A Synthetic and Computational Investigation into the Direct Synthesis of α -Hydroxymethylated Enones from β-keto Phosphonates

Article abstract of DOI:10.1071/CH09301

The synthesis of a range of -hydroxymethylated enones has been achieved using the Villiras modification of the HornerWadsworthEmmons (HWE) reaction. Scope, limitations, and mechanistic aspects of this reaction were investigated using a combination of synthetic and computational studies. These investigations support a SchlosserCorey type reaction mechanism that is balanced between two pathways with the outcome influenced by the steric environment of the substrate. CSIRO 2009.

Full text of DOI:10.1071/CH09301

RESEARCH FRONT
Full Paper
CSIRO PUBLISHING
Aust. J. Chem. 2009, 62, 720–727
www.publish.csiro.au/journals/ajc
A Synthetic and Computational Investigation into
the Direct Synthesis of α-Hydroxymethylated
Enones from β-Keto Phosphonates
Sarah J. Ryan,A Christopher D. Thompson,A and David W. LuptonA,B
ASchool of Chemistry, Monash University, Clayton, Vic. 3800, Australia.
BCorresponding author. Email: david.lupton@sci.monash.edu.au
The synthesis of a range of α-hydroxymethylated enones has been achieved using theVilliéras modification of the Horner–
Wadsworth–Emmons (HWE) reaction. Scope, limitations, and mechanistic aspects of this reaction were investigated
using a combination of synthetic and computational studies. These investigations support a Schlosser–Corey type reaction
mechanism that is balanced between two pathways with the outcome influenced by the steric environment of the substrate.
Manuscript received: 20 May 2009.
Final version: 11 June 2009.
O
O
P
O
O
P
(i)
Introduction
OEt
OEt
HO
R
OEt
OEt
Interception of the Wittig reaction by low temperature deproto-
nation and alkylation was introduced separately by Schlosser[1a]
and Corey[1b] as a method for the assembly of allylic alcohols
in a stereochemically well defined fashion.[1] As part of studies
directed towards the synthesis of limonoid natural products,[2]
a range of α-hydroxymethylated enones (e.g., 1) were required.
While these compounds can be prepared from the enone (e.g., 3)
by a Morita–Baylis–Hillman (MBH) reaction,[3] more conve-
nient is their preparation using a Schlosser–Corey type reaction
withanappropriateHorner–Wadsworth–Emmons(HWE)orsta-
bilizedWittig reagent (Scheme 1).[4] This approach was reported
byVilliéras for ester-containing HWE reagents (2, R = OR)[5a–c]
and has been exploited in several synthetic applications.[5d–g]
The analogous approach for ketone-containing HWE reagents
has received little attention,[5c] while the mechanism of this reac-
tion is yet to be probed. As a result of our synthetic needs, the
well documented utility of MBH adducts,[6] and problems asso-
ciated with their preparation,[3c,7] it was decided to investigate
the scope and mechanism of the Villiéras approach to MBH
adducts.
2a R ϭ Ph, 55%
b
c
d
e
R ϭ mBrC6H4, 42%
R ϭ mNO2C6H4, 39%
R ϭ pMeOC6H4, 57%
R ϭ 2-Furyl, 42%
O
O
O
O
P
P
OEt
EtO
8
OEt
OEt
2l, 42%
Scheme 2. Reagents and conditions: (i) MgCl2 (2.5 mol equiv.), Et3N
(4 mol equiv.), CH3CN 18◦C, and then RCOCl (1 mol equiv.), 0 → 18◦C.
O
O
O
P
(i)
P
OMe
OMe
OMe
OMe
R
2f R ϭ PhCHϭCH, 37%
g
h
i
R ϭ 2-MeOC6H4CHϭCH, 50%
R ϭ CHϭCH2, 13%
R ϭ (CH3)3C, 42%
k
R ϭ (CH2)7CH3, 42%
Scheme 3. Reagents and conditions: (i) n-BuLi (1 mol equiv.), CuI
(1 mol equiv.) THF, −78◦C → −40◦C, and then RCOCl (1 mol equiv.),
−40◦C → 18◦C.
Results and Discussion
In order to undertake the proposed studies, access to a range
of β-keto phosphonates 2 was required.[8] In our hands, the
known acylation decarboxylation of (diethoxyphosphoryl)acetic
acid (Scheme 2) proved convenient for the preparation of known
(2a–e) and new (2l) phosphonates.[8b]
prepared from dimethyl methylphosphonate by its correspond-
ing cuprate (Scheme 3).[8c]
Initial studies into the Villiéras reaction explored the
use of reaction conditions analogous to those reported for
α-phosphonoesters.[5a] When phosphonate 2a was reacted under
these conditions the expected product 1a formed in 41% yield
accompanied by enone 3a and phosphonate 4a in a 5:2:1 ratio
(Table 1, entry 1). While this is the first report of these by-
products in aVilliéras–HWE reaction, their formation seemed in
keeping with the reaction conditions, which may provide enone
3a by a conventional HWE reaction and phosphonate 4a from
While this approach was operationally simple, it failed to pro-
vide β-keto phosphonates 2f–i and k. These compounds were
Morita–Baylis–
Hillman reaction
Villiéras HWE
reaction
O
O
OH
O
O
P(OR)2
R
R
R
3
1
2
Scheme 1.
© CSIRO 2009
10.1071/CH09301
0004-9425/09/070720

RESEARCH FRONT
Direct Synthesis of α-Hydroxymethylated Enones from β-Keto Phosphonates
721
Table 1. Selected optimization of Villiéras–HWE reaction of β-keto phosphonates
X CO , CH O
(4 equiv.), solvent
temperature
2
3
2
O
O
P
O
OH
O
O
OEt
Ph
ϩ
Ph
Ph
ϩ Ph
O
P
(Eqn 1)
OEt
OEt
2a
1a
3a
4a
OEt
O
Ph
Entry
X2CO3
[mol-%]
Solvent
H2O
t
TA
RatioB
1a:3a:4a
Yield 1aC
[h]
[◦C]
[%]
1D
2
3F
4
5
6
100 (K)
100 (K)
100 (K)
100 (K)
10 (K)
70 (K)
70 (Na)
70 (Li)
2
2
2
1
2
2
2
2
rt
rt
rt
0
rt
rt
rt
rt
5:2:1
24:1:37
38:1:0
5:2:1
12:1:0
14:1:0
14:1:0
14:1:1
41
33
16
45
20
66
65
52
DMSO/H2OE
H2O
H2O
H2O
H2O
H2O
7
8
H2O
Art = room temperature.
