Efficient synthesis of 1α-fluoro A-ring phosphine oxide, a useful building block for vitamin D analogues, from (S)-carvone via a highly selective palladium-catalyzed isomerization of dieneoxide to dieneol

Article abstract of DOI:10.1021/jo015788y

The 1α-fluoro A-ring phosphine oxide 1, a useful building block for fluorinated vitamin D analogues, was synthesized from (S)-carvone in 13 synthetic steps, and only five isolations, in 22% overall yield. In the key synthetic step, a highly selective palladium-catalyzed isomerization of dieneoxide 18 to dieneol 20 was achieved using an appropriately selected fluorinated alcohol as a catalytic proton source.

Full text of DOI:10.1021/jo015788y

J . Org. Chem. 2001, 66, 6141-6150
6141
Efficien t Syn th esis of 1r-F lu or o A-Rin g P h osp h in e Oxid e, a Usefu l
Bu ild in g Block for Vita m in D An a logu es, fr om (S)-Ca r von e via a
High ly Selective P a lla d iu m -Ca ta lyzed Isom er iza tion of Dien eoxid e
to Dien eol
Marek M. Kabat, Lisa M. Garofalo, Andrzej R. Daniewski, Stanley D. Hutchings, Wen Liu,
Masami Okabe, Roumen Radinov,* and Yuefen Zhou
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 May 23, 2001
The 1R-fluoro A-ring phosphine oxide 1, a useful building block for fluorinated vitamin D analogues,
was synthesized from (S)-carvone in 13 synthetic steps, and only five isolations, in 22% overall
yield. In the key synthetic step, a highly selective palladium-catalyzed isomerization of dieneoxide
18 to dieneol 20 was achieved using an appropriately selected fluorinated alcohol as a catalytic
proton source.
In tr od u ction
proceeds in seventeen steps from (S)-carvone and in an
overall yield of 3.9%.6 More recently, 1 has been prepared
in eleven steps from vitamin D3 in a lower yield of 2.4%.7
Finally, an alternative approach from (S)-carvone has
been developed, using a more effective fluorination
procedure but, overall, resulted in a much longer and
lower-yielding process.8
A number of vitamin D analogues containing the 1R-
fluorosubstituted A-ring fragment1 have been shown to
possess interesting biological activities.2 Earlier studies
have shown that 1R-fluoro-25-hydroxyvitamin D3 has
strong phagocytic activity toward human promyelocytic
leukemia cells.3 More recently, modified D-ring and side
chain analogues, such as Ro 26-9228 (3), were reported
to increase bone mineral density, while showing reduced
calciuria and calcemia in rats compared to 1,25-dihy-
droxyvitamin D3,4 and have been under evaluation for
the treatment of osteoporosis. Closely related compounds
were found to stimulate HL-60 cell differentiation, and
could be of use for the treatment of leukemia.5
The A-ring phosphine oxide 16 (Scheme 1) is a useful
building block for the synthesis of 1R-fluoro vitamin D
analogues, such as 3, via Wittig-Horner coupling with
the appropriate CD-ring portion 2, followed by desilyla-
tion.
In the first synthesis (Scheme 2),6 the conversion of
the 1R-epoxide 5 to allylic alcohol 9 required 10 steps.
Four steps were needed for the formal isomerization of
the dieneoxide moiety in 5 with inversion of configuration
at C-1 to obtain the 1â-epimeric allylic alcohol with
exocyclic double bond in 9. The remaining six steps from
7 to 8 were necessary for the oxidative degradation of
the isopropenyl substituent to the corresponding silylated
alcohol, which also involved the sequential protection of
the other two hydroxyl groups in the process.
We envisioned that the appropriate dieneoxide moiety,
such as in the 1â-epimer 11, could be isomerized directly
to the requisite allylic alcohol 12 with an exocyclic double
bond using palladium catalysis. In addition, this ap-
proach obviates the extensive protecting group manipu-
lation, as no intermediary protection of the alcohol group
would be needed.
Three syntheses of 1 have been reported thus far. The
first synthesis, and the most efficient of the three,
* Corresponding author. Tel: (973) 235-3292. Fax: (973) 235-7239.
(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) Vickery, B. H.; Avnur, Z.; Uskokovic´, M.; Peleg, S. Proc.
Workshop Vitam. D 2000, 11th (Vitam. D: Endocrine System), 887-92.
(b) Kobayashi, Y.; Taguchi, T. Proc. Workshop Vitam. D 1988, 7th
(Vitam. D: Mol., Cell. Clin. Endocrinol.), 3-11.
(3) (a) Oshima E.; Sai, H.; Takatsuto, S.; Ikekawa, N.; Kobayashi,
Y.; Tanaka, Y.; DeLuca H. F. Chem. Pharm. Bull. 1984, 32, 3525. (b)
Oshima E.; Takatsuto, S.; Ikekawa, N.; DeLuca H. F. Chem. Pharm.
Bull. 1984, 32, 3518.
(4) (a) Nestor, J . J ., J r.; Manchand, P. S.; Uskokovic´, M. R.; Vickery,
B. H. US Patent US 5872113 1999; CAN 130 168545 1999. (b)
Manchand, P. S.; Nestor, J . J ., J r.; Uskokovic´, M. R.; Vickery, B. H.
Eur. Pat. Appl. EP 808832 1996; CAN 128 34927 1997.
(5) Iacobelli, J . A.; Uskokovic´, M. R. US Patent US 6030963 2000;
CAN 132 175819 2000.
Noyori et al. have reported that simple alkyl-substi-
tuted 1,3-diene monoepoxides (dieneoxides) can isomerize
(7) Kiegiel, J .; Wovkulich, P. M.; Uskokovic´, M. R. Tetrahedron Lett.
1991, 32, 6057.
(6) Shiuey, S.-J .; Kulesha, I.; Baggiolini, E. G.; Uskokovic´, M. R. J .
Org. Chem. 1990, 55, 243.
(8) Barbier, P.; Mohr, P.; Muller, M.; Masciadri, R. J . Org. Chem.
1998, 63, 6984.
10.1021/jo015788y CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/11/2001

6142 J . Org. Chem., Vol. 66, No. 18, 2001
Kabat et al.
Sch em e 1
Sch em e 2
to dieneols, or to enones, under the influence of a catalytic
amount of tetrakis(triphenylphosphine)palladium [Pd-
(PPh3)4].9 The course of such rearrangements was typi-
cally dependent on the structure of the substrates, and
to some extent on the ligands used.10 However, no clear
rationale for selectivity control is available in order to
direct the reaction of a given substrate to the desired
dieneol product.
We wish to report here the complete selectivity control
of the isomerization of 11 to 12, and its application for a
highly efficient synthesis of 1, aimed at minimizing
intermediate isolation and purification steps.
