Regioselective synthesis and structural studies of substituted γ-hydroxybutenolides with use of a tandem baylis-hillman/singlet oxygenation reaction

Article abstract of DOI:10.1021/jo702762u

(Chemical Equation Presented) The scope and limitations of a regioselective synthesis of either α- or β-substituted γ-hydroxybutenolides from 3-furfural and enones are described. The sequence features a Baylis-Hillman reaction followed by singlet oxygen o

Full text of DOI:10.1021/jo702762u

Regioselective Synthesis and Structural Studies of Substituted
γ-Hydroxybutenolides with Use of a Tandem Baylis-Hillman/Singlet
Oxygenation Reaction
Santoshkumar N. Patil and Fei Liu*
Department of Chemistry & Biomolecular Sciences, Macquarie UniVersity, Sydney, NSW, 2109 Australia
fliu@alchemist.chem.mq.edu.au
ReceiVed December 28, 2007
The scope and limitations of a regioselective synthesis of either R- or ꢀ-substituted γ-hydroxybutenolides
from 3-furfural and enones are described. The sequence features a Baylis-Hillman reaction followed by
singlet oxygen oxidation with use of either an amine base or TBAF as the regioselectivity switches.
Structural studies in solution on some of the resulting γ-hydroxybutenolides are also reported.
SCHEME 1
Introduction
Bioactive natural products frequently contain a γ-hydroxy-
butenolide moiety that is substituted at either the R or ꢀ position.
Many R-substituted γ-hydroxybutenolides are known anti-
microbial compounds, and ꢀ-substituted γ-hydroxybutenolides
are often used as anticancer or anti-inflammatory agents.1 There
has been on-going interest in developing facile and selective
methods for synthesizing diverse and highly functionalized
analogues of this class of structures from readily available
starting materials.2 Approaches such as photooxidative conver-
sion of functionalized furans to R- or ꢀ-substituted γ-hydroxy-
butenolides with regioselectivity have been frequently employed
particularly for practical convergent synthesis (Scheme 1).3
We have reported the use of a simple reagent to confer
regioselectivity for either R- or ꢀ-substituted γ-hydroxybuteno-
lides in this photooxidative transformation as a versatile strategy
to access rapidly butenolides from readily available synthons
such as furfural and acrylates.4 The furans are functionalized
by using a Baylis-Hillman reaction with acrylates and then
subjected to endoperoxide formation under photochemical
conditions. Mechanistically, the two deprotonation pathways of
the endoperoxide intermediate (3), a and b, can be differentiated
through consideration of steric and neighboring group effects
(Scheme 2). When the functionalized furans are treated with a
bulky Hu¨nig’s base, pathway b is preferred to provide the
ꢀ-substituted γ-hydroxybutenolides. When TBAF is used instead
of Hu¨nig’s base, pathway a is preferred to provide the
(1) For recent examples of R- or ꢀ-substituted, bioactive butenolides found
in nature, see: (a) Wright, A. D.; de Nys, R.; Angerhofer, C. K.; Pezzuto, J. M.;
Gurrath, M. J. Nat. Prod. 2006, 69, 1180–1187. (b) Keyzers, R. A.; Davies-
Coleman, M. T. Chem. Soc. ReV. 2005, 34, 355–365. (c) Grossmann, G.;
Poncioni, M.; Bornand, M.; Jolivet, B.; Neuburger, M.; Sequin, U. Tetrahedron
2003, 59, 3237–3251. (d) Charan, R. D.; McKee, T. C.; Boyd, M. R. J. Nat.
Prod. 2001, 64, 661–663.
(2) For recent reviews of synthetic routes to butenolides, see:(a) Bru¨ckner,
R. Curr. Org. Chem. 2001, 5, 679–718. (b) Carter, N. B.; Nadany, A. E.;
Sweeney, J. B. J. Chem. Soc., Perkin Trans. 2002, 1, 2324–2342.
(3) For initial work in photooxidative conversion of furans to γ-hydroxy-
butenolides, see: (a) Grimminger, W.; Kraus, W. Liebigs Ann. Chem. 1979, 10,
1571–1576. (b) Brownbridge, P.; Chan, T. H. Tetrahedron. Lett. 1980, 21, 3431–
3434. (c) Katsumura, S.; Hori, K.; Fugiwara, S.; Isoe, S. Tetrahedron. Lett. 1985,
26, 4625–4628. (d) Kernan, M. R.; Faulkner, D. J. J. Org. Chem. 1988, 53,
2773–2776. For a recent summary of the use of singlet oxygen in biomimetic
syntheses, see: (e) Margaros, I.; Montagnon, T.; Tofi, M.; Pavlakos, E.;
Vassilikogiannakis, G. Tetrahedron 2006, 62, 5308–5317. For a very recent
example of regioselective synthesis of γ-hydroxybutenolides, see: (f) Aquino,
M.; Bruno, I.; Riccio, R.; Gomez-Paloma, L. Org. Lett. 2006, 8, 4831–4834.
(4) (a) Patil, S. N.; Liu, F. Org. Lett. 2007, 9, 195–198. (b) Patil, S. N.; Liu,
F. J. Org. Chem. 2007, 72, 6305–6308.
4476 J. Org. Chem. 2008, 73, 4476–4483
10.1021/jo702762u CCC: $40.75 2008 American Chemical Society
Published on Web 05/20/2008

Studies of Substituted γ-Hydroxybutenolides
SCHEME 2. Regioselective Synthesis of r- or ꢀ-Substituted
γ-Hydroxybutenolides from 3-Furfural and Acrylates
The 3′-hydroxy group of the BH adducts (7a-e) was
protected by using a bulky TBS group. The protection is not
necessary for the singlet oxygenation but does enhance the
regioselectivity for the formation of ꢀ-substituted γ-hydroxy-
butenolides. When the functionalized furan has a bulky side
chain, such as 7f, this protection is not required for complete
regioselectivity in our method. This protection step was
problematic in the case of 7a where the use of imidazole with
a highly activated enone led to a complex mixture of reaction
products. However, an optimized method with TBSOTf and 2,6-
lutidine was able to circumvent this problem. Reactions were
monitored carefully as prolonged reaction time may result in
lower yields.