BRatio calculated by integration of suitably separated peaks within the 1H NMR spectrum of the reaction mixture.
CIsolated yield following flash column chromatography.
DCH2O formed in situ from paraformaldehyde.
E1:1 ratio of DMSO and H2O.
FEight equivalents of formaldehyde.
O
O
P
O
O
P
O
O
P
of allylic alcohol 1a the stoichiometry of formaldehyde was
increased, which resulted in a good ratio of 1a:3a:4a (38:1:0),
but lower overall yield (Table 1, entry 3). Cooling the reaction
mixture had little bearing on either the ratio of products, or the
yield (Table 1, entry 4). Finally, attention was turned to the stoi-
chiometry of K2CO3. When 10 mol-% of K2CO3 was used, the
ratio of 1a:3a:4a was 12:1:0, although the yield was a mere 20%
(Table 1, entry 5). Increasing the amount of base to 70 mol-%
allowed the formation of allylic alcohol 1a to be achieved in 66%
yield (Table 1, entry 6) with only a trace of by-product 3a. The
use of Na2CO3 and Li2CO3 was also investigated and found to
have little bearing on the reaction outcome (Table 1, entries 7
and 8).
These studies indicate that with ketone-containing HWE
reagents the formation of by-products 3a and 4a occur readily,
but can be suppressed through the control of the stoichiometry
of base. From a mechanistic standpoint, the formation of com-
pounds 3a and 4a bring to light several possibilities in addition
to the Schlosser–Corey type dialkylation originally proposed
by Villiéras. In order to clarify the mechanism of this reac-
tion and ascertain the generality of these reaction conditions,
an investigation into the reaction scope was undertaken. When
electron-poor aromatic, electron-rich aromatic, and heteroaro-
matic phosphonates were subjected to the optimized reaction
conditions (Table 1, entries 2–5) they all smoothly provided
products 1b–e in acceptable yields, in each case accompanied
by a small amount of by-product 3b–e, identifiable by 1H NMR
spectroscopy.
Pathway A
CH2O
OEt
OEt
R
R
OEt
R
OEt
O
OEt
OH
OEt
5
5؅
cb2
CH2O
O
ϪEtO2PO2
O
O
O
P
cb2
OEt
R
R
OEt
OH
3
6
ϪEtO2PO2
O
O
O
R
O
P
R
R
O
P
OEt
O
OH
CH2O
؊cb2
OEt
OEt
R
OEt
O
1
Pathway B
7
R
R
O
Hϩ
ϪHO
O
O
O
P
OEt
O
R
O
P
OEt
O
H2O
Pathway C
OEt
OEt
8
4
R
R
Scheme 4. Proposed reaction mechanisms.
the degradation of allylic alcohol 1a. Alternately, the formation
of such by-products may be explained by the initial formation of
enone 3a, which is subsequently converted into allylic alcohol
1a by a MBH reaction catalyzed by phosphonate 2a.[9] In this
case, the formation of 4a would arise as a result of dehydration
of a MBH intermediate, rather than elimination of phosphonate
2a (Scheme 4).
If the reaction proceeds by pathway B/C it was reasoned that
conversion of enone 3a into allylic alcohol 1a may be enhanced
using a dimethyl sulfoxide/H2O solvent mix known to accelerate
the rate of MBH reactions.[10,11] In the event, this decreased the
formation of enone 3a, but favoured the formation of phospho-
nate 4a (Table 1, entry 2). In order to increase the formation
When cinnamyl phosphonate 2f was subjected to the reac-
tion conditions, the expected product 1f formed in an excellent
yield of 92% (Table 2, entry 6) without any by-product 3f. Inter-
estingly when 2-methoxycinnamyl phosphonate 2g (Table 2,
entry 7) was allowed to react the yield was reduced, and once
more by-product 3g was identified. Unfortunately, when acrolyl
chloride-derived phosphonate 2h was subjected to the optimized
reaction conditions, decomposition was observed (Table 2,
entry 8).

RESEARCH FRONT
722
S. J. Ryan, C. D. Thompson, and D. W. Lupton
Table 2. Scope of the tandem olefination MBH reaction
Entry
1
SM
Product
YieldA [%]
O
OH
O
O
P
66
C6H5
C6H5
OEt
OEt
2a
1a
OH
O
O
P
O
mBrC6H4
OEt
2
3
4
5
6
7
63
65
65
56
92
50
mBrC6H4
OEt
2b
1b
OH
O
O
P
O
mNO2C6H4
pMeOC6H4
OEt
mNO2C6H4
OEt
2c
1c
OH
O
O
P
O
OEt
pMeOC6H4
OEt
2d
1d
O
O
O
OH
P
O
O
OEt
OEt
2e
1e
OH
O
O
O
P
C6H5
C6H5
OMe
OMe
2f
1f
OH
O
O
O
P
2-MeOC6H4
2-MeOC6H4
OMe
OMe
2g
1g
O
O
P
B
OEt
8
–
–
OEt
2h
O
O
P
C
OEt
9
–
–
OEt
2i
O
OH
O
O
P
10
96
60
C6H5
C6H5
OEt
OEt
2j
1j
O
O
P
O
OH
H3C
O
H3C
OEt
11
12
6
6
OEt
2k
1k
OH
O
O
O
P
OH
O
O
P
EtO
OEt
57
8
OEt
OEt
8
2l
1l
AIsolated yield following flash column chromatography.
BDecomposition of starting material 2h.
CNo reaction of starting material 2i.

RESEARCH FRONT
Direct Synthesis of α-Hydroxymethylated Enones from β-Keto Phosphonates
723
O
O
O
O
OH
Ph
4a
O
P
Ph
8a
O
P
Ph
O
P
K2CO3,
CH2O, H2O
K2CO3, H2O
OEt
Ph
1a:4a (1:3.5)
(Eqn 3)
OEt
OEt
(Eqn 2)
OEt
O
OEt
O
OEt
O
Ph
Ph
Ph
4a:8a (1:2.9)
reacted with two equivalents of nitro phosphonate 2c (Eqn 4)
was undertaken. This experiment produced none of product 1a,
which indicates that β-keto phosphonate 2c does not undergo
conjugate addition into enone 3a.A similar crossover experiment
was undertaken using phosphonate 2d. However, as with phos-
phonate 2c, compound 1a was not identified (Eqn 5). In both of
these experiments hybrid intermediates of type 8 or 4 were never
observed, which clearly indicates that the anions of the β-keto
phosphonates 2c and 2d are incapable of undergoing conjugate
addition with enone 3a.
