Synthesis of 7a-substituted Hajos-Wiechert ketone analogues

Article abstract of DOI:10.1021/jo800638s

(Chemical Equation Presented) A general and efficient route to 2-substituted 1,3-cyclopentadiones 3 has been developed. This operationally simple, two-step procedure is amenable to multigram scale preparations of these useful synthetic intermediates. These compounds are then transformed to previously unknown, higher analogues of the Hajos-Parrish-Eder-Sauer-Wiechert ketone (enone 1, R = Me) following an enantioselective Robinson annulation.

Full text of DOI:10.1021/jo800638s

SCHEME 1. Retrosynthesis of HW-ketones
Synthesis of 7a-Substituted Hajos-Wiechert
Ketone Analogues
Jason W. J. Kennedy,† Sophia Vietrich, Hilmar Weinmann,
and Dominic E. A. Brittain*
Medicinal Chemistry, Bayer Schering Pharma AG,
13342 Berlin, Germany
chemists,5 and is a rare example of the catalytic formation of a
stereogenic quaternary carbon center.6 In view of this long
history and ongoing interest, it is surprising that very few
analogues have been reported where the angular substituent is
not a methyl group.7
dominic.brittain@bayerhealthcare.com
ReceiVed March 20, 2008
These bicyclic diketones 1 are generally prepared via the
proline-catalyzed aldol cyclization of triketone 2, which in turn
comes from the Michael addition of cyclic 2-substituted 1,3-
dione 3 with methyl vinyl ketone 4 (Scheme 1). Key to this
process is the availability of the 2-substituted dione 3. Although
these substituted diones are very useful synthetic building blocks
for steroid and many other cyclopentanoid natural products,8
their preparation presents some significant challenges9 and their
availability is limited.10 Therefore, the first problem that needed
to be solved in the development of a route to HW-ketone
analogues was the establishment of a general route to 2-sub-
stituted 1,3-cyclopentadiones.
These diones are theoretically available by reaction of 1,3-
cyclopentadione 3b with electrophiles. However, yields for these
additions are generally quite low because the anion from 3b
preferentially undergoes O- rather than C-alkylation.8,11 While
a solution to this problem for 2-substituted 1,3-cyclohexadiones
was recently reported by Ramachary and Kishor in their
preparation of analogues of the Wieland-Miescher ketone,12
there remains no general synthetic approach to 2-substituted 1,3-
cyclopentadiones.
A general and efficient route to 2-substituted 1,3-cyclopen-
tadiones 3 has been developed. This operationally simple,
two-step procedure is amenable to multigram scale prepara-
tions of these useful synthetic intermediates. These com-
pounds are then transformed to previously unknown, higher
analogues of the Hajos-Parrish-Eder-Sauer-Wiechert
ketone (enone 1, R ) Me) following an enantioselective
Robinson annulation.
TheHajos-Parrish-Eder-Sauer-Wiechertketone1a(Scheme
1, R ) Me, henceforth referred to as the HW-ketone), first
reported independently by researchers at Schering AG1 and
Hoffmann La Roche Inc.2 over three decades ago, has found
extensive use in organic synthesis. This ketone is of biological
interest because it resembles the CD-ring system of steroids,
and is of synthetic interest because it can be prepared in
enantiomerically pure form from a remarkably simple organo-
catalyzed process.3 Consequently, it has been used as a chiral
building block in several syntheses.4 Furthermore, its synthesis
involves the first enantioselective aldol reaction developed by
Although simple alkylation of 1,3-cyclopentadione 3b is not
an effective reaction, an alternative approach starts with the
Knoevenagel addition of the dione to aldehydes (Figure 1).
However, this route is also not without its problems. The product
from these additions is the excellent Michael acceptor ene-dione
5, which, among other side reactions can then rapidly add a
(5) List, B.; Hoang, L.; Martin, H. J. Proc. Natl. Acad. Sci. 2004, 101, 5839–
5842.