Stereoselective aldol addition of ethyl acetate to (S)-
carvone (4) promoted by cerium(III) chloride has been
described,11 in which 2 equivalents of both its lithium
enolate and the expensive cerium salt were used. We
found that the amounts of these reagents could be
reduced to 1.3 equivalents each and complete conversion
of 4 to inter alia 13 still occurred when using tert-butyl
acetate as the reagent. However, further reduction in the
equivalents of the cerium(III) chloride, while using 1.3
equiv of the enolate, resulted in incomplete reaction; ca.
50% conversion with 0.5 equiv of CeCl3, while no reaction
occurred without this additive. Nevertheless, to our
pleasant surprise, complete conversion to 13 was achieved
when 0.5 equiv of the relatively inexpensive cerium(III)
fluoride was used instead. It was subsequently found that
even 0.25 equiv of CeF3 was enough to drive the reaction
to completion. Using tert-butyl acetate instead of ethyl
acetate in this first step allowed for easy purification of
subsequent intermediates 15 and 16 by crystallization.
The regio- and stereoselective directed epoxidation of
crude tertiary allylic alcohol 13 with tert-butyl hydro-
peroxide, catalyzed by vanadyl acetylacetonate VO(acac)2,
then afforded epoxide 14. The epoxidation of 13 in
dichloromethane was sluggish, even in the presence of
molecular sieves and with a large excess of reagents, and
Resu lts a n d Discu ssion
The requisite substrate for the rearrangement, di-
eneoxide 18, was readily obtained in 7 synthetic steps
from (S)-carvone as shown in Scheme 3. The correct 1â-
configuration of the epoxide was initially established in
14, by transfer of chirality from the existing stereocenter
in (S)-carvone, using a two-step aldol addition/directed
epoxidation sequence via the tertiary allylic alcohol 13.
(9) Suzuki, M.; Oda, Y.; Noyori, R. J . Am. Chem. Soc. 1979, 101,
1623.
(10) (a) Tsuda, T.; Tokai, M.; Ishida, T.; Saegusa, T. J . Org. Chem.
1986, 51, 5216. (b) Visentin, G.; Piccolo, O.; Consiglio, G. J . Mol. Catal.
1990, 61, L1.
(11) Liu, H.-J .; Zhu, B.-Y. Can. J . Chem. 1991, 69, 2008.

1R-Fluoro A-Ring Phosphine Oxide
J . Org. Chem., Vol. 66, No. 18, 2001 6143
Sch em e 3
gave only a modest yield of 14. A clean conversion and
good yield of 14 was realized by carrying out the reaction
in refluxing cyclohexane with a Dean-Stark condenser.
This procedure resulted in complete reaction after 5 h
using 1.2 mol % of VO(acac)2 and 1.2 equiv of the tert-
butyl hydroperoxide as a 5-6 M nonane solution. Epoxide
14 was relatively unstable and was used directly in the
next step, after extractive workup, as a mixture with
nonane.
oxidation was achieved using a 3:1 mixture of hexanes-
EtOAc. The resulting acetate intermediate was then
hydrolyzed with a catalytic amount (15 mol %) of sodium
methoxide in MeOH at 0 °C, and alcohol 16 was isolated
by crystallization from EtOAc-hexane in 77% yield from
15.
After selective protection of the secondary hydroxyl
over the tertiary hydroxyl group in 16, the â-elimination
of the tertiary alcohol in ester 17 was readily achieved.
Thus, after silylation of 16 using tert-butyldimethylsilyl
chloride (TBSCl) and imidazole in THF, the salts were
removed by filtration, and the THF solution of 17 was
subjected to dehydration with thionyl chloride in pyridine
to give (E)-dieneoxide 18 exclusively. Adding the THF
solution of 17 to a preformed, cold thionyl chloride/
pyridine mixture minimized the formation of chloride 19,
which was otherwise a major byproduct. Although 19
could be converted to 18 separately upon subsequent
treatment with DBU, the desired dehydrochlorination did
not occur under the reaction conditions described above.
Oxidative cleavage of the isopropenyl group was then
achieved in a two-step ozonolysis/Baeyer-Villiger se-
quence, without protection of the tertiary hydroxyl group.
Thus, crude epoxide 14 in MeOH was ozonized at -70
°C, in the presence of sodium bicarbonate. The hydro-
peroxide intermediate was reduced in situ with dimethyl
sulfide to obtain ketone 15, in 76% yield from 4, after
crystallization from EtOAc-hexane.12
The Baeyer-Villiger oxidation of 15 was attempted
under standard conditions using m-CPBA in dichlo-
romethane,13 but the reaction was very slow producing
a large amount of byproducts. Other reagents examined
were either ineffective or resulted in complete decomposi-
tion of the substrate. It was subsequently found that the
m-CPBA oxidation could be accelerated in less polar
solvents. Thus, using 2 equivalents of m-CPBA in com-
mon solvents, the following conversions of 15 were
obtained after 16 h at room temperature: toluene (80%
+ byproducts) > 1:1 hexanes-EtOAc (67%) > EtOAc
(55%) > tert-butyl methyl ether (51%) > acetonitrile (20%
+ byproducts) > DMF (10%). Finally, fast and clean
Now the success of the synthesis hinged upon the
selective isomerization of dieneoxide 18 to dieneol 20.
Before attempting the palladium-catalyzed isomerization,
several other methods known to effect the rearrangement
of epoxides to allylic alcohols were tried, but failed to give
even a trace amount of the desired product. However,
when dieneoxide 18 was treated with Pd(PPh3)4 in THF
at 65 °C, a mixture of the desired allylic alcohol 20 and
the isomeric R,â-unsaturated ketone 21 was obtained in
the unfavorable ratio of 1:3.
(12) Alternatively, the hydroperoxide intermediate could be isolated
from the ozonolysis, acetylated, and the isolated peracetate subjected
to Criegee rearrangement; see: Schreiber, S. L.; Liew, W.-F. Tetrahe-
dron Lett. 1983, 24, 2363. However, this procedure gave alcohol 16 in
low yield after chromatographic separation, and ketone 15 as the major
byproduct.
(13) For a review on the Baeyer-Villiger oxidation, see: Krow, G.
R. Org. React. 1993, 43, 251.

6144 J . Org. Chem., Vol. 66, No. 18, 2001
Kabat et al.
Ta ble 1. p Ka of Ad d itives vs Selectivity
pKa
(relative to selectivity %
alcohol additive
mol %
water)
20 vs 21
none
tert-BuOH
25
25
26
32
91
92
95
100
10
10
10
10
10
19
MeC(CF3)2OH (22a )
(CF3)2CHOH (22b)
PhC(CF3)2OH (22c)
1,3-C6H4[C(CF3)2OH]2 (22d )
(CF3)3COH (22e)
9.51
9.13
8.52
8.48
5.18
Eventually, the ratio was improved to 3:1 when tris-
(2-furyl)phosphine ligand was used instead of Ph3P, but
20 mol % of the palladium catalyst was still necessary
for complete conversion. Further ligand screening re-
vealed, that an identical selectivity could be achieved
with only 1.5 mol % of a palladium/1,2-bis(diphenylphos-
phino)ethane catalyst. Still, the selectivity could not be
improved any further using other ligands.
