Facile preparation of perfluoro-tert-butyl ethers by the Mitsunobu reaction

Full text of DOI:10.1021/jo951515p

J . Org. Chem. 1996, 61, 361-362
361
Sch em e 1
F a cile P r ep a r a tion of P er flu or o-ter t-bu tyl
Eth er s by th e Mitsu n obu Rea ction
David P. Sebesta,* Sarah S. O’Rourke, and
Wolfgang A. Pieken
NeXstar Pharmaceuticals, Inc., 2860 Wilderness Place,
Boulder, Colorado 80301
Sch em e 2
Received August 16, 1995
In the course of our efforts toward incorporating func-
tionally modified nucleotide triphosphates in the genera-
tion of oligonucleotide libraries for screening,1 we were
in need of a fluorinated hydroxylamine derivative such
as 4 (Scheme 1). Two recent reports from the Falck
laboratory2 describing the preparation of fluoroalkyl
ethers by a modified Mitsunobu reaction have prompted
us to disclose our complementary findings in this area.
Although perfluoro-tert-butyl (PFTB) alcohol (3) has
recently become commercially available (Aldrich), a
review of the literature reveals only a single contribution
describing alkylations of this alcohol to form the corre-
sponding PFTB ethers.3 The protocol calls for generation
of the sodium salt of the alcohol and subsequent treat-
ment with alkyl halides in refluxing methyl ethyl ketone
for several days. In this way, moderate to good yields of
the corresponding PFTB ethers were achieved. In our
hands, however, treatment of tosylate 2 with NaOC(CF3)3
(in the presence of KI) under these conditions resulted
in a mere 11% yield of the desired ether. In light of this
low yield and the cumbersome reaction conditions, we
were prompted to consider alternative etherification
methodologies.
The triphenylphosphine/diethyl azodicarboxylate (TPP/
DEAD) protocol, known as the Mitsunobu reaction, has
been employed to alkylate a variety of nucleophiles with
primary and secondary alcohols.4 Treatment of an
alcohol with TPP/DEAD results in the formation of a
reactive triphenylphosphonium intermediate which is
displaced by the conjugate base of any of a number of
acids added subsequently to the reaction mixture. Car-
boxylic and phosphoric acids, phenols, imides and N-
hydroxyimides, halide and azide salts, heterocyclic amides,
and activated carbon acids have successfully been em-
ployed as Mitsunobu nucleophiles, however, with the
exception of intramolecular etherifications to form 3- to
7-membered rings, alcohols are typically not suitable.
This is presumably due to the relative pKa’s of most
alcohols and the carboxyhydrazine anion intermediate
which deprotonates the conjugate acid of the nucleophile.
Typically, alcohols (pKa ) 16-19)5 are not acidic enough
Sch em e 3
to be deprotonated to any appreciable extent and are not
sufficiently reactive to displace the phosphonium leaving
group to form ethers (except in the kinetically favorable
case of ring formation). We predicted, however, that
PFTB alcohol, substantially more acidic (pKa ) 9.5)6 than
aliphatic alcohols due to the electron-withdrawing effect
of the nine fluorine atoms, would be deprotonated under
the Mitsunobu reaction conditions and that the resultant
alkoxide would displace the oxophosphonium leaving
group under milder conditions than the previous alkyl-
ation protocol. In fact, treatment of a 0 °C THF solution
of primary alcohol 1 with 1.5 equiv of TPP and 1.5 equiv
of DEAD, followed after 5 min by 1.5 equiv of PFTB
alcohol, resulted in clean conversion to the desired PFTB
ether 4 within less than 30 min (Scheme 2). Chromato-
graphic purification of 4 afforded analytically pure mate-
rial in 84% yield.
Perfluoro-tert-butyl ethers of two of other alcohol
substrates were similarly prepared (Scheme 3). Benzo-
dioxan derivative 5 was converted to PFTB ether 6 in
80% yield, while geraniol (7) was converted to ether
derivative 8 in 59% yield.
(1) (a) Tuerk, C.; Gold, L. Science 1990, 249, 505-510. (b) Lin, Y.;
Qiu, Q.; Gill, S. C.; J ayasena, S. D. Nucleic Acid Res. 1994, 22, 5229-
5234. (c) McGee, D. P. C.; Vargeese, C.; Zhai, Y.; Kirschenheuter, G.
P.; Settle, A.; Siedem, C. R.; Pieken, W. A. Nucleosides Nucleotides
1995, 1329-1339.
(2) (a) Falck, J . R.; Yu, J .; Cho, H.-S. Tetrahedron Lett. 1994, 35,
5997-6000. (b) Cho, H.-S.; Yu, J .; Falck, J . R. J . Am. Chem. Soc. 1994,
116, 8354-8355.
(3) Pavlik, F. J . Perfluorotertiaryalkyl Ethers. U.S. Patent no.
3,981,928.
The few examples described here, along with the
complementary fluoroalkyl etherifications reported by the
Falck group,7 suggest that the Mitsunobu fluoroetheri-
fication represents a potentially useful methodology for
the formation of a synthetically challenging class of
compounds.