(6) (a) Corey, E. J.; Guzman-Perez, A. Angew. Chem., Int. Ed. 1998, 37,
388–401. (b) Denissova, I.; Barriault, L. Tetrahedron 2003, 59, 10105–10146.
(7) For some notable exceptions, see refs 1 and 2, as well as the following: (a)
Smith, H.; Hughes, G. A.; Douglas, G. H.; Hartley, D.; McLoughlin, J.; Siddal,
J. B.; Wendt, G. R.; Buzby, G.C., Jr.; Herbst, D. R.; Ledig, K. W.; McMenamin,
J. R.; Pattison, T. W.; Suida, J.; Tokolics, J.; Edgren, R. A.; Jansen, A. B. A.;
Gadsby, B.; Watson, D. H. R.; Phillips, P. C. Experientia 1963, 19, 394–396.
(b) Blazejewski, J. C. J. Fluorine Chem. 1990, 46, 515–519. (c) Lacoste, E.;
Vaique, E.; Berlande, M.; Pianet, I.; Vincent, J.-M.; Landais, Y. Eur. J. Org.
Chem. 2007, 167–177.
† Current address: Bayer CropScience AG, IOP Industrialization, Geba¨ude
A729, 41538 Dormagen, Germany.
(1) Eder, U.; Sauer, G.; Wiechert, R. Angew. Chem., Int. Ed. Engl. 1971,
10, 496–497.
(2) Hajos, Z. G.; Parrish, D. R. J. Org. Chem. 1974, 39, 1615–1621.
(3) For a recent review, see Section 2.21 in: Mukherjee, S.; Yang, J. W.;
Hoffmann, S.; List, B. Chem. ReV. 2007, 107, 5471–5569.
(8) Schick, H.; Eichhorn, I. Synthesis 1989, 47, 7–492, and references cited
therein.
(4) Selected examples: (a) Paquette, L. A.; Wang, T.-Z.; Sivik, M. R. J. Am.
Chem. Soc. 1994, 116, 2665–2666. (b) Enev, V. S.; Petrov, O. S.; Neh, H.;
Nickisch, K. Tetrahedron 1997, 53, 13709–13718. (c) Overman, L. E.; Rucker,
P. V. Tetrahedron Lett. 1998, 39, 4643–4646. (d) Renneberg, D.; Pfander, H.;
Leumann, C. J. J. Org. Chem. 2000, 65, 9069–9079. (e) Molander, G. A.;
Quirmbach, M. S.; Silva, L.F., Jr.; Spencer, K. C.; Balsells, J. Org. Lett. 2001,
3, 2257–2260. (f) Lepage, O.; Stone, C.; Deslongchamps, P. Org. Lett. 2002, 4,
1091–1094. (g) Lugar, C. W., III; Magee, D.; Adrian, M. D.; Shetler, P.; Bryant,
H. U.; Dodge, J. A. Bioorg. Med. Chem. Lett. 2003, 13, 4281–4284. (h) Chochrek,
P.; Kurek-Tyrlik, A.; Michalak, K.; Wicha, J. Tetrahedron Lett. 2006, 47, 6017–
6020.
(9) For example, Beilstein only lists three syntheses for benzyl-type analogues
of 3, which are either not general or unsuitable for large-scale preparations: (a)
Hiraga, K. Chem. Pharm. Bull. 1965, 13, 1359–1361. (b) Tolstikov, G. A.;
Ismailov, S. A.; Velder, Ya. L.; Miftakhov, M. S. J. Org Chem. USSR (Engl.
Transl.) 1991, 27, 72–78. (c) Poss, A. J.; Belter, R. K. J. Org. Chem. 1988, 53,
891–891.
(10) The MDL Available Chemicals Directory reveals that only the methyl
and ethyl analogues of 3 are commercially available.
(11) McIntosh, J. M.; Beaumier, P. M. Can. J. Chem. 1973, 51, 843–847.
(12) Ramachary, D. B.; Kishor, M. J. Org. Chem. 2007, 72, 5056–5068.
