An efficient route for the preparation of a 21-fluoro progestin-16 alpha,17 alpha-dioxolane, a high-affinity ligand for PET imaging of the progesterone receptor.

Article abstract of DOI:10.1021/jo020190r

Two different synthetic routes were explored for the synthesis of fluoro furanyl norprogesterone (FFNP) 1, a high-affinity ligand for the progesterone receptor (PgR) that is being developed as a PET imaging agent for PgR-positive breast cancer. Both approaches proceed through a key intermediate, triol 5. The first approach, starting from keto-ketal 2, employed a dioxenyl group as a synthon for installing a corticosteroid side chain in keto-alcohol 4. The second approach, starting from propargylic acetate 12b, involved the application of a two-step method, a Pd(II)-catalyzed oxidative rearrangement followed by a base-catalyzed acetate rearrangement of the intermediate unsaturated acetate 13b, to generate the requisite corticosteroid side chain in keto-acetate 14b. This intermediate was further elaborated to the final product 1 via efficient dihydroxylation with potassium permangnate, furan acetalization with scandium triflate, and mesylation and fluorination reactions. The palladium-catalyzed route is considerably more efficient than the dioxene approach for the synthesis of key intermediate triol 5, and the scandium triflate-catalyzed acetalization, in particular, led to a considerable improvement in the overall yield of the endo furan acetal alcohol 16a. This route provides a major improvement in the overall yield of the final progestin target, FFNP 1.

Full text of DOI:10.1021/jo020190r

An Efficien t Rou te for th e P r ep a r a tion of a 21-F lu or o
P r ogestin -16r,17r-Dioxola n e, a High -Affin ity Liga n d for P ET
Im a gin g of th e P r ogester on e Recep tor
Dange Vijaykumar, Wang Mao, Karen S. Kirschbaum, and J ohn A. Katzenellenbogen*
Department of Chemistry, University of Illinois at Urbana Champaign, 600 South Mathews Avenue,
Urbana, Illinois 61801
jkatzene@uiuc.edu
Received March 18, 2002
Two different synthetic routes were explored for the synthesis of fluoro furanyl norprogesterone
(FFNP) 1, a high-affinity ligand for the progesterone receptor (PgR) that is being developed as a
PET imaging agent for PgR-positive breast cancer. Both approaches proceed through a key
intermediate, triol 5. The first approach, starting from keto-ketal 2, employed a dioxenyl group as
a synthon for installing a corticosteroid side chain in keto-alcohol 4. The second approach, starting
from propargylic acetate 12b, involved the application of a two-step method, a Pd(II)-catalyzed
oxidative rearrangement followed by a base-catalyzed acetate rearrangement of the intermediate
unsaturated acetate 13b, to generate the requisite corticosteroid side chain in keto-acetate 14b.
This intermediate was further elaborated to the final product 1 via efficient dihydroxylation with
potassium permangnate, furan acetalization with scandium triflate, and mesylation and fluorination
reactions. The palladium-catalyzed route is considerably more efficient than the dioxene approach
for the synthesis of key intermediate triol 5, and the scandium triflate-catalyzed acetalization, in
particular, led to a considerable improvement in the overall yield of the endo furan acetal alcohol
16a . This route provides a major improvement in the overall yield of the final progestin target,
FFNP 1.
In tr od u ction
It has been more difficult to develop PET imaging agents
for PgR, however. 21-[18F]Fluoro-16R-ethyl-19-norproges-
terone, which showed excellent target tissue-selective
distribution in rodents,8 proved to be ineffective in
humans because of its rapid reduction to the 20-dihydro
progestin by a 20-dehydrogenase.9,10
To develop an imaging agent for PgR that would be
less prone to the action of 20-dehydrogenases, we exam-
ined a number of progestin 16R,17R-ketals and acetals
of aromatic carbonyl compounds,11,12 known to be potent
progestins in vivo.11,13,14 We thought that their bulky
substituents near the C-20 carbonyl group would slow
Breast cancer is the number one cause of cancer-
related deaths in middle-aged women,1 yet many breast
cancers can be successfully treated with the antiestrogens
tamoxifen and raloxifene. Currently, monitoring the
presence of estrogen receptor (ER) and progesterone
receptor (PgR) in tumors is the best method for assessing
the hormone responsiveness of breast cancer.2-4 However,
the methods currently used to determine the presence
and concentration of ER and PgR in tumors (radioligand
binding and immunoassays5) require tissue biopsy and
are thus invasive. Determining the levels of these recep-
tors noninvasively through radioligand imaging using
positron emission tomography (PET) would have a num-
ber of advantages.
(7) Dehdashti, F.; Flanagan, F. L.; Mortimer, J . E.; Katzenellenbo-
gen, J . A.; Welch, M. J .; Siegel, B. A. Eur. J . Nucl. Med. 1999, 26,
51-56.
(8) Pomper, M. G.; Katzenellenbogen, J . A.; Welch, M. J .; Brodack,
J . W.; Mathias, C. J . J . Med. Chem. 1988, 31, 1360-1363.
(9) Verhagen, A.; Studeny, M.; Luurtsema, G.; Visser, G. M.; de
Goeij, C. C. J .; Sluyser, M.; Nieweg, O. E.; Van der Ploeg, E.; Go, K.
G.; Vaalburg, W. Nucl. Med. Biol. 1994, 21, 941-952.
(10) Dehdashti, F.; McGuire, A. H.; Van Brocklin, H. F.; Siegel, B.