Synthesis of r- or ꢀ-Substituted γ-Hydroxybutenolides
from the BH Adducts. With a diverse collection of BH adducts,
the chemo- and regioselectivity scope of the singlet oxygenation
reaction was examined. Following our previous studies of the
photooxidation reaction of the 3-hydroxyacrylate furans (Scheme
2), these functionalized furans were treated with either Hu¨nig’s
base or TBAF to investigate the generality of these regioselec-
tivity switches (Tables 2 and 3).
The TBS-protected BH adducts, 8a-8c and 8e, upon Hu¨nig’s
base-mediated singlet oxygen oxidation, returned ꢀ-substituted
γ-hydroxybutenolides with complete regioselectivity (entries
1-3 and 5, Table 2). The reaction with 8d resulted in some
formation of 9d indicated by 1H NMR spectroscopy of the crude
reaction mixture. However, successful purification of 9d could
not be achieved. While methanol is the typical solvent to use,
DCM was used in some cases for better solubility and
purification outcomes. Overall the singlet oxygen oxidation is
chemoselective with good to excellent isolated yields. The
electron-deficient double bond in the substrates is inert in the
oxidation, and regioselectivity remained uncompromised even
for substrates with small side chains such as 8a and 8e (entries
1 and 5, Table 2). For 7f, where the side chain is large, complete
regioselectivity was achieved even without TBS protection of
the 3′-hydroxy group of the furan (entry 6, Table 2). When
TBAF was used instead of Hu¨nig’s base in singlet oxygen
oxidation, complete reversal of selectivity for the R-substituted
γ-hydroxybutenolides was observed (Table 3). The regioselec-
tivity was unaffected even for 7f, where the bulky side chain
may potentially interfere with the deprotonation step. This mild
and facile approach may find general applicability in the
synthesis of natural products that contain both R- and ꢀ-sub-
stituted γ-hydroxybutenolide cores.
Strucutral Studies with NMR and Optical Rotation
Spectroscopy. The γ-hydroxybutenolide core is known as the
key pharmacophore of this class of compounds that usually exist
as a mixture of epimers due to the γ-carbon center.9 Normally,
this epimeric mixture is readily indicated by NMR spectroscopy
through the manifestation of doubled peaks. Occasionally,
questions arose regarding the possibility of selectivity for one
epimer in this equilibrium. The likely reasons are due to either
conformational bias or trapping of the γ-hydroxy group, such
as the case with ꢀ-substituted γ-hydroxybutenolide natural
products dysidiolide or dictyodentrillolide.10,11 In the case of
R-substituted γ-hydroxybutenolides due to H-bonding between
the fluoride anion and the neighboring 3′-hydroxy group.
Effectively, either R- or ꢀ- substituted γ-hydroxybutenolides
can be readily synthesized by applying a different regioselec-
tivity switch to the same furan intermediate by using the same
method.
Herein we report the scope and limitation of this approach
using 3′-furfural and a variety of enones. In addition, studies
have also been performed with solution-based techniques, such
as NMR or optical rotation spectroscopy, to address structural
issues of representative butenolides given the presence of an
epimerizable center at the γ-position.
Results and Discussion
Functionalization of 3-Furfural with the Baylis-Hillman
(BH) Reaction. The Baylis-Hillman reaction was employed as
an economic and mild method for synthesizing funtionalized
furans with use of a variety of enones in moderate to good yields
(50-70%). The BH reaction, although efficient, can be limited
in scope. Hence optimization of conditions was performed for
each substrate. Methods using DBU as the base5 were successful
for the cyclic enones (entries 2 and 3, Table 1), but failed for
methyl vinyl ketone (6a). Changing DBU to DABCO did not
improve the outcome. Finally, 7a was successfully synthesized
by using imidazole and proline as cocatalysts (entry 1, Table
1).6 Aqueous trimethylamine was used for the synthesis of 7e
(entry 5, Table 1).7 The yield of 7d synthesized with use of
DABCO in a mixture of dioxane and water (1:1) was low (26%),
but improved significantly when the reaction was conducted in
neat condition with DABCO and phenol as cocatalysts (entry
4, Table 1).8 In general, DBU was found to be a BH catalyst
with more generality and can be used for diverse alkenes such
as electron deficient enones, sterically demanding acrylates (e.g.,
cholesterol acrylate), and cyclic enones.
(9) Glaser, K. B.; De Carvalho, M. S.; Jacobs, R. S.; Kernan, M. R.; Faulkner,
D. J. Mol. Pharm. 1989, 36, 782–788.
(5) (a) Aggarwal, V. K.; Mereu, A. Chem. Ind. 2001, 82, 595–600. (b)
Aggarwal, V. K.; Mereu, A. Chem. Commun. 1999, 2311–2312.
(6) Shi, M.; Jiang, J.-K.; Li, C.-Q. Tetrahedron Lett. 2001, 43, 127–130.
(7) Cai, J.; Zhou, Z.; Zhao, G.; Tang, C. Org. Lett. 2002, 4, 4723–4725.
(8) Faltin, C.; Fleming, E. M.; Connon, S. J. J. Org. Chem. 2004, 69, 6496–
6499.
(10) Gunasekera, S. P.; McCarthy, P. J.; Kelly-Borges, M.; Lobkovsky, E.;
Clardy, J. J. Am. Chem. Soc. 1996, 118, 8759–8760.
(11) (a) Cambie, R. C.; Bergquist, P. R.; Karuso, P. J. Nat. Prod. 1988, 51,
1014–1016. (b) Cambie, R. C.; Craw, P. A.; Bergquist, P. R.; Karuso, P. J. Nat.
Prod. 1988, 51, 331–334.
