Thermodynamically favored aldol reaction of propargyl or allenyl esters: Regioselective synthesis of carbinol allenoates

Article abstract of DOI:10.1002/anie.200704221

Retroaldol regains control: A countercation may exert an important effect in promoting thermodynamic control in the TBAF-mediated aldol reaction of propargyl or allenyl esters (or a mixture of both) to produce, under mild conditions, carbinol allenoates exclusively (see scheme; TBAF = tetra-n-butylammonium fluoride). (Chemical Equation Presented)

Full text of DOI:10.1002/anie.200704221

Angewandte
Chemie
DOI: 10.1002/anie.200704221
Aldol Reactions
Thermodynamically Favored Aldol Reaction of Propargyl or Allenyl
Esters: Regioselective Synthesis of Carbinol Allenoates**
Bo Xu and Gerald B. Hammond*
Dedicated to Professor David F. Wiemer
The aldol reaction is one of the cornerstones of synthetic
organic chemistry, and as such it has been the subject of
considerable optimization and multiple applications.[1] If a
double bond is appended to the enolate precursor through
conjugation, it gives rise to the so-called vinylogous aldol
reaction depicted in Scheme 1 (top).[2] The inherent prefer-
occur, maximum control over the regioselectivity of this
reaction must be exercised to favor the thermodynamic
product.
During our research on the synthesis and reactivity of
fluorinated allenes,[5] we became interested in investigating
whether the nature of the countercation could influence the
outcome of base-mediated processes. Herein we report a
practical and highly regioselective, TBAF-mediated synthesis
(TBAF = tetra-n-butylammonium fluoride) of carbinol alle-
noates from either 1 or 2 (Scheme 2, X = OEt) by a
thermodynamically controlled alkynylogous aldol reaction.
Scheme 1. Regioselectivity considerations in the aldol reaction of
extended enolates.
ence for a-functionalization (kinetic product) is attributed to
the higher electron density at the a-carbon atom compared to
the g-carbon atom. The synthetic importance of the vinyl-
ogous aldol reaction has spurred the development of methods
that address the regiochemistry problem but at the cost of
increasing the experimental demands of the reaction or
restricting the scope of substrates that can be utilized. A
synthetically attractive yet unprecedented strategy is the aldol
reaction of an enolate conjugated with a triple bond
(Scheme 1, bottom). This reaction could be regarded as an
alkynylogous[3] aldol reaction, and, as in the case of the
vinylogous aldol reaction, this reaction is invariably prone to
form the kinetic a-product. The reaction of an alkynylenolate
with an aldehyde could lead to an interesting building block,
namely, a substituted g-carbinol allenoate (X = OEt), for
which no general synthetic methodologies exist.[4] For this to
Scheme 2. Preparation of silylalkynylketene acetal 3. HMDS=hexame-
thyldisilazide.
The synthetic potential of the carbinol allenoates was
showcased by their efficient conversion to dihydrofurans or
g-lactones. Further, a silylalkynylketene acetal reacted under
Mukaiyama conditions to yield hydroxyalkynoates or carbi-
nol allenoates in highly regioselective fashion, or with an
electrophilic fluorinating reagent to produce a a-fluoropro-
pargyl ester.
In theory, if 1 or 2 were treated with a strong enough base,
the deprotonation of the a-hydrogen atom of 1 or the g-
hydrogen atom of 2 should occur, furnishing alkynylenolate
anion A (Scheme 2, top).[4a,e] Ab initio geometry optimization
calculations (DFT/B3LYP/6.311G*) show that A is an
energetically favored anionic species regardless of which
hydrogen atom was abstracted. Thus, in principle, either 1 or 2
could be used as a synthetic precursor of A. The flexibility in
the choice of starting material is an added advantage of
conducting an aldol reaction with alkynylenolate A, because
the propargyl and allenyl ester precursors 1 and 2 are readily
prepared using standard protocols.[6]
[*] B. Xu, Prof. Dr. G. B. Hammond
Department of Chemistry
University of Louisville
Louisville, KY 40292 (USA)
Fax: (+1)508-852-3899
E-mail: gb.hammond@louisville.edu
[**] Financial support of the National Science Foundation (CHE-
0513483) is gratefully acknowledged.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2008, 47, 689 –692
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
689

Communications
The alkynylenolate anion A was generated under kinetic
The reaction of fluoropropargyl ester 1a with p-chloro-
benzaldehyde (Table 1) was used as model for our studies.
