Conformational fixation of enolates by intramolecular metal...Fluorine interaction

Article abstract of DOI:10.1021/ol990821c

(equation presented) Methyl acetates with fluorine-containing auxiliaries at their 2 position were demonstrated to react smoothly with various electrophiles in a re face preferential manner (up to 90% de). This was interpreted as the result of an intermed

Full text of DOI:10.1021/ol990821c

ORGANIC
LETTERS
1999
Vol. 1, No. 6
905-908
Conformational Fixation of Enolates by
Intramolecular Metal‚‚‚Fluorine
Interaction
,†
Takashi Yamazaki,* Makoto Ando,† Tomoya Kitazume,† Toshio Kubota,‡ and
Masao Omura‡
Department of Bioengineering, Tokyo Institute of Technology, 4259 Nagatsuta-cho,
Midori-ku, Yokohama 226-8501, Japan, and Department of Materials Science, Ibaraki
UniVersity, Nakanarusawa 4-12-1, Hitachi, Ibaraki 316-8511, Japan
tyamazak@bio.titech.ac.jp
Received July 16, 1999
ABSTRACT
Methyl acetates with fluorine-containing auxiliaries at their 2 position were demonstrated to react smoothly with various electrophiles in a re
face preferential manner (up to 90% de). This was interpreted as the result of an intermediary enolate constructing a bicyclo[3.3.0] system by
the concomitant intramolecular chelation of a metal with r-oxygen and fluorine, resulting in the auxiliary effectively blocking attack from the
opposite si face.
Stereoselective introduction of electrophiles R to a carbonyl
moiety has been extensively investigated.1 For instance, esters
or amides possessing stereogenic center(s) at the correspond-
ing alcohol or amine, respectively, sometimes demonstrate
excellent diastereoselectivities by successful formation of
conformationally rigid enolates leading to ready discrimina-
tion of two enolate π faces. On the other hand, while it seems
to be an intriguing and promising route to stereoselectively
construct a new carbon-carbon bond with the aid of an
auxiliary directly connected to the reaction site, only a small
number of examples have appeared in the literature thus far.2
This led us to investigate the diastereoinducing ability of
tertiary alkoxy auxiliaries at the 2 position of methyl acetate.
Our present plan is based on the expectation that fluorine in
1 (RdCH3-nFn, n ) 1-3) would successfully participate in
fixation of enolate conformations3 by intramolecular F‚‚‚metal
interaction.4,5 If this is really the case, the environment around
two enolate π faces becomes more or less different, which
consequently results in diastereofacially discriminating reac-
tions with appropriate electrophiles. To inspect such ability
in detail, three types of fluorinated substrates 1a, 1b, and
(2) R-R*2N substrates: (a) Panek, J. S.; Masse, C. E. J. Org. Chem.
1998, 63, 2382. (b) Blaser, D.; Ko, S.-Y.; Seebach, D. J. Org. Chem. 1991,
56, 6230. (c) Tabcheh, M.; El Achqar, A.; Pappalardo, L.; Roumestant,
M.-L.; Viallefont, P. Tetrahedron 1991, 47, 4611. (d) Chaari, M.; Jenhi,
A.; Lavergne, J.-P.; Viallefont, P. Tetrahedron 1991, 47, 4619. (e) McIntosh,
J. M.; Leavitt, R. K.; Mishra, P.; Cassidy, K. C.; Drake, J. E.; Chadha, R.
J. Org. Chem. 1988, 53, 1947. R-R*O substrates: (a) Fujita, M.; Laine´, D.
Ley, S. V. J. Chem. Soc., Perkin Trans. 1 1999, 1647. (b) d'Angelo, J.;
Page`s, O.; Maddaluno, J.; Sumas, F.; Revial, G. Tetrahedron Lett. 1983,
24, 3951. See also Touzin, A. M. Tetrahedron Lett. 1975, 1477.
(3) (a) Ooi, T.; Kagoshima, N.; Uraguchi, D.; Maruoka, K. Tetrahedron
Lett. 1998, 39, 7105. (b) Ooi, T.; Kagoshima, N.; Maruoka, K. J. Am. Chem.
Soc. 1997, 119, 5754. (c) Ooi, T.; Kondo, Y.; Miura, T.; Maruoka, K.
Tetrahedron Lett. 1997, 38, 3951.
(4) (a) Yamazaki, T.; Kitazume, T. In Enantiocontrolled Synthesis of
Fluoro-Organic Compounds: Stereochemical Challenges and Biomedicinal
Targets; Soloshonok, V. A., Ed.; John Wiley & Sons: New York, 1999;
Chapter 19. (b) Murphy, E. F.; Murugavel, R.; Roesky, H. W. Chem. ReV.