J. Org. Chem. Vol. 73, No. 12, 2008 4477

Patil and Liu
TABLE 1. Synthesis of Functionalized Furans
a 6a (3 equiv), imidazole (0.3 equiv), proline (0.3 equiv), DMF, rt, 3 days.4 b DBU (1 equiv), neat, rt, 12 h. c 3-Furfural (1.5 equiv), DABCO (1
equiv), 6d (1 equiv), phenol (1 equiv), 45°C, 3 days. d Trimethylamine (1 equiv) with 20% recovered furfural. e DBU (1 equiv), DCM. f TBSOTf (2
equiv), 2,6-lutidine (3 equiv), anhydrous DCM 0 °C to rt. g TBSCl (1.3 equiv), imidazole (2.5 equiv).
dysidiolide, broad peaks were obtained by NMR spectroscopy
for this natural product, and X-ray crystallography provided the
structure of a single epimer. However, as selective crystallization
of one epimer cannot be excluded, it is not possible for this to
be used as definitive evidence for a single epimer in solution
with peak broadening due to conformational equilibrium. In the
case of dictyodentrillolide, where the γ-hydroxy group is
acetylated, NMR spectroscopy indicated the presence of one
epimer with one set of broadened peaks, although it is unclear
whether this broadening is due to long-range coupling or some
conformational equilibrium. Most recently, Miles and co-
workers reported the fast epimerization of γ-hydroxybutenolides
such as manoalide by an alkyl amine.12 With a catalytical
amount of an amine base, the two γ-epimers exchange rapidly
to provide only one set of peaks in the 1H or 13C NMR spectrum.
These reports summarize the ambiguities frequently encountered
in interpreting NMR spectroscopic data where processes can
be obscured by either the inherent time scale or traces of
impurity below the detection limit of this solution technique.
We have observed that our ꢀ-substituted acrylyl γ-hydroxy-
butenolides appeared as one set of peaks in the 1H or 13C NMR
spectra when using Hu¨nig’s base in the oxygenation reaction
but not triethylamine. Two possibilities can explain the apparent
NMR pattern: one epimer (racemic) appearing as one set of
peaks or two epimers (both racemic) in fast exchange (Scheme
3) catalyzed by a base. To ascertain which possibility is
applicable requires solution techniques that are both time-
dependent and time-independent for comparison to reach a
definitive conclusion. With a facile method available to access
(12) Miles, W. H.; Duca, D. G.; Selfridge, B. R.; Palha De Sousa, C. A.;
Hamman, K. B.; Goodzeit, E. O.; Freedman, J. T. Tetrahedron Lett. 2007, 48,
7809–7812.
4478 J. Org. Chem. Vol. 73, No. 12, 2008

Studies of Substituted γ-Hydroxybutenolides
TABLE 2. Synthesis of ꢀ-Substituted γ-Hydroxybutenolides
a MeOH was the solvent instead of DCM. b A complex mixture was obtained. NMR spectrscopy of the crude mixture indicated characteristic signals
of 9d but purification was not successful. c From 7f.
SCHEME 3. Epimerization of ꢀ-Siloxyacrylyl
γ-Hydroxybutenolide (Racemic)
they clearly separated to two sets of peaks at -40 °C (stack 5).
The starting set of peaks can be recovered by increasing the
temperature again, showing no temperature-dependent degrada-
tion of the sample (stacks 6-9). However, the two sets of
butenolide peaks in the spectrum at -40 °C (stack 5) did not
1
align with those in the H spectrum of 12a obtained at room
temperature from a sample of 12a that was treated with a protic
acid to ensure its complete epimerization. As there may be
complications from potential conformational populations at
lower temperatures, this demonstrates the difficulty in reaching
a definitive conclusion on the epimerization state of 12a with
use of NMR.
ꢀ-substituted γ-hydroxybutenolides, we proceeded to a com-
parative structural study using NMR and optical rotation
spectroscopy.
ꢀ-Substituted γ-hydroxybutenolide 12a, prepared by singlet
oxygenation with Hu¨nig’s base, was subjected to a temperature-
variant NMR study (Figure 1). The starting point was a room
We then proceeded to prepare the 3′(S) isomer of 12a. In the
absence of an enantiomeric specific method to synthesize
chemically the BH precursor of 12a, we resorted to esterase
enzyme-catalyzed resolution of BH products.13 While this
method failed to apply to 12a due to presumably the inability
1
temperature H spectrum (stack 1, Figure 1) showing one set
of peaks between 5 and 7 ppm for the butenolide. As the
temperature was lowered, the peaks continued to broaden until
J. Org. Chem. Vol. 73, No. 12, 2008 4479

Patil and Liu
TABLE 3. Synthesis of r-Substituted γ-Hydroxybutenolides
The epimerization state of 3′(S)-γ-hydroxybutenolide 17 was
next investigated by optical rotation studies. If 17 existed as one
epimer at the γ-hydroxy carbon, then its optical rotation would be
a nonzero number for this enantiomerically pure sample while
displaying one set of peaks in the NMR spectra. The epimerization
of 17 from one to two after acid treatment will lead to an optical
rotation value very different from that of the pure enantiomer along
with two sets of peaks in the NMR spectra (possibility A).14 If 17
is already a mixture of two epimers in rapid exchange at the
γ-hydroxy carbon, then its optical rotation will not change before
and after acid treatment and irrespective of changes of its NMR
signals (possibility B). We have proposed previously that the
observation of one set of peaks in the NMR of these butenolide
using Hu¨nig’s base and neutral silica gel may favor one epimer of
the γ-hydroxybutenolide (possibility A) without definitive evidence
to exclude either possibility. The optical rotation studies now offer
the opportunity to exclude one of the two possibilities with
certainty, independent of the apparent NMR appearances. As shown
in Figure 2, 17, prepared by using Hu¨nig’s base in the singlet
oxgenenation reaction and purified by neutral silica gel, recorded
a specific optical rotation of 9.1° while presenting as one set of
1
peaks in its H NMR spectrum. Treatment of 17 under acidic
conditions to ensure its epimerization led to two sets of peaks in
1
the H NMR spectrum, yet its specific optical rotation remained
the same. This observation is consistent with only possibility B,
not A. Furthermore, after a catalytic amount of Hu¨nig’s base (0.1
mol %) was added to the epimerized sample of 17 according to
the Miles protocol, the 1H NMR spectrum reverted back from two
sets of peaks to one set of peaks, yet its specific optical rotation
again remained unchanged. This clearly indicated that the base-
catalyzed fast epimerization of the γ-center is operable here with
exclusion of possibility A.15 This also demonstrates that optical
rotation can be used as a time-independent technique in solution
to definitively clarify potentially ambiguous NMR interpretations
for these γ-hydroxybutenolides.16
In conclusion, this work illustrates the scope and limitation
of using simple regioselectivity switches to access either R- or
ꢀ-substituted γ-hydroxybutenolides from enone functionalized
furans by singlet oxygenation. Studies in solution using NMR
and optical rotation spectroscopy of some of the γ-hydroxy-
butenolides also provided further information regarding the
epimerization process at the γ-hydroxy carbon center of these
butenolides.
a A complex mixture resulted without evidence of the presence of
10d.