While these studies indicate that pathway B/C does not give
rise to the formation of allylic alcohol 1a, the origin of enone 4a
and ketone 8a remained to be explained.As they are not forming
from the reaction of enone 3a, it appeared likely that these mate-
rials arise as a result of degradation of allylic alcohol 1a. When a
2:1 mixture of β-keto phosphonate 2a and alcohol 1a were mixed
in the presence of water and K2CO3 after 30 min the quantity of
compounds 2a and 1a were found to have decreased and inter-
mediate 8a was now observed (Scheme 5).As this reaction lacks
formaldehyde, it is not possible to form enone 4a. However, this
reaction indicates that allylic alcohol 1a is not stable under the
optimized reaction conditions and can lead to the formation of
degradation products.
Finally, the reactions of alkyl-substituted phosphonates
were investigated. While pivaloyl chloride-derived β-keto
phosphonate 2i failed to provide alcohol 1i (Table 2, entry 9);
dihydrocinnamyl phosphonate 2j gave the expected product 1j in
high yield, without the formation of by-product 3j (Table 2, entry
10). Similarly alkyl phosphonates 2k and bisphosphonate 2l pro-
vided allylic alcohols 1k and 1l without by-products (Table 2,
entries 11 and 12).
Based on optimization and scope studies, three plau-
sible reaction mechanisms for the Villiéras–HWE reaction
can be proposed (Scheme 3). Based on a Schlosser–Corey
type mechanism, as originally proposed by Villiéras,[5a]
hydroxymethylation of the starting phosphonate affords alko-
xide 5, which then undergoes proton transfer to form
5ꢀ. This intermediate subsequently reacts with a second
molecule of formaldehyde to give intermediate 6,¶ which
fragments to provide compound 1 (pathway A). Alter-
natively, intermediate 5 may directly fragment to afford
olefin 3, which then undergoes MBH reaction catalyzed by
β-keto phosphonate 2 to provide alcohol 1 directly (pathway B),
or by enone 4 (pathway C).[9] The failure of the least hindered
substrates 2f, j, k, and l to afford any by-products would appear
to suggest that the reaction occurs by a mechanism sensitive
to sterics, which supports the originally proposed pathway A.
Potentially, pathways B and C become operative in cases where
steric crowding slows the formation of intermediate 6. The iso-
lation of enone 3a clearly demonstrates that the first step in
pathway B/C proceeds, however, the viability of this pathway to
afford allylic alcohol 1a remains in question.
Investigation into these mechanistic possibilities was under-
taken using synthetic and computational methods. It was hoped
that a kinetic analysis of the reaction could be undertaken to
clarify the involvement of these pathways, however, because of
the rapid nature of the reaction and its heterogeneity this was
not possible. Thus, the interconversion of isolated intermedi-
ates, specifically enone 3a, phosphonate 4a, and ketone 8a† was
investigated. Initial studies focussed on demonstrating that phos-
phonate 8a and 4a could be converted into allylic alcohol 1a
(Eqn 2 and 3). When phosphonate 8a was exposed to formalde-
hyde and K2CO3 after 2 h 25% conversion to compound 4a was
observed (Eqn 2), however allylic alcohol 1a failed to form.
Similarly, when enone 4a was dissolved in water and reacted
with K2CO3 allylic alcohol 1a formed (Eqn 3), but only as the
minor product with most of the material remaining as unreacted
starting material.
Calculations
While synthetic studies suggest that allylic alcohol 1a forms by
pathway A, and that by-products arise as a result of competing
reactions (in the case of enone 3a) or decomposition (in the case
of phosphonate 4a), it was decided to investigate the reaction
using computational approaches to gain further insight into the
mechanism. The use of computational techniques to study the
Wittig and HWE reaction is well documented[12] with notable
work byAndo.[12c] Using a similar approach the energy profile of
the reaction was calculated at B3LYP/6-31+G∗ level of theory,
using the software package Gaussian 03.[13] Incorporation of the
polarization continuum solvent model (PCM, ε = 78.69) allowed
the simulation of the reaction in water.[14] All structures were
derived from the conjugate base of 2a (Et = Me) [namely cb2a
(Et = Me)], two formaldehyde molecules, and a potassium ion.
Local minima and transition states were confirmed by zero or
one negative frequency, respectively, and have been calculated
for all species on pathway A and on all species up until enone
3a on pathway B/C. As in the work of Ando, inclusion of the
cation is essential – in its absence the hydroxymethylation does
not proceed. Gas phase calculations predict that the barrier to
forming enone 3a is lower than that for the formation of enone
1a. However, incorporation of the solvation model inverts these
energies and provides results more consistent with the synthetic
studies (Fig. 1, and with computationally derived structures in
Figure S1 of theAccessory Publication). In the PCM/DFT calcu-
lations divergence of pathwaysA and B/C occurs at intermediate
5. The difference in energy between the two pathways at the rate
While these experiments prove that intermediates 8a and 4a
can give rise to allylic alcohol 1a, considering the extended
reaction time required to achieve low levels of conversion, their
significance appears to be minimal. Subsequent investigations
focussed on the conversion of enone 3a into allylic alcohol 1a
by a MBH reaction catalyzed by HWE reagent 2. In order to
establish this, a crossover experiment in which enone 3a was
¶The dihydroxymethylated derivative of phosphonate 2 (R = OEt) is indicated as an intermediate in the hydroxymethylation reported in Ref. [5a].
†Intermediate 8a was prepared as described in the experimental section as a part of other optimization studies.