Serendipitously, it was then found, that addition of a
fluorinated alcohol, 1,1,1,3,3,3-hexafluoro-2-phenyl-2-pro-
panol (22c), increased formation of allylic alcohol 20
versus the undesired enone 21 and simultaneously
improved the turnover frequency of the original pal-
ladium-triphenylphosphine catalyst. Remarkably, even
a catalytic amount of 10 mol % of 22c was sufficient to
increase the selectivity for allylic alcohol 20 formation
to 10:1. Increasing the amount of this additive further
to 50 mol % and 100 mol % gave an improved 16:1 and
19:1 ratio of 20:21, respectively.
This interesting effect was explored using other alcohol
additives, and the selectivity was found to correlate with
the pKa of these additives. Indeed, the more acidic
fluorinated alcohols14 with pKa < 9 were particularly
effective. As shown in Table 1 and Figure 1, a sharp
increase in selectivity for the desired allylic alcohol 20
was observed when the pKa of the additive decreased
from 9.1 to 8.5. The ‘titration curve’ shown in Graph 1
strongly suggests that the reaction pathways leading to
either 20 or 21 diverge in a protonation step of a common
intermediate of comparable basicity. Interestingly, only
the fluorinated alcohols performed well in this reaction.
Other more common proton sources, such as aliphatic
alcohols, phenols and carboxylic acids, resulted in no or
incomplete reaction and no effect on selectivity.
F igu r e 1. pKa of additives vs selectivity.
Although perfluoro-tert-butyl alcohol (22e) gave a
better selectivity (95:5) than the less acidic 22c and 22d
(91:9 and 92:8, respectively), the reactions run with the
latter compounds were cleaner, than that with 22e. Thus,
22c and 22d were studied in more detail in different
solvents to determine their catalytic efficiency and the
results are shown in Figure 2. Diol 22d gave better
turnover frequency than 22c, particularly so in less polar
solvents. Moreover, diol 22d is a monomer commonly
used for the preparation of fluoropolymers,15 and it is
available in bulk at a relatively low cost.16 The reaction
using this additive was, therefore, optimized further.
Ultimately, virtually complete selectivity (>99:1) for
the desired 20 was obtained by carrying out the reaction
with 1 mol % of the palladium catalyst, prepared in situ
from 0.5 mol % Pd2dba3(CHCl3) and 5 mol % of triph-
enylphosphine, and 2 mol % of 22d in a less polar solvent,
F igu r e 2. Catalytic effect of fluorinated alcohols on selectivity
at 65 °C.
toluene, and at the lower temperature of 35 °C. This
reaction temperature also resulted in an increased purity
of the product. It was found that two major polar
byproducts were formed on prolonged heating of the
product 20 in concentrated solutions, or even upon
allowing the solidified 20 to stand at ambient tempera-
ture overnight. These byproducts were found to be
(14) (a) Filler, R.; Schure, R. M. J . Org. Chem. 1967, 32, 1217. (b)
Chang, I. S.; Price, J . T.; Tomlinson, A. J .; Willis, C. J . Can. J . Chem.
1972, 50, 512.
(15) Cassidy, P. E.; Aminabha, T. M.; Reddy, S.; Fitch, J . W., III.
Eur. Polym. J . 1995, 31, 353.
(16) 22d is available from PCR, Inc. at $295/500 g, compared to $98/
25 g for 22c.

1R-Fluoro A-Ring Phosphine Oxide
J . Org. Chem., Vol. 66, No. 18, 2001 6145
Sch em e 4
isomers resulting from the Diels-Alder dimerization of
20, and the structure of one of them was subsequently
determined unequivocally as 23 by X-ray crystallography.
The remarkable catalytic effect of a proton source on
the chemoselectivity of â-hydride elimination may be
explained as follows. The palladium(0) catalyst formed
in situ reacts initially with the dieneoxide to form a
zwitterionic π-allyl palladium alkoxide complex 24. With-
out a suitable proton source to protonate the alkoxide,
the hydrogen next to the alkoxide migrates preferentially
to the cationic palladium to give the enolate 26, which
then collapses to give the undesired enone 21 as the
major product. This selectivity of the H-shift comes from
the increased electron density of the C-H bond, due to
electron-donation from the alkoxide. When a proton
source of appropriate pKa is present, the zwitterionic
intermediate 24 can be protonated, possibly via preco-
ordination, to give Pd-complex 25 as an ion-pair with the
corresponding conjugate base of the proton source used.17
Now the inductive effect of the resulting hydroxyl group
in 25, as compared to the electron-donating ability of the
alkoxide in 24, prevents the formation of enolate 26.
Thus, â-hydride elimination from the methyl group
occurs, preferentially, instead to give the desired allylic
alcohol 20 as the major product, regenerating the proton
source and palladium catalyst. Protonation of 24 is
facilitated in nonpolar solvents such as toluene, as
opposed to THF, which explains the solvent effect ob-
served.
-95 °C.6 In our hands under similar conditions (0.07 M
dichloromethane solution at -90 °C), fluorination of tert-
butyl ester 20 with 4 equivalents of DAST gave a mixture
of the desired Sn2 product 27 and the corresponding Sn2′
product 30 in a ratio of 2.5:1. The former was then
isolated in a modest 52% yield by chromatography.
To improve the isolated yield of 27, many solvents were
investigated. Trichloroethylene was found to be the best
solvent, both in terms of a slightly better isomer ratio
(2.9:1 at -78 °C in a 0.07 M solution of 20), and the
highest isolated yield of 27 (70% from pure 20). Modified
DAST reagents (dimethylaminosulfur trifluoride, bis(2-
methoxyethyl)aminosulfur trifluoride,18 and morpholi-
nosulfur trifluoride) were examined under the same
conditions, but gave inferior results (27:30 ratios of 2.5:
1, 2.2:1, and 2.0:1, respectively).
Concentration and excess reagent were also found to
affect selectivity. With 4 equivalents of DAST, inferior
ratio were obtained both at a higher (0.60 M; ratio 1.7:1)
and lower (0.016 M, ratio 2.4:1) concentration. Further-
more, at a substrate concentration of 0.22 M, 27:30 ratios
of 2.2:1, 2.5:1, and 2.6:1 were obtained with 1.5, 2.0 and
2.5 equivalents of DAST, respectively.
As high dilution and a large excess of the expensive
reagent are not desirable for scale-up, larger scale
reactions were carried out at concentration of 0.22 M of
20 in trichloroethylene using 2 equivalents of DAST,
although these conditions were slightly inferior in terms
of selectivity. From these reactions fluoride 27 was
consistently obtained in 55% yield over the four steps
from alcohol 16 without purification of intermediates (i.e.
17, 18, and 20).