SCHEME 4. Synthesis of 3′(S)-Butenolide 17
Experimental Section
Preparation of Cholesterol Acrylate 6f ((8S,9S,10R,13R,14S,17R)-
10,13-Dimethyl-17-((R)-6-methylheptan-2-yl)-2,3,4,7,8,9,10,11,12,
13,14,15,16,17-tetradecahydro-1H-cyclopenta[a]phenanthren-3-yl
Acrylate). Acryloyl chloride was added dropwise to a stirred solution
of cholesterol and triethylamine at 0 °C under nitrogen. The reaction
(14) The scenario of the two epimers with two stereocenters having identical
optical rotation values is highly unlikely here, although it is possible for complex
natural products with many stereocenters to exhibit indistinguishable optical
rotation values for their epimers. For a recent example, see: Cecil, A. R. L.; Hu,
Y.; Vicent, M. J.; Duncan, R.; Brown, R. C. D. J. Org. Chem. 2004, 69, 3368–
3374.
of the reported esterase enzyme to use 12a as a substrate, another
BH adduct 13, the ethyl ester analogue of 12a, was accepted
by the enzyme to give 15 in 97% ee with 14 being the hydrolysis
product. Furan 15 was then converted to butenolide 17, with
an S configuration at the 3′-carbon, using singlet oxygen and
Hu¨nig’s base (Scheme 4).
(15) For a detailed study on the rates and mechanisms of acid- or base-
catalyzed cyclic hemiacetal formation between an oxygen nucleophile and an
aldehyde, see: Harron, H.; McClelland, R. A.; Thankachan, C.; Tidwell, T. T. J.
Org. Chem. 1981, 46, 903–910.
(16) CD spectroscopy indicated a positive cotton effect for 17 both before
and after acid treatment and no difference was discernable (spectra in the
Supporting Information). See: Soriente, A.; Crispino, A.; De Rosa, M.; De Rosa,
S.; Scettri, A.; Scognamiglio, G.; Villano, R.; Sodano, G. Eur. J. Org. Chem.
2000, 6, 947–953.
(13) Bhuniya, D.; Narayanan, S.; Lamba, T. S.; Krishna Reddy, K. V. S. R.
Synth. Commun. 2003, 33, 3717–3726.
4480 J. Org. Chem. Vol. 73, No. 12, 2008

Studies of Substituted γ-Hydroxybutenolides
1
FIGURE 1. Stacked H NMR spectra of 12a.
mixture was allowed to warm to room temperature and stirred for
5 h to affod cholesterol acrylate as a white solid. Yield 82%; H
63.5, 109.4, 126.7, 140.2, 143.9, 147.3, 159.6, 210.1; ESI [M +
Na+] 201.0530, calcd for (C10H10O3Na)+ 201.0528.
1
NMR (400 MHz, CDCl3) δ 0.65 (s, 3H), 0.84 (d, J ) 6.6 Hz, 3H),
0.84 (d, J ) 6.6 Hz, 3H), 0.89 (d, J ) 6.6 Hz, 3H), 2.5-1.0 (m,
31H), 5.35 (d, J ) 4.6 Hz, 2H), 5.75 (dd, J ) 10.42, 1.5 Hz, 1H),
6.07 (dd, J ) 17.2, 10.4 Hz, 1H), 6.35 (dd, J ) 17.27, 1.5 Hz,
1H); 13C NMR (100 MHz, CDCl3) δ 12.3, 19.2, 19.8, 21.5, 23.0,
23.3, 24.3, 24.7, 28.2, 28.4, 28.7, 32.3, 32.3, 50.4, 56.6, 57.1, 74.5,
123.2, 129.4, 130.7, 139.9, 166.0.
General Procedure for the Preparation of Baylis-Hillman
Adducts. DBU (1.04 mmol) was added to a mixture of an activated
alkene (1.04 mmol) and 3-furfural (1.04 mmol) being stirred at 0
°C. The reaction mixture was allowed to warm to room temperature
and stirred under nitrogen at room temperature until completion.
Upon finishing, the reaction mixture was adsorbed onto silica and
purified by flash column chromatography using petroleum ether/
ethyl acetate (2:1) to afford the Baylis-Hillman adduct. For 7f,
DCM (400 µL per 1 mmol of cholesterol acrylate) was added to
the reaction mixture to improve solubility.
3-(Furan-3-yl(hydroxy)methyl)but-3-en-2-one (7a). Yield 61%;
1H NMR (400 MHz, CDCl3) δ 2.34 (s, 3H), 3.31 (br s, -OH),
5.55 (s, 1H), 6.04 (s, 1H), 6.15 (s, 1H), 6.29-6.30 (m, 1H),
7.33-7.34 (m, 2H); 13C NMR (100 MHz, CDCl3) δ 26.9, 66.5,
109.5, 126.8, 140.2, 143.6, 149.8, 200.8; ESI [M + Na+] 189.0523,
calcd for (C9H10O3Na)+ 189.0528.
2-(Furan-3-yl(hydroxy)methyl)cyclopent-2-enone (7b). Yield
70%; 1H NMR (400 MHz,CDCl3) δ 2.47-2.49 (m, 2H), 2.60-2.64
(m, 2H), 3.35 (d, J ) 4.7 Hz, -OH), 5.53 (s, 1H), 6.39 (s, 1H),
7.36-7.45 (m, 3H); 13C NMR (100 MHz, CDCl3) δ 27.1, 35.8,
2-(Furan-3-yl(hydroxy)methyl)cyclohex-2-enone (7c). Yield
75%; 1H NMR (400 MHz, CDCl3) δ 1.88-1.99 (m, 2H), 2.30-2.42
(m, 4H), 3.7 (br s, -OH), 5.45 (s, 1H), 6.27 (t, J ) 1.3 Hz, 1H),
6.88 (t, J ) 4.1 Hz, 1H), 7.29-7.33 (m, 2H); 13C NMR (100 MHz,
CDCl3) δ 22.8, 25.9, 38.8, 66.0, 109.5, 127.2, 139.9, 140.5, 143.4,
147.4, 200.7; ESI [M + Na+] 215.0677, calcd for (C11H12O3Na)+
215.0678.
