosphonoacetate (43.3 mL, 216 mmol) in THF (45 mL) was
heated slowly to 55 °C (Caution! Exotherm), and then the
heating bath was removed. An exotherm ensued, which raised
the temperature of the mixture to 69 °C over 5 min. The
temperature of the mixture slowly decreased to 66 °C over 55
(1S,4S,6S)-4-(1-Acetylh yd r op er oxy-1-m eth oxyeth yl)-1-
m eth yl-7-oxa bicyclo[4.1.0]h ep ta n -2-on e (10). A solution of

1586 J . Org. Chem., Vol. 67, No. 5, 2002
Daniewski et al.
min, and a clear solution resulted. Approximately 25 mL of
THF was then removed by distillation at 50-55 °C under a
slightly reduced pressure.24 After the resulting mixture was
cooled to 3 °C with an ice water bath, crude 12 as obtained
above (25.2 g, 98.4 mmol in theory) was added in one portion
with the aid of THF (15 mL). The mixture was stirred at 5-6
°C for 90 min, at 11 °C for 18 h, and at 24 °C for 2 h. TLC
analysis indicated complete reaction. Then, the mixture was
diluted with 8:1 v/v hexanes-EtOAc (100 mL), washed with
water (3 × 36 mL), and concentrated to dryness at 38 °C under
reduced pressure. The residue was dissolved in hexane (115
mL) and filtered through silica gel (50 g). The silica plug was
then washed with 8:1 v/v hexanes-EtOAc (191 mL), and the
combined filtrate and washes were concentrated to dryness
at 37 °C under reduced pressure. The residue was further dried
under high vacuum for 1 h to give 24.4 g (76.1%) of crude 7 as
(2E)-2-[(3S,5R)-3,5-Bis[[(1,1-d im eth yleth yl)d im eth ylsi-
lyl]oxy]-2-m eth ylen ecycloh exyliden e]-eth an ol ((E)-6). (E)-5
(27.7 g, 62.9 mmol) was dissolved in hexane (250 mL). After
the mixture was cooled to -75 °C, a solution of 1 M DIBALH
in hexanes (157 mL, 157 mmol) was added dropwise over 45
min while maintaining the temperature of the reaction mixture
below -65 °C. This resulting mixture was then stirred under
nitrogen at -70 °C for 30 min. TLC analysis indicated
complete reaction. The cooling bath was removed, and the
reaction mixture was allowed to warm to 0 °C. While the
mixture was cooled with an ice-water bath, cold water (380
mL) was added slowly over 5 min (Caution! Gas evolution).
The cooling bath was then removed, and the mixture was
stirred at room temperature for 1.5 h. The resulting suspension
was diluted with EtOAc (300 mL), and the mixture was
washed with 0.25 N HCl (2 × 300 mL), causing most of the
solids to dissolve. The combined aqueous washes were diluted
with 0.5 N HCl (120 mL), decreasing the pH from 5 to 2, and
then extracted with EtOAc (300 mL). The organic phase and
extract were combined, washed with saturated aqueous NaCl
solution (300 mL), dried over MgSO4, and concentrated to
dryness at 35 °C under reduced pressure to give 27 g of the
crude product. This material was dissolved in hexane (25 mL)
and filtered through silica gel (25 g). The silica plug was then
washed with 10-100% EtOAc-hexane, and the combined
filtrate and washes were concentrated to dryness at 35 °C
under reduced pressure. The residue was further dried under
high vacuum to give 23.7 g (94.5%) of (E)-6 as a colorless solid.
TLC (9:1 hexanes-EtOAc; short-wave UV detection and PMA
stain): Rf (E)-5 ) 0.6 and Rf (E)-6 ) 0.2.
(2Z)-2-[(3S,5R)-3,5-Bis[[(1,1-d im eth yleth yl)d im eth ylsi-
lyl]oxy]-2-m eth ylen ecycloh exylid en e]-eth a n ol ((Z)-6).13
A 500 mL photoreaction vessel equipped with a cooling jacket,
Pyrex immersion well, nitrogen inlet tube (immersed deep into
the reaction mixture), and outlet bubbler was charged with
crude (E)-6 as obtained above (23.7 g, 59.4 mmol), tert-butyl
methyl ether (100 mL), and a solution of 9-fluorenone-4-
carboxylic acid (1.33 g, 5.94 mmol) in tert-butyl methyl ether
(150 mL). While water was rapidly passed through the well
and the reaction vessel was immersed in a 15 °C water bath,
the above solution was irradiated with a 450 W medium-
pressure mercury lamp through a uranium filter. During the
photolysis, nitrogen was continuously bubbled through the
reaction solution via the inlet tube. The reaction was monitored
by HPLC analysis, which indicated that a photostationary
state was reached at 98.6% conversion after 6 h of irradiation.