RESEARCH FRONT
724
S. J. Ryan, C. D. Thompson, and D. W. Lupton
OCH2
OCH2
OMe
ϩ
ϩ
K
Ϫ
K
50
25
O
O
O
O
OMe
OMe
P
P
Ph
Ph
O
OMe
O
9
Ϫ
cb2a
5
OCH2
0
cb2a؅
ϩ
6
K
CH2O
OCH2
OMe
ϩ
10
Ϫ25
Ϫ50
O
ϩ
K
O
ϩ
OMe
K
P
K
O
O
O
5؅
ϩ
Ϫ
Ϫ
OMe
Ph
12
CH2O
Ph
O
P
O
O
O
O
P
OMe
Ph
OMe
OMe
K
P
Ph
OMe
OH
Ph
HO
Ϫ
O
O
Ϫ
OMe
O
Ϫ75
P
OMe
Ϫ
OMe
OH
Ϫ100
Ϫ125
Ϫ150
Ϫ175
Ϫ200
Ϫ225
O
ϩ
11
K
Ph
3a
O
O
OMe
OMe
P
Ϫ
3a؅
1a؅
Ph
HO
O
1a
Potassium
Carbon
OH
O
Pathway A
Ph
Phosphorus
Oxygen
Pathway B/C
Reaction coordinate
Fig. 1. Reaction coordinate calculated using PCM/B3LYP/6-31+G∗.
determining step is 36.15 kJ mol−1 and corresponds to the tran-
sition states for the formation of oxetane 9 in pathway B/C and
oxetane 10 in pathway A. Only small barriers (∼4 kJ mol−1) for
P–C and C–O bond cleavage isolate the oxetanes 9 and 10 from
the final products 1a and 3a.
O
O
P
O
OH
O
OEt
Ph
Ph
Ph
O
P
OEt
OEt
OEt
O
2a
1a
8a
Ph
These computational results concur with the synthetic studies
and indicate that the formation of allylic alcohol 1a occurs by a
lower energy pathway than the formation of enone 3a. The mag-
nitude of this difference is somewhat larger than the synthetic
studies suggest, however, this may be a result of the limitations of
our computational model, which simplifies the solvent mixture.
:
:
:
–
2
1
Ratio
K2CO3, H2O
1.5
0.5
:
0.5
Scheme 5.
indicate that when enone 3 forms it is unlikely to react further
to provide allylic alcohol 1; and that phosphonates 4a and 8a
arise from the degradation of product 1a. These results and the
DFT/PCM studies indicate that the formation of allylic alco-
hol 1a is occurring by a Schlosser–Corey type mechanism as
originally proposed by Villiéras (pathway A).
Through the course of these studies it has been demonstrated
that the Villiéras modification of the HWE reaction is a viable
method for the synthesis of known and new MBH adducts.
Conclusions
Based on DFT studies, and the observed by-product formation in
the experimental studies, it appears that formation of the allylic
alcohol 1 is more energetically favourable than the formation
of enone 3. With substrates less sterically demanding than a
phenyl group (R = cinnamyl, hydrocinnamyl, etc.) this differ-
ence in transition-state energy is large enough that the allylic
alcohol 1 is the only product formed. Crossover experiments
K2CO3,
O
O
OH
OH
O
OH
O
O
H2O,
P
CH2O
ϩ
OEt
mNO2C6H4
Ph
Ph
mNO2C6H4
2c
(Eqn 4)
(Eqn 5)
OEt
1a
1c
63%A
3a
1 equiv.
0%A
2 equiv.
K2CO3,
H2O,
CH2O
O
P
O
O
OH
O
O
ϩ
OEt
pOMe2C6H4
Ph
Ph
pOMe2C6H4
OEt
1d
1a
3a
1 equiv.
2d
2 equiv.
0%A
65%A
AYields were determined based on 1H NMR analysis.

RESEARCH FRONT
Direct Synthesis of α-Hydroxymethylated Enones from β-Keto Phosphonates
725
The reaction proceeds rapidly in all cases, and when unhin-
dered HWE reagents are used, gives rise to the desired product in
high yield and without any by-products. By employing synthetic
and theoretical studies the first direct evidence for the reaction
mechanism proposed by Villiéras has been obtained.
1715.6s, 1394.6m, 1368.8m, 1254.9s br, 1025.2s br. δH
(500 MHz) 3.98–4.05 (8H, m), 2.95 (4H, d, J 23.0), 2.48 (4H, t, J
7.0), 1.42–1.45 (4H, m), 1.21 (12H, t, J 7.0), 1.13–1.15 (8H, m).
δC (125 MHz) 202.2 (2C), 62.5 (d, J 6.5, 4C), 44.0 (2C), 42.3 (d,
J 126, 2C), 29.1 (2C), 28.8 (2C), 23.3 (2C), 16.3 (d, J 6.0, 4C).
m/z (ESI) 471.2 (70%) [M + H]+, 493.1 (100) [M + Na]+.
Experimental
General Method for α-Hydroxymethylation Reaction
Unless otherwise specified, proton (1H) and carbon (13C) NMR
spectra were recorded in CDCl3 on a Varian Mercury 300 spec-
trometeroperatingat300 MHzforprotonand75 MHzforcarbon
nuclei or a Varian DRX 500 spectrometer operating at 500 MHz
for proton and 125 MHz for carbon. Signals arising from the
residual protio-forms of the solvent were used as the internal
standard. Infrared spectra (νmax) were recorded on a Perkin–
Elmer RXI FTIR Spectrometer. Samples were analyzed as thin
films on NaCl plates and IR band intensities are expressed
as s = strong, m = medium, w = weak, and br = broad. Low-
resolution mass spectrometry (ESI) was performed on a Micro-
mass Platform QMS spectrometer. High-resolution mass spectra
(HRMS) were recorded on a Bruker BioApex 47e FTMS fitted
with an Analytical electrospray source using NaI for accurate
mass calibration. Low-resolution (EI) mass spectra were also
recorded on a VG Micromass 70/70F mass spectrometer with
an ion source temperature of 200◦C and electron impact energy
of 70 eV. Melting points were measured on a Stuart hot-stage
microscope apparatus. Flash column chromatography was per-
formed on silica gel (Davisil LC60A, 40–63 µm silica media)
using compressed air or nitrogen. TLC was performed using
aluminium-backedplates coated with 0.2 mmsilica(Merck, DC-
Platten, Kieselgel; 60 F254 plates). Eluted plates were visualized
using a 254 nm UV lamp and/or by treatment with a suitable
stain followed by heating.