10.1021/jo800638s CCC: $40.75
Published on Web 06/10/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 5151–5154 5151

FIGURE 2. Preparation of 2-substituted 1,3-cyclohexadiones 8. Com-
bined yields for both the Knoevenagel and hydrogenation steps are
shown.
FIGURE 3. Preparation of the triketones 2.
than reactions with branched aldehydes. This result is in line
with the observation that the Knoevenagel addition with
unbranched aldehydes is not efficient.13 A significant advantage
of this process is that the product from both the Knoevenagel
and hydrogenation reactions is obtained simply by precipitation
and filtration in sufficiently pure form to be used in subsequent
reactions without any further treatment. Multigram amounts of
these products are thus readily available from this methodology.
The overall efficiency of this preparation would be even
further increased if the two steps could be performed as a one-
pot, tandem process. In a preliminary experiment, stirring
benzaldehyde with 1,3-cyclopentadione 3b, pyrrolidine, and
palladium catalyst in trifluoroethanol under hydrogen yielded
the expected product 3c in essentially the same yield as from
the two-step process. Unfortunately, the reaction with benzal-
dehyde appears to be an exceptional case, and this tandem
strategy could not be successfully applied to other aromatic or
aliphatic aldehydes.
FIGURE 1. Yields for the Mannich bases 6 and diones 3.
second molecule of cyclodione to give the bis-adduct 7.
However, Bolte and co-workers reported that 5 can be ef-
fectively trapped in the presence of pyrrolidine as the stable
and easily isolated Mannich base 6.13,14 We reasoned that under
reducing conditions in a weakly acidic solvent, Mannich base
6 would slowly regenerate ene-dione 5, which would then
rapidly undergo hydrogenation before unwanted side reactions
could occur. In line with this theory, we were pleased to find
that palladium-catalyzed hydrogenation of 6 in 2,2,2-trifluoro-
ethanol afforded the desired substituted cyclopentadione 3 in
good overall yields. 1H NMR spectra of these compounds
revealed that they exist exclusively in their enol forms. The
use of trifluoroethanol is crucial for success in this hydrogena-
tion, since hydrogenations in ethanol returned the Mannich base
unchanged.
We were pleased to see that our methodology for preparing
2-substituted 1,3-diones is not limited to cyclopentadione. Both
1,3-cyclohexadione and dimedone afforded the expected 2-sub-
stituted products 8 in similar yields following our established
protocol (Figure 2).12 As with the 5-membered analogues,
sequences with unbranched aliphatic aldehydes gave low yields
in the Knoevenagel condensation.
Our attempts to prepare HW-ketone analogues started via the
classical approach outlined in Scheme 1, where the triketone 2
is first isolated and then cyclized in a separate step. The requisite
triketone was prepared via treatment of diketone 3 with MVK
4 in refluxing THF in the presence of NEt3 (Figure 3).16 Other
procedures for this transformation were generally unsuccessful.17
A variety of 1,3-cyclopentadiones bearing aromatic and
aliphatic substituents were subsequently prepared with this
methodology (Figure 1). Sterically demanding aldehydes,
heterocyclic aldehydes, and aldehydes bearing boronic esters
all undergo effective transformation to dione 3. Similarly, the
chlorinated compound 3h can be prepared if the reduction is
performed in the presence of Ph2S.15 Overall yields for this two-
step process were generally good, although reactions with
unbranched aliphatic aldehydes were noticeably lower yielding
(13) Bolte, M. L.; Crow, W. D.; Yoshida, S. Aust. J. Chem. 1982, 35, 1421–
1429.
(14) The enedione 5 can also be trapped by using thiophenol: Fuchs, K.;
Paquette, L. A. J. Org. Chem. 1994, 59, 528–532.
(15) Mori, A.; Miyakawa, Y.; Ohashi, E.; Haga, T.; Maegawa, T.; Sajiki, H.
Org. Lett. 2006, 8, 3279–3281.
(16) Zhuang, Z.-P.; Zhou, W.-S. Tetrahedron 1985, 41, 3633–3641.