The reaction mixture was concentrated to dryness at 30 °C
under reduced pressure and then dried under high vacuum
for 30 min. The residue was dissolved in 9:1 hexanes-EtOAc
(50 mL) and filtered through silica gel (50 g). The silica plug
was then washed with 9:1 EtOAc-hexane (750 mL), and the
combined filtrate and washes were concentrated to dryness
at 35 °C under reduced pressure. The residue was further dried
under high vacuum to give 21.2 g (89.3%) of crude (Z)-6 as an
off-white solid. This material was dissolved in acetonitrile (22
mL) at reflux, and the resulting solution was stored in a freezer
at -20 °C overnight. The resulting solid was collected by
filtration, washed with ice-cold acetonitrile (3 × 3 mL), and
dried by suction and then under vacuum to give 18.3 g (77.2%)
of (Z)-6 as a colorless solid, mp 68-70 °C. HPLC analysis
indicated that this material contained 0.36% of (E)-6 and
0.24% of the 1â-epimer. HPLC (column ) ES SI 3 µm, 5 ×
150 mm; mobile phase ) 7% THF in heptane at 1 mL/min;
detection at 220 nm): Rt (Z)-6 ) 10.5, Rt (E)-6 ) 11.9, Rt )
13.3 (1â-epimer) min.
1
a yellow oil. H NMR analysis indicated this material to be a
8.5:1 mixture of E/ Z epimers. This material was used directly
in the next step without further purification. TLC (3:1 dichlo-
romethane-hexane; short-wave UV detection and PMA
stain): Rf 12 ) 0.55, Rf (E)-7 ) 0.45, and Rf (Z)-7 ) 0.35.
Eth yl (2E)-2-[(3S,5R)-3,5-Bis[[(1,1-d im eth yleth yl)d im -
eth ylsilyl]oxy]-2-m eth ylen e-cycloh exyliden e]acetate ((E)-
5).13 Tris(dibenzylideneacetone)dipalladium(0)-chloroform ad-
duct (388 mg, 0.375 mmol) and triphenylphosphine (985 mg,
3.75 mmol) were combined in a reaction flask. The flask was
evacuated and refilled with nitrogen three times, and then
toluene (23 mL) was added via a syringe. The resulting deep
purple mixture was stirred at ambient temperature for 1 h to
give a yellow slurry (Note: It is critical to allow enough time,
usually from 30 min to 1 h, for the formation of the active
catalyst before proceeding with the reaction. A color change
from deep purple to yellow, as well as disappearance of purple
particles, indicates complete catalyst formation. Preparation
in a more dilute solution is not recommended as it makes it
more difficult for the active catalyst to form). Then, 1,3-bis-
(1,1,1,3,3,3-hexafluoro-2-hydroxypropyl)benzene (0.37 mL, 1.5
mmol) was added. The slurry became red-orange. After three
minutes of stirring at ambient temperature (19 °C), a solution
of crude 7, as obtained above (24.4 g, 74.9 mmol in theory) in
toluene (100 mL) prepared under nitrogen, was added to the
catalyst solution via cannula using a slight positive nitrogen
pressure. After 10 minutes of stirring at ambient temperature,
under a slight positive pressure of nitrogen, the reaction
mixture was heated for 16 h at 40 °C. TLC analysis indicated
complete reaction. The mixture was concentrated on a rotary
evaporator at <40 °C under reduced pressure to remove most
of the toluene. The resulting brown oil25 was dissolved in DMF
(80 mL), and the resulting solution was cooled in an ice-water
bath; then, imidazole (6.12 g, 89.8 mmol) followed by tert-
butylchlorodimethylsilane (13.5 g, 89.8 mmol) was added. After
10 min, the cooling bath was removed and stirring was
continued at room temperature for 16 h. TLC analysis indi-
cated complete reaction. The reaction mixture was diluted with
hexane (300 mL) and washed with water (2 × 150 mL). The
combined aqueous washes were extracted with hexane (2 ×
100 mL), and the combined extracts were also washed with
water (2 × 50 mL). The organic phase and extracts were then
combined, dried over MgSO4, and concentrated to dryness
under reduced pressure. The residue was dissolved in hexane
(100 mL) and filtered through silica gel (200 g). The silica plug
was then washed with 98:2 v/v hexanes-EtOAc (1.5 L), and
the combined filtrate and washes were concentrated to dryness
under reduced pressure. The residue was further dried under
high vacuum for 1 h to give 27.7 g (84.0%) of (E)-5 as a colorless
oil. TLC (3:1 petroleum ether-Et2O; short-wave UV detection
and PMA stain): Rf 8 ) 0.45, Rf 5 ) 0.9, Rf 7 ) 0.85, Rf 19 )
0.6, and Rf of dba ) 0.7.