To a stirred solution of phosphonate 2 (0.5 mmol) in aqueous
formaldehyde (2.0 mmol in 2 mL of H2O), K2CO3 (0.7 mL of
a 0.5 M aqueous solution, 0.35 mmol) was added dropwise. The
reaction mixture was allowed to stir for 1 h, at which time the
mixture was extracted with diethyl ether (3 × 5 mL). The com-
bined organic extracts were washed with brine (3 × 5 mL), dried
(MgSO4), and concentrated under vacuum.The resulting residue
was purified by flash column chromatography to afford, after
concentration of the appropriate fractions, the title compound.
The reaction mixture derived from the reaction of phosphonate
2a when separated by column chromatography (silica gel, 50%
EtOAc/hexane) afforded enone 3a (Rf 0.7) along with the major
product allylic alcohol 1a (Rf 0.4) in a ratio of 14:1.
2-(Hydroxymethyl)-1-phenylprop-2-en-1-one (1a)
Obtained as a clear oil (54 mg, 66%). Rf 0.4 (50% EtOAc/
hexane). νmax/cm−1 3417.4s br, 1652.9s, 1597.3m, 1447.6m,
1321.8m, 978.5m. δH (500 MHz) 7.75–7.76 (2H, m), 7.55–7.58
(1H, m), 7.44–7.48 (2H, m), 6.12 (1H, s), 5.83 (1H, s), 4.51
(2H, s). δC (125 MHz) 196.4, 146.7, 137.6, 132.8, 129.8, 128.6,
127.5, 63.1.
1-(3-Bromophenyl)-2-(hydroxymethyl)prop-
2-en-1-one (1b)
Starting materials and reagents were purchased from Sigma–
Aldrich and were used as supplied or, in the case of some liquids,
distilled. Tetrahydrofuran (THF) was distilled from sodium
benzophenone ketyl. Concentration under reduced pressure was
performedonarotaryevaporatorwiththewaterbathtemperature
not exceeding 40◦C.
Obtained as
a white solid (74 mg, 63%). Rf 0.3
(35% EtOAc/hexane). Mp 41.4–49.8◦C. (Found: 238.9709
[M − H•]+. C10H9BrO2 requires: 238.9708 [M − H•]+.)
νmax/cm−1 3105.8m br, 1645.1s, 1625.6m, 1562.2m, 1419.2w,
1078.0m. δH (500 MHz) 7.86 (1H, t, J 2.0), 7.64–7.69 (2H, m),
7.33 (1H, t, J 8.0), 6.19 (1H, t, J 0.5), 5.81 (1H, s), 4.49 (2H, s).
δC (125 MHz) 196.5, 146.3, 139.4, 135.7, 132.5, 130.2, 128.2,
128.2, 122.8, 63.0. m/z (EI) 239 (17%) [M − H•]+, 185 (42),
161 (100), 155 (34), 115 (13), 85 (8).
Experimental Procedures for the Synthesis of Phosphonates
(E)-Dimethyl 4-(2-Methoxyphenyl)-2-oxobut-
3-enylphosphonate (2g)
2-(Hydroxymethyl)-1-(4-methoxyphenyl)prop-
Using a modification to the procedure reported by Nes-
terov et al.[8c] dimethyl methylphosphonate was acylated with
2-methoxy cinnamyl chloride to yield the title compound as a
yellow oil (1.45 g, 50%). Rf 0.3 (EtOAc). (Found: 307.0656
[M + Na]+. C20H40O8P2 requires: 307.0711 [M + Na]+.)
νmax/cm−1 3473.0br, 2956.1m, 1682.6m, 1651.8m, 1597.3s,
1248.9s, 1027.9s. δH (500 MHz) 7.97 (1H, d, J 17), 7.56 (1H,
dd, J 7.5, 1.5), 7.38 (1H, tm, J 7.5), 6.97 (1H, t, J 7.5), 6.92 (1H,
dd, J 7.5, 1.5), 6.91 (1H, d, J 17), 3.90 (3H, s), 3.81 (6H, d, J 11),
3.36 (2H, d, J 23). δC (125 MHz) 191.5 (d, J 6.4), 158.8, 140.5,
132.4, 129.1, 126.3, 123.1, 120.9, 111.3, 55.7, 53.2 (d, J 6.5),
39.4 (d, J 129). m/z (ESI) 307.0 (100%) [M + Na]+, 285.1 (8)
[M + H]+.
2-en-1-one (1d)
Obtained as a clear oil (63 mg, 65%). Rf 0.3 (35%
EtOAc/hexane). (Found: 193.0863 [M − H•]+. C10H12O3
requires: 193.0865 [M − H•]+.) νmax/cm−1 3434.2s br,
2935.6m, 1647.0m, 1600.0s, 1509.3m, 1309.1m, 1260.3s,
1166.5s. δH (500 MHz) 7.80 (2H, d, J 9.0), 6.93 (2H, d, J 9.0),
6.05 (1H, t, J 1.0), 5.73 (1H, s), 4.47 (2H, s), 3.86 (3H, s). δC
(125 MHz) 197.0, 163.6, 146.5, 132.2, 130.0, 125.7, 113.8, 63.7,
55.7. m/z (ESI) 215.1 (100%) [M + Na]+.
1-(Furan-2-yl)-2-(hydroxymethyl)prop-2-en-1-one (1e)
Obtained as a clear oil (43 mg, 56%). Rf 0.2 (35% hex-
ane/EtOAc). (Found: 175.0366 [M + Na]+. C8H8O3 requires:
175.0371 [M + Na]+.) νmax/cm−1 3412.6s br, 2872.0m,
1646.3s, 1618.0m, 1560.8m, 1465.5s, 1391.1m, 1324.3m. δH
(500 MHz) 7.65 (1H, dd, J 1.5, 1.0), 7.20 (1H, dd, J 4.0, 1.0),
6.54 (1H, dd, J 4.0, 1.5), 6.20 (1H, d, J 0.5), 6.06 (1H, s), 4.45
(2H, s). δC (125 MHz) 183.5, 152.0, 147.5, 146.3, 125.6, 120.5,
112.4, 63.3. m/z (ESI) 175.0 (100%) [M + Na]+.
Tetraethyl 2,11-dioxododecane-1,12-diyldiphosphonate (2l).
Using the procedure described by Kim et al.,[8b] however
with 0.5 equivalents of sebacoyl chloride and 1 equivalents
of all other reagents the title compound was prepared as
a yellow oil (1.98 g, 42%). Rf 0.2 (10% MeOH:EtOAc).
