Synthesis of α-CF3-substituted carbonyl compounds with relative and absolute stereocontrol using electrophilic CF3-transfer reagents

Article abstract of DOI:10.1021/ol2023328

Evans-type chiral lithium imide enolates undergo diastereoselective α-trifluoromethylation with a hypervalent iodine-CF₃ reagent with up to 91% combined isolated yield and 97:3 dr. The resulting isolated diastereopure products can be further transformed into valuable products without racemization.

Full text of DOI:10.1021/ol2023328

ORGANIC
LETTERS
2011
Vol. 13, No. 21
5762–5765
Synthesis of r-CF3-Substituted Carbonyl
Compounds with Relative and Absolute
Stereocontrol Using Electrophilic CF3-
Transfer Reagents
†
Vaclav Matousek, Antonio Togni,* Vincent Bizet, and Dominique Cahard*
,†
‡
,‡
ꢀ
ꢁ
Department of Chemistry and Applied Biosciences, Swiss Federal Institute of
Technology, ETH Zu€rich, CH-8093 Zu€rich, Switzerland, and UMR 6014 CNRS
ꢀ
COBRA, Universite et INSA de Rouen, 76130 Mont-Saint-Aignan, France
togni@inorg.chem.ethz.ch; dominique.cahard@univ-rouen.fr
Received August 28, 2011
ABSTRACT
Evans-type chiral lithium imide enolates undergo diastereoselective R-trifluoromethylation with a hypervalent iodineꢀCF3 reagent with up to 91%
combined isolated yield and 97:3 dr. The resulting isolated diastereopure products can be further transformed into valuable products without
racemization.
The remarkable utility of organofluorine compounds in
many branches of the fine chemicals industry and in life
and materials science is growing in importance.1 Indeed,
the presence of highly fluorinated alkyl groups often
triggers significant changes in the physical and chemical
properties of molecules, such as solubility, lipophilicity,
increased metabolic stability, and better bioavailability.2
The introduction of a trifluoromethyl group, the most-
sought fluorinated alkyl group, into a selected target in a
reliable and effective fashion undoubtedly belongs among
the most fundamental and desired transformations in
fluorine chemistry.3 Of specific interest are methodologies
appropriateforlate-stagetrifluoromethylationinmultistep
syntheses, which renders them a valuable tool when juxta-
posed with the classical “building block approach”.4 Early
work in this field by Iseki and co-workers demonstrated
that the well-established technique of diastereoselective
functionalization of chiral imide enolates, originally devel-
oped by Evans,5 can be successfully extended to radical
trifluoromethylation.6 Althoughhighdiastereomericratios
up to 93:7 were obtained for radical trifluoromethylation,6
a large excess of trifluoroiodomethane (a gaseous and
† ETH Z€urich.
‡
ꢀ
UMR 6014 CNRS COBRA, Universite et INSA de Rouen.
(1) For recent reviews and monographs, see: (a) Isanbor, C.; O’Hagan,
€
D. J. Fluorine Chem. 2006, 127, 303. (b) Muller, K.; Faeh, C.; Diederich,
F. Science 2007, 317, 1881. (c) Purser, S.; Moore, P. R.; Swallow, S.;
Gouverneur, V. Chem. Soc. Rev. 2008, 37, 320. (d) Furuya, T.; Kamlet,
A. S.; Ritter, T. Nature 2011, 473, 470. (e) Tomashenko, O. A.; Grushin,
V. V. Chem. Rev. 2011, 111, 4475. (f) Kirsch, P. Modern Fluoroorganic
Chemistry: Synthesis, Reactivity, Applications; Wiley-VCH: Weinheim,
Germany, 2004. (g) Uneyama, K. Organofluorine Chemistry; Blackwell
Publishing: Oxford, U.K., 2006.
(2) (a) Thayer, A. M. Chem. Eng. News. 2006, 84, 15. (b) Leroux,
F. R.; Manteau, B.; Vors, J.-P.; Pazenok, S. Beilstein J. Org. Chem. 2008,
4, 13. (c) The Pesticide Manual: A World Compendium, 13th ed., version
3.2; Tomlin, C. D. S., Ed.; British Crop Production Council: Hampshire,
U.K., 2005. (d) Kleemann, A.; Engel, J.; Kutscher, B.; Reichert, D. Phar-
maceutical Substances, electronic version; Thieme Chemistry: Stuttgart,
Germany, 2006.
(3) For recent reviews about stereoselective introduction of fluori-
nated groups, see: (a) Ma, J.-A.; Cahard, D. Chem. Rev. 2008, 108, PR1.