(4) (a) For a comprehensive review of the Mitsunobu Reaction, see:
Hughes, D. L. Org. React. 1992, 42, 335-656. (b) For recent applica-
tions in carbon-carbon bond formation, see: Falck, J . R.; Yu, J .
Tetrahedron Lett. 1992, 33, 6723-6726. Yu, J .; Cho, H.-S.; Falck, J .
R. J . Org. Chem. 1993, 58, 5892-5894. (c) For a recent application in
the formation of aryl glycosides, see: Roush, W. R.; Lin, X. J . Org.
Chem. 1991, 56, 5740-5742.
(5) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in
Organic Chemistry; Harper and Row: New York, 1987; Third Ed, p
300.
(6) Henne, A. L.; Francis, W. C. J . Am. Chem. Soc. 1953, 75, 991-
992.
0022-3263/96/1961-0361$12.00/0 © 1996 American Chemical Society

362 J . Org. Chem., Vol. 61, No. 1, 1996
Notes
4 days. The reaction mixture was concentrated in vacuo and
the residue dissolved in a minimum amount of dichloromethane
and purified on a column of 250 mL of silica gel packed in
hexanes (eluting with 250 mL each of 0, 10, 20, 30, 40, and 50%
EtOAc/hexanes) to afford 0.27 g (11%) of 4 as a white chalk.
b. Mitsu n obu Tech n iqu e. To a stirred solution of 2.75 g
(13.3 mmol) of 1 in 150 mL of THF was added 5.2 g (19.8 mmol)
of Ph3P. The solution was cooled to 0 °C and treated with 3.12
mL (19.2 mmol) of DEAD. After several minutes, 2.95 mL (21.2
mmol) of perfluoro-tert-butyl alcohol (3) was added. The cooling
bath was removed, and the mixture was stirred for 15 min, after
which time TLC analysis showed complete conversion of 1. The
reaction mixture was concentrated in vacuo and the residue
applied to a column of 600 mL of silica gel packed in hexanes.
The product was eluted with 500 mL each of 0, 10, 20, 30, 40,
and 50% EtOAc/hexanes to afford 4.66 g (83%) of 4 as a white
chalk: 1H NMR (DMSO) δ 7.87 (br s, 4 H), 4.46-4.42 (m, 4 H);
13C NMR (CDCl3) δ 164.21, 135.44, 129.5, 124.32, 120.84 (q, J
) 293.3 Hz), 80.3 (m), 76.25, 68.53; 19F NMR (DMSO) δ 54.46.
Exp er im en ta l Section
Reactions were performed in flame-dried glassware using
anhydrous solvents. Proton, 19F, and 13C NMR spectra were
measured in the indicated solvents at 300, 282, and 75 MHz,
respectively. Proton and 13C NMR spectra were calibrated
against internal standards (TMS and/or solvent), and 19F NMR
spectra were calibrated against an external standard (CF2Cl2).
O-(2-Hyd r oxyeth yl)-N-h yd r oxyp h th a lim id e (1).8 To a
stirred solution of 40.3 g (0.25 mol) of N-hydroxyphthalimide in
500 mL of DMF was added 22.3 g (0.27 mol) of NaOAc. The
dark red solution was stirred for 15 min, and then 19.1 mL (0.27
mol) of 2-bromoethanol was added via syringe. The mixture was
heated to 60 °C, stirred for 16 h, and then concentrated in vacuo.
The residue was dissolved in EtOAc and washed five to seven
times with NaHCO3 solution until none of the red to yellow color
of the N-hydroxyphthalimide anion remained in the aqueous
phase. The organic phase was dried over Na2SO4 and concen-
trated to afford 25.2 g (49%) of 1 as a white chalk: 1H NMR
(DMSO) δ 7.87 (br s, 4H), 4.82 (t, J ) 4.2 Hz, OH), 4.17 (t, J )
4.8 Hz, 2H), 3.69 (q, J ) 4.8 Hz, 2H); 13C NMR (CDCl3) δ 165.20,
135.52, 129.32, 124.40, 80.17, 59.71. Anal. Calcd for C10H9-
NO4: C, 57.95; H, 4.38; N, 6.76. Found: C, 57.89; H, 4.50; N,
6.60.
O-(2-(p-Tolu en esu lfon yloxy)eth yl)-N-h yd r oxyp h th a lim -
id e (2). To a stirred, 0 °C solution of 8.6 g (0.042 mol) of 1 in
80 mL of pyridine was added 9.52 g (0.05 mol) of TsCl. The
mixture was left in the refrigerator at 4 °C for 48 h and then
concentrated in vacuo. The residue was dissolved in EtOAc and
washed twice with HCl solution (1:1 concentrated HCl:H2O). The
organic phase was dried over Na2SO4 and concentrated. The
pale pink solid was triturated with Et2O to afford 8.5 g (56%)
of 2 as a white solid: 1H NMR (DMSO) δ 7.87 (s, 4H), 7.79
(d, J ) 8.4 Hz), 7.48 (d, J ) 8.1 Hz), 4.36-4.31 (m, 4H), 2.42
(s, 3H); 13C NMR (DMSO) δ 163.91, 145.95, 135.56, 132.80,
130.95, 129.29, 128.41, 123.99, 75.31, 68.73, 21.20. Anal. Calcd
for C17H15NO6S: C, 56.50; H, 4.18; N, 3.88. Found: C, 56.52;
H, 4.31; N, 3.98.