Fluoride 27 also undergoes Diels-Alder dimerization
similar to alcohol 20, especially in the solid state, but in
contrast to 20 the reaction is highly regio- and stereose-
lective to give dimer 31 exclusively. The structure of 31
The allylic alcohol 20, containing an exocyclic double
bond, thus obtained, was then converted to the 1R-fluoro
A-ring phosphine oxide 1 in 5 steps as shown in Scheme
4.
The reaction of the corresponding ethyl ester analogue
of 20 (i.e. alcohol 9, see Scheme 2 above) with DAST had
been previously described to give the desired fluorinated
product 10 in 60% yield, when carried out in 0.025 M
dichloromethane solution with 8 equivalents of DAST at
(17) Alternatively, direct protonation of epoxide 18 is also possible,
see, e.g.: Das, U.; Crousse, B.; Kesavan, V.; Bonnet-Delpon, D.; Be´gue´,
J .-P. J . Org. Chem. 2000, 65, 6749. However, only the combination of
an appropriate fluoro alcohol with palladium gave the desired reaction.
(18) Deoxo-Fluor was made available courtesy of Air Products &
Chemicals, Inc.

6146 J . Org. Chem., Vol. 66, No. 18, 2001
Kabat et al.
was also confirmed by X-ray crystallography.19 Thus,
concentration of solutions containing 27 were carried out
as rapidly as possible at, or below, room temperature,
and this product was then stored in a freezer to prevent
its cyclodimerization.
39 were formed in a 1:1 ratio under these conditions.
Alternative reducing agents proved ineffective.
Alternative methods of fluorination were also stud-
ied: displacement of the corresponding mesylate with
fluoride,20 reactions with perfluoro-1-butanesulfonyl fluo-
ride/DBU,21 N,N-diisopropyl-1-fluoro-2-methylpropena-
mine,22 etc. None of those methods, however, proved
fruitful.
Fluorination of related compounds (Z)-ester 32, (E)-
allyl acetate 34, and (Z)-allyl acetate 35 with DAST were
also examined. The former gave lactone 33 exclusively.
The allyl acetates, 34 and 35, reacted to give 2.5:1 and
4:1 ratios of the desired Sn2 product and the undesired
Sn2′ product, respectively. Although a better ratio was
obtained from the latter, its preparation requires two
additional steps and, therefore, it was deemed impracti-
cal.
It was subsequently found, however, that ester 27 was
smoothly reduced to 28 when the reaction was run in
THF,24 at -40 °C for 5 h, leading to the formation of only
a trace amount of 38.
A solution of the crude (E)-allylic alcohol 28 in tert-
butyl methyl ether was then irradiated with a 450W
medium-pressure mercury lamp through uranium filter,
in the presence of 10 mol % of 9-fluorenone25 as the
photosensitizer. The uranium filter was essential for
clean conversion, as previously reported.6 After chroma-
tography to remove the sensitizer, the known6 (Z)-allylic
alcohol 29 was obtained in 90% yield over the two steps
from 27.
Finally, alcohol 29 was converted to the desired phos-
phine oxide 1 in 76% yield, through an improved two-
step procedure, by a substitution reaction using triph-
osgene/pyridine, directly followed by displacement of the
resulting allylic chloride with diphenylphosphine oxide
sodium salt.26
In summary, we have developed a new efficient process
for the preparation of 1R-fluoro A-ring phosphine oxide
1, which is useful for the synthesis of 3 and other vitamin
D analogues bearing the same A-ring fragment. The new
process, which compares very favorably to the previous
syntheses, gave the key precursor fluoride 27 in 9 steps
and 32% overall yield, with only three isolations (two
crystallizations and one chromatography), and without
intermediate purification. Using this process, phosphine
oxide 1 was ultimately prepared, via 27, in 13 steps and
22% overall yield from (S)-carvone, with only two ad-
ditional chromatographic purifications, bringing the total
number of isolations to five. Further synthetic applica-
tions of the selective isomerization of dieneoxides using
the catalytic system described herein are being investi-
gated in our laboratories and will be reported in due
course.
The reduction of the tert-butyl ester in 27 turned out
to be surprisingly difficult, presumably due to the steric
bulk of the tert-butyl group. With DIBALH in hexane,
toluene, or dichloromethane at -70 °C, a relatively stable
complex, presumably 36, was formed, which gave alde-
hyde 37 upon hydrolysis. With 2.5 equiv of DIBALH at
0 °C, a mixture of the desired 28, defluorinated 38, and
isobutylated 3923 was obtained. In hexane, only 38 and
(19) For a recent study of stereoselectivity in Diels-Alder cycload-
ditions of allylic fluorides, see: Gre´e, D.; Laurent, V.; Gre´e, R.; Toupet,
L.; Washington, I.; Pelicier, J .-P.; Villacampa, M.; Pe´rez, J .-M.; Houk,
K. N. J . Org. Chem. 2001, 66, 2374. The thermal instability of the
corresponding methyl ester analogue of 27 has been previously noted
(see ref 8), but has been attributed to polymerization.
(20) Pilcher, A. S.; Ammon, H. L.; DeShong, P. J . Am. Chem. Soc.
1995, 117, 5166.
(21) Bennua-Skalmowski, B.; Vorbru¨ggen, H. Tetrahedron Lett.
1995, 36, 2611.
(22) Munyemana, F.; Frisque-Hesbain, A.-M.; Devos, A.; Ghosez, L.
Tetrahedron Lett. 1989, 30, 3077.
(24) Commercial 1 M DIBALH solution in THF (Aldrich) was found
less effective in this reaction than a freshly prepared solution of
DIBALH in THF.
(25) Fluorene, used previously as a sensitizer,6 was less effective,
requiring a large excess and prolonged irradiation.
(26) For an improved preparation of A-ring phosphine oxides from
allylic alcohols, see: Daniewski, A. R.; Garofalo, L. M.; Kabat, M. M.
Synth. Commun., accepted for publication.
(23) For related chemistry, see: Ooi, T.; Uraguchi, D.; Kagoshima,
N.; Maruoka, K. Tetrahedron Lett. 1997, 38, 5679.

1R-Fluoro A-Ring Phosphine Oxide
J . Org. Chem., Vol. 66, No. 18, 2001 6147
reaction flask, then the condenser in order to prevent the hot
vapors from reaching the addition funnel, and (3) an explosion
shield was placed in front of the reaction vessel. A solution of
crude 13 as obtained above (207 g, 776 mmol) and vanadyl
acetylacetonate (3.09 g, 11.7 mmol) in cyclohexane (770 mL)
was heated to gentle reflux and tert-butyl hydroperoxide, 5.0-
6.0 M in nonane (170 mL, 850-1020 mmol) was added
(Caution!) over 90 min. The green solution turned deep red
upon initiation of the addition and a mild exotherm ensued.