The presence of fluorine in substrate 1a allowed us to monitor
the reaction using 19F NMR spectroscopy with minimal steric
disruption. As predicted, under kinetic conditions the a-
isomer 4a is obtained in very high yields (entries 1–2,
Table 1).
conditions by deprotonation of 2a and trapped as its
silylalkynylketene acetal 3 in good to very good yields
(Scheme 2, bottom). Figure 1 illustrates the highest occupied
molecular orbital (HOMO) of alkynylenolate anion A (R1 =
R2 = Me) and its corresponding silylalkynylketene acetal 3
According to computational calculations
(Gaussian 03, DFT/B3LYP/6.311G*), the g-prod-
uct 5a is 18 kJmolÀ1 more stable than 4a. This
finding means that under purely thermodynamic
control, the allenylic alcohol 5a should predom-
inate. In general, longer reaction times and higher
temperatures will favor the thermodynamic prod-
uct through the intermediacy of a retroaldol
reaction, but those conditions may also lead to
dehydration and other side reactions. Under
standard thermodynamic conditions (entries 3–6,
Table 1), the hydroxyalkynoate 4a was the only
Figure 1. HOMO of optimized alkynylenolate anion A (R1 =R2 =CH3, X=OEt) and
corresponding silyl ether 3 (R=TMS).
product isolated. We hypothesized that the size
and Lewis acidity of the countercation could
influence the rate of the retroaldol reaction.
When the size of the counterion is enlarged and
along with the charge densities at C2 and C4. It is clear that
these two carbon atoms correspond to the most nucleophilic
sites in the molecule, but in an aldol reaction the higher
electron density at C2 (a-position) should favor the kinetic
product 4 (Table 1), in spite of the fact that the g-adduct 5 is
thermodynamically more stable, owing to its extended con-
jugation.
its Lewis acidity is diminished, a chelated intermediate cannot
be formed as easily, and an open alkoxide intermediate might
be formed instead. This alkoxide intermediate may be
energetically disfavored because of charge separation and
loss of chelation, which in turn should enhance the rate of the
retroaldol process, thus causing the reaction to fall under
thermodynamic control. Experimentally, we observed that
bases containing a rather large tetrabutylammonium cation
tend to favor formation of the thermodynamic g-product 5a.
TBAF (commercial solution in THF) gave the best yield and
selectivity in THF (entry 9, Table 1). Alcoholic or predom-
inantly aqueous media had a deleterious effect (entries 13–15,
Table 1). Traditional kinetic bases like LDA or sodium
hexamethyldisilazide, when complexed with HMPA or
crown ether to increase the size of the counterion and
reduce its complexing ability, augmented the production of
the thermodynamic g-product 5a (compare entries 1 and 2
with entries 7 and 16, Table 1).
Table 1: Screening conditions for thermodynamic control.
Entry Base/additives
Equiv
T
[8C]
t [h] 4a
[%][a]
5a
1
[%][a]
[%][a]
1
2
3
4
5
6
7
8
LDA
Na[(TMS)2N]
KF
DBU
CsF
1.5
1.5
2
0.4
2
2
1
2
2
0.2
2
2
2
2
À50
À50
RT
RT
RT
1
1
3
3
3
82
90
89
61
70
44
73
10
0
0
0
0
0
0
27
8
83
3
0
0
0
0
A crossover experiment that supported the occurrence of
a retroaldol reaction is depicted in Equation (1). When the
two diastereoisomers of the kinetic product 4a were treated
with one equivalent of chlorobenzaldehyde, conversion to the
5
16
27
0
71
0
17
100
100
100
100
2
Et3N
RT 48
LDA/HMPA
[(Bu)4N][AcO]
TBAF
À50
1
3
0
0
0
9
1.5 trace
1.5
10
11
12
13
14
15
16
TBAF
70
0
0
0
0
Na2CO3
pyridine
TBAF/H2O
TBAF[b]
RT 12
RT 12
RT 12
RT 12
0
À50
0
0
0
5
thermodynamic product 5a occurred in high yield. This
outcome is only possible if a mechanism involving a
thermodynamically favored retroaldol reaction is invoked.