(b) Shibata, N.; Mizuta, S.; Kawai, H. Tetrahedron: Asymmetry 2008,
19, 2633. (c) Shibata, N.; Matsnev, A.; Cahard, D. Beilstein J. Org.
ꢀ
Chem. 2010, 6, 65. (d) Valero, G.; Companyo, X.; Rios, R. Chem.;Eur.
J. 2011, 17, 2018.
(4) Nie, J.; Guo, H.-C.; Cahard, D.; Ma, J.-A. Chem. Rev. 2011, 111,
455.
(5) Evans, D. A.; Ennis, M. D.; Mathre, D. J. J. Am. Chem. Soc. 1982,
104, 1737.
(6) (a) Iseki, K.; Nagai, T.; Kobayashi, Y. Tetrahedron Lett. 1993, 34,
2169. (b) Iseki, K.; Nagai, T.; Kobayashi, Y. Tetrahedron: Asymmetry
1994, 5, 961.
r
10.1021/ol2023328
Published on Web 10/03/2011
2011 American Chemical Society

rather costly reagent) had tobe employed, and O2 had to be
present to initiate the radical process. Later, Umemoto and
co-workers showed that ate complexes obtained from
potassium ketone enolates and stochiometric amounts of
atropisomeric chiral boroles allow enantioselective trifluoro-
methylation by means of S-CF3 dibenzothiophenium
salts in up to 45% ee.7 A related strategy was employed
by Cahard and co-workers for the electrophilic trifluoro-
methylation of ammonium enolates of β-keto esters gen-
erated by means of cinchona alkaloids, which afforded
products with up to 71% ee.8 Even more recent work by
Shibata, Cahard, and co-workers revealed that chiral
guanidines also lead to enantiodiscrimination of the en-
olate stereofaces through H-bonding to β-keto esters.9
Chiral imidazolidinone organocatalysts that transiently
form enamines from primary aldehydes yield highly en-
antiopure trifluoromethylated products (in up to 99% ee)
upon reaction with a large excess of trifluoroiodomethane
under radical photolytic conditions.10 In addition, a nearly
equal range of substrates undergo enantioselective electro-
philic trifluoromethylation with hypervalent iodine CF3-
transfer reagents under Lewis acid cocatalysis with prac-
tically the same ee values. However, an excess of the
substrate had to be employed.11
We focused our early efforts on the diastereoselective
trifluoromethylation of chiral imide enolates by varying
the nature of the metal. We began our study by reacting 1
or 2 with the (Z)-boron enolate of 3a obtained using
€
dibutyl boron triflate and Hunig’s base. Disappointingly,
we could not detect any desired trifluoromethylation. The
zinc enolate generated from diethylzinc and 3a anal-
ogously did not furnish any trifluoromethylated product.
We speculated that a more reactive enolate might even-
tuallyserve asa moreeffectivenucleophile intheattempted
trifluoromethylation with 1 or 2. To our delight, lithiation
of 3a with 1.1 equiv of LDA in THF at ꢀ78 °C followed by
addition of 1.1 equiv of 1 led to the formation of the R-
trifluoromethylated diastereomers4a in 60% isolated yield
and 89:11 dr (Table 1). Encouraged by this promising
Table 1. Screening of Reaction Conditions
conv
dr
entry
1
conditions
(%)a,b of 4aa
1.1 equiv of LDA, THF, ꢀ78 °C, 30 min,
then 1.1 equiv of 1 at ꢀ78 °C to rt over 4 h
same as 1 but with 1.1 equiv of LiHMDS
same as 1 but with 1.1 equiv of NaHMDS
same as 1 but with 1.1 equiv of KHMDS
same as 2 but with transmetalation using
1.1 equiv of CuI at 0 °C 30 min prior to
addition of 1
100 (60) 89:11
2
3
4
5
100 (82) 90:10
90
50
ꢀ
88:12
75:25
ꢀ
Figure 1. The most prominent examples of hypervalent iodineꢀ
6
same as 5 but with transmetalation
using 1.1 equiv of ZnCl2 at ꢀ78 °C 30 min
prior to addition of 1
ꢀ
ꢀ
CF3 reagents.
a Determined by 19F NMR analysis. b Isolated yields are shown in
parentheses.