A.; Andriole, D. P.; Griffeth, L. K.; Pomper, M. G.; Katzenellenbogen,
J . A.; Welch, M. J . J . Nucl. Med. 1991, 32, 1532-1537.
(11) Kochanny, M. J .; VanBrocklin, H. F.; Kym, P. R.; Carlson, K.
E.; O’Neil, J . P.; Bonasera, T. A.; Welch, M. J .; Katzenellenbogen, J .
A. J . Med. Chem. 1993, 36, 1120-7.
We have demonstrated that 16R-[18F]fluoroestradiol
(FES) is an effective breast tumor ER imaging agent.6,7
* To whom correspondence should be addressed. Phone: (217) 333-
6310. Fax: (217) 333-7325.
(1) Chevallier, B.; Mosseri, V.; Dauce, J . P.; Bastit, P.; J ulien, J . P.;
Asselain, B. Br. J . Cancer 1990, 61, 436-40.
(2) Muss, H. B. Breast Cancer Res. Treat. 1992, 21, 15-26.
(3) Rayter, Z. Br. J . Surg. 1991, 78, 528-35.
(4) Allegra, J . C.; Lippman, M. E.; Thompson, E. B.; Simon, R.;
Barlock, A.; Green, L.; Huff, K. K.; Do, H. M.; Aitken, S. C. Cancer
Res. 1979, 39, 1447-54.
(12) Buckman, B. O.; Bonasera, T. A.; Kirschbaum, K. S.; Welch,
M. J .; Katzenellenbogen, J . A. J . Med. Chem. 1995, 38, 328-37.
(13) Fried, J .; Sabo, E. F. In Hormonal Steroids, Biochemistry,
Pharmacology and Therapeutics: Proceedings of the First International
Congress on Hormonal Steroids; Academic Press: New York, 1965; Vol.
2, p 15.
(5) Merkel, D. E.; Osborne, C. K. Hematol. Oncol. Clin. North Am.
1989, 3, 641-52.
(6) Katzenellenbogen, J . A. In Estrogens, Progestins, and Their
Antagonists; Pavlik, E. J ., Ed.; Birkha¨user: Boston, 1996; pp 197-
242.
(14) Fried, J .; Sabo, E. F.; Grabowich, P.; Lerner, L. J .; Kessler, W.
B.; Brennan, D. M.; Borman, A. Chem. Ind. (London) 1961, 465.
10.1021/jo020190r CCC: $22.00 © 2002 American Chemical Society
Published on Web 06/14/2002
4904
J . Org. Chem. 2002, 67, 4904-4910

An Efficient Route to a PET Imaging Agent
SCHEME 1
SCHEME 2
their reduction by the serum 20-dehydrogenases. The
most promising progestin derivative for PET imaging we
have found is Fluoro Furanyl Norprogesterone (FFNP)
1,12 and in preparation for tumor imaging trials in
humans, a considerable quantity of this material was
needed for sterility, activity, and toxicity studies.
The synthetic method that we originally reported for
the preparation of FFNP 111 was very challenging and
unfortunately proved to be difficult to reproduce reliably
and in satisfactory yields (Scheme 1). A particular
problem was the conversion of keto-ketal 2 to toluene-
sulfonyl isocyanide 3 using a tosylmethyl isocyanide
(TosMIC) method,11,15-18 en route to the generation of the
corticosteroid side chain in the intermediate keto-alcohol
4. Despite several modifications that we made to improve
the yield, we found that it was impractical to use this
route to prepare FFNP 1 in large quantities.
In this report, we describe two different synthetic
routes toward triol 5, a key intermediate, that we
explored to develop an alternative and more efficient
synthetic route to FFNP 1. We also cover significant
improvements in other steps of the synthesis.
Resu lts a n d Discu ssion
2,3-Dih yd r o-1,4-d ioxen e Ap p r oa ch to Key Tr iol 5.
To synthesize triol 5, the key intermediate for the
synthesis of the FFNP 1, the 1,4-dioxene addition ap-
proach of Fetizon et al.,19 was investigated (Scheme 2),
starting from commercially available dione 6. The 3-ke-
tone group in dione 6 was protected as the ethylene glycol
ketal in three steps to furnish compound keto-ketal 2.20
The reaction of keto-ketal 2 with 1,4-dioxenyl-lithium at
low temperatures furnished tertiary alcohol 7 in ∼50%
yield. The fact that nearly half of the starting material
was recovered after this reaction suggested that the
D-ring carbonyl group was undergoing enolization, a
common problem with cyclopentanones. Efforts to in-
(15) Van Leusen, D.; Van Leusen, A. M. Recl. Trav. Chim. Pays-
Bas 1991, 110, 393-401.
(16) Van Leusen, D.; Van Leusen, A. M. Tetrahedron Lett. 1984,
25, 2581-4.
(17) Van Leusen, A. M.; Van Leusen, A. M. (Gist-Brocades N. V.,
Neth.). Eur. Pat. Appl., EP 123734, 1984; p 24.
(18) Van Leusen, A. M.; Van Leusen, A. M. (Gist-Brocades, N. V.,
Neth.). Eur. Pat. Appl. EP 123736, 1984; p 25.
(19) Fetizon, M.; Goulaouic, P.; Hanna, I. J . Chem. Soc., Perkin
Trans. 1 1990, 4, 1107.
(20) Saha, N. N. Steroids 1968, 12, 735-47.
J . Org. Chem, Vol. 67, No. 14, 2002 4905

Vijaykumar et al.