After completion of the addition, the resulting orange-green
solution was heated to reflux for 3 h with constant removal of
water by a Dean-Stark condenser. The volume of the water
in the trap increased by ca. 4 mL. TLC analysis indicated the
presence of only a small amount of starting material. After
cooling to below room temperature with an ice-water bath, 1
M sodium bisulfite solution (77 mL) and saturated aqueous
NaHCO3 solution (150 mL) were added. After 5 min, an iodine-
starch paper test indicated no peroxide to be present. The
organic layer was separated, then washed with saturated
aqueous NaHCO3 solution (3 × 150 mL) and saturated
aqueous NaCl solution (150 mL), dried over Na2SO4, and
concentrated under reduced pressure at <30 °C bath temp.
Further drying at room temperature under high vacuum for
2 h gave 247 g (overweight) of crude 14, containing nonane,
as a pale yellow solid. TLC 3:1 hexanes-EtOAc; PMA stain;
Rf 13 ) 0.80 and Rf 14 ) 0.55.
ter t-Bu tyl 2((4S,1R,2R,6R)-4-Acetyl-2-h yd r oxy-1-m eth -
yl-7-oxa bicyclo[4.1.0]h ep t-2-yl)a ceta te (15). Crude 14 as
obtained above (247 g, ca. 776 mmol) and NaHCO3 (24 g, 286
mmol) in MeOH (1.8 L) was cooled with a dry ice/acetone bath,
the nitrogen inlet-tube was replaced with a gas dispersion tube
with a porous fritted glass tip (25-50 µ), 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 -70 °C. The reaction turned pale blue after
5 h. After ozonized air was passed for an additional 15 min
through the mixture at a reduced flow rate of 1 LPM, excess
ozone was removed by purging with air (4 LPM) for 25 min.
The resulting white suspension was treated with dimethyl
sulfide (75 mL, 1.02 mol) and allowed to warm to room
temperature overnight. An iodine-starch paper test indicated
no peroxide to be present. The insoluble inorganic salts were
removed by filtration and washed with EtOAc (100 mL). The
combined filtrate and washes were concentrated under reduced
pressure (bath temperature e 30 °C) to remove essentially all
of the MeOH. The resulting yellow, milky residue was parti-
tioned between EtOAc (1 L) and water (300 mL). The aqueous
layer was separated and extracted with EtOAc (50 mL). The
combined organic extracts were washed with saturated aque-
ous NaCl solution (300 mL), dried over Na2SO4, and concen-
trated under reduced pressure (bath temperature e 35 °C).
The resulting pale yellow oil was dissolved in EtOAc (150 mL)
and hexane (600 mL) was added. The resulting suspension was
then stored in the refrigerator overnight. The solids were
collected by filtration, washed with cold hexanes-EtOAc (4:
1) mixture (400 mL), and dried by suction, then under high
vacuum at room temperature to give 155.9 g (68.5% over 3
steps) of 15 as a white solid (mp 92-94 °C). The combined
mother liquor and washes were washed with saturated aque-
ous NaHCO3 solution (3 × 100 mL) and saturated aqueous
NaCl solution (100 mL), dried over Na2SO4, and concentrated
under reduced pressure (bath temperature e 35 °C). The
residue was diluted with EtOAc (40 mL) and hexane (280 mL)
was added. The resulting slightly cloudy solution was stored
in the refrigerator over the weekend. The resulting solid was
collected by filtration, washed with a cold hexanes-EtOAc (7:
1) mixture (4 × 40 mL), and dried by suction, then under high
vacuum at room temperature to give to give 18.3 g (8.0% over
3 steps) of a second crop of 15 as a white solid (mp 91-93 °C).
The two crops were combined to give a total yield of 174 g
(76.5% over 3 steps) of 15. TLC 1:1 hexanes-EtOAc; PMA
stain; Rf 14 ) 0.70 and Rf 15 ) 0.50.
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, 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 was used to generate ozonized air (shell pressure
6 PSIG; flow rate 4 LPM; 110V). Commercial grade reagents
and solvents were used without purification, except as indi-
cated. m-CPBA (57-86% Aldrich) was dried under high
vacuum at room temperature for 24 h (ca. 24% weight loss
upon drying) and the ratio of m-CPBA to m-chlorobenzoic acid
(m-CBA) was determined by NMR analysis in CDCl3. Diphen-
ylphosphine oxide was prepared according to the literature
procedure27 with minor modifications as follows: Chlorodiphen-
ylphosphine (92 g, Fluka 97%) was added to 1 N HCl (500 mL)
and the mixture was stirred at room temperature overnight.
Then, the mixture was extracted with methylene chloride and
the organic layer was washed with saturated aqueous NaHCO3
solution and saturated aqueous NaCl solution, dried over Na2-
SO4, and concentrated to give 74.6 g (88.5%) of diphenylphos-
phine oxide as a white solid of mp 55-56 °C.
ter t-Bu t yl 2-[(5S,1R)-1-H yd r oxy-2-m et h yl-5-(1-m et h -
ylvin yl)cycloh ex-2-en yl]a ceta te (13). A THF (400 mL)
solution of diisopropylamine (136 mL, 1.04 mol) was cooled to
-40 °C. Butyllithium, 2.5 M in hexanes (416 mL, 1.04 mol)
was added over 7 min, while maintaining the reaction tem-
perature at -40 ( 10 °C. After stirring at that temperature
for 5 min and then cooling to -70 °C, tert-butyl acetate (140
mL, 1.04 mol) was added over 8 min, while maintaining the
temperature of the reaction mixture below -60 °C. After
stirring at -70 °C for 20 min, cerium(III) fluoride (78.8 g, 0.40
mol) was added in one portion. Three minutes later, (S)-
carvone (4) (125 mL, 0.80 mol) was added over 7 min, while
maintaining the temperature of the reaction mixture below
-60 °C. After stirring for 5 min at -70 °C, TLC analysis
indicated complete reaction. The reaction was quenched by the
addition of a 1:1 mixture of AcOH and EtOH (240 mL). An
exotherm ensued that raised the temperature of the mixture
to -30 °C. After the exotherm had subsided, the cooling bath
was removed, and the resulting thick suspension was diluted
with EtOH (300 mL) and EtOAc (400 mL), and allowed to
warm to ambient temperature. The solid was removed by
filtration through a pad of Celite and washed with EtOAc (3
× 300 mL). The combined filtrate and washes were washed
sequentially with water (400 mL), 0.5 N HCl (400 mL),
saturated aqueous NaHCO3 solution (400 mL), and saturated
aqueous NaCl solution (400 mL), dried over Na2SO4, and
concentrated to dryness. Further drying at 70 °C/5 Torr gave
207 g (97.0%) of 13 as a pale yellow oil. TLC 4:1 hexanes-
EtOAc; short-wave UV detection for 4 and PMA stain for 13,
both Rf ) 0.7. Since both 4 and 13 have the same Rf value,
but the starting material 4 strongly absorbs short-wave UV
light, the reaction was judged complete when TLC analysis
showed only a faint spot for 4 under UV254, and a strong one
for 13 by staining with PMA. This material was used directly
in the next step without further purification.
ter t-Bu tyl 2-[(4S,1R,2R,6R)-2-Hyd r oxy-1-m eth yl-4-(1-
m et h ylvin yl)-7-oxa b icyclo[4.1.0]h ep t -2-yl]a cet a t e (14).