1997, 97, 7, 3425. (c) Plenio, H. Chem. ReV. 1997, 97, 3363.
(5) Although such interaction has been computationally expected, very
little experimental evidence has been reported on the metal‚‚‚fluorine
interaction. See ref 4 for details.
† Tokyo Institute of Technology.
‡ Ibaraki University.
(1) (a) Evans, D. A. In Asymmetric Synthesis; Morrison, J. D., Ed.;
Academic Press: New York, 1984; Vol. 3, p 1. (b) Chiral Auxiliaries and
Ligands in Asymmetric Synthesis; Seyden-Penne, J., Ed.; John Wiley &
Sons: New York, 1995; p 157.
10.1021/ol990821c CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/17/1999

1c, were prepared as their racemic forms along with
nonfluorinated 1d with an isopropyl group, sometimes
referred to as a steric equivalent to a CF3 moiety,6 as a
reference.
as a base7 in THF at -78 °C along with the addition of 1.2
equiv of an appropriate electrophile to the resultant potassium
enolate.
Reactions of the trifluorinated substrate 1a were carried
out with various active electrophiles under the conditions
depicted above (entries 7-13),8,9 and it was observed as the
general trend that alkylation furnished the products 2a to 6a
in good to excellent yields with diastereoselectivities (DS)
in the range of 68-90% de. Addition of HMPA affected
DS values of products only slightly (entries 8 and 9), and
sufficiently increased the enolate reactivity to form the adduct
7a with ethyl iodide which could not be obtained at all
without this additive (entry 14).
First, trifluoromethyl-containing ester 1a was subjected to
conditions under various combinations of bases and solvents,
and the resultant enolate solution was then treated with
benzyl bromide as a representative electrophile (Scheme 1,
Scheme 1
Table 2 describes the results (yields and steric factors8,10
)
when esters 1a to 1d are independently reacted with such
typical electrophiles as methyl iodide, benzyl bromide, and
allyl iodide. A stepwise decrease of the number of fluorine
atoms resulted in lower alkylation yields while the DS values
of the products did not seem to be remarkably influenced as
long as substrates contained at least one fluorine atom. In
quite a sharp contrast, the nonfluorinated counterpart 1d was
found to be the less potent nucleophile especially in terms
of diastereoselectivity, 30-40% lower than the others.
Relative stereochemistries of 3a and 3b were unambiguously
clarified as 1′S*,2R* by X-ray crystallographic analyses after
their derivatization into the corresponding benzyl amides 8a
and 8b,8,11,12 respectively, with complete retention of the
original DS values (Scheme 1).
Table 1). Although bases with lithium as a countercation
provide low yields of product 2a (entries 1 and 2), KHMDS
proved to work quite efficiently to form 2a in 79% isolated
yield with 80% diastereomeric excess (entry 4). Use of Et2O
instead of THF resulted in a decrease of enolate reactivity
(entry 5). Further study allowed us to conclude that the best
condition would be employment of 1.2 equiv of KHMDS
Table 2. Representative Reaction Results
isolated yield (%) [DS, % de]
steric factorsa
R
MeI
PhCH2Br
allyl-I
Es
Es′
CF3 (1a )
67 [90]
58 [88]
50 [82]
50 [56]
79 [80]
45 [84]
41 [84]
31 [48]
82 [68]
61 [72]
49 [62]
77 [28]
-1.16
-0.67
-0.24
-0.47
-0.78
-0.32
-0.20
-0.48
CHF2 (1b)
CH2F (1c)
i-Pr (1d )
Table 1. Reaction of 1a with Various Electrophilesa
a Es: Taft steric constant with reference to a methyl group (Me: 0.00).
Es': revised Taft steric constant by Dubois and co-workers. Me: 0.00. Ph:
-2.31.