SCHEME 3
crease the yield in the 1,2-addition reaction by adding
anhydrous cerium chloride to enhance the electrophilicity
of the C-17 carbonyl were unsuccessful.
approach based on a palladium-catalyzed rearrangement
of propargyl esters23,24 to be particularly attractive,
because it generates the requisite corticosteroid side
chain in just three steps, starting from readily available
propargylic alcohol precursors. This convenient method-
ology obviates the need for protection of the 3-keto group
(Scheme 3).
Thus, the commercially available steroid propargylic
alcohol 11 (norethindrone) was converted to the corre-
sponding propargylic isobutyrate ester 12a ,25 quantita-
tively, employing standard conditions, i.e., isobutyric
anhydride, pyridine, and a catalytic amount of DMAP.
Treatment of this propargylic isobutyrate ester 12a with
5 mol % potassium tetrabromopalladate and 2 equiv of
water in DME in an oxygen atmosphere at 65 °C resulted
in the formation of unsaturated isobutyrate 13a in 85%
yield. NMR analysis revealed the presence of an 80:20
mixture of geometrical isomers, but no effort was made
to separate them, because both can be converted to the
corticosteroid side chain.
Treatment of the unsaturated isobutyrate 13a with
DBU at 55 °C generated the corticosteroid side chain,
presumably via a 1,2-acyl transfer on the dienolate
intermediate, thereby giving the keto-isobutyrate 14a in
85% yields. Thus, starting from propargylic alcohol 11,
we could obtain the keto-isobutyrate 14a , having the
necessary side chain, in three steps and in excellent yield.
This contrasts with the earlier dioxene approach, wherein
the requisite keto-alcohol 4 was obtained from dione 6
in eight steps and in low overall yield (Scheme 2).
Dehydration of tertiary alcohol 7 furnished diene 8 in
47% yield. The more electron-rich double bond on the
dioxenyl ring was then selectively epoxidized with 1 equiv
of m-chloroperbenzoic acid (mCPBA) and the labile
epoxide spontaneously solvolyzed to the hemiacetal-ketal
9. This intermediate was reduced to the ketal 10, which
was hydrolyzed with p-toluenesulfonic acid in moist
acetone, to obtain the desired product keto-alcohol 4
(Scheme 2). Dihydroxylation of the ∆16(17) double bond in
keto-alcohol 4 with a stoichiometric amount of osmium
tetroxide in pyridine11 furnished the requisite key inter-
mediate triol 5 in 55% yield. The less hindered ∆16(17)
double bond undergoes oxidation in preference to the ∆4(5)
double bond.
Despite some improvements over the original approach
based on the TosMIC route,11 there are still several
drawbacks with this synthetic route, namely, the protec-
tion and deprotection sequence of the 3-ketone group in
dione 6, the poor yield in the dioxene 1,2-addition and
subsequent elimination reactions, and finally, the low-
yielding conversion of diene 8 to keto-alcohol 4. Thus, for
the purpose of producing the desired imaging agent
FFNP 1 in larger quantities, we looked further for a
shorter and more efficient synthesis.
P d (II)-Ca ta lyzed Oxid a tive Rea r r a n gem en t Ap -
p r oa ch to Key Tr iol 5. There are a number of alterna-
tive approaches for introducing the 21-hydroxy-16-ene-
20-keto (corticosteroid) side chain in steroidal systems
that are described in the literature.21-24 We found the
We were disconcerted to find, however, that hydrolysis
of the isobutyrate ester in keto-isobutyrate 14a under a
(21) Bakos, T.; Lokos, M.; Vincze, I.; Volford, J . Steroids 1993, 58,
115-8.
(24) Kataoka, H.; Watanabe, K.; Miyazaki, K.; Tahara, S.; Ogu, K.;
Matsuoka, R.; Goto, K. Chem. Lett. 1990, 1705-8.
(25) Watson, T. G.; Hosking, M.; Herz, J . E.; Torres, J . V.; Muller,
J .; Murillo, A.; Cruz, S.; Shafiee, A.; Vossoghi, M. Steroids 1983, 41,
255.
(22) Hessler, E. J .; VanRheenen, V. H. (Upjohn Co., USA). Ger.
Offen. DE 2920290, 1979; p 80.
(23) Kataoka, H.; Watanabe, K.; Goto, K. Tetrahedron Lett. 1990,
31, 4181-4.
4906 J . Org. Chem., Vol. 67, No. 14, 2002

An Efficient Route to a PET Imaging Agent
TABLE 1. F u r a n Aceta liza tion of Dih yd r oxy Aceta te
15b or Tr iol 5
variety of conditions (both basic and acidic) failed to
furnish keto-alcohol 4 in a clean manner and led to
extensive decomposition of the starting material. As a
result, the required keto-alcohol 4 could be obtained only
in very poor yields (<20%) from this precursor.
To circumvent this problem, we repeated the entire
reaction sequence with the corresponding acetate, in the
hope that it would be easier to hydrolyze this less
hindered ester. It is of note that literature reports
indicate that the palladium-catalyzed propargyl ester
rearrangements do not proceed as effectively with ac-
etates as they do with isobutyrates.23,24 Nevertheless, we
converted propargylic alcohol 11 to the corresponding
propargylic acetate ester 12b26 and subjected it to the
same reaction sequence described for propargylic isobu-
tyrate ester 12a .
We found, in fact, that the Pd(II)-catalyzed rearrange-
ment of the propargylic acetate ester 12b gave unsatur-
ated acetate 13b in 80% yields, and the base-catalyzed
acetate transfer also furnished keto-acetate 14b in
similar yields. Furthermore, we were gratified to find
that in contrast to the keto-isobutyrate 14a , the keto-
acetate 14b could be hydrolyzed to keto-alcohol 4 under
fairly mild conditions, using potassium carbonate and
moist methanol, in 85% yield. Having thus successfully
achieved an efficient synthesis of this important inter-
mediate, we turned our attention to improving the yields
of subsequent steps.