2-(Furan-3-yl(hydroxy)methyl)acrylamide (7d). Yield 64%; 1H
NMR (400 MHz, CD3OD) δ 5.54 (s, 1H), 5.68 (s, 1H), 5.91 (s,
1H), 6.38 (s, 1H), 7.39-7.43 (m, 2H); 13C NMR (100 MHz,
CD3OD) δ 68.0 111.2, 121.1, 129.2, 141.9, 145.2, 148.1, 173.1;
ESI [M + Na+] 190.0482, calcd for (C8H9NO3Na)+ 190.0480.
2-(Furan-3-yl(hydroxy)methyl)acrylonitrile (7e). Yield 68%;
1H NMR (400 MHz, CDCl3) δ 2.60 (br s, -OH), 5.27 (s, 1H),
6.04 (s, 1H), 6.10 (s, 1H), 6.40-6.42 (m, 1H), 7.42 (s, 1H),
7.47-7.49 (m, 1H); 13C NMR (400 MHz, CDCl3) δ 67.6, 108.9,
117.4, 125.2, 126.0, 130.7, 140.8, 144.5; ESI [M + Na+] 172.0376,
calcd for (C8H7NO2Na)+ 172.0374.
(8S,9S,10R,13R,14S,17R)-10,13-Dimethyl-17-((R)-6-methyl-
heptan-2-yl)-2,3,4,7,8,9,10,11,12,13,14,15,16,17-tetradecahydro-
1H-cyclopenta[a]phenanthren-3-yl 2-(Furan-3-yl(hydroxy)me-
thyl)acrylate (7f). Yield 16%; 1H NMR (400 MHz, CDCl3) δ 0.67
(s, 3H), 0.86 (d, J ) 6.6 Hz, 3H), 0.86 (d, J ) 6.6 Hz, 3H), 0.91
(d, J ) 6.6 Hz, 3H), 0.92 (s, 1H), 1.02 (s, 3H), 0.92-1.70 (m, 25
H), 1.70-2.05 (m, 2H), 3.15 (t, J ) 6.2 Hz, 1H), 5.35-5.40 (m,
2H), 5.50 (d, J ) 6.4 Hz, 1H), 5.83 (dd, J ) 2.3, 2.3 Hz, 1H),
6.35-6.37 (m, 2H), 7.36-7.40 (m, 2H); 13C NMR (100 MHz,
J. Org. Chem. Vol. 73, No. 12, 2008 4481

Patil and Liu
CDCl3) δ -4.6, -4.5, 18.7, 26.2, 26.9, 35.9, 63.3, 109.4, 128.3,
139.6, 143.3, 149.8, 158.7, 207.9; ESI [M + Na+] 315.1400, calcd
for (C16H24O3SiNa)+ 315.1392.
2-((tert-Butyldimethylsilyloxy)(furan-3-yl)methyl)cyclohex-2-
1
enone (8c). Yield 81%; H NMR (400 MHz, CDCl3) δ 0.01 (s,
3H), 0.01 (s, 3H), 0.88 (s, 9H), 1.87-2.04 (m, 2H), 2.34-2.42
(m, 4H), 5.70 (s, 1H), 6.26-6.27 (m, 1H), 7.15 (t, J ) 4.22 Hz,
1H), 7.24-7.26 (m, 1H), 7.29-7.30 (m, 1H); 13C NMR (100 MHz,
CDCl3) δ -4.6, 18.7, 23.2, 26.1, 26.3, 38.9, 63.7, 109.5, 129.4,
139.6, 142.6, 142.9, 145.4, 198.5; ESI [M + Na+] 329.153939,
calcd for (C17H26O3SiNa)+ 329.154342.
2-((tert-Butyldimethylsilyloxy)(furan-3-yl)methyl)acryla-
mide (8d). Yield 36%; 1H NMR (400 MHz, CDCl3) 0.10 (s, 3H),
0.10 (s, 3H), 0.91 (s, 9H), 5.52 (s, 1H), 5.59 (s, 1H), 6.09 (d, J )
1.3 Hz, 1H), 6.24-6.25 (m, 1H), 6.7 (br s, -NH2), 7.30-7.31 (m,
1H), 7.32 (t, J ) 1.74 Hz, 1H); 13C NMR (400 MHz, CDCl3) -4.7,
-4.5, 18.6, 26.1, 70.2, 109.4, 123.1, 128.0, 139.6, 143.7, 144.5
168.3; ESI [M + Na+] 304.1342, calcd for (C14H23NO3SiNa)+
304.1345.
2-((tert-Butyldimethylsilyloxy)(furan-3-yl)methyl)acryloni-
trile (8e). Yield 85%; 1H NMR (400 MHz, CDCl3) δ 0.02 (s, 3H),
0.10 (s, 3H), 0.91 (s, 9H), 5.24 (s, 1H), 5.95-5.97 (m, 1H),
6.03-6.05 (m, 1H), 6.34-6.73 (m, 1H), 7.39 (t, J ) 1.8 Hz, 1H),
7.42-7.44 (m, 1H); 13C NMR (100 MHz, CDCl3) -4.6, -4.5, 18.6,
26.1, 68.4, 108.9, 117.5, 126.1, 127.5, 129.2, 140.3, 144.1; ESI
[M + Na+] 286.1244, calcd for (C14H21NO2SiNa)+ 286.1239.
General Procedure for the Singlet Oxygen Oxidation
Reaction with Hu¨nig’s Base. To a mixture of a Baylis-Hillman
adduct (0.30 mmol) and rose bengal (3 mg, 0.003 mmol) in
methanol or DCM (70 mL) was added Hu¨nig’s base (0.36 mmol).
The reaction mixture was then exposed to singlet oxygen (generated
from compressed air with a 150 W flood light) at -78 °C. The
reaction mixture was then monitored for disappearance of starting
material. Upon completion the reaction mixture was removed from
the bath, and the solvent was removed under vacuum at room
temperature and purified by flash column chromatography to afford
the butenolide product.