entry
EX
time (h)
yieldb (%) DS (% de)
1c
2c
3c
4
5d
6e
7
8f
9f
10
11
12
13
14f
PhCH2Br
PhCH2Br
PhCH2Br
PhCH2Br
PhCH2Br
PhCH2Br
MeI
2.0
1.0
2.0
2.0
14.0
1.5
1.0
1.0
1.0
1.5
2.0
0.5
2.0
2.0
(12)
(trace)
(41)
(85)
(9)
86
76
80
83
75
90
80
76
68
70
80
76
76
79
67
A similar diastereofacial preference was demonstrated
when benzaldehyde was employed as an electrophile. Thus,
as shown in Scheme 2, 1a furnished all possible stereoiso-
meric mixtures of 98 in a ratio of 57:32:8:3 which were
chromatographically separated into two groups consisting of
the first and third compounds (2R*,3S*-9 and 2S*,3R*-9,
respectively) and the second and fourth isomers (2R*,3R*-9
and 2S*,3S*-9, respectively). Independent oxidation of these
two groups by PDC did not affect the original isomeric ratios,
(37)
(80)
(69)
(40)
(85)
(58)
(68)
(61)
(37)
MeI
MeI
CH2dCHCH2I
CH2dCHCH2Br
CHtCCH2Br
(E)-PhCHdCHCH2I
EtI
82
32
60
58
37
a Unless otherwise noted, 1.2 equiv of KHMDS was added to 1a in THF
at -78 °C, and after 0.5 h, an appropriate electrophile (1.2 equiv) was added.
b The yield determined by 19F NMR is shown is parentheses. c LDA,
LHMDS, and NaHMDS were employed instead of KHMDS in entries 1,
2, and 3, respectively. d In Et2O. e Reaction was conducted at 0 °C. f 3.0,
10.0, or 1.0 equiv of HMPA (entries 8, 9, or 14) was added prior to addition
of KHMDS, respectively.
(6) (a) Weseloh, G.; Wolf, C.; Ko¨nig, W. A. Chirality 1996, 8, 441. (b)
Bott, G.; Field, L. D.; Sternhell, S. J. Am. Chem. Soc. 1980, 102, 5618.
See also: Nagai, T.; Nishioka, G.; Koyama, M.; Ando, A.; Miki, T.;
Kumadaki, I. Chem. Pharm. Bull. 1991, 39, 233.
(7) KHMDS was used after evaporation of the original solvent, toluene,
under reduced pressure.
(8) All new products showed satisfactory spectral as well as analytical
data.
906
Org. Lett., Vol. 1, No. 6, 1999

Scheme 2a
and BCl3-mediated removal of the CF3-containing auxiliary
crystallographic analysis.12 These three independent trans-
formations led to unambiguous confirmation of relative
stereochemical relationship of 9 as (2R*,3R*), (2S*,3R*),
and (2S*,3S*) for the second to fourth isomers, respectively.
Consequently, this result demonstrated that the enolate
diastereofacial selectivity was 89:11, again in favor of the
re face attack.
To obtain further mechanistic insight, 1a was transformed
into the corresponding ketene silyl acetal and NMR analysis
of the crude product indicated the formation of a single
stereoisomer.14 Considering the fact that 2-alkoxy esters in
general furnish Z-enolates in a preferential fashion,15,16 the
present diastereoselectivity would be the reflection of the
exclusive formation of Z-potassium enolate from 1a, fol-
lowed by the preferential reaction at its re face.
from the latter combination yielded a single anti isomer, anti-
12.13 Moreover, the relative stereostructure of bis-p-nitroben-
zoate 10 derived from the major diastereomer (2R*,3S*)-9
was unambiguously determined as (1′S*,2R*,3S*) by X-ray
(9) A solution of 1a (0.5 mmol) in THF (1.5 mL) was added dropwise
to a solution of potassium bis(trimethylsilyl)amide (0.5 M in toluene
(remoVed prior to use), 1.2 mL, 0.6 mmol) in THF (0.5 mL) at -78 °C.
After 30 min, methyl iodide (0.60 mmol) was added, and the mixture was
stirred at that temperature. After the reaction was quenched with saturated
aqueous NH4Cl, extraction was carried out with ethyl acetate three times,
and the combined organic layers were washed with saturated aqueous NH4-
Cl, dried over MgSO4, and concentrated. The desired material was obtained
after purification by silica gel chromatography (hexane:AcOEt ) 9:1):
combined yield 79% (a 90:10 diastereoisomer mixture); 1H NMR major
isomer δ 1.68-1.72 (3 H, m), 3.01 (1 H, dd, J ) 5.5, 13.5 Hz), 3.08 (1 H,
dd, J ) 8.0, 14.0 Hz), 3.67 (3 H, s), 3.96 (1 H, dd, J ) 5.5, 8.5 Hz),
7.11-7.35 (5 H, m); minor isomer (representative peaks are shown) δ 1.79-
1.81 (3 H, m), 3.49 (3 H, s), 4.22 (1 H, dd, J ) 5.5, 8.0 Hz); 13C NMR δ
17.48, 40.05, 51.58, 74.23, 80.67 (q, J ) 28.5 Hz), 124.56 (q, J ) 283.1
Hz), 126.77, 127.75, 127.85, 128.00, 128.19, 128.83, 129.60, 134.87, 135.71,
172.97; 19F NMR major isomer δ 82.05 (s); minor isomer δ 83.44 (s); IR
(neat) ν 3060, 3030, 2950, 1760 cm-1. Anal. Calcd for C19H19O3F3: C,
64.77; H, 5.43. Found: C, 64.40; H, 5.20.