Because the dihydroxylation of the unprotected keto-
alcohol 4 with osmium tetroxide gave the triol 2 in only
∼50% yield (Scheme 2), we attempted to improve the
efficiency of the ∆16(17) olefin dihydroxylation, by carrying
it out on the protected keto-alcohol, i.e., keto-acetate 14b.
Thus, treatment of keto-acetate 14b with the cheaper
dihydroxylating agent, potassium permanganate in
acetone,27-29 furnished dihydroxy acetate 15b in 85%
yields. Subsequent acetate hydrolysis to give triol 5 was
also efficient (84% yield). On the other hand, efforts to
hydrolyze isobutyrate ester in dihydroxy isobutyrate 15a
were not so successful (∼30% yield). Thus, by this five-
step sequence we were able to obtain triol 5 in ∼40%
overall yield, starting from commercially available pro-
pargylic alcohol 11. By employing the palladium-medi-
ated oxidative rearrangement approach, as well as by
using potassium permanganate as a dihydroxylating
agent, we obtained nearly a 10-fold increase in the overall
yield of the triol 5 compared to the dioxene approach.
compd
R
acid catalyst
endo:exoa
yieldb
15b
5
15b
5
Ac
H
Ac
H
0.023 M 70% HClO4
0.023 M 70% HClO4
1 mol % Sc(OTf)3
1 mol % Sc(OTf)3
10 mol % Sc(OTf)3
1:2 (17a :17b)
1:1 (16a :16b)
1:6 (17a :17b)
1:2 (16a :16b)
1:1 (16a :16b)
55%
45%
75%
75%
80%
5
H
a
b
Based on the 1H NMR. Isolated yield.
diastereomeric ratio but also improving the overall yield
of the reaction, we attempted furan acetalization with a
catalytic amount of 70% perchloric acid on the protected
dihydroxy acetate 15b. Unfortunately, this method led
to only a very negligible improvement in the yield (45%)
and also resulted in the formation of a less favorable 1:2
diastereomeric mixture of endo 17a and exo 17b furan
acetal acetates.
We imagined that the hydrous Bro¨nsted acid utilized
in the acetalization reaction was perhaps responsible for
the low yields. Considering this, we wondered whether
the use of an anhydrous Lewis acid such as scandium
triflate [Sc(OTf)3], a reagent known to be a good catalyst
for acetalization,30 might prove to be more efficient in this
reaction. Thus, the furan acetalization was carried out
on the protected dihydroxy acetate 15b with 1 mol %
scandium triflate as the catalyst. Unfortunately, this
resulted in the formation of a very undesireable 1:6 ratio
of endo and exo products 17a and 17b, respectively,
though the isomeric acetals were obtained in 75% yield.
On the other hand, Sc(OTf)3-catalyzed furan acetalization
of unprotected triol 5 resulted in the formation of a more
favorable 1:2 mixture of endo and exo isomers 16a and
16b furan acetal alcohols, respectively, also in similar
yield.
It is known that the endo acetal isomer is the thermo-
dynamically more stable product,13,14 and dioxolone for-
mation should be reversible. Thus, furan acetalization
was again attempted on triol 5, but now with a 10-fold
higher concentration of Sc(OTf)3 (10 mol %). Under these
conditions, we obtained a 1:1 mixture of endo and exo
furan acetal diastereomers in 80% yields. These diaster-
eomers were separated by reversed-phase HPLC to
obtain diastereomerically pure endo furan-acetal alcohol
16a . The undesired exo isomer 16b, which was also
obtained in this purification, could again be equilibrated
to a 1:1 mixture of endo and exo furan acetal diastereo-
mers by re-treatment with the Sc(OTf)3 (10 mol %) and
furfuraldehyde. Thus, this catalytic route also provided
a convenient method for recycling exo furan acetal alcohol
16b to the desired endo isomer 16a .
Im p r ovem en ts in th e F lu or in e Su bstitu tion Re-
a ction . Finally, we turned our attention toward improv-
ing the final fluorination step. It was observed previously
that conversion of the endo furan acetal alcohol 16a to
the corresponding triflate 18a was very inefficient (∼30%),
as was conversion of the triflate 18a to the FFNP 1
(overall yield of 9%).12 A major problem in this conversion
was the fact that the triflate 18a is very unstable and
thus difficult to isolate and purify. Although it is less
reactive, we decided to investigate whether the meth-
Im p r ovem en ts in F u r a n Aceta liza tion . After hav-
ing addressed the problem of the efficient synthesis of
triol 5, we turned our attention toward improving the
yields of furan acetalization (Table 1). From prior work,12
it was known that triol 5, upon treatment with furfural
and a catalytic amount of 70% perchloric acid, gave a 1:1
diastereomeric mixture of endo and exo furan acetal
alcohols 16a and 16b, respectively, in ∼40% yields
(Scheme 4). With the objective of not only improving this
(26) Leung, S. L.; Karunanithy, R.; Becket, G.; Yeo, S. H. Steroids
1985, 46, 639-47.
(27) Hydorn, A. E.; Korzun, J . N.; Moetz, J . R. Steroids 1964, 3, 493-
504.
(28) Van Leusen, D.; Van Leusen, A. M. J . Org. Chem. 1994, 59,
7534-8.