(Found: 493.2093 [M + Na]+. C20H40O8P2 requires: 493.2096
[M + Na]+.) νmax/cm−1 3474.5br, 2982.7s, 2931.5s, 2855.8m,

RESEARCH FRONT
726
S. J. Ryan, C. D. Thompson, and D. W. Lupton
(E)-4-(Hydroxymethyl)-1-phenylpenta-1,4-dien-
3-one (1f)
7.53–7.57 (3H, m), 7.43–7.49 (3H, m), 7.33–7.36 (2H, m), 5.99
(1H, s), 5.65 (1H, s), 4.67 (1H, ddd, J 22.5, 9.5, 5.0), 3.77 (3H,
d, J 11.0), 3.75 (3H, d, J 11.0), 3.19–3.25 (2H, m). δC (75 MHz)
197.9, 195.9, 144.0 (d, J 13.1), 137.5 (d, J 13.8), 133.8, 132.5,
130.4, 129.6, 129.0, 128.9, 128.7, 128.4, 53.6 (d, J 6.9), 53.5 (d,
J 6.9), 45.2 (d, J 128.2), 31.4 (d, J 4.3). m/z (ESI) 395.2 (100%)
[M + Na]+.
Obtained as a white solid (86.6 mg, 92%). Rf 0.4 (35%
EtOAc/hexane). Mp 83.4–85.6◦C. (Found: 189.0914 [M + H]+,
211.0729 [M + Na]+. C12H12O2 requires: 189.0916 [M + H]+,
211.0735 [M + Na]+.) νmax/cm−1 3194.57w br, 1654.20m,
1596.27s, 1449.35m. δH (500 MHz) 7.70 (1H, d, J 15.5), 7.59–
7.60 (2H, m), 7.40–7.42 (3H, m), 6.29 (1H, d, J 15.5), 6.21 (1H,
s), 6.08 (1H, t, J 1.5), 4.45 (2H, d, J 6.0), 2.66 (1H, t, J 6.0, OH).
δC (125 MHz) 191.9, 148.0, 144.7, 134.8, 130.9, 129.2, 128.7,
125.0, 121.2, 63.2. m/z (ESI) 211.2 (73%) [M + Na]+, 193.9 (5),
181.2 (6), 159.0 (8).
Diethyl 1,5-Dioxo-1,5-diphenylpentan-
2-ylphosphonate (8a)
Formaldehyde (0.15 mL of a 10% aqueous solution,
0.5 mmol) was added over 1 h using a syringe pump to a stirred
solution of phosphonate 2a (128.1 mg, 0.5 mmol) in K2CO3
(1 mL of a 2 M aqueous solution, 0.5 mmol). The reaction
mixture was diluted with brine (5 mL), extracted into ether
(3 × 5 mL) and dried (MgSO4). Following filtration and con-
centration under reduced pressure, purification by flash column
chromatography (1/1, v/v, EtOAc in hexane) afforded ketone 8a
(99.0 mg, 51%yield)asaclearoil. Rf 0.15(50%hexane/EtOAc).
(Found: 411.1335 [M + Na]+. C21H25O5P requires: 411.1337
[M + Na]+.) νmax/cm−1 3409.0br, 2926.3m, 1679.7s, 1596.9m,
1448.8m, 1259.8m, 1035.4br. δH (500 MHz) 8.00–8.03 (2H, m),
7.89–7.92 (2H, m), 7.52–7.59 (2H, m), 7.41–7.49 (4H, m), 4.41
(1H, ddd, J 23.0, 9.0, 5.0), 4.03–4.16 (4H, m), 3.24 (1H, dt, J
10.5, 7.0), 2.95 (1H, dt, J 10.5, 7.0), 2.55–2.62 (1H, m), 2.41–
2.48 (1H, m), 1.24 (3H, t, J 7.5), 1.18 (3H, t, J 7.5). δC (75 MHz)
199.5, 196.5 (d, J 5.5), 137.6, 136.9, 133.8, 133.5, 129.1, 128.9,
128.3, 63.1 (d, J 4.1), 62.9 (J 4.1), 46.3 (d, J 127.9), 36.4 (d, J
11.9), 22.4 (d, J 4.5), 16.5 (d, J 6.4), 16.4 (d, J 6.4) (1 signal
overlapping). m/z (ESI) 411.2 (100%) [M + Na]+.
2-(Hydroxymethyl)-5-phenylpent-1-en-3-one (1j)
Obtained as a clear oil (91 mg, 96%). Rf 0.3 (35% hex-
ane/EtOAc). (Found: 191.1069 [M + H]+. C12H14O2 requires:
191.1072 [M + Na]+.) νmax/cm−1 3436.4s br, 2930.5m,
1673.4s, 1495.9m, 1453.6m. δH (500 MHz) 7.26–7.30 (2H, m),
7.19–7.21 (3H, m), 6.10 (1H, s), 6.00 (1H, t, J 1.5), 4.32 (2H,
s), 3.04 (2H, tm, J 8.0), 2.94 (2H, t, J 8.0). δC (125 MHz) 201.6,
147.0, 141.2, 128.7, 128.6, 126.4, 125.5, 62.7, 39.8, 30.2. m/z
(ESI) 213.1 (100%) [M + Na]+.
2-(Hydroxymethyl)dec-1-en-3-one (1k)
Obtained as a clear oil (55 mg, 60%). Rf 0.4 (35%
EtOAc/hexane). (Found: 183.1394 [M − H•]+. C11H20O2
requires: 183.1385 [M − H•]+.) νmax/cm−1 3167.6m br,
2927.5s, 2849.0m, 1663.3s, 1467.7m, 1401.1w. δH (500 MHz)
6.09 (1H, s), 5.98 (1H, t, J 1.0), 4.29 (2H, s), 2.68 (2H, t, J
7.5), 1.59 (2H, t, J 7.0), 1.22–1.30 (8H, m), 0.86 (3H, t, J 7.0).
δC (125 MHz) 203.1, 147.0, 125.1, 62.8, 38.1, 31.9, 29.4, 29.3,
24.5, 22.8, 14.3. m/z (EI) 183 (2%) [M − H•]+, 100 (36), 85
(100), 81 (20), 57 (20).