FIGURE 2. A comparative study of 17 with NMR and optical rotation
spectroscopy: (a) 1H NMR spectrum of 17 prepared by using Hu¨nig’s
base and purified by neutral silica gel ([R]D 9.1 at 23.2 °C) and (b) 1H
NMR spectrum of epimerized 17 ([R]D 8.5 at 22.6 °C).
4-(1-(tert-Butyldimethylsilyloxy)-2-methylene-3-oxobutyl)-5-
hydroxyfuran-2(5H)-one (9a). Yield 91%, colorless oil; 1H NMR
(400 MHz, CDCl3) δ 0.01 (s, 3H), 0.05 (s, 3H), 0.88 (s, 9H), 2.36
CDCl3) δ 12.3, 19.2, 19.8, 21.5, 23.0, 23.3, 24.3, 24.8, 26.23, 3.15,
28.5, 28.7, 32.3, 32.4, 36.6, 37.1, 37.4, 38.4, 38.4, 39.9, 40.2, 42.8,
50.4, 56.6, 57.1, 67.5, 75.5, 109.8, 123.4, 126.1, 127.2, 139.8, 139.8,
140.2, 142.2, 143.8, 166.3; ESI [M + Na+] 559.3741, calcd for
(C35H52O4Na)+ 559.3763.
(s, 3H), 5.50 (s, 1H), 5.82-6.04 (m, 2H), 6.30-6.31 (m, 2H); 13
C
NMR (100 MHz, CDCl3) δ -4.8, -4.6, 18.5, 26.1, 26.4, 26.7,
65.3, 66.4, 98.3, 118.6, 118.8, 128.5, 129.1, 148.5, 169.3, 170.3,
171.0, 171.4, 199.4, 199.9; ESI [M + H+] 313.1469, calcd for
(C15H25O5Si)+ 313.1471.
4-((tert-Butyldimethylsilyloxy)(5-oxocyclopent-1-enyl)methyl)-
5-hydroxyfuran-2(5H)-one (9b). Yield 91%, colorless oil; 1H
NMR (400 MHz, CDCl3) δ 0.04 (s, 3H), 0.08 (s, 3H), 0.89 (s,
9H), 1.94 (br s, -OH), 2.50-2.53 (m, 2H), 2.67-2.70 (m, 2H),
5.23-5.40 (m, 1H), 5.89-6.09 (m, 2H), 7.76 (s, 1H); 13C NMR
(100 MHz, CDCl3) δ -4.6, 18.6, 26.1, 27.6, 35.9, 64.5, 64.6, 97.9,
98.8, 118.4, 120.4, 146.3, 147.1, 162.5, 163.1, 168.2, 169.7, 170.5,
170.8, 208.8, 209.6; ESI [M + Na+] 347.1284, calcd for
(C16H24O5SiNa)+ 347.1291.
4-((tert-Butyldimethylsilyloxy)(6-oxocyclohex-1-enyl)methyl)-
5-hydroxyfuran-2(5H)-one (9c). Yield 88%, colorless oil; 1H NMR
(400 MHz, CDCl3) δ 0.02 (s, 3H), 0.08 (s, 3H), 0.89 (s, 9H),
1.98-2.04 (m, 2H), 2.45-2.49 (m, 4H), 5.48 (s, 1H), 5.91 (s, 1H),
6.03 (s, 1H), 7.24-7.26 (m, 1H); 13C NMR (100 MHz, CDCl3) δ
-4.7, -4.5, 18.6, 22.9, 26.1, 26.3, 38.6, 64.9, 98.7, 118.6, 139.4,
149.7, 170.9, 199.7; ESI [M + Na+] 361.1444, calcd for
(C17H26O5SiNa)+ 361.1441.
2-((tert-Butyldimethylsilyloxy)(2-hydroxy-5-oxo-2,5-dihydro-
furan-3-yl)methyl)acrylonitrile (9e). Yield 95%, solid; 1H NMR
(400 MHz, CD3OD) δ 0.15 (s, 6H), 0.95 (s, 9H), 5.27 (s, 1H),
5.96-6.22 (m, 4H); 13C NMR (100 MHz, CD3OD) δ -4.2, -4.0,
19.8, 26.9, 71.9, 100.4, 117.8, 121.2, 125.9, 134.7, 169.6, 172.6;
ESI [M + H+] 296.1313, calcd for (C14H22NO4)+ 296.1318.
General Procedure for the Preparation of Protected
Baylis-Hillman Adducts. TBSOTf (2.0 mmol) (or TBSCl (1.3
mmol)) was added to a stirring solution of Baylis-Hillman adducts
(1.0 mmol) and 2,6-lutidine (3 mmol) (or imidazole (2 mmol)) in
anhydrous DCM (100 µL) under nitrogen at 0 °C. The reaction
mixture was allowed to warm to room temperature and stirred under
nitrogen. The reaction mixture was monitored by TLC for disap-
pearance of starting material. In general, reactions were complete
in 45 min with use of TBSOTf. Prolonged reaction time in the
TBSOTf method may result in silylation of the enone in the case
of 8a. Upon finishing the reaction mixture was purified by passing
through silica gel with petroleum ether/ethyl acetate (9:1) to afford
the product as a colorless oil.
3-((tert-Butyldimethylsilyloxy)(furan-3-yl)methyl)but-3-en-2-
1
one (8a). Yield 87%; H NMR (400 MHz, CDCl3) δ -0.01 (s,
3H), 0.03 (s, 3H), 0.89 (m, 9H), 2.30 (s, 3H), 5.72 (s, 1H), 6.10 (s,
1H), 6.23-6.24 (m, 1H), 6.27-6.28 (m, 1H), 7.27 (s, 1H),
7.29-7.30 (m, 1H); 13C NMR (100 MHz, CDCl3) δ -4.6, -4.5,
18.7, 26.3, 26.9, 64.9, 109.5, 124.7, 128.7, 139.8, 143.0, 151.9,
199.1; ESI [M + Na+] 303.1386, calcd for (C15H24O3SiNa)+
303.1392.