(29) Annen, K.; Hofmeister, H.; Laurent, H.; Wiechert, R. Liebigs
Ann. Chem. 1982, 966-72.
(30) Ishihara, K.; Karumi, Y.; Kubota, M.; Yamamoto, H. Synlett
1996, 839-841.
J . Org. Chem, Vol. 67, No. 14, 2002 4907

Vijaykumar et al.
SCHEME 4
anesulfonate (mesylate) 18b might not be a more suitable
intermediate.
permanganate), thereby improving the yield of this step
as well. Further improvement in the furan acetalization
was obtained by the use of a sufficient quantity of
anhydrous scandium triflate as a catalyst. Finally, the
efficiency of the fluorination reaction was improved using,
in place of the unstable triflate, the more stable mesylate.
This latter route enables the synthesis of the precursor
for the F-18-labeled FFNP 1 in sufficient quantities for
medical imaging studies.
Thus, we converted the hydroxy group in endo furan
acetal alcohol 16a to the corresponding mesylate deriva-
tive 18b in more than 90% yield, by treatment with
methanesulfonyl chloride and triethylamine. Nucleophilic
displacement of the mesylate 18b with tetrabutylammo-
nium fluoride (TBAF) in DMSO furnished the requisite
fluoro compound 1 in only modest yield (ca. 30%, based
on the fluoride ion).31 Nevertheless, this fluorination
method constitutes a significant improvement in the
overall yield for the conversion of endo furan acetal
alcohol 16a to FFNP 1.
Exp er im en ta l Section
Gen er a l Meth od s. Dioxene,32 progargylic esters 12a 25 and
12b,26 and keto-ketal 220 were prepared according to the
reported procedures. Reaction progress was monitored by
analytical thin-layer chromatography. Silica gel used in flash
chromatography33 was 32-63 mm. TLC plates were visualized
using short-wave UV light (254 nm), potassium permanganate,
or phosphomolybdic acid. Melting points were uncorrected.
Con clu sion
To address the shortcomings of an earlier route,11 two
different approaches for the synthesis of a precursor for
FFNP (1), a potential PET imaging agent for progester-
one receptors in breast cancer, have been explored. In
the first, less efficient approach, a dioxene unit was
utilized as the synthon for the requisite progestin side
chain. In the second, more efficient approach, a palla-
dium(II)-catalyzed oxidative rearrangement was em-
ployed to generate the progestin side chain from a
propargylic acetate 12b. This synthetic route does away
with the protection and deprotection protocol of the
3-ketone group, and more importantly, the requisite 21-
hydroxy-20-keto-∆16(17) system is installed in only three
steps starting from commercially available steroid pro-
pargylic alcohol 11. Finally, the acetate-protected 21-
hydroxy group in keto-acetate 14b enables the use of a
less toxic, cheaper dihydroxylating agent (potassium
1
Standard H NMR and 13C NMR spectra were recorded at 500
and 125 MHz, respectively, and the chemical shifts in the NMR
spectra are given in the δ scale. Infrared spectra were recorded
on a FT-IR spectrometer. MS and HRMS were recorded with
either an electron impact (70 eV) or a FAB mass spectrometer.
HPLC was performed isocratically (70:30 water-acetonitrile)
using a 5 µm ODS partisil column at a 5 mL/min flow rate.
Dry THF and CH2Cl2 were dried using an alumina column
solvent dispensing system. Experimental details for the com-
pounds shown in Scheme 2 (7, 8, 4) and Scheme 4 (17a and
17b) are given in Supporting Information.
3-Keto-20-isobu ta n oyloxy-19-n or p r egn a -4,17(20)-d ien -
21-a l (13a ). To a stirred solution of propargylic isobutyrate
ester 12a (1.4 g, 3.8 mmol), water (7.6 mmol, 0.137 mL), and
DME (18 mL) was added potassium tetrabromopalladate
(96 mg, 0.19 mmol). The reaction was heated at 65 °C under
an oxygen atmosphere for 3 h. Upon completion of the reac-
tion, the solvent was evaporated, and flash chromatography
(3:7 EtOAc-hexane) gave unsaturated isobutyrate 13a as
a yellow viscous oil (1.24 g) in 85% yield. Both 1H and 13C
(31) For the ultimate radiosynthesis of high specific activity [18F]-
FFNP 1, typical alcohol fluorination reagents such as DAST cannot
be used. In fact, fluoride ion is the only high specific activity fluorine-
18 precursor that is routinely available. Also, because radiofluorination
reactions are done with only tracer quantities of (18F) fluoride ion, we
have run the displacement reaction with fluoride ion as the limiting
reagent, to model the radiolabeling conditions.
(32) Moss, R. D.; Paige, J . N. J . Chem. Eng. Data 1967, 12, 452-4.
(33) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43,
2923-5.