Ca u tion ! Since tert-butyl hydroperoxide is a strong oxidizing
agent, the following precautions were taken: (1) an addition
funnel was placed at a height to prevent the hot cyclohexane
vapor from heating the hydroperoxide solution, (2) a stream
of nitrogen was passed from the addition funnel, through the
(27) Quin, L. D.; Montgomery, R. I. J . Org. Chem. 1963, 28, 3315.

6148 J . Org. Chem., Vol. 66, No. 18, 2001
Kabat et al.
ter t-Bu t yl 2((4S,1R,2R,6R)-4-Acet yloxy-2-h yd r oxy-1-
m eth yl-7-oxa bicyclo[4.1.0]h ep t-2-yl)a ceta te. 15 (82.2 g,
289 mmol), m-CPBA, 91% (115 g, 606 mmol) and a 3:1
hexanes-EtOAc mixture (840 mL) were combined in a reaction
flask. Upon mixing, an endotherm ensued and the temperature
of the mixture decreased from 20 °C to 13 °C. The white
suspension was stirred at room temperature (ca. 20 °C) for 3
days (Note: Reaction temperatures higher than 25 °C result
in increased formation of byproducts). NMR analysis indicated
∼98% conversion and a m-CPBA/m-CBA ratio of ∼2:3. The
reaction was cooled to 5 °C in an ice-water bath and a 2.5 M
K2CO3 solution (145 mL, 435 mmol) was added dropwise to
the mixture at e 12 °C over 8 min. Then, a 2 M sodium sulfite
solution (180 mL, 360 mmol) was added over 25 min, while
maintaining the temperature of the mixture below 12 °C. The
cold bath was removed and the mixture was stirred at ambient
temperature for 90 min. NMR analysis of the organic layer
indicated the presence of a 1:4 mixture of m-CPBA to product.
Then, dimethyl sulfide (6 mL, 82 mmol) was added and the
resulting thin suspension was stirred for 15 min. An iodine-
starch paper test then indicated complete reduction. The solid
was removed by filtration and washed with EtOAc (100 mL).
The filtrate and washes were combined and the phases were
separated. The organic phase was washed with a 10% KHCO3
solution (30 mL) and dried over magnesium sulfate. The
aqueous phases were combined and back-extracted with EtOAc
(200 mL), and the organic extract was washed with 10%
KHCO3 solution (20 mL) and dried over magnesium sulfate.
The back-extraction process was repeated two more times. All
of organic extracts were combined and concentrated at e 30
°C under reduced pressure. The residue was dried under high
vacuum at room temperature overnight to give 81.3 g (93.6%)
of the crude acetate as a colorless oil. TLC 1:1 hexanes-EtOAc;
PMA stain; Rf 15 ) 0.50 and acetate intermediate Rf ) 0.55.
ter t-Bu tyl 2-((4S,1R,2R,6R)-2,4-Dih yd r oxy-1-m eth yl-7-
oxa bicyclo[4.1.0]h ep t-2-yl)a ceta te (16). The crude product
obtained above (81.3 g, 270 mmol) was dissolved in MeOH (270
mL). The solution was cooled for 30 min with an ice water
bath, then a 25% NaOMe solution in MeOH (9.3 mL, 40.5
mmol) was added dropwise over 10 min. After stirring at 0 °C
for 4 h, TLC analysis indicated complete reaction. The reaction
mixture was quenched with AcOH (3.0 mL, 52.6 mmol), then
concentrated at e 30 °C under reduced pressure. The resulting
milky residue was dried under high vacuum at room temper-
ature for 30 min, then partitioned between EtOAc (500 mL)
and a 5% KHCO3 solution (50 mL). The layers were separated,
and the organic phase was washed with a 5% KHCO3 solution
(50 mL) and saturated aqueous NaCl solution (50 mL). The
combined aqueous phases were extracted with EtOAc (2 × 100
mL). The organic extracts were combined, dried over MgSO4
and concentrated at e 35 °C under reduced pressure. The
resulting pale yellow oil (ca. 76 g) was dissolved in EtOAc (70
mL) and crystallization was induced by the addition of a seed
crystal. Then, hexane (350 mL) was gradually added, and the
resulting suspension was allowed to stand at room tempera-
ture overnight. The solids were collected by filtration, washed
with a 5:1 hexane-EtOAc mixture (2 × 70 mL) and dried by
suction, then under high vacuum at room temperature to give
54.8 g (78.4%) of 16 as a white solid (mp 91-92 °C). The
combined mother liquor and washes were diluted with hexane
(300 mL) and stored in the freezer overnight. The supernatant
was removed by decantation, and the residue was dissolved
in EtOAc (100 mL). The organic solution was washed with a
5% KHCO3 solution (20 mL) and saturated aqueous NaCl
solution (20 mL), dried over MgSO4 and concentrated under
reduced pressure (bath temperature e 35 °C). The residue (4.3
g) was dissolved in EtOAc (5 mL), and crystallization was
induced by the addition of a seed crystal. Hexane (25 mL) was
then gradually added, and the resulting suspension was
allowed to stand at room temperature overnight. The solids
were collected by filtration, washed with a 5:1 hexanes-EtOAc
mixture (12 mL) and dried by suction, then under high vacuum
at room temperature to give 2.5 g (3.6%) of a second crop of
16 as an off-white solid (mp 91-92 °C). The two crops were
combined to give a total yield of 57.3 g (76.7% over 2 steps) of
16. TLC 1:1 hexanes-EtOAc; PMA stain; Rf 16 ) 0.25.
ter t-Bu tyl 2-[(4S,1R,2R,6R)-2-Hyd r oxy-1-m eth yl-7-oxa -
4-[[(1,1-d im eth yleth yl)d im eth ylsilyl]oxy]-bicyclo[4.1.0]-
h ep t-2-yl]a ceta te (17). 16 (28.6 g, 111 mmol), imidazole (20.5
g, 301 mmol), tert-butylchlorodimethylsilane (19.6 g, 130
mmol), and THF (170 mL) were combined in a reaction flask.