Because both 1 and 2 generate alkynylenolate A upon
deprotonation, it should make no difference whether we use
propargyl derivative 1, allene compound 2, or a mixture of
both (Table 2). Indeed, the TBAF-mediated aldol reaction
TBAF[c]
2
2
3
1
71
68
Na[(TMS)2N]/
[15]c-5
28
0
[a] Yields were determined by 19F NMR spectroscopy using PhCF3 as
internal standard. [b] Ethanol as solvent. [c] tert-Butyl alcohol as solvent.
LDA=lithium diisopropyl amide, DBU=1,8-diazabicyclo[5.4.0]undec-7-
ene, HMPA=hexamethyl phosphoramide.
690
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 689 –692

Angewandte
Chemie
Table 2: Alkynylogous aldol reactions from compound 1 and 2.
Entry
R1
R2
R3
EWG
1:2
5 [%][a]
1
2
3
4
5
6
7
8
nC6H13
nC6H13
nC6H13
Ph
Ph
Ph
F
F
Bn
p-Cl-C6H4
p-Cl-C6H4
p-Cl-C6H4
p-Cl-C6H4
p-NO2-C6H4
n-C5H11
(E)-cinnamyl
Ph
p-Cl-C6H4
p-NO2-C6H4
o-Br-C6H4
COOEt
COOtBu
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
100:0
100:0
4.5:1
5:95
5:95
5:95
5:95
5:95
5:95
5:95
5:95
5a, 56
5b, 60
5c, 56
5d, 68
5e, 71
5 f, 51
5g, 71
5h, 65
5i, 90
5j, 70
5k, 77
Scheme 3. Synthetic transformations of carbinol allenoate 5h.
DIBALH=diisobutylaluminum hydride.
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
Ph
nC6H13
nC6H13
nC6H13
nC6H13
9
10
11
[a] Yields of isolated compound. In all cases, traces or no compound 4
was detected. EWG=electron-withdrawing group, Bn=benzyl.
produced 5 regardless of the isomeric purity of 1 or 2
(compare entries 1, 3, and 9, Table 2). Not needing pure
allenyl (1) or propargyl esters 2 as starting materials is an
important feature of our procedure, because it is well-known
that 1 and 2 tend to interconvert in the presence of base
through a prototropic rearrangement.[7] Compounds 1 and 2
with alkyl or aryl substituents have been used successfully
with hexanal, cinnamaldehyde, and other aromatic aldehydes
in yields ranging from good to excellent (Table 2).
Scheme 4. Synthetic transformations of silylalkynylketene acetal 3.
NFSI=N-fluorobenzenesulfonimide.
react with other electrophiles, as exemplified by the synthesis
of a hitherto inaccessible quaternary a-fluoropropargyl ester
10 (Scheme 4, bottom).
g-Disubstituted allenoate 6 undergoes a facile a-aldoliza-
tion under our standard conditions, yielding the correspond-
ing a-carbinol allenoate 7 in satisfactory yields [Eq. (2)].
In summary, we have described an extended aldol reaction
of alkynylenolates with aldehydes in the presence of TBAF, in
À
which the C C bond formation takes place at the g-position,
thus yielding difficult-to-obtain carbinol allenoates under
mild conditions. The latter were further converted into
attractive structural motifs such as g-lactones or dihydrofur-
ans. Under Mukaiyama-type conditions, a silylalkynylketene
acetal furnished a- or g-aldol products depending on the
choice of Lewis acid. The broader implications of this reaction
in organic synthesis, including its asymmetric variant, are
currently under investigation.
Experimental Section
3b: LDA ( 2m in THF, 7.5 mL, 15 mmol) was added slowly over 5 min
to a mixture of dry THF (50 mL) and HMPA (3.58 g, 20 mmol) at
À808C. The reaction mixture was stirred at this temperature for
20 min, then a solution of 2a (2.10 g, 10 mmol) in THF (10 mL) was
introduced slowly over 5 min. The resulting solution was stirred for
10 min at À808C. tert-Butyldimethylsilyl chloride (TBDMSCl, 1.81 g,
12 mmol) in THF (6 mL) was then added by cannula. The reaction
mixture was allowed to warm to 08C and was stirred for 30 min at
08C. The yellow solution was diluted with cold hexane (50 mL) and
washed with cold saturated sodium bicarbonate solution (3 ” 30 mL).