4908 J . Org. Chem., Vol. 67, No. 14, 2002

An Efficient Route to a PET Imaging Agent
NMR indicated the presence of an 80:20 mixture of geometrical
isomers. Major isomer. 1H NMR (CDCl3) δ 9.59 (1 H, s, CHO),
5.84 (1 H, br s, olefinic H-4), 1.29 (3 H, d, J ) 7 Hz) and 1.27
(3 H, d, J ) 7 Hz) CH(CH3)2, 1.01 (3 H, s, C-18 -CH3). 13C
NMR (CDCl3): δ 199.7, 175, 165.7, 161.2, 124.8, 124.7, 71.9,
52.8, 48.9, 47.0, 42.3, 39.5, 36.4, 35.2, 34.7, 33.7, 30.8, 26.9,
26.5, 26.3, 24.3, 19.0, 18.9. MS (EI): m/z 384 (M+). HRMS:
calcd for C24H32O4, 384.2300; found, 384.2300.
powder. Crystallization (dichloromethane and hexane) gave
colorless crystals, mp 167-168 °C. H NMR (CDCl3): δ 6.62
1
(1 H, br s, olefinic H-16), 5.65 (1 H, br s, olefinic H-4), 4.87 (1
H, d, J ) 16 Hz, H-21), 4.72 (1 H, d, J ) 16 Hz, H-21), 1.99 (3
H, s, acetyl methyl), 0.79 (3 H, s, C-18 methyl). 13C NMR
(CDCl3): δ 199.15, 190.0, 169.8, 165.8, 151.1, 143.5, 124.2, 65.2,
54.1, 49.3, 46.3, 42.0, 37.9, 36.0, 34.8, 33.8, 31.9, 30.4, 26.0,
25.7, 20.0, 15.5. MS (EI): m/z 356 (M+). HRMS: calcd for
C22H28O4, 356.1987; found, 356.1982. Anal. Calcd for C22H28O4:
C, 74.13; H, 7.92; O, 17.95. Found: C, 73.80; H, 7.94; O, 18.26.
21-Isob u t a n oyloxy-19-n or p r e gn -4,16-d ie n e -3,20-d i-
on e (14a ). To a stirred solution of unsaturated isobutyrate
13a (0.9 g, 2.33 mmol) in dry ethyl acetate (5 mL) was added
DBU (1.16 mmol, 0.17 mL), and the reaction was heated at
60 °C overnight and then cooled. The organic solvent was
removed under vacuum. Flash chromatography (2:3 EtOAc-
hexane) furnished keto-isobutyrate 14a (760 mg) as a white
amorphous powder in 85% yield. Crystallization (ethyl acetate
and hexane) gave colorless crystals, mp 120-121 °C. IR
(neat): 1742 (ester CdO), 1673 (unsaturated CdO), 1616
(olefinic), 1586 (olefinic). 1H NMR (CDCl3): δ 6.67 (s, 1 H,
olefinic H-16), 5.71 (s, 1 H, olefinic H-4), 4.77 (d, 1 H, J ) 16
Hz, H-21), 4.93 (s, 1 H, J ) 16.7 Hz), 1.08-1.19 (m, 6H,
21-Hyd r oxy-19-n or p r egn -4,16-d ien e-3,20-d ion e (4). To
a solution of keto-acetate ester 14b (500 mg, 1.4 mmol) in
MeOH (7 mL), THF (7 mL), and water (2 drops) was added
potassium carbonate (177 mg, 1.4 mmol). The mixture was
stirred at room temperature for 15 min. The solvent was then
evaporated and the reaction mixture diluted with ethyl acetate
(20 mL) and washed with water (2 × 5 mL). The organic layer
was washed with brine and dried over sodium sulfate. Evapo-
ration of the organic layer followed by flash chromatography
(2:3 EtOAc-hexane) gave keto-alcohol 4 as an off-white
crystalline powder (360 mg) in 85% yield, whose spectral data
(1H and 13C NMR) were identical to those reported in the
literature.11 Crystallization of the product (dichloromethane
and ethyl ether) gave colorless crystals, mp 212-216 °C.11
21-Acyloxy-16r,17r-d ih yd r oxy-19-n or p r egn -4-en e-3,20-
d ion e (15b). To the stirred solution of keto-acetate 14b (2.1
g, 5.5 mmol) and formic acid (14 mmol, 0.523 mL) in acetone
(145 mL) at 0 °C was gradually added a solution of potassium
permangnate (0.869 g, 5.5 mmol) in water (21.6 mL) and
acetone (122 mL). Workup and purification were as described
for compound 15a , furnishing 1.79 g of dihydroxy-acetate 15b
as a white powder in 85% yield. Crystallization (dichlo-
romethane and hexane) gave colorless crystals, mp 191-193
°C. 1H NMR (CDCl3): δ 5.82 (1 H, br s, olefinic H-4), 5.04 (1H,
d, J ) 17.8 Hz, H-21), 4.98 (1 H, dd, J ) 9.2, 2.14 Hz, H-16),
3.87 (1 H, d, J ) 17.8 Hz, H-21), 3.94 (1 H, s, OH), 2.16 (3 H,
s, COCH3), 0.73 (3 H, s, H-18). 13C NMR (CDCl3): δ 206, 200.2,
170.8, 166.5, 127.7, 88.5, 73.0, 68.6, 48.7, 48.4, 48.3, 42.4, 39.9,
36.4, 35.3, 33.4, 31.0, 30.3, 26.4, 25.5, 20.5, 14.3. MS (EI): m/z
390.3 (M+). HRMS: calcd for C22H30O6, 390.2042; found,
390.2038.
isopropyl methyl group), 0.87 (s, 3 H, C-18 methyl group). 13
C
NMR (CDCl3): δ 199.5, 190.4, 176.1, 166.0, 151.5, 143.4, 124.4,
65.1, 54.4, 49.5, 46.5, 42.3, 39.2, 36.2, 35.0, 34.0, 33.4, 32.0,
30.6, 26.2, 25.2, 25.9, 18.7 (2C), 15.7. MS (EI): m/z 384.3 (M+).
HRMS: calcd for C24H32O4, 384.2300; found, 384.2294. Anal.
Calcd for C24H32O4: C, 74.97; H, 8.39; O, 16.64. Found: C,
74.82; H, 8.60; O, 16.58.