5152 J. Org. Chem. Vol. 73, No. 13, 2008

previously reported in the study of intramolecular cyclizations
of a triketone similar to 2.22
These optimized conditions were then applied to the prepara-
tion of HW-ketone analogues (Figure 4). All of these compounds
were prepared in good yield (except for the sterically demanding
1g) and moderate to excellent enantiopurity as determined by
chiral HPLC. X-ray crystallography of enone 1c confirmed that
the (R)-isomer was formed by reaction with (S)-proline 10a, as
would be expected from the generally accepted transition state
model for these reactions.5
Despite this success, a limitation of the process soon became
apparent. We were unable to procure triketone 2 from three of
the diones listed in Figure 1: the p-Cl-, p-CF3-, and the boronic
ester-substituted 3h, 3i, and 3j, respectively. The product
obtained following standard treatment of these compounds with
MVK and NEt3 was a complex mixture that sometimes
contained the desired triketone, but also the enone 1 along with
other suspected aldol products. Although milder reaction condi-
tions did lead to the isolation of triketone 2, the formation of
enone 1 suggested that the aldol cyclization with these substrates
was a relatively facile process. Indeed, a tandem Michael
addition-aldol cyclization approach to the HW-ketone 1a with
a stoichiometric amount of proline has already been reported,23
and a catalytic version to the Wieland-Miescher ketone has
been developed by Barbas.24
Building on our experience from the triketone cyclization
optimization, we were delighted to see that this tandem reaction
is indeed a viable process when applied to the substituted diones
3. Treatment of a solution of MVK and dione 3 in DMSO with
(S)-proline 10a led to a slow formation of triketone 2 with
concomitant conversion to hydroxyketone 9 and sometimes
small amounts of enone 1. Purification of this intermediate 9
followed by acid-catalyzed dehydration successfully generated
enones 1 in good yields and excellent enantioselectivities (Figure
5). This new tandem method was applied not only to the
substrates that were problematic under the first method but also
to other representative diones.
FIGURE 4. HW-ketone analogues 1 from triketone 2. The superscript
“a” indicates hydroxyketone 9 was not isolated. Yields reported for
enone 1 are combined yields for the cyclization and dehydration of
triketone 2.
We next sought to optimize the conditions for the asymmetric
aldol cyclization. With use of benzyl-substituted 2c as a test
substrate, several solvents and catalysts were screened for their
suitability in this cyclization.18 In organocatalyzed reactions with
the triketones 2, cyclization to the aldol product 9 takes place
over a couple of days, but dehydration to the enone 1 does not
readily occur under the reaction conditions (Figure 4). Subse-
quent treatment of the hydroxyketone 9 with acid then causes
rapid elimination to the enone 1. Although this dehydration can
be performed without isolation of the hydroxyketone 9 by adding
acid to the mixture after consumption of 2, best results are
obtained by isolating and purifying the intermediate alcohol
first.19
While both DMSO and DMF gave good yields and high
levels of enantioselectivity, other solvents led either to poor
levels of enantioselectivity (MeCN, EtOH, THF, CHCl3, DCE)
or to no conversion (water, toluene). Proline 10a turned out to
be the best catalyst for the cyclization, generally affording
reasonable yields and enantioselectivities greater than 95:5.
Reactions with the sulfonamide and tetrazole proline analogues20
were notably faster than those with proline, albeit with slightly
lower yields and levels of enantioselectivity. Very low enanti-
oselectivity was seen with phenylalanine as catalyst, and no
reaction was observed with MacMillan’s tert-butyl imidazoli-
dinone catalyst.21 During these cyclizations the formation of
diketone 3 was often observed, which accounts for the modest
yields in some reactions. This retro-Michael reaction has been
To explore the scalable nature of this process, 1,3-cyclopen-
tadione 3b was converted without complications to the benzyl
diketone 3c in an 80% yield on a 100 mmol scale. Although
the tandem Michael addition-enantioselective aldol cyclization
at this scale still requires some optimization, dione 3c was
converted to enone 1c in reasonable overall yield and purity
and with an ee of >98%.