The organic layer was dried over Na2SO4 and filtered, and the filtrate
was concentrated in vacuo. The residue was purified by short-path
distillation (85–958C, 1 Torr) to yield 3b as a colorless oil (2.64 g,
81%, Z/E 8:92). (E) isomer: 1H NMR (500 MHz, CDCl3, 258C,
TMS): d = 0.16 (s, 6H; CH3), 0.88 (t, 3J(H,H) = 7.0 Hz, 3H; CH3), 0.94
(s, 3 H; CH3), 1.22–1.27 (m, 2H; CH2), 1.27 (t, 3J(H,H) = 7.5 Hz, 3H;
CH3), 1.38–1.41 (m, 6H), 1.48–1.52 (m, 2H; CH2), 1.65 (s, 3H; CH3),
Carbinol allenoates have been reported to be good
cycloaddition partners.[8] They also have been used in the
synthesis of lactones through conjugate nucleophilic addition
to the central carbon atom of the allene moiety.[4c] To
showcase the synthetic potential of carbinol allenoate 5, we
conducted short and high-yielding syntheses of dihydrofuran
8 and g-lactone 9, both of which contain structural features
found in numerous natural products (Scheme 3).
Finally, the silylalkynylketene acetal 3 can undergo a
Mukaiyama aldol-type reaction[9] to produce a- or g-aldol
products 4 or 5 with excellent regioselectivity with the
appropriate choice of Lewis acids, albeit in moderate yields
thus far (Scheme 4, top). Silylalkynylketene acetal 3 can also
Angew. Chem. Int. Ed. 2008, 47, 689 –692
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
691

Communications
3
3
2.31 (t, J(H,H) = 7.5 Hz, 2H; CH2), 4.06 ppm (q, J(H,H) = 7.0 Hz,
2H; CH2); 13C NMR (125 MHz, CDCl3, 258C, TMS): d = À4.3, 13.9,
15.0, 16.2, 17.9, 19.7, 22.5, 25.5, 28.5, 29.1, 31.4, 66.6, 78.8, 79.5, 90.5,
158.0 ppm. GC–MS (EI) m/z = 324 [M+], 296, 268, 227, 149, 93.
5h: A solution of TBAF (1m in THF, 1.0 mL, 1 mmol) was added
dropwise to a mixture of 2a (105 mg, 0.5 mmol) and THF (0.5 mL) at
08C and the reaction mixture was stirred for 3 h at 08C. Then,
saturated NH4Cl solution (10 mL) was added to quench the reaction.
After stirring for 5 min at about 08C, the resulting aqueous mixture
was extracted with ether (3 ” 15 mL). The extract was washed with
brine (3 ” 10 mL) and dried over Na2SO4. The solvent was removed
under reduced pressure, and the crude product was purified by flash
silica gel chromatography (10–20% ethyl acetate in hexane) to give
5h (102 mg, 65%) as a colorless oil. 5h was obtained as a mixture of
two diastereomers in a ratio of 3.1:1. IR (neat): n˜ = 3418, 2927, 1959,
1707, 1455, 1267, 1124 cmÀ1; 1H NMR (500 MHz, CDCl3, 258C, TMS)
Major isomer: d = 0.85 (t, 3J(H,H) = 7.5 Hz, 3H; CH3), 1.17–1.22 (m,
6H), 1.30–1.39 (m, 6H), 1.87–1.94 (m, 1H), 1.92 (s, 3H; CH3), 2.72 (d,
Christmann), Wiley-VCH, Weinheim, 2006, p. 105; c) B. Bazµn-
Tejeda, G. Bluet, G. Broustal, J.-M. Campagne, Chem. Eur. J.
2006, 12, 8358, and references therein; d) M. Kalesse, Natural
Products Synthesis II Targets, Methods, Concepts, Vol. 244,
Springer, New York, 2005, p. 43; e) S. E. Denmark, J. R. Heem-
stra, G. L. Beutner, Angew. Chem. 2005, 117, 4760; Angew. Chem.
Int. Ed. 2005, 44, 4682; f) H. Akikawa, K. Ishihara, S. Saito, H.
Yamamoto, Synlett 2004, 732; g) G. Casiraghi, F. Zanardi, G.
Appendino, G. Rassu, Chem. Rev. 2000, 100, 1929; h) J. Kruger,
E. M. Carreira, J. Am. Chem. Soc. 1998, 120, 837; i) M. Sugiura, Y.
Yagi, S.-Y. Wei, T. Nakai, Tetrahedron Lett. 1998, 39, 4351; j) K.