In the context of the newly expanding trifluoromethyla-
tion chemistry of hypervalent CF3ꢀiodine compounds
developed in one of our groups,12 we reasoned that an
enantiopure enolate wouldundergodiastereoselectiveelec-
trophilic trifluoromethylation after treatment with either
of the reagents 1 or 2 (Figure 1).13
result, we continued our screening of reaction conditions,
varying parameters such as the solvent, base, counterion,
and additive. As is evident from Table 1, lithium enolates
show excellent reactivity and afford high dr values of the
R-trifluoromethylated product 4a. It is noteworthy that
LiHMDS gave better isolated yields than LDA. Switching
to the corresponding sodium enolate maintains the origi-
nal dr value, but a decline in reactivity starts to become
apparent. The potassium enolate shows both the worst
reactivity and diastereoselectivity of all the tested alkali
metal enolates. Attempted trifluoromethylation, after
transmetalation of lithium enolates with Cu(I) or Zn(II)
salts, resulted in the decomposition of 1, while 2 showed
itself to be practically inert toward the basic metal enolates
listed in Table 1. Simple treatment of 3aand 2 in DCM at rt
(7) Umemoto, T.; Adachi, K. J. Org. Chem. 1994, 59, 5692.
(8) Ma, J.-A.; Cahard, D. J. Fluorine Chem. 2007, 128, 975.
(9) Noritake, S.; Shibata, N.; Nomura, Y.; Huang, Y.; Matsnev, A.;
Nakamura, S.; Toru, T.; Cahard, D. Org. Biomol. Chem. 2009, 7, 3599.
(10) Nagib, D. A.; Scott, M. E.; MacMillan, D. W. C. J. Am. Chem.
Soc. 2009, 131, 10875.
(11) Allen, A. E.; MacMillan, D. W. C. J. Am. Chem. Soc. 2010, 132,
4986.
(12) (a) Eisenberger, P.; Gischig, S.; Togni, A. Chem.;Eur. J. 2006,
12, 2579. (b) Eisenberger, P.; Kieltsch, I.; Armanino, N.; Togni, A.
Chem. Commun. 2008, 1575. (c) Stanek, K.; Koller, R.; Togni, A. J. Org.
Chem. 2008, 73, 7678. (d) Koller, R.; Stanek, K.; Stolz, D.; Aardoom, R.;
Niedermann, K.; Togni, A. Angew. Chem., Int. Ed. 2009, 48, 4332. (e)
Fantasia, S.; Welch, J. M.; Togni, A. J. Org. Chem. 2010, 75, 1779. (f)
Koller, R.; Huchet, Q.; Battaglia, P.; Welch, J. M.; Togni, A. Chem.
€
Commun. 2009, 5993. (g) Niedermann, K.; Fruh, N.; Vinogradova, E.;
in the presence of 30 mol % CuBr Me2S resulted only in
3
Wiehn, M. S.; Moreno, A.; Togni, A. Angew. Chem., Int. Ed. 2011, 50,
1059. (h) Kieltsch, I.; Eisenberger, P.; Togni, A. Angew. Chem., Int. Ed.
€
2007, 46, 754. (i) Capone, S.; Kieltsch, I.; Flogel, O.; Lelais, G.; Togni, A.
(13) For recent reviews about hypervalent iodine-mediated R-func-
ꢀ
Helv. Chim. Acta 2008, 91, 2035. (j) Santschi, N.; Togni, A. J. Org. Chem.
2011, 76, 4189. (k) Wiehn, M. S.; Vinogradova, E. V.; Togni, A. J.
Fluorine Chem. 2010, 131, 951.
tionalization, see: (a) Brand, J. P.; Gonzalez, D. F.; Nicolai, S.; Waser, J.
Chem. Commun. 2011, 47, 102. (b) Merritt, E. A.; Olofsson, B. Synthesis
2011, 517.
Org. Lett., Vol. 13, No. 21, 2011
5763

partial decomposition of 2 to CF3Br but no detectable
trifluoromethylation. The high reactivity and selectivity of
lithium enolates is explained by the strong Lewis acidity of
the counterion, which is considered to lead to tighter
chelation and possibly to partial CF3-reagent activation.