In conclusion, we have developed a general and readily
scalable method to prepare 2-substituted 1,3-cyclopentadiones,
allowing easy access to these synthetically useful compounds.
These compounds were then readily transformed into previously
unknown analogues of the HW-ketone. We expect that these
steroid subunits will find wide use as valuable chiral building
blocks.
Experimental Section
(17) (a) Hajos, Z. G.; Parrish, D. R. Organic Syntheses; Wiley: New York,
1990; Collect. Vol. VII, pp 363-368. (b) Rajamannar, T.; Balasubramanian,
K. K. Synth. Commun. 1994, 24, 279–292. (c) Keller, E.; Feringa, B. L.
Tetrahedron Lett. 1996, 37, 1879–1882.
General Procedure for the Preparation of the Diones 3. A
solution of aldehyde (15 mmol) and pyrrolidine (0.83 mL, 10 mmol)
in CH2Cl2 (25 mL) was treated with a slurry of cyclopentadione
3b (986 mg, 10.0 mmol) and pyrrolidine (0.41 mL, 5 mmol) in
(18) Full details of the screening study can be found in the Supporting
Information.
(19) Small amounts of triketone 2 are difficult to detect under the reaction
conditions and are not readily separated from enone 1.
(20) Cobb, A. J. A.; Shaw, D. M.; Longbottom, D. A.; Gold, J. B.; Ley,
S. V. Org. Biomol. Chem. 2005, 3, 84–96.
(22) Collins, D. J.; Fallon, G. D.; Skene, C. E. Aust. J. Chem. 1994, 47,
623–648.
(23) Rajagopal, D.; Narayanan, R.; Swaminathan, S. Tetrahedron Lett. 2001,
42, 4887–4890.
(21) Austin, J. F.; MacMillan, D. W. C. J. Am. Chem. Soc. 2002, 124, 1172–
1173.
(24) Bui, T.; Barbas, C. F., III. Tetrahedron Lett. 2000, 41, 6951–6954.
J. Org. Chem. Vol. 73, No. 13, 2008 5153

2.93 (s, 2H), 2.49 (m, 4H), 2.10 (s, 3H), 2.00 (m, 4H); 13C NMR
(75 MHz, CDCl3) δ 217.2, 207.8, 135.1, 129.7, 128.6, 127.4, 61.1,
42.9, 37.8, 36.3, 29.9, 28.4; IR (KBr, cm-1) 3017, 2939, 1722, 1708,
759, 708; MS (EI, m/z (%)) 258 (M+, 21), 172 (25), 117 (28), 91
(100), 43 (22); HRMS (ES, m/z) calcd for C16H19O3 259.1334, found
259.1337.
General Procedure for the Preparation of the Enones 1
via Enantioselective Aldol Cyclization of the Triketones 2. A
solution of triketone 2 (1.0 mmol) in either DMF (1 mL) or DMSO
(10 mL) was degassed through 2 freeze-thaw cycles, treated with
(S)-proline (30 mol %), and stirred under Ar in the dark. Aqueous
workup and flash chromatography after 5-7 d afforded hydroxy-
ketone 9. Acid-catalyzed dehydration of 9 with either 1 M H2SO4/
DMF or 1.25 M HCl/MeOH led to the desired enone 1. Repre-
sentative characterization data for the benzyl-substituted enone 1c:
TLC (50% EtOAc/hexanes, UV/KMnO4) 0.50; [R]20 (+) 347.9
D
(c 0.49, CHCl3); 1H NMR (400 MHz, CDCl3) δ 7.26 (m, 3H), 7.07
(m, 2H), 6.05 (d, J ) 2.2 Hz, 1H), 3.07 (AB, J ) 13.1 Hz, 1H),
3.04 (AB, J ) 13.1 Hz, 1H), 2.51 (m, 3H), 2.25 (m, 3H), 2.04 (m,
1H), 1.83 (dt, J ) 13.6 Hz, 6.1 Hz, 1H); 13C NMR (100 MHz,
CDCl3) δ 217.7, 198.2, 169.3, 135.6, 129.6, 128.7, 127.6, 125.2,
54.2, 42.4, 37.0, 32.9, 29.3, 28.2; IR (KBr, cm-1) 3030, 2921, 1740,
1670, 1601, 1205, 763, 713; MS (EI, m/z (%)) 240 (M+, 100), 181
(23), 169 (24), 119 (35); HRMS (EI, m/z) calcd for C16H16O2
240.115030, found 240.113514.