(12) (1′S*,2R*)-8a: C19H20F3NO2, M ) 351.37, colorless prism, orthor-
hombic, a ) 17.215(4) Å, b ) 22.638(4) Å, c ) 9.416(3) Å, V ) 3669(1)
Å3, T ) 299.0 K, space group Pbca (no. 61), Z ) 8, R ) 0.036, Rw
)
0.036, GOF ) 1.75. (1′S*,2R*)-8b: C19H21F2NO2, M ) 333.38, colorless
prism, monoclinic, a ) 9.913(4) Å, b ) 22.067(5) Å, c ) 9.277(4) Å, â )
117.8988°, V ) 1793.7(10) Å3, T ) 296.0 K, space group P21/a (no. 14),
Z ) 4, R ) 0.031, Rw ) 0.031, GOF ) 1.67. (1′S*,2R*,3S*)-10:
C32H25F3N2O9, M ) 638.55, colorless prism, monoclinic, a ) 8.701(4) Å,
b ) 11.891(3) Å, c ) 28.513(4) Å, â ) 98.48(2)°, V ) 2917(1) Å3, T )
296.0 K, space group P21/n (no. 14), Z ) 4, R ) 0.041, Rw ) 0.042, GOF
) 1.44.
(10) For Es value, see: Taft, R. W., Jr. In Steric Effects in Organic
Chemistry; Newman, M. S., Ed.; John Wiley & Sons: New York, 1956; p
556. For the revised Es value, Es', see: MacPhee, J. A.; Panaye, A.; Dubois,
J.-E. Tetrahedron 1978, 34, 3553.
(11) 2 N aqueous NaOH (0.75 mL) was added to a solution of 2a (0.36
g, 1.0 mmol) in MeOH (1.0 mL) at 0 °C, and the reaction mixture was
stirred at room temperature for 1 d. Crude carboxylic acid thus obtained
was treated with (COCl)2 (0.18 mL, 2.0 mmol) and a catalytic amount of
DMF, and the solution was stirred overnight and then reacted with 2.5 mmol
of benzylamine. Usual workup and chromatographic purification furnished
8a in 47% total yield (a 90:10 diastereoisomer mixture): mp 128-129 °C;
1H NMR major isomer δ 1.39 (3 H, d, J ) 6.8 Hz), 1.68-1.72 (3 H, m),
3.87 (1 H, q, J ) 6.8 Hz), 4.45 (1 H, dd, J ) 5.9, 14.7 Hz), 4.50 (1 H, dd,
J ) 6.1, 14.7 Hz), 6.87-6.92 (1 H, m), 7.26-7.53 (10 H, m); minor isomer
(representative peaks are shown) δ 1.46 (3 H, d, J ) 6.8 Hz), 1.79-1.81
(3 H, m); 13C NMR δ 16.96, 20.49, 42.86, 71.47, 80.65 (q, J ) 28.3 Hz),
124.60 (q, J ) 282.7 Hz), 127.39, 127.50, 127.62, 128.17, 128.57, 129.21,
135.61, 137.79, 173.41; 19F NMR major isomer δ 81.49 (s); minor isomer
δ 82.58 (s); IR (KBr) ν 3240, 3090, 2950, 1650, 1570 cm-1. Anal. Calcd
for C19H20F3NO2: C, 64.95; H, 5.74; N, 3.99. Found: C, 64.95; H, 5.35;
N, 3.92.
(13) Matthews, B. R.; Jackson, W. R.; Jacobs, H. A.; Watson, K. G.
Aust. J. Chem. 1990, 43, 1195.
(14) Spectroscopic data of the crude mixture was as follows: 1H NMR
δ 0.25 (9 H, s), 1.75-1.78 (3 H, m), 3.37 (3 H, s), 5.06 (1 H, s), 7.29-
7.45 (3 H, m), 7.48-7.57 (2 H, m); 13C NMR δ 0.09, 18.10, 55.17, 80.97
(q, J ) 27.7 Hz), 102.69, 124.96 (q, J ) 282.9 Hz), 127.67, 128.06, 128.63,
136.33, 152.66; 19F NMR δ 82.20 (s).