Anal. Calcd for
C14H8NO4F9: C, 39.54; H, 1.90; N, 3.29.
Found: C, 39.54; H, 2.03; N, 3.28.
2-((P er flu or o-ter t-bu tyloxy)m eth yl)-1,4-ben zod ioxa n (6).
To a stirred, 0 °C solution of 0.65 g (3.9 mmol) of 5 in 40 mL of
THF was added 1.54 g (5.9 mmol) of Ph3P, followed by 0.92 mL
(5.9 mmol) of DEAD. After 2 min, 0.87 mL (6.2 mmol) of
perfluoro-tert-butyl alcohol (3) was added. The cooling bath was
removed, and the mixture was stirred for 5 h and concentrated
in vacuo. The residue was triturated with hexanes, and the
filtrate was concentrated to afford 2 g of a yellow oil. This
material was purified on a column of 100 mL of silica gel (eluting
with 0, 5, 10, and 15% EtOAc in hexanes) to afford 1.2 g (80%)
of 6 as a faint yellow wax: 1H NMR (CDCl3) δ 6.92-6.83 (m,
4H), 4.42-4.40 (m, 1H), 4.32-4.09 (m, 4H); 13C NMR (CDCl3) δ
142.9, 142.5, 122.0, 121.8, 120.2 (q, J ) 291 Hz), 117.35, 79.7
(m, J ) 30 Hz), 70.83, 67.42, 64.39; 19F NMR (CDCl3) δ 74.7.
Anal. Calcd for C13H9O3F9: C, 40.62; H, 2.36. Found: C, 40.81;
H, 2.44.
Ger a n yl P er flu or o-ter t-bu tyl Eth er (8). To a stirred
solution of 0.52 mL (3 mmol) of geraniol (7) in 15 mL of THF
was added 1.18 g (4.5 mmol) of Ph3P. The solution was cooled
to 0 °C and treated with 0.71 mL (4.5 mmol) of DEAD. After
several minutes, 0.67 mL (4.8 mmol) of perfluoro-tert-butyl
alcohol (3) was added and the cooling bath removed. The
mixture was allowed to warm to ambient temperature overnight.
The reaction mixture was concentrated in vacuo and the residue
applied to a column of 100 mL of silica gel packed in hexanes.
The product was eluted using 250 mL each of hexanes and 10%
and 20% EtOAc in hexanes to afford 0.65 g (59%) of 8 as a
colorless oil: 1H NMR (400 MHz, CDCl3) δ 5.34 (dd, J ) 7.0, 1.1
Hz, 1H), 5.09-5.05 (m, 1H), 4.51 (d, J ) 7.0 Hz, 2H), 2.18-2.05
(m, 4H), 1.68 (s, 6H), 1.59 (s, 3H); 19F NMR (CDCl3) δ 70.22.
Anal. Calcd for C14H17OF9: C, 45.17; H 4.60. Found: C, 46.77;
H 5.08.
O-(2-(P er flu or o-ter t-bu tyloxy)eth yl)-N-h yd r oxyp h th a l-
im id e (4): a . Alk oxid e Alk yla tion Tech n iqu e.3 To a stirred,
0 °C slurry of 0.13 g (5.53 mmol) of NaH in 10 mL of THF was
added 0.77 mL (5.53 mmol) of perfluoro-tert-butyl alcohol (3).
After 15 min, the mixture was concentrated in vacuo. The
residue was dissolved in 20 mL of methyl ethyl ketone and this
solution treated with 2.0 g (5.53 mmol) of 2 and 92 mg (0.55
mmol) of KI. The mixture was heated to reflux and stirred for
(7) Professor Falck reports etherifications using (CF3)2CHOH,
CF3(CF2)6CH2OH, (CF3)2PhCOH, and CF3CH2OH as acidic alcohol
nucleophiles under Mitsunobu (TPP/ DEAD) and modified Mitsunobu
(Bu3P, 1,1′-(azodicarbonyl)dipiperidine) conditions.2a,b We also observed
alkylation of CF3CF2CH2OH and C6F5CH2OH with alcohol 1; however,
under our unoptimized conditions, these etherifications were not
particularly efficient (21-36% yields). Falck describes the necessity
for heating reactions employing less acidic fluoro alcohols (pKa 11-
13),2a and these data may suggest a pKa threshold above which acids
are not deprotonated to a sufficient extent for efficient Mitsunobu
reaction to occur at ambient temperatures.
Ack n ow led gm en t. The authors would like to ex-
press their appreciation to Mr. J ames Reed for his
assistance with NMR and HPLC instrumentation.
(8) For N-hydroxyphthalimide alkylation, see: Rougny, A.; Daudon,
M. Bull. Soc. Chim. Fr. 1976, 833-838.
J O951515P