Solvent screening was limited to ethereal solvents in
which alkali metal enolates are commonly prepared. The
use of diethyl ether under identical conditions revealed
serious solubility problems both of substrate 3a and re-
agent 1, leading to improper mixing and resulting in
sluggish reactions and low conversions. Increasing the
amount of 1 to 1.4 equiv led to an 86% isolated yield of
diastereomers that were separable by column chromatog-
raphy. Under identical reaction conditions, Umemoto’s
reagent (S-CF3 dibenzothiophenium triflate) also proved
reactive toward the Li enolate of 3a, with almost the same
diastereoselectivity (89:11). However, incomplete conver-
sion and a lower isolated yield of 4a was observed. Further
increasing the amount of LiHMDS to 1.4 equiv and the
amount of 1 to2 equivraisedthe combined isolatedyield of
4a to 93%. As an acceptable compromise in terms of yield
and reagent consumption, we identified the optimal con-
ditions to be lithiation of the substrate with 1.1 equiv of
Table 2. Substrate Scope
R
product
yield (%)
dr
Bn
Ph
4a
5
86
83
40
46
75
91
83
74
ꢀ
90:10
88:12
77:23
81:19
96:4
94:6
97:3
86:14
ꢀ
2-naphthyl
Me
6
7
i-Pr
8
c-hex
9
t-Bu
10
11
12
13
OBn
2-thiophenyl
N-phth
ꢀ
ꢀ
A survey of the substrates depicted in Table 2 delineates
two evident trends: (1) With increasing steric bulk of the
acyl group attached to the oxazolidinone, the isolated yield
increases, culminating with the cyclohexylacetyl substitu-
ent. (2) A similar tendency can be observed in the depen-
dence of diastereoselectivity on the steric bulk of the acyl
group, as clearly demonstrated in the aliphatic series of
acylgroups (7ꢀ10). Thehighest drvalue wasobtainedwith
the most bulky tert-butylacetyl-substituted oxazolidinone
10, whereas the lowest dr value was linked to the smaller
propanoyl-substituted oxazolidinone 7. We attribute the
lower observed dr for 6 to the well-known propensity of
arylacetic acids and their derivatives to undergo base-
induced epimerization.15 Indeed, R-trifluoromethylated
product 5 was obtained with 88:12 dr under our optimized
reaction conditions, but longer standing of the reaction
mixture at 0 °C led to significant erosion of the product dr
and partial product decomposition. Importantly, under
the same reaction conditions, 4a did not show any notice-
able decrease in dr. Instead of clean R-trifluoromethyla-
tion, the attempted synthesisof 12led toa complex mixture
arising from extensive heteroaromatic trifluoromethyla-
tion. Access to compound 13 also appeared proble-
matic because the metalation step suffered from severe
complications, presumably because the freshly generated
enolate can undergo facile decomposition. Attempts at
R-trifluoromethylation of tertiary acylcarboxamides did
not meet with any success. In all the cases where trifluoro
methylation could be accomplished, the resulting diaste-
reomeric R-trifluoromethylated products could be purified
by either preparative silica gel column chromatography or
crystallization, delivering diastereopure compounds. X-ray
analysis of the major diastereomer of 4a (see Figure 3)
unequivocally ascertained the absolute configuration of
Figure 2. Screening of various chiral auxiliaries.
LiHMDS at ꢀ78 °C in THF followed by addition of
1.4 equiv of 1 all at once and slow warming to rt over 4 h.
Next, a screening of various chiral auxiliaries (Figure 2)
led us to conclude that auxiliaries containing aromatic
groups, preferably with a rigidified skeleton (4d) offer the
best performance with regardto dr. The cis-aminoborneol-
derived auxiliary proved to be detrimental with respect to
diastereofacial discrimination of the corresponding Li
enolate. Although the cis-aminoindanol-derived auxiliary
gave the best dr, we decided to investigate the scope of
(4R)-4-phenyl-1,3-oxazolidin-2-one, as in 4a, because it
can be more readily accessed from relatively cheap
(R)-phenylglycine in two synthetic steps.14 (4R)-4-Phenyl-1,
3-oxazolidin-2-one constitutes a compromise between dia-
stereoselectivity and easy access.
Having identified a suitable chiral auxiliary, we turned
our efforts to exploring the substrate scope of the reaction.
(14) (a) McKennon, M. J.; Meyers, A. I.; Drauz, K.; Schwarm, M.
J. Org. Chem. 1993, 58, 3568. (b) Ito, Y.; Sasaki, A.; Tamoto, K.;
Sunagawa, M.; Terashuma, S. Tetrahedron 1991, 47, 2801.
(15) (a) Xie, Y. C.; Liu, H. Z.; Chen, J. Y. Biotechnol. Lett. 1998, 20,
455. (b) Yuchun, X.; Huizhou, L.; Chen, J. Y. Int. J. Pharm. 2000,
196, 21.