An initial mild exotherm (10 to 12 °C) subsided quickly. The
mixture was then stirred overnight. TLC analysis indicated
complete reaction. The solids were removed by filtration using
a sintered glass funnel and washed thoroughly with THF (200
mL). The combined, colorless filtrate and wash were concen-
trated under reduced pressure at 25 °C, then under high
vacuum for 30 min to yield 48.7 g (overweight) of crude 17 as
a white solid. TLC 1:1 hexanes-EtOAc; PMA stain; Rf 16 )
0.16 and Rf 17 ) 0.79. 1H NMR analysis indicated the presence
of ca. one equivalent of protonated imidazole. This material
was used directly in the next step without further purification.
ter t-Bu t yl (2E)-2-[(1S,4R,6R)-1-Met h yl-7-oxa -4-[[(1,1-
d im et h ylet h yl)d im et h ylsilyl]oxy]-b icyclo[4.1.0]h ep t -2-
ylid en e]a ceta te (18). Pyridine (136 mL, 1.68 mmol, water
content e 0.02%) was cooled in an ice water bath and thionyl
chloride (13.6 mL, 186 mmol) was added. The initial exotherm
to 27 °C was allowed to subside and the solution was stirred
at ambient temperature for 40 min. The resulting yellow
solution was then cooled to -34 °C and a solution of crude 17
as obtained above (48.7 g, 111 mmol in theory) in THF (86
mL) was added dropwise over 1 h at a rate to maintain the
reaction temperature below -25 °C. The reaction mixture was
allowed to warm to 0 °C over 100 min, then poured into a
mixture of saturated aqueous NaHCO3 solution (700 mL) and
hexanes (350 mL). The resulting mixture was stirred for 30
min until no noticeable gas evolution was observed. The
hexane layer was separated, washed with 1 M citric acid (350
mL), dried over Na2SO4, and concentrated to dryness under
reduced pressure to yield 40.7 g (overweight) of 18 as a
colorless oil. 1H NMR analysis indicated that this material
contained some silyl byproducts, and was approximately 90%
pure. TLC 9:1 hexanes-EtOAc; short-wave UV detection and
PMA stain; Rf 17 ) 0.04 and Rf 18 ) 0.21. This material was
used directly in the next step without further purification.
ter t-Bu t yl (2E)-2-[(3R,5R)-5-[[(1,1-Dim et h ylet h yl)d i-
m eth ylsilyl]oxy]-3-h ydr oxy-2-m eth ylen e-cycloh exyliden e]-
a ceta te (20). Tris(dibenzylideneacetone)dipalladium(0)-chlo-
roform adduct (570 mg, 0.551 mmol) and triphenylphosphine
(1.45 g, 5.55 mmol) were combined in a reaction flask. The
flask was evacuated and refilled with nitrogen three times,
then toluene (35 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 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 indicate 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 (0.54 mL, 2.18 mmol) was
added. The slurry became red-orange. After three minutes of
stirring at ambient temperature (19 °C), a solution of crude
18 as obtained above (40.7 g, 110 mmol in theory) in toluene
(160 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 to 32 °C overnight, then to 35 °C for 2 h (Note: While
3% of the starting material 18 was still detected by HPLC, see
below, reaction at 35 °C overnight gave complete conversion).
The reaction mixture was rapidly concentrated on a rotary
evaporator at 25 °C (bath temperature), under reduced pres-
sure, and the residue was dried under high vacuum for 30 min
to give 44.8 g (overweight) of crude 20 as a reddish oil (Note:
The product should not be allowed to stand at ambient
temperature any longer than necessary as dimerization via
Diels-Alder reaction occurs upon standing in a concentrated
solution or, even more rapidly, in solid state. In the procedure

1R-Fluoro A-Ring Phosphine Oxide
J . Org. Chem., Vol. 66, No. 18, 2001 6149
described, the concentration and drying process took ca. 90 min.
The product may be stored in the freezer, where it partially
solidifies, without deterioration). TLC 3:1 hexanes-EtOAc;
short-wave UV detection and PMA stain; Rf 18 ) 0.74, Rf 20
) 0.45 and Rf 21 ) 0.50. HPLC analysis was performed on a
Nucleosil 5µm, 4.6 × 2500 mm column with 2% 2-propanol in
hexane at 0.5 mL/min. The percentages given are the area
percentages of the corresponding peaks detected at 220 nm:
Rt 18 ) 7.6 min, Rt 21 ) 8.8 min, Rt ) 8.9 min (dibenzylide-
neacetone), Rt 20 ) 12.1 min, and Rt 23 ) 18 min. HPLC
analysis indicated this material, 20, to be ca. 87% pure with
ca. 3% of the starting material 18, less than 1% of the ketone
byproduct 21 and ca. 3% of the dimer 23 present. This material
was used immediately in the next step without further
purification.
ter t-Bu t yl (2E)-2-[(3S,5R)-5-[[(1,1-Dim et h ylet h yl)d i-
m eth ylsilyl]oxy]-3-flu or o-2-m eth ylen e-cycloh exylid en e]-
a ceta te (27). Diethylaminosulfur trifluoride (33 mL, 0.25 mol)
was dissolved in trichloroethylene (290 mL). After cooling to
-70 °C, a solution of crude 20 as obtained above (44.7 g, 111
mmol in theory) in trichloroethylene (210 mL) was added over
80 min, while maintaining the temperature of the reaction
mixture below -70 °C. After stirring at -70 °C for 15 min,
the mixture was allowed to warm to -50 °C and MeOH (30
mL) was added in one portion to quench the reaction. The
resulting mixture was diluted with hexane (850 mL), washed
with water (2 × 300 mL) and saturated aqueous NaHCO3
solution (300 mL), dried over Na2SO4, and rapidly concentrated
on a rotary evaporator at e 30 °C (bath temperature) (Note:
The product should not be allowed to stand at ambient
temperature any longer than necessary as dimerization via
Diels-Alder reaction occurs upon standing in a concentrated
solution or, even more rapidly, in the solid state). The residue
was dried under high vacuum for 30 min, then purified by
chromatography on silica gel 60 (1.5 kg), eluting with 4:1-2:1
hexane:dichloromethane. The appropriate fractions were com-
bined and concentrated on a rotary evaporator at e 30 °C (bath
temperature). The residue was immediately placed in a freezer
while it crystallized. Since the crystallization is exothermic,
it should be carried out at a low temperature in order to
prevent the Diels-Alder dimerization. The resulting solid was
dried under high vacuum at room temperature for 1 h to give
21.6 g (54.7% yield over 4 steps from 16) of 27 as a white solid.
TLC 8:1 hexanes-EtOAc; short-wave UV detection and PMA
stain; Rf 20 ) 0.3, Rf 27 ) 0.8 and Rf 30 ) 0.75.
Pyrex immersion well, nitrogen inlet tube (immersed deep into
the reaction mixture) and outlet-bubbler, was charged with
crude 28 as obtained above (8.33 g, 28.9 mmol), tert-butyl
methyl ether (450 mL), and 9-fluorenone (524 mg, 2.91 mmol).