21-Isobu ta n oyloxy-16r,17r-d ih yd r oxy-19-n or p r egn -4-
en e-3,20-d ion e (15a ). To a stirred solution of keto-isobutyrate
14a (500 mg, 1.3 mmol) and formic acid (3.3 mmol, 0.126 mL)
in acetone (34 mL) at 0 °C was added gradually a solution of
potassium permanganate (0.205 g, 1.3 mmol) in water (5.1 mL)
and acetone (29 mL). Upon completion of the addition, the
reaction was stirred for 5 min and quenched with a 10%
aqueous solution of KHSO3 (5 mL). Acetone was removed
under vacuum, and the aqueous layer was extracted with ethyl
acetate (2 × 15 mL). The organic layer was washed with brine
and dried over anhydrous sodium sulfate. Evaporation of the
solvent followed by flash chromatography (4:1 EtOAc-hexane)
furnished dihydroxy isobutyrate 15a as white amorphous solid
(490 mg) in 90% yield. Recrystallization (dichloromethane/
hexane) gave colorless crystals, mp 170-172 °C. IR (CHCl3
16r,17r,21-Tr ih yd r oxy-19-n or p r egn -4-en e-3,20-d ion e
(5): F r om Dih yd r oxy Aceta te 15b. To a solution of dihy-
droxy acetate 15b (500 mg, 1.28 mmol) in MeOH (7 mL), THF
(7 mL), and water (2 drops) was added potassium carbonate
(193 mg, 1.28 mmol), and the mixture was stirred at room
temperature for 15 min. Upon completion of the reaction, the
solvent was evaporated and the residue was diluted with
acetone and passed through a pad of silica gel to give triol 5.
Crystallization (acetone and hexane) yielded a white amor-
1
film): 3413 (OH), 1738 (CdO), 1724 (CdO), 1660 (CdO). H
NMR (CDCl3): δ 5.8 (br s, 1H, olefinic H-4), 4.95 (d, 1 H, J )
20 Hz), 4.85 (d, 1 H, J ) 20 Hz), 4.85 (t, 1 H, J ) 9 Hz, H-16).
13C NMR (CDCl3): δ 206.5, 200.5, 176.9, 166.9, 124.4, 88.5,
72.9, 68.4, 48.7, 48.4, 42.3, 39.8, 36.3, 35.3, 33.6, 33.3, 30.9,
30.2, 26.3, 25.5, 18.9, 18.8. MS (FAB): m/z 419 [M + H]+.
HRMS (FAB): calcd for C24H32O4 [M + H]+, 419.2433; found,
419.2436.
11
phous powder (360 mg) in 84% yield, mp 191-193 °C.
3-Keto-20-a cetoxy-estr a -4,17(20)-d ien -21-a l (13b). To a
stirred solution of propargylic acetate ester 12b (4.2 g, 12.3
mmol), water (24.6 mmol, 0.442 mL), and DME (75 mL) was
added potassium tetrabromopalladate (619 mg, 1.23 mmol).
The mixture was heated at 65 °C under an oxygen atmosphere
for 3 h. Workup and purification were the same as those
described for compound 13a , yielding 3.5 g (80%) of the
Spectral data (1H and 13C NMR) of compound 5 were identical
to those reported in the literature.11 F r om Dih yd r oxy Isobu -
tyr a te 15a . To a solution of dihydroxy isobutyrate 15a (500
mg, 1.28 mmol) in MeOH (7 mL), THF (7 mL), and water (2
drops) was added potassium carbonate (193 mg, 1.28 mmol),
and the mixture was stirred at room temperature for 15 min.
Workup and purification were as described above, furnishing
triol 5 in 30% yield, whose spectral data (1H and 13C NMR)
were identical to those reported in the literature.11
16r,17r-[(R)-(1′-r-Fu r ylm eth yliden e)dioxy]-21-h ydr oxy-
19-n or p r egn -4-en e-3,20-d ion e (En d o, 16a ) a n d 16r,17r-
[(S)-(1′-â-F u r ylm et h ylid en e)d ioxy]-21-h yd r oxy-19-n or -
p r egn -4-en e-3,20-d ion e (Exo, 16b): Sc(OTf)3 Meth od . To
a solution of triol 5 (1 mmol) in furfural (6 mL) was added
anhydrous magnesium sulfate (50 mg) and a catalytic amount
of Sc(OTf)3 (1 or 10 mmol %), and the mixture was stirred at
room temperature for 24 h. Evaporation of the solvent in a
vacuum followed by flash chromatography (40-80% ethyl
acetate in hexane) furnished a mixture of endo and exo
hydroxy furan acetal 16a and 16b in 75-80% yields. Both 1H
and 13C spectral data were identical to those reported in the
1
unsaturated acetate ester 13b as a viscous oil. H NMR gave
an 80:20 mixture of geometrical isomers. Major isomer. 1H
NMR (CDCl3): δ 9.56 (1 H, br s, CHO), 5.8 (1 H, br s, olefinic
H-4), 3.1-2.75 (2 H, m, H-16), 2.19 (3 H, s, COCH3), 0.97 (3H,
s, C-18 methyl group). 13C NMR (CDCl3): δ 199.6, 183.9, 168.9,
165.7, 161.6, 138.5, 124.6, 52.7, 48.9, 46.9, 42.2, 39.3, 36.3, 35.1,
34.6, 30.7, 26.8, 26.4, 26.7, 24.2, 20.3, 16.1. MS (EI): m/z 356
(M+). HRMS: calcd for C22H28O4: 356.1988; found: 356.1982.