[[(1R,3S,5Z)-5-(2-Ch lor oet h ylid en e)-4-m et h ylen e-1,3-
cycloh exa n ed iyl]bis(oxy)]bis(1,1-d im eth yleth yl)d im eth -
ylsila n e.13 Caution! This reaction should be performed in a
well-ventilated hood. (Z)-6 (18.2 g, 45.6 mmol), hexanes (250
mL), and triphosgene (6.76 g, 22.8 mmol) were combined in a
reaction flask. The resulting solution was then cooled with an
ice-water bath, and after a clear solution resulted, triethy-
lamine (22.3 mL, 160 mmol) was added dropwise over 10 min
with vigorous stirring. After the mixture was stirred at 5 °C
for 20 min, the cooling bath was removed and the resulting
(24) When the reaction was carried out without removal of the THF,
the subsequent reaction was slow and the E/Z ratio of the product was
lower (<7:1).
(25) This material contained, by HPLC analysis, (E)-8 (Rt ) 17.6
min), ca. 11% of the enone byproduct 19 (Rt ) 10.2 min), and ca. 1.4%
of the 1â-epimer of (E)-8 (Rt ) 13.4 min).

Efficient Synthesis of the A-Ring Phosphine Oxide
J . Org. Chem., Vol. 67, No. 5, 2002 1587
thick suspension was stirred at room temperature for 1 h. TLC
analysis indicated complete reaction. Then, the reaction
mixture was diluted with hexanes (150 mL) and washed
consecutively with ice-cold 0.25 N HCl (2 × 250 mL) and water
(2 × 250 mL). The combined aqueous washes were back-
extracted with hexanes (2 × 100 mL). The organic extracts
were combined, washed with saturated aqueous NaCl solution
(150 mL), dried over MgSO4, and concentrated to dryness at
30 °C under reduced pressure. The residual mixture was then
purged with nitrogen for 15 min to give 19.2 g (overweight) of
crude product as a slightly hazy, yellow oil. This material
solidified upon being stored overnight in the freezer and was
directly used to the next step without further purification. This
product is relatively unstable at room temperature and,
therefore, should be stored in the freezer. TLC (9:1 hexanes-
EtOAc; short-wave UV detection and PMA stain): Rf (Z)-6 )
0.2 and chloride product Rf ) 0.6.
[(2Z)-2-[(3S,5R)-5-Bis[[(1,1-d im eth yleth yl)d im eth ylsi-
ly l]o x y ]-2-m e t h y le n e c y c lo h e x y lid e n e ]-e t h y l]d ip h e -
n ylp h osp h in e Oxid e (1).13 DMF (170 mL) and sodium
hydride, 60% dispersion in mineral oil (2.02 g, 50.6 mmol),
were combined in a reaction flask. Then, diphenylphosphine
oxide (10.2 g, 50.6 mmol) was added in one portion. Gas
evolution was observed, and a mild exotherm raised the
temperature of the mixture to 28 °C. The mixture was stirred
for 50 min at room temperature to give a slightly cloudy yellow
solution. After the solution was cooled to -45 °C with a dry
ice-acetone bath, a solution of the crude product from the
previous step (19.2 g, 45.2 mmol in theory) in DMF (70 mL)
was added dropwise over 25 min while maintaining the
temperature of the reaction mixture below -35 °C. The
reaction mixture was stirred for 1.5 h at temperatures from
-30 to -35 °C and then allowed to warm to 0 °C and stirred
at that temperature for 30 min. TLC analysis indicated
complete reaction. The reaction mixture was diluted with
diethyl ether (500 mL) and washed with water (2 × 200 mL).
The combined aqueous washes were extracted with diethyl
ether (2 × 150 mL), and these extracts were also washed with
water (2 × 200 mL). The organic phase and extracts were then
combined, dried over MgSO4, and concentrated at 35 °C under
reduced pressure to give 26.2 g of a cloudy yellow oil. This
material was dissolved in hexanes (50 mL), and the resulting
solution was filtered through silica gel (150 g). The silica plug
was then washed consecutively with hexane (200 mL), 9:1
hexanes-EtOAc (1 L), 8:2 hexanes-EtOAc (1 L), and 7:3
hexanes-EtOAc (1 L). The appropriate fractions were com-
bined and concentrated to dryness at 35 °C under reduced
pressure. The residue was further dried under high vacuum
to give 22.3 g (83.7% over two steps) of 1 as a colorless foam.
TLC (1:1 hexanes-EtOAc; short-wave UV detection and PMA
stain): Rf chloride substrate ) 0.95 and Rf 1 ) 0.45.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra of
nine compounds, obtained as indicated in the Experimental
Section. This material is available free of charge via the
Internet at http://pubs.acs.org.
J O0161577