While water and a 20 °C coolant were rapidly passed through
the well and the cooling jacket, respectively, the above solution
was irradiated through a uranium filter with a 450 W medium
pressure mercury lamp. During the photolysis, nitrogen was
continuously bubbled through the reaction solution via the
inlet tube. After 6 h of irradiation, NMR analysis indicated
complete reaction. The reaction mixture was concentrated to
dryness to give 9.13 g of crude product as a yellow oil, which
was subjected to chromatography on silica gel 60 (300 g). First,
the column was eluted with 95:5 hexanes-EtOAc (2.5 L) to
remove the sensitizer. Then, the column was further eluted
with 9:1 hexanes-EtOAc (1 L), 8:2 hexanes-EtOAc (1 L), and
7:3 hexanes-EtOAc (1 L). The appropriate fractions (the 8:2
hexanes-EtOAc eluate) were combined and concentrated
under reduced pressure. Further drying under high vacuum
for 1 h gave 7.43 g (89.8%) of 29 as a colorless oil. This product
is relatively unstable at room temperature and should be
stored in the freezer. TLC 4:1 hexanes-EtOAc; short-wave UV
detection and PMA stain; 28 and 29 both Rf ) 0.3, and
fluorenone Rf ) 0.5.
[[(1R,3Z,5S)-3-(2-ch lor oeth ylid en e)-5-F lu or o-4-m eth yl-
e n e c y c lo h e x y l]o x y ](1,1-d im e t h y le t h y l)-d im e t h y ls i-
la n e.6 Ca u tion ! This reaction should be performed in a well
ventilated hood. 29 (8.07 g, 28.2 mmol), hexanes (150 mL), and
triphosgene (4.18 g, 14.1 mmol) were combined in a reaction
flask. The resulting solution was then cooled to 0 °C with an
ice-acetone bath and a solution of pyridine (4.50 mL, 55.6
mmol) in hexanes (20 mL) was added over 30 min. After
stirring at 0 °C for 30 min, the cooling bath was removed and
the resulting pale-yellow reaction mixture was stirred at room
temperature for 30 min. Then, the reaction mixture was
diluted with hexanes (250 mL) and washed with saturated
copper (II) sulfate solution (3 × 200 mL). The combined
aqueous washes were back-extracted with hexanes (2 × 100
mL). The organic extracts were combined, dried over MgSO4
and concentrated to dryness on a rotary evaporator to give
9.0 g (overweight) of crude product as a pale yellow oil. This
material was used immediately in the next step without
further purification. This product is relatively unstable at room
temperature and should be stored in the freezer. TLC 4:1
hexanes-EtOAc; short-wave UV detection and PMA stain; Rf
29 ) 0.3, and chloride product Rf ) 0.9.
(2E)-2-[(3S,5R)-5-[[(1,1-d im et h ylet h yl)d im et h ylsilyl]-
oxy]-3-Flu or o-2-m eth ylen ecycloh exyliden e]-eth a n ol (28).
27 (10.3 g, 28.9 mmol) was dissolved in THF (70 mL), freshly
distilled from sodium/benzophenone ketyl. After cooling to -70
°C, a solution (freshly prepared at -10 °C) of DIBALH, neat
(13.4 mL, 75.2 mmol) in THF (75 mL) was added over 20 min
via a cannula, while maintaining the temperature of the
reaction mixture between -68 and -70 °C. This mixture was
then stirred under nitrogen at -60 to -50 °C for 1.5 h and at
-50 to -40 °C for 1h, then at -42 °C for 5 h. After this time,
TLC analysis indicated the presence of very small amounts of
27 and the defluorinated byproduct 38. The reaction mixture
was quenched by the slow addition of water (15 mL) at -40
°C, then allowed to warm to 25 °C (occasional cooling is
required to prevent the temperature of the mixture from
exceeding 25 °C). After stirring at room temperature for 10
min, a gel formed. Thus, the mixture was diluted with EtOAc
(100 mL), then stirred at room temperature for 1.5 h. Solids
were removed by filtration using a pad of Celite and washed
thoroughly with EtOAc (350 mL). The combined filtrate and
washes were concentrated under reduced pressure at e 30 °C
(bath temperature) and the residue was dried under high
vacuum for 1 h to give 8.33 g (100.3%) of crude 28 as a colorless
oil. This product is relatively unstable at room temperature
and should be stored in the freezer. TLC 4:1 hexanes-EtOAc;
short-wave UV detection and PMA stain; Rf 27 ) 0.9, Rf 28 )
0.3, Rf 37 ) 0.6 and Rf 38 ) 0.33.
[(2Z)-2-[(3S,5R)-5-[[(1,1-d im eth yleth yl)d im eth ylsilyl]-
oxy]-3-Flu or o-2-m eth ylen ecycloh exyliden e]-eth yl]diph en -
ylp h osp h in e oxid e (1).6 DMF (50 mL) and sodium hydride,
60% dispersion in mineral oil (1.33 g, 33.1 mmol) were
combined in a reaction flask. While cooling with a water bath
(10 °C), diphenylphosphine oxide (6.70 g, 33.1 mmol) was
added in small portions over 15 min. The water bath was then
removed and the resulting yellow solution was stirred at room
temperature for 30 min. After cooling to -60 °C with a dry
ice acetone bath, a solution of the crude product from the
previous step (9.0 g, 28.2 mmol in theory) in DMF (20 mL)
was added dropwise, via
a syringe, over 15 min, while
maintaining the temperature of the reaction mixture below
-50 °C. The reaction mixture was stirred at -60 °C for 2 h,
then allowed to warm to room temperature over 1 h. The
reaction mixture was diluted with diethyl ether (600 mL) and
washed with water (600 mL). The combined aqueous washes
were extracted with diethyl ether (200 mL). The organic
extracts were combined, dried over MgSO4 and concentrated
under reduced pressure to give a white solid. This crude
product was recrystallized from diisopropyl ether (25 mL). The
resulting solid was collected by filtration, washed with cold
diisopropyl ether (5 mL) and dried under high vacuum to give
7.93 g (59.8%) of 1 as a white solid. The mother liquor was
concentrated and the residue was subjected to chromatography
on silica gel, eluting with hexanes-EtOAc (7:3 to 1:1). The
appropriate fractions were combined and concentrated to
dryness to give an additional 2.22 g (16.7%) of 1. Thus, the
total yield of 1 was 10.1 g (76.5% overall yield from 29). TLC
(2Z)-2-[(3S,5R)-5-[[(1,1-d im et h ylet h yl)d im et h ylsilyl]-
oxy]-3-Flu or o-2-m eth ylen ecycloh exyliden e]-eth an ol (29).6
A 500 mL photoreaction vessel equipped with a cooling jacket,

6150 J . Org. Chem., Vol. 66, No. 18, 2001
Kabat et al.
1:1 hexanes-EtOAc; short-wave UV detection and PMA stain;
Rf 29 ) 1.0 and Rf 1 ) 0.28.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra of
13 compounds, obtained as indicated in the Experimental
Section. This material is available free of charge via the
Internet at http://pubs.acs.org.
Ack n ow led gm en t. We thank Mr. Michael P. Lanyi
of the Roche Physical Chemistry Department for pKa
determinations.
J O015788Y