21-Acetoxy-19-n or p r egn a -4,16-d ien e-3,20-d ion e (14b).
To the stirred solution of unsaturated acetate ester 13b (1 g,
2.8 mmol) in dry ethyl acetate (25 mL) was added DBU (1.4
mmol, 0.2 mL), and the mixture was heated at 60 °C overnight.
Workup and purification were as described for compound 14a ,
yielding 790 mg (80%) of keto-acetate ester 14b as white
J . Org. Chem, Vol. 67, No. 14, 2002 4909

Vijaykumar et al.
literature. 12 Isom er iza tion of Exo F u r a n Aceta l 16b. To a
solution of pure exo 21-hydroxy furan acetal 16b (40 mg, 0.11
mmol), which was obtained after reversed-phase HPLC puri-
fication of a 1:1 mixture of endo and exo furan acetal alcohols
16a and 16b,12 in furfural (1.5 mL) was added anhydrous
magnesium sulfate (25 mg) and a catalytic amount of Sc(OTf)3
(10 mmol %). The reaction was stirred at room temperature
for 12 h. Evaporation of the solvent in a vacuum followed by
flash chromatography (40-80% ethyl acetate in hexane)
furnished 30 mg of a 1:1 mixture of endo and exo furan acetal
alcohols 16a and 16b in 75% yield. Both 1H and 13C spectral
data were identical to those reported in the literature.12
16r,17r-[(R)-(1′-r-F u r ylm eth ylid en e)d ioxy]-21[(m eth -
yl)su lfon yl]oxy-19-n or pr egn -4-en e-3,20-dion e (En do, 18b).
To a stirred solution of endo furan acetal alcohol 16a (50 mg,
0.115 mmol) in methylene chloride (1 mL) at -78 °C was added
triethylamine (46 mg, 0.46 mmol) followed by methanesulfonyl
chloride (16 mg, 0.14 mmol). The reaction mixture was stirred
for 20 min, quenched with methanol (0.1 mL), and extracted
with ethyl acetate (2 × 10 mL). The organic layer was washed
with water (5 mL) and brine and dried over anhydrous sodium
sulfate. Evaporation of the organic layer and rapid flash
chromatography (30-60% ethyl acetate in hexane) furnished
endo furan acetal methanesulfonate 18b (55 mg) as a white
powder in 95% yield. Recrystallization (methanol) gave color-
THF (0.5 mL) was added TBAF (15 µL, 0.014 mmol, 1 M
solution in THF), and the reaction mixture was heated in a
sealed reaction vial at 80 °C for 10 min. The reaction mixture
was diluted with water and extracted with ethyl acetate. The
organic layer was washed with water and brine and dried over
sodium sulfate. Evaporation of the solvent and purification of
the residue by flash chromatography (30-50% ethyl acetate
in hexane) furnished fluorinated compound 1 (2.3 mg) in 30%
yield (with respect to TBAF, the limiting reagent). The
spectroscopic properties (both 1H and 13C) were identical to
those reported previously.12
Ack n ow led gm en t. We are grateful for support of
this research through a grant from the Department of
Energy (DE FG02 86ER60401). NMR spectra were
obtained in the Varian Oxford Instrument Center for
Excellence in NMR Laboratory. Funding for this instru-
mentation was provided in part from the W. M. Keck
Foundation, the National Institutes of Health (PHS 1
S10 RR104444-01), and the National Science Founda-
tion (NSF CHE 96-10502). Mass spectra were obtained
on instruments supported by grants from the National
Institute of General Medical Sciences (G 27029), the
National Institutes of Health (RR 01575), and the
National Science Foundation (PCM 8121494).
1
less crystals, mp 193-195 °C. H NMR (CDCl3): δ 7.47 (1 H,
dd, J ) 1.7 and 0.8 Hz), 6.53 (1 H, dd, J ) 3.2 and 0.64 Hz),
6.4 (1 H, dd, J ) 3.4 and 1.9 Hz), 5.63 (1 H, s, acetal-H), 5.23
and 5.0 (2 H, d, J ) 18.3 Hz, H-21), 5.05 (1 H, J ) 5.8 Hz,
H-16), 3.25 (3 H, s, sulfonyl-CH3), 0.77 (3 H, s, C-18 CH3). 13C
NMR (CDCl3): δ 202.5, 199.7, 165.7, 148.1, 143.8, 124.8, 110.4,
110.3, 98.2, 82.8, 71.6, 48.7, 47.9, 47.0, 42.2, 39.4, 36.4, 35.2,
33.1, 31.3, 31.0, 26.5, 25.7, 14.6. MS (EI): m/z 505 (M+).
HRMS: calcd for C26H32O8S, 505.1896; found, 505.1898.
16r,17r-[(R)-(1′-r-F u r ylm eth ylid en e)d ioxy]-21[flu or o]-
oxy-19-n or p r egn -4-en e-3,20-d ion e (En d o, 1). To a stirred
solution of the methanesulfonate 18b (10 mg, 0.019 mmol) in
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails for the synthesis of compounds keto-alcohol 4, tertiary
alcohol 7, diene 8, furan acetal alcohols 16a and 16b, furan
acetal acetates 17a and 17b, and methanesulfonate 18a are
described and 1H NMR spectra of 7, 8, 13a , 15a , 13b, 15b, a
mixture 17a and 17b, and 18b are provided. This material is
available free of charge via the Internet at http://pubs.acs.org.
J O020190R
4910 J . Org. Chem., Vol. 67, No. 14, 2002