(15) In this text, E,Z nomenclature for enolates refers to the relative
relationship between OR moieties at the 2 position and OM (M: metal)
groups.
(16) Since Kanemasa and co-workers reported that methyl 2-(1,1-
dimethylethoxy)acetate yielded the ketene silyl acetal in favor of the Z
isomer (E/Z ) 17/83), our ketene silyl acetal from 1a, obtained as a single
stereoisomer, was believed to have the same stereoisomeric preference.
See: Kanemasa, S.; Nomura, M.; Wada, E. Chem. Lett. 1991, 1735.
Org. Lett., Vol. 1, No. 6, 1999
907

CdC-O-C-CF3 dihedral angle of 119.1°.4 The second
most stable conformer also included similar Li‚‚‚F and ‚‚‚
R-O chelation likewise causing steric hindrance around the
si face, but the auxiliary of the third isomer proved to shield
the opposite re face of the enolate. Since the energy
difference from the most stable form was 0.041 and 0.858
kcal/mol, respectively, the diastereofacial selectivity was
anticipated to be as high as 94.6:5.4 at -78 °C, which, at
least qualitatively, supported the present stereochemical
results.19
The present paper describes the high potency of fluorine-
containing auxiliaries effectively attaining good to excellent
diastereoselectivities where, as we expected, metal‚‚‚fluorine
intramolecular interaction is most likely to successfully fix
the enolate conformations. Although extremely high selectiv-
ity has not been attained yet, it is especially worthwhile to
note that although the second smallest element, fluorine,
merely imposes minimum steric modification, introduction
of eVen one fluorine atom to the prototypical achiral donor
1 (R ) CH3) realizes effectiVe discrimination between enolate
π faces in a ratio of up to 92:8.
Since the auxiliaries could be removed with complete
retention of the original stereochemistry as shown in Scheme
2 (see the conversion of a mixture of 2R*,3R*- and
2R*,3R*-9 to anti-12) and the structurally similar trifluori-
nated tertiary alcohol was already resolved by O’Hagan’s
group,20 the present method could be readily extended to
the construction of optically active molecules. Further
application of this metal‚‚‚fluorine interaction along with the
judicious design of auxiliaries for improvement of diastereo-
facial selectivities is underway in our laboratory
Figure 1. Most stable conformation of the model lithium enolate.
Representative bond lengths (Å) and Mulliken charges are shown.
We have also performed ab initio computation17 of the
model enolate of 1a. The Chem3D model of the most stable
conformer is depicted in Figure 1. It was quite apparent from
this model that, along with the expected intramolecular
interaction of R-oxygen and lithium, one of the fluorine
atoms in the CF3 group formed a strong chelation with Li.
This was supported by the Li‚‚‚F distance of 1.879 Å, 0.043
Å shorter than the Li‚‚‚O distance.18 Such interaction led to
significant elongation of the C-F bond length at the same
time (0.072-0.078 Å). The bicyclo[3.3.0] system formed
allowed the auxiliary to effectively cover the si face with a
OL990821C
(17) Because of computational limitation, we have considered the lithium
enolate from 1a with substitution of an ester methoxy group for hydrogen.
This metal change from potassium (actual species) to lithium might be valid
as long as this model is employed for the explanation of the diastereofacial
selectivity because of their same propensity in the enolate π-facial selection
despite poor yield (see entries 1 and 4 in Table 1). All structures were
fully optimized with the ab initio software Mulliken implemented in CAChe
Worksystem (SONY/Tektronix Corporation) on an IBM RS-6000-3CT
workstation at the asdfsadfasMB3LYP/3-21G level of theory.
(19) As shown in Table 1, when HMPA was added, the DS values were
slightly decreased with the same sense of diastereofacial preference.
Computation of the “naked” enolate was also performed to determine that
the auxiliary of the most stable conformer covered the si face, but the
opposite face was shielded by the second most stable isomer, 2.76 kcal/
mol higher in energy. These results led to the anticipation that the alkylation
would preferentially occur from the re face even in the presence of HMPA.
Details will be published elsewhere.
(18) Since the van der Waals radius of fluorine is 0.05 Å smaller than
that of oxygen, interaction of lithium with fluorine would be roughly
estimated to be as strong as that with oxygen.
(20) O’Hagan, D.; Zaidi, N. A. J. Chem. Soc., Perkin Trans. 1 1992,
947.
908
Org. Lett., Vol. 1, No. 6, 1999