Tomooka, A. Nagasawa, S.-Y. Wei, T. Nakai, Tetrahedron Lett.
1996, 37, 8899; k) W. C. Black, A. Giroux, G. Greidanus,
Tetrahedron Lett. 1996, 37, 4471.
[3] For a reference to the term “alkynylogous” see: K. Mikami, A.
Yoshida, Y. Matsumoto, Tetrahedron Lett. 1996, 37, 8515.
[4] a) S. D. Lepore, Y. He, P. Damisse, J. Org. Chem. 2004, 69, 9171;
b) N. Krause, A. S. Hashmi, Modern Allene Chemistry, Wiley-
VCH, Weinheim, 2004; c) J. G. Knight, S. W. Ainge, C. A. Baxter,
T. P. Eastman, S. J. Harwood, J. Chem. Soc. Perkin Trans. 1 2000,
3188; d) S. Tsuboi, H. Kuroda, S. Takatsuka, T. Fukawa, T. Sakai,
M. Utaka, J. Org. Chem. 1993, 58, 5952; e) N. A. Petasis, K. A.
Teets, J. Am. Chem. Soc. 1992, 114, 10328.
3J(H,H) = 4.5 Hz, 1H), 4.14–4.24 (m, 2H; CH2), 5.15 (d, J(H,H) =
4.5 Hz, 1H), 7.27–7.36 (m, 3H), 7.42–7.46 ppm (m, 2H); minor
isomer: d = 0.85 (t, J(H,H) = 7.5 Hz, 3H; CH3), 1.17–1.22 (m, 6H),
3
3
1.30–1.39 (m, 6H), 1.87–1.94 (m, 1H), 1.89 (s, 3H; CH3), 2.78 (d, 3J(H,
H) = 4.0 Hz, 1H), 4.14–4.24 (m, 2H), 5.26 (d, 3J(H,H) = 4.0 Hz, 1H),
7.27–7.36 (m, 3H), 7.42–7.46 ppm (m, 2H); GC–MS (EI) m/z = 300,
253, 181, 149, 107; Elemental analysis (%) calcd. for C20H28O3:
C 75.91, H 8.92; found: C 75.97, H 9.09.
[5] a) M. Mae, J. A. Hong, B. Xu, G. B. Hammond, Org. Lett. 2006, 8,
479; b) B. Xu, G. B. Hammond, Angew. Chem. 2005, 117, 7570;
Angew. Chem. Int. Ed. 2005, 44, 7404.
[6] a) H. M. L. Davies, T. A. Boebel, Tetrahedron Lett. 2000, 41,
8189; b) R. W. Lang, H.-J. Hansen, Org. Synth. 1990, Coll. Vol. 7,
232.
Received: September 13, 2007
Published online: December 6, 2007
[7] a) M. J. Sleeman, G. V. Meehan, Tetrahedron Lett. 1989, 30, 3345;
b) J. C. Clinet, G. Linstrumelle, Synthesis 1981, 875; c) M. C.
Croudace, N. E. Schore, J. Org. Chem. 1981, 46, 5357.
[8] a) J. D. Winkler, K. J. Quinn, C. H. MacKinnon, S. D. Hiscock,
E. C. McLaughlin, Org. Lett. 2003, 5, 1805; b) R. H. Schlessinger,
C. P. Bergstrom, J. Org. Chem. 1995, 60, 16; c) M. E. Jung, N.
Nishimura, J. Am. Chem. Soc. 1999, 121, 3529; d) K. A. Parker,
S. M. Ruder, J. Am. Chem. Soc. 1989, 111, 5948.
Keywords: aldol reaction · alkynes · isomers · propargyl esters ·
thermodynamic control
.
[1] Modern Aldol Reactions, Vol. 1 and 2 (Ed.: R. Mahrwald), Wiley-
VCH, Weinheim, 2004.
[2] a) T. Y. Liu, H. L. Cui, J. Long, B. J. Li, Y. Wu, L. S. Ding, Y. C.
Chen, J. Am. Chem. Soc. 2007, 129, 1878; b) M. Kalesse, J.
Hassfeld in Asymmetric Synthesis—The Essentials (Eds.: M.
[9] S. E. Denmark, J. R. Heemstra, J. Am. Chem. Soc. 2006, 128, 103.
692
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 689 –692