FIGURE 5. HW-ketone analogues 1 via tandem Michael addition-aldol
cyclization. Yields reported for enone 1 are combined yields for the
three steps from diketone 3.
General Procedure for the Preparation of the Enones 1
via Tandem Michael Addition-Aldol Cyclization of Dione 3.
A solution of dione 3 (1.0 mmol) and MVK 4 (188 µL, 2.0 mmol)
in DMSO (10 mL) was degassed through 2 freeze-thaw cycles,
treated with (S)-proline (30 mol %), and stirred under Ar in the
dark. Aqueous workup and flash chromatography after 5-7 d
afforded hydroxyketone 9. Acid-catalyzed dehydration of 9 with
either 1 M H2SO4/DMF or 1.25 M HCl/MeOH led to the desired
enone 1. Enones prepared via this tandem procedure were spec-
troscopically identical with those prepared via the enantioselective
aldol cyclization of triketone 2.
CH2Cl2 (25 mL). The resulting tan solution was stirred at rt for 3 h
and then concentrated to give a brown residue. Addition of cold
EtOAc caused precipitation of the desired Mannich base 6, which
was then isolated by filtration. A 0.1 M solution of this Mannich
base in 2,2,2-trifluoroethanol was then treated with 10% palladium
on charcoal (5 mol %) and hydrogenated under a balloon of H2.
After 23 h the catalyst was removed via filtration and the residue
was concentrated. Treatment of the resulting oil with 1 M HCl
caused precipitation of the product, which was filtered off and
washed with water to give the desired dione 3. Representative
characterization data for the benzyl-substituted dione 3c: TLC (10%
1
MeOH/CH2Cl2, UV/KMnO4) 0.54; H NMR (400 MHz, CD3OD)
Acknowledgment. We gratefully acknowledge the technical
staff of Bayer Schering Pharma, including Oliver Schenk and
Ingrid Pissarczyk for HPLC analyses, and Siegfried Ba¨sler for
the X-ray structure of 1c.
δ 7.17 (m, 5H), 3.40 (s, 2H), 2.48 (s, 4H); 13C NMR (100 MHz,
CD3OD) δ 199.5, 138.5, 128.7, 128.4, 128.0, 108.0, 68.2, 53.6,
32.4, 23.7; IR (KBr, cm-1) 2920, 2580, 1580, 1370; MS (CI, m/z
(%)) 189 (M + H+, 100), 173 (18), 113 (10); HRMS (ES, m/z)
calcd for C12H13O2 189.0916, found 189.0919.
Supporting Information Available: Complete experimental
and characterization data for all compounds reported in this
paper, details of the optimization studies for the aldol cyclization
and the large scale preparation of enone 1c, and X-ray crystal
structure data for enone 1c. This material is available free of
charge via the Internet at http://pubs.acs.org.
General Procedure for the Preparation of the Triketones
2. A solution of dione 3 (1.0 mmol), MVK 4 (200 µL, 2.0 mmol),
and NEt3 (650 µL, 4.8 mmol) in THF (15 mL) was refluxed at
85 °C under Ar for 20 h. Removal of the solvent afforded an oil,
which was then purified by flash chromatography to afford the
desired triketone 2. Representative characterization data for the
benzyl-substituted triketone 2c: TLC (30% EtOAc/hexane, KMnO4)
1
0.34; H NMR (300 MHz, CDCl3) δ 7.21 (m, 3H), 7.01 (m, 2H),
JO800638S
5154 J. Org. Chem. Vol. 73, No. 13, 2008