2-((tert-Butyldimethylsilyloxy)(furan-3-yl)methyl)cyclopent-
2-enone (8b). Yield 36%; 1H NMR (400 MHz, CDCl3) δ 0.01 (s,
3H), 0.04 (s, 3H), 0.89 (s, 9H), 2.36-2.49 (m, 2H), 2.51-2.67
(m, 2H), 5.50 (s, 1H), 6.34-6.35 (m, 1H), 7.30 (t, J ) 1.7 Hz,
1H), 7.35-7.36 (m, 1H), 7.59-7.61 (m, 1H); 13C NMR (100 MHz,
4482 J. Org. Chem. Vol. 73, No. 12, 2008

Studies of Substituted γ-Hydroxybutenolides
(8S,9S,10R,13R,14S,17R)-10,13-Dimethyl-17-((R)-6-methyl-
heptan-2-yl)-2,3,4,7,8,9,10,11,12,13,14,15,16,17-tetradecahydro-
1H-cyclopenta[a]phenanthren-3-yl 2-(Hydroxy(2-hydroxy-5-
oxo-2,5-dihydrofuran-3-yl)methyl)acrylate (11f). Yield 93%; 1H
NMR (400 MHz, CDCl3) δ 0.67 (s, 3H), 0.85 (d, J ) 6.5 Hz, 3H),
0.86 (d, J ) 6.6 Hz, 3H), 0.91(d, J ) 6.5 Hz, 3H), 1.02 (s, 3H),
0.92-2.5 (m, 31H), 4.2-3.8 (br s, -OH), 4.75-4.60 (m, 1H), 5.34
(s,1H), 5.39 (d, J ) 4.4 Hz, 1H), 5.9-6.2 (m, 3H), 6.42(s, 1H);
13C NMR (100 MHz, CDCl3) δ 12.3, 19.2, 19.8, 21.5, 23.0, 23.3,
24.3, 24.8, 28.1, 28.5, 28.7, 32.3, 32.4, 36.3, 36.7, 37.0, 37.3, 38.4,
39.9, 40.2, 42.8, 50.4, 56.6, 57.1, 67.2, 76.2, 98.5, 119.3, 123.7,
139.2, 139.5, 165.8, 168.9, 171.1; ESI [M + Na+] 591.3683, calcd
for (C35H52O6Na)+ 591.3662.
23.3, 24.3, 24.8, 28.2, 28.5, 28.7, 32.3, 32.4, 36.7, 37.1, 37.4, 38.5,
40.0, 40.2, 42.8, 50.5, 56.6, 57.2, 67.0, 75.8, 97.6, 123.6, 123.8,
128.4, 138.3, 138.8, 139.7, 146.1, 165.8, 170.3; ESI [M + Na+]
591.3661, calcd for (C35H52O6Na)+ 591.3662.
Deprotection of the Silylated Butenolides. To the TBS
protected butenolide (0.3 mmol) in anhydrous DCM (100 µL) was
added tetrabutylammonium fluoride, 1.0 M solution in tetrahydro-
furan (0.3 mmol). The mixture was stirred under nitrogen and at
room temperature for 6 h. The solvent was removed under vacuum,
and the residue was then passed through silica gel to remove
impurities (silica gel 60 Å 0.06-0.2 mm, 70-230 mesh) and further
purified by flash column chromatography to afford the butenolide
as a colorless oil.
General Procedure for the Singlet Oxygen Oxidation
Reaction with TBAF. Tetrabutylammonium fluoride, 1.0 M
solution in tetrahydrofuran (0.36 mmol), was added to the mixture
of a Baylis-Hillman adduct (0.30 mmol) and rose bengal (3 mg,
0.003 mmol) in dichloromethane (70 mL). Methanol (70 mL) was
used as the solvent for polar starting materials. The reaction mixture
was then exposed to singlet oxygen (generated from compressed
air with a 150 W flood light) at -78 °C for 5 h. The reaction
mixture was then removed from the bath, and the solvent was
removed under vacuum at room temperature. The crude mixture
was dissolved in acetonitrile (2 mL), passed through Poly-Prep
columns containing prefilled AG50W-X8 (H+) resin of ion ex-
change capacity of 3.4 (nominal mequiv/2 mL of resin), and flushed
with 8 mL of acetonitrile. The acetonitrile was removed and the
residue, after the protonation step, was filtered through silica gel
to remove trace impurities (silica gel 60 Å 0.06-0.2 mm, 70-230
mesh) and further purified by flash column chromatography to
afford the butenolide as a colorless oil. In the case of polar
butenolides, more than one protonation step may be necessary.
5-Hydroxy-3-(1-hydroxy-2-methylene-3-oxobutyl)furan-2(5H)-
5-Hydroxy-4-(1-hydroxy-2-methylene-3-oxobutyl)furan-2(5H)-
one (11a). Yield 60%; 1H NMR (400 MHz, CDCl3) δ 2.41 (s, 3H),
5.35 (s, 1H), 6.0 (s, 1H), 6.19 (s, 1H), 6.29 (s, 1H), 6.36 (s, 1H);
13C NMR (100 MHz, CDCl3) δ 26.7, 68.1, 100.0, 119.8, 129.7,
146.9, 167.7, 170.4, 200.6; ESI [M + Na+] 221.0422, calcd for
(C9H10O5Na)+ 221.0426.
5-Hydroxy-4-(hydroxy(5-oxocyclopent-1-enyl)methyl)furan-
1
2(5H)-one (11b). Yield 80%; H NMR (400 MHz, CD3OD) δ
2.41-2.5 (m, 2H), 2.64-2.74 (m, 2H), 4.6 (br s, -OH), 5.29 (s,
1H), 5.99 (br s, 1H), 6.11 (br s, 1H), 7.77 (t, J) 2.8 Hz, 1H); 13
C
NMR (100 MHz, CD3OD) δ 28.8, 36.7, 64.0, 100.7, 119.5, 146.6,
165.2, 172.6, 173.6, 210.8; ESI [M + Na+] 233.0427, calcd for
(C10H10O5Na)+ 233.0426.
5-Hydroxy-4-(hydroxy(6-oxocyclohex-1-enyl)methyl)furan-
2(5H)-one (11c). Yield 52%; 1H NMR (400 MHz, CD3OD) δ
1.96-2.05 (m, 2H), 2.40-2.51 (m, 4H), 4.6 (br s, -OH), 5.42 (s,
1H), 5.86-6.11 (m, 2H), 7.16 (s, 1H); 13C NMR (100 MHz,
CD3OD) δ 24.5, 27.1, 27.8, 40.0, 65.8, 66.5, 100.7, 119.2, 140.3,
15 0.9, 152.4, 172.6, 173.8, 200.6, 201.2; ESI [M + Na+] 247.0575,
calcd for (C11H12O5Na)+ 247.0576.