Conversion of Phosphonate 8a into Enone 4a (Eqn 2)
To a stirred solution of phosphonate 8a (144 mg, 0.5 mmol)
in aqueous formaldehyde (2.0 mmol in 2 mL of H2O), K2CO3
(0.7 mL of a 0.5 M aqueous solution, 0.35 mmol) was added
dropwise. The reaction mixture was allowed to stir for 2 h,
at which time the mixture was extracted with diethyl ether
(3 × 5 mL). The combined organic extracts were washed with
brine (3 × 5 mL), dried (MgSO4), and concentrated under vac-
uum.Analysis of the residue by 1H NMR spectroscopy indicated
the presence of enone 4a and starting material 8a in a ratio of
1:2.9.
2,13-Bis(hydroxymethyl)tetradeca-1,13-diene-
3,12-dione (1l)
Obtained as a white solid (81 mg, 57%). Rf 0.3 (EtOAc).
Mp 88.7–91.5◦C. (Found: 305.1726 [M + Na]+. C16H26O4
requires: 305.1729 [M + Na]+.) νmax/cm−1 3198.6br, 2909.9s,
2848.0m, 1662.7s, 1465.3m, 1408.4w. δH (500 MHz) 6.10 (2H,
s), 5.99 (2H, s), 4.31 (4H, s), 2.69 (4H, t, J 7.5), 1.59–1.62 (4H,
m), 1.30 (8H, s). δC (125 MHz) 203.1, 147.0, 125.3, 63.0, 38.0,
29.4, 29.3, 24.4. m/z (ESI) 305.3 (100%) [M + Na]+.
Conversion of Enone 4a into Allylic Alcohol 1a (Eqn 3)
During the optimization studies phosphonate 2a (0.5 mmol)
wasreactedasdetailedinTable1andproduced, alongwithallylic
alcohol 1a, enone 3a, and phosphonate 4a. For the ratio of these
by-products in each specific optimization see Table 1.
A solution of enone 4a (200 mg, 0.5 mmol) in K2CO3 (0.7 mL
of a 0.5 M aqueous solution, 0.35 mmol) was allowed to stir for
2 h, after which the mixture was extracted with diethyl ether
(3 × 5 mL). The combined organic extracts were washed with
brine (3 × 5 mL), dried (MgSO4), and concentrated under vac-
uum.Analysis of the residue by 1H NMR spectroscopy indicated
the presence of allylic alcohol 1a and enone 4a in a ratio of 1:3.5.
1-Phenylprop-2-en-1-one (3a)
Obtained as a clear oil. Rf 0.7 (1/1, v/v, EtOAc in hexane).
νmax/cm−1 2925.3m, 1671.9s, 1608.5m, 1447.9m, 1403.9m. δH
(500 MHz) 7.95(3) (1H, d, J 8.0), 7.95(1) (1H, d, J 8.0), 7.58
(1H, t, J 8.0), 7.49 (2H, t, J 8.0), 7.17 (1H, dd, J 17, 11), 6.44
(1H, dd, J 17, 2.0), 5.94 (1H, dd, J 11, 2.0, 1H). δC (75 MHz)
191.5, 137.7, 133.4, 132.8, 130.6, 129.1, 129.0.
Reaction of Phosphonate 2c and 2d in the Presence of
Enone 3a (Eqn 4 and 5)
A solution of enone 3a (67 mg, 0.5 mmol) and phosphonate 2c
(301 mg, 1 mmol) or 2d (286 mg, 1 mmol) in aqueous formalde-
hyde (2.0 mmol in 2 mL of H2O) and K2CO3 (2.1 mL of a
0.5 M aqueous solution, 1.05 mmol) were allowed to stir for
2 h, after which the mixture was extracted with diethyl ether
(3 × 5 mL). The combined organic extracts were washed with
brine (3 × 5 mL), dried (MgSO4), and concentrated under vac-
uum.Analysis of the residue by 1H NMR spectroscopy indicated
Dimethyl 4-Benzoyl-1-oxo-1-phenylpent-4-en-
2-ylphosphonate (4a)
Obtained as a clear oil. Rf 0.2 (50% hexane/EtOAc).
(Found: 373.1202 [M + H]+. C20H21O5P requires: 373.1205
[M + H]+.) νmax/cm−1 3409.0br, 2926.3m, 1679.7s, 1596.9m,
1448.8m, 1259.8m, 1035.4br. δH (500 MHz) 7.97–7.98 (2H, m),

RESEARCH FRONT
Direct Synthesis of α-Hydroxymethylated Enones from β-Keto Phosphonates
727
the absence of any allylic alcohol 1a and the formation of enones
1c and 1d, respectively.
[5] (a) J. Villiéras, M. Rambaud, Synthesis 1982, 924. doi:10.1055/
S-1982-29998
(b) J. Villiéras, M. Rambaud, Org. Synth., Coll. Vol. 1988, 8, 265.
(c) J. Villiéras, M. Rambaud, Synthesis 1983, 300. doi:10.1055/
S-1983-30311
(d) S. J. Danishefsky, N. Mantlo, J. Am. Chem. Soc. 1988, 110, 8129.
doi:10.1021/JA00232A027
(e) J. Robertson, J. W. P. Dallimore, P. Meo, Org. Lett. 2004, 6, 3857.
doi:10.1021/OL048396X
(f) R. Takagi, T. Nerio, Y. Miwa, S. Matsumura, K. Ohkato, Tetra-
hedron Lett. 2004, 45, 7401. doi:10.1016/J.TETLET.2004.08.068
(g) J. Qi, J. A. Porco, J. Am. Chem. Soc. 2007, 129, 12682.
doi:10.1021/JA0762339
Addition of Phosphonate 2a to Allylic Alcohol 1a
A mixture of phosphonate 2a (512 mg, 2.0 mmol) and allylic
alcohol 1a (162 mg, 1 mmol) was treated with K2CO3 (1 mL
of a 2 M aqueous solution, 0.5 mmol) and stirred magnetically
for 0.5 h. The reaction mixture was diluted with brine (5 mL),
extracted into ether (3 × 5 mL), and dried (MgSO4). Follow-
ing filtration and concentration, the residue was analyzed by
1H NMR spectroscopy and phosphonates 2a, allylic alcohol 1a,
and ketone 8a identified in a ratio of 3:1:1.