5764
Org. Lett., Vol. 13, No. 21, 2011

Figure 3. Putative mechanism of trifluoromethylation and X-ray structure of the major diastereomer of 4a.
the newly created stereocenter to be R, resulting from the
kineticallyfasterattack of the trifluoromethylating reagent
1 at the sterically less shielded re stereoface of the corre-
sponding Li enolate.
tion-free synthetic transformations of diastereopure R-
CF3-acyloxazolidinones (Scheme 1). Reduction of pure
It is interesting to note that there have already been
attempts in our group to trifluoromethylate Li enolates of
ketones and carboxylic acid esters. However, no appro-
priate conditions for successful R-trifluoromethylation
with either 1 or 2 could be found. Evans-type acyloxazo-
lidinones represent in this context a rather welcome excep-
tion. In an attempt to find a rationale for the success of this
class of enolizable carbonyl compounds, we assume the
following speculative mechanism: the Li enolate of acyl
oxazolidinone 3a is stereoselectively attacked by 1, which
leads to exchange of the internal carboxylate ligand
at the hypervalent iodine center, affording R-iodanyl
intermediate 14 (Figure 3) by analogy to arylation of
malonates16 and acetoxylation17 and tosyloxylation18
of ketones, although an alternative formation of an OꢀI
intermediate cannot be conclusively ruled out.19 The ox-
azolidinone moiety is believed to function as a weak,
neutral monodentate oxygen-centered ligand, increasing
the stability of intermediate 14, which subsequently under-
goes stereoretentive ligand coupling to give 4a and lithium
2-iodobenzoate. The beneficial role of the weakly coordi-
nating oxazolidinone20 is also believed to lie first in in-
creasing the steric bulk around the iodine center (thus
significantly decreasing the rate of ligand exchange at the
hypervalent iodine center) and second in protecting the
activated R-position of the carbonyl substrate in 14 from
attack by external nucleophiles. The oxazolidinone moiety
thus promotes selective monosubstitution at the hyperva-
lent iodine center, leading to productive R-trifluoromethy-
lation and suppressing unproductive side reactions.
As an extension of the presented trifluoromethylation
protocol, we explored the potential of further racemiza-
Scheme 1. Acces to Enantiopure R-CF3 Alcohol and Acid
Derivatives
4awith NaBH4 inTHF/water cleanly gavethe correspond-
ing product (2R)-2-benzyl-3,3,3-trifluoropropan-1-ol (15)
in 94% yield with 99þ% ee. Similarly, treatment of 4a with
H2O2 in the presence of 1.6 equiv of LiOH gave, after
acidification, the corresponding product (2R)-2-benzyl-
3,3,3-trifluoropropanoic acid (16) in 86% yield with
99þ% ee, indicating full conservation of stereointegrity.
To conclude, we have successfully developed an efficient
protocol, mechanistically distinct from Iseki’s radical ap-
proach, for trifluoromethylation of chiral imide enolates
using electrophilic CF3-transfer reagents that affords iso-
lated yields of up to 91% with diastereomeric ratios up to
97:3. The “acid reagent” appeared to show somewhat
higher reactivity toward chiral lithium imide enolates than
Umemoto’s reagent. The oxazolidinone substituent is
believed to act beneficially by suppressing the extent of
unproductive side reactions, leading to clean R-trifluoro-
methylation. The resulting isolated diastereopure products
could be further transformed into valuable products with-
out racemization.
(16) Oh, C. H.; Kim, J. S.; Jung, H. H. J. Org. Chem. 1999, 64, 1338.
(17) Ochiai, M.; Takeuchi, Y.; Katayama, T.; Sueda, T.; Miyamoto,
K. J. Am. Chem. Soc. 2005, 127, 12244.
€
Acknowledgment. R. Aardoom (ETH Zurich) is ac-
knowledged for carrying out the X-ray crystal structure
(18) Altermann, S. M.; Richardson, R. D.; Page, T. K.; Schmidt,
R. K.; Holland, E.; Mohammed, U.; Paradine, S. M.; French, A. N.;
Richter, C.; Bahar, A. M.; Witulski, B.; Wirth, T. Eur. J. Org. Chem.
2008, 5315.
(19) Norrby, P.-O.; Petersen, T. B.; Bielawski, M.; Olofsson, B.
Chem.;Eur. J. 2010, 16, 8251.
(20) (a) Harvey, J. S.; Simonovich, S. P.; Jamison, C. R.; MacMillan,
D. W. C. J. Am. Chem. Soc. 2011, 133, 13782. (b) Bigot, A.; Williamson,
A. E.; Gaunt, M. J. J. Am. Chem. Soc. 2011, 133, 13778.
study of compound 4a. This work was supported by ETH
€
€
Stael programme for financial support to V.M. and V.B.
Zurich. Wethank the bilateral FrenchꢀSwiss Germaine de
Supporting Information Available. Experimental details
and copies of 1H, 13C, and 19F spectra. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
Org. Lett., Vol. 13, No. 21, 2011
5765