2-(Hydroxy(2-hydroxy-5-oxo-2,5-dihydrofuran-3-yl)methy-
l)acrylonitrile (11e). Yield 54%; 1H NMR (400 MHz, CD3OD) δ
4.51 (br s, -OH), 5.09 (br s, 1H), 6.09-6.19 (m, 4H); 13C NMR
(100 MHz, CD3OD) δ 68.7, 98.7, 116.3, 119.1, 123.9, 132.8, 167.7,
171.2; ESI [M + Na+] 204.0265, calcd for (C8H7NO4Na)+
204.0273.
Preparation of 3′(S)-17. The esterase enzyme (11 mg, g15 units/
mg solid) was added to a mixture of phosphate buffer (pH 7.1, 41
mL), 13 (94 mg, 0.48 mmol), and DMSO (3 mL). The solution
was stirred at 37-39 °C. After 4 h, HPLC (analytical column: (S,S)
Welko, 25 cm × 4.6 cm; mobile phase: isopropanol/hexane (3:
97); flow rate: 0.3 mL/min) indicated 97% ee for 15 (unhydrolyzed
13). The reaction mixture was then acidified to pH 2 and extracted
with ethyl acetate. Removal of ethyl acetate followed by flash
column chromatography afforded 15 in 97% ee (32 mg, 34% yield).
The synthesis and characterization of 3′(S)-17 was followed as
previously reported.4
1
one (10a). Yield 82%; H NMR (400 MHz, CD3OD) δ 2.37 (s,
3H), 4.53-4.72 (br s, -OH), 5.38 (s, 1H), 6.19 (s, 1H), 6.39 (s,
1H), 7.07 (s, 1H); 13C NMR (100 MHz, CD3OD) δ 27.1, 98.8,
128.9, 140.7, 148.6, 150.5, 172.8, 201.2; ESI [M + Na+] 221.0418,
calcd for (C9H10O5Na)+ 221.0426.
5-Hydroxy-3-(hydroxy(5-oxocyclopent-1-enyl)methyl)furan-
1
2(5H)-one (10b). Yield 91%; H NMR (400 MHz, CD3OD) δ
2.40-2.48 (m, 2H), 2.61-2.70 (m, 2H), 5.15-5.19 (m, 1H), 6.10
and 6.11 (t, J ) 0.9, 1.0, Hz, 1H), 7.16 and 7.18 (t, J ) 1.3, 1.4
Hz, 1H), 7.69-7.73 (m, 1H); 13C NMR (100 MHz, CD3OD) δ 28.6,
36.8, 62.2, 62.4, 99.8, 139.7, 139.9, 146.9, 147.1, 148.9, 149.0,
164.4, 172.6, 172.7, 210.9; ESI [M + Na+] 233.0419, calcd for
(C10H10O5Na)+ 233.0426.
5-Hydroxy-3-(hydroxy(6-oxocyclohex-1-enyl)methyl)furan-
2(5H)-one (10c). Yield 90%; 1H NMR (400 MHz, CD3OD) δ
1.97-2.05 (m, 2H), 2.37-2.50 (m, 4H), 5.32 (S, 1H), 5.88 (s, 1H),
7.06-7.13 (m, 2H); 13C NMR (100 MHz, CD3OD) δ 24.5, 27.7,
40.0, 64.3, 64.3, 105.0, 105.1, 140.5, 140.6, 141.6, 141.7, 146.7,
146.8, 150.6, 150.7, 172.2, 172.3, 200.7, 200.8; ESI [M + Na+]
247.0573, calcd for (C11H12O5Na)+ 247.0582.
Acknowledgment. We thank Professor Miles for his helpful
discussions. We also thank Professor Packer and her group from
the Australian Proteome Analysis Facility for help on mass
spectrometry of some of the butenolides. This work is supported
by an Australian Research Council Discovery Grant to F.L.
(ARC-DP055068). S.P. is supported by a predoctoral research
fellowship from Macquarie University.
2-(Hydroxy(5-hydroxy-2-oxo-2,5-dihydrofuran-3-yl)methyl)-
1
acrylonitrile (10e). Yield 95%; H NMR (400 MHz, CD3OD) δ
5.06 (s, 1H), 6.04-6.24 (m, 4H), 7.32 (s, 1H); 13C NMR (100 MHz,
CD3OD) δ 68.6, 68.8, 99.9, 100.0, 118.2, 118.3, 126.1, 126.3, 134.1,
134.2, 134.6, 138.3, 149.9, 150.1, 172.1; ESI [M + Na+] 204.0273,
calcd for (C8H7NO4Na)+ 204.0273.
(8S,9S,10R,13R,14S,17R)-10,13-Dimethyl-17-((R)-6-methyl-
heptan-2-yl)-2,3,4,7,8,9,10,11,12,13,14,15,16,17-tetradecahydro-
1H-cyclopenta[a]phenanthren-3-yl 2-(Hydroxy(5-hydroxy-2-
oxo-2,5-dihydrofuran-3-yl)methyl)acrylate (10f). Yield 80%; 1H
NMR (400 MHz, CDCl3) δ 0.67 (s, 3H), 0.87 (d, J ) 5.9 Hz, 3H),
0.87 (d, J ) 5.9 Hz, 3H), 0.91 (d, J ) 5.8 Hz, 3H), 1.02 (s, 3H),
0.92-2.5 (m, 28 H), 4.2-3.6 (br s, -OH), 4.68 (s, 1H), 5.31 (s,
1H), 5.39 (s, 1H), 6.00 (s, 1H), 6.12 (s, 1H), 6.40 (s, 1H), 7.11 (s,
1H); 13C NMR (100 MHz, CDCl3) δ 12.4, 19.2, 19.8, 21.5, 23.1,
Supporting Information Available: NMR spectra of com-
pounds 6a,7a-f, 8a-e, 9a-c, 9e, 10a-c, 10e-f, 11a-c, and
11e-f, chiral HPLC chromatogram of 15, CD spectra of 17,
1
and full H NMR spectra of 12a at various temperatures. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JO702762U
J. Org. Chem. Vol. 73, No. 12, 2008 4483