[6] (a) For example see: S. N. Patil, F. Liu, J. Org. Chem. 2008, 73, 4476.
doi:10.1021/JO702762U
Accessory Publication
(b) B. M. Trost, H.-C. Tsui, F. D. Toste, J. Am. Chem. Soc. 2000, 122,
3534. doi:10.1021/JA994326N
(c) S. Matsumoto,Y. Okubo, K. Mikam, J. Am. Chem. Soc. 1998, 120,
4015. doi:10.1021/JA972393G
Accessory publication contains 1H NMR spectra (1l, 2l, 4a, and
8a) and 13C NMR spectra (1l, 2l, and 8a) of indicated structures.
Also available are documents detailing the computational meth-
ods used along with the Cartesian coordinates of all computed
structures, and vibrational frequencies of minima and transition
states. The documents are available from the Journal’s website.
[7] For the use of surfactants to enhance the MBH reaction see:
A. Porzelle, C. M. Williams, B. D. Schartz, I. R. Gentle, Synlett 2005,
19, 2963.
[8] (a) D. Y. Kim, M. S. Kong, K. Lee, J. Chem. Soc., Perkin Trans. 1
1997, 1361 thus we followed. doi:10.1039/A607683A
(b) D. Y. Kim, M. S. Kong, T. H. Kim, Synth. Commun. 1996, 26,
2487. doi:10.1080/00397919608004561
Acknowledgements
The authors thank Yasmean Kalam for preliminary studies. The authors
are grateful for the financial support of the Australian Research Council
through the discovery program (DP0881137), Monash University through
the ECR program and computing support from VPAC and APAC (NCI).
D.W.L. acknowledges the provision of equipment from Roche Palo Alto.
(c) For the use of cuprates see for example: V. V. Nesterov, O. I.
Kolodiazhnyi, Tetrahedron 2007, 63, 6720 and references therein.
doi:10.1016/J.TET.2007.04.101
[9] The conjugate addition of HWE reagents is well-documented see for
example: (a) B. J. English, R. M. Williams, Tetrahedron Lett. 2009,
50, 2713. doi:10.1016/J.TETLET.2009.03.145
References
[1] (a) M. Schlosser, K.-F. Christmann, Synthesis 1969, 38.
(b) E. J. Corey, H. Yamamoto, J. Am. Chem. Soc. 1970, 92, 226.
doi:10.1021/JA00704A052
(b) M. G. Edwards, M. N. Kenworthy, R. R. A. Kitson, M. S. Scott,
R. J. K. Taylor, Angew. Chem. Int. Ed. 2008, 47, 1935. doi:10.1002/
ANIE.200705329
(c) M. Schlosser, K.-F. Christmann,Angew. Chem. Int. Ed. Engl. 1966,
5, 126. doi:10.1002/ANIE.196601261
(d) E. J. Corey, J. A. Katzenellenbogen, G. H. Posner, J. Am. Chem.
Soc. 1967, 89, 4245. doi:10.1021/JA00992A065
[10] For the MBH reaction with DMSO see: K. E. Price, S. J. Broadwater,
B. J. Walker, D. T. McQuade, J. Org. Chem. 2005, 70, 3980.
doi:10.1021/JO050202J
[11] (a) For rate enhancement of the MBH reaction in H2O see:
V. K. Aggarwal, D. K. Dean, A. Mereu, R. Williams, J. Org. Chem.
2002, 67, 510. doi:10.1021/JO016073Y
[2] (a) D.A. Mulholland, B. Parel, P. H. Coombes, Curr. Org. Chem. 2000,
4, 1011. doi:10.2174/1385272003375941
(b) X.-D. Luo, S.-H. Wu, Y.-B. Ma, D.-G. Wu, Phytochemistry 2000,
55, 867. doi:10.1016/S0031-9422(00)00247-8
(b) J. Cai, Z. Zhou, G. Zhao, C. Tang, Org. Lett. 2002, 4, 4723.
doi:10.1021/OL027197F
(c) T. Amagata, K. Minoura, A. Numata, J. Nat. Prod. 2006, 69, 1384.
doi:10.1021/NP0600189
(c) S. Luo, P. G. Wang, J.-P. Cheng, J. Org. Chem. 2004, 69, 555.
doi:10.1021/JO035345P
[3] (a)A. B. Baylis, M. E. D. Hillman, Offenlegungsschrift 2155113, U.S.
Patent 3,743,669 1972, Chem. Abstr. 1972, 77, 34174q.
(b) K. Morita, Z. Suzuki, H. Hirose, Bull. Chem. Soc. Jpn. 1968, 41,
2815. doi:10.1246/BCSJ.41.2815
[12] For leading computational studies on the Wittig reaction see:
(a) H. Yamataka, S. Nagase, J. Am. Chem. Soc. 1998, 120, 7530.
doi:10.1021/JA974237F
(b) R. I. Robiette, J. Richardson, V. K. Aggarwal, J. N. Harvey, J. Am.
Chem. Soc. 2006, 128, 2394. doi:10.1021/JA056650Q
(c) For the HWE reaction see: K. Ando J. Org. Chem. 1999, 64, 6815.
doi:10.1021/JO9909150
(d) J. Motoyoshiya, T. Kusaura, K. Kokin, S.-I. Yokoya,
Y. Takaguchi, Tetrahedron 2001, 57, 1715. doi:10.1016/
S0040-4020(01)00007-2
(c) For a review see: D. Basavaiah, A. J. Rao, T. Satyanarayana, Chem.
Rev. 2003, 103, 811. doi:10.1021/CR010043D
(d) For computational studies see: R. Robiette, V. K. Aggarwal,
J. N. Harvey, J. Am. Chem. Soc. 2007, 129, 15513. doi:10.1021/
JA0717865
[4] (a) W. S. Wadsworth, W. D. Emmons, J. Am. Chem. Soc. 1961, 83,
1733. doi:10.1021/JA01468A042
[13] M. J. Frisch, G. W.Trucks, H. B. Schlegel, G. E. Scuseria, M.A. Robb,
J. R. Cheeseman, J. Montgomery, et al., Gaussian 03 2004 (Gaussian,
Inc., Wallingford, CT).
(b) L. Horner, H. Hoffmann, H. G. Wippel, Chem. Ber. 1958, 91, 61.
doi:10.1002/CBER.19580910113
(c) For a review see: B. E. Maryanoff, A. B. Reitz, Chem. Rev. 1989,
89, 863. doi:10.1021/CR00094A007
[14] S. Mierts, E. Scrocco, J. Tomasi, Chem. Phys. 1981, 55, 117.
doi:10.1016/0301-0104(81)85090-2