Titanium-mediated synthesis of 1,4-diketones from grignard reagents and acyl cyanohydrins

Article abstract of DOI:10.1002/anie.201003923

Double duty: In the presence of titanium isopropoxide, Grignard reagents were found to react with acyl cyanohydrins to give substituted 5-hydroxy-1,4-diketones (see scheme). This new reaction involves a formal addition of a 1,2-dianion equivalent to both the ester and nitrile moieties.

Full text of DOI:10.1002/anie.201003923

Angewandte
Chemie
DOI: 10.1002/anie.201003923
Synthetic Methods
Titanium-Mediated Synthesis of 1,4-Diketones from Grignard
Reagents and Acyl Cyanohydrins**
Paul Setzer, Alice Beauseigneur, Morwenna S. M. Pearson-Long, and Philippe Bertus*
1,4-Diketones are important synthetic intermediates for the
preparation of a wide range of five-membered cyclic com-
pounds, including cyclopentenones,[1] furans, thiophenes, and
pyrroles.[2] Various methods are used for the synthesis of
1,4-diketones,[3] including the conjugate addition of acyl anion
equivalents to a,b-enones (Scheme 1, path a),[4] and the less
convenient reverse strategy involving the addition of homo-
enolate equivalents to acid derivatives (Scheme 1, path b).[5]
Scheme 2. The Kulinkovich reaction.
homoenolate equivalent, and the addition of a second acid
derivative would afford 1,4-dicarbonyl derivatives in one step
(Scheme 2, path b).[12] To our knowledge, no sequential
addition of two different carboxylic acid derivatives to
titanacyclopropanes leading to 1,4-dicarbonyl compounds
has been reported to date. We present herein our solution
to this issue, which is the one-step formation of 1,4-diketones
compounds from acyl-protected cyanohydrins.
As part of our investigations related to the titanium-
mediated synthesis of cyclopropylamines from nitriles,[10] we
have demonstrated that cyanoesters 1 were particularly good
substrates, affording spirocyclopropane lactams 2 with a
catalytic amount of Ti(OiPr)4 (Scheme 3a).[13] Since amino-
alcohols such as 4 are precursors of aminocyclopropanecar-
Scheme 1. Strategies for the preparation of 1,4-diketones.
An alternative approach, which involves the addition of
1,2-dianion equivalents to two carboxylic acid derivatives
(Scheme 1, path c) is unknown to date, principally because of
the problems associated with the generation of 1,2-dianions
and the subsequent selective addition to two distinct carbox-
ylic acid derivatives.
Group IV metal/(h2-alkene) complexes are generally best
represented as metallacyclopropanes,[6] suggesting the pres-
ence of two polar carbon–metal bonds.[7] The resulting
1,2-dianion reactivity pattern was observed with the
(iPrO)2Ti/(h2-ethylene) complex A and various acid deriva-
tives,[8–11] as exemplified by the Kulinkovich reaction
(Scheme 2). In this reaction, carboxylic esters are converted
into cyclopropanols by the use of EtMgBr and Ti(OiPr)4,
through a formal double nucleophilic addition of A to the
same electrophilic carbon atom (Scheme 2, path a). The
metallacycle intermediate B could be considered as a
[*] P. Setzer, Dr. A. Beauseigneur, Dr. M. S. M. Pearson-Long,
Prof. Dr. P. Bertus
Scheme 3. Titanium-mediated cyclopropanation of cyanoesters.
Unitꢀ de Chimie Organique Molꢀculaire et Macromolꢀculaire
(UCO2M UMR 6011), CNRS and Universitꢀ du Maine
Avenue O. Messiaen, 72085 Le Mans cedex 9 (France)
Fax: (+33)343-83-39-02
boxylic acid (ACC) derivatives, we turned our attention to
using carboxylate-protected cyanohydrins 3 with the aim to
selectively provide acyl-protected cyclopropylamines 4 by
cyclopropanation and subsequent acyl transfer (Scheme 3b).
In an initial experiment, cyanomethyl benzoate 3a,
prepared in one step from benzoic acid and chloroacetoni-
trile, was reacted under the conditions (EtMgBr (2 equiv),
E-mail: philippe.bertus@univ-lemans.fr
[**] We are grateful to the CNRS and the Region Pays-de-la-Loire for a
PhD fellowship (to P.S.). A.B. also thanks the Agence Nationale de
la Recherche.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201003923.
Angew. Chem. Int. Ed. 2010, 49, 8691 –8694
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8691

Communications
Table 1: Optimization of the reaction conditions.
Table 2: Preparation of 1,4-dicarbonyl derivatives.
Entry[a] Substrate
Product
Yield
[%][b]
1
2
3
4
3b
3c
3d
3e
5b 65
Entry[a]
Ti(OiPr)4
[equiv]
EtMgBr
[equiv]
Solvent
4a
[%]
5a
[%][b]
1
2
3
4
5
6
7
8
0.2
1.1
1.1[c]
1.1
1.1
1.1
1.1[d]
0
2
Et2O
Et2O
Et2O
Et2O
Et2O
THF
THF
Et2O
12
10
5
6
3
28
0[e]
0[e]
45
67
71 (69)
70
17
48
44
0
5c 60
5d 60
5e 62
2.1
2.1
3
1.5
2.1
2.1
2.1
[a] Unless otherwise mentioned, reactions were carried out with 3a
(1 mmol), Ti(OiPr)4, EtMgBr (solution in Et2O) in the indicated solvent
(5 mL). [b] NMR yield. Yield of isolated product given in parentheses.
[c] EtMgBr was added at 08C. [d] [Cp2ZrCl2] was used instead of
Ti(OiPr)4. [e] By-products resulting from the direct addition of EtMgBr
to 3a were obtained.
5
3 f
3g
3h
3i
5 f 36
5g 50
5h 62
5i 55
5j 35
5k 42
5l 45
5m 56
5n 40
6
7
Ti(OiPr)4 (0.2 equiv), Et2O) used for the synthesis of 2
(Table 1). The expected cyclopropylamine 4a was obtained in
only 12% yield. The main product was identified as the
1,4-diketone 5a, that resulted from a formal addition of
1,2-ethylene dianion to both ester and nitrile moieties of the
cyanoester. When the reaction was run with a stoichiometric
amount of Ti(OiPr)4, a better yield of 5a was obtained
(Table 1, entry 2). The yield was again slightly increased by
mixing the reagents at 08C and subsequently warming the
reaction mixture to room temperature (Table 1, entry 3).
Interestingly, an excess of EtMgBr did not affect the reaction
outcome (Table 1, entry 4). In contrast, lowering the amount
of the Grignard reagent dramatically reduced the conversion,
showing clearly the need of at least two equivalents of
EtMgBr (Table 1, entry 5). When THF was used instead of
Et2O as the solvent, the ratio of 4a/5a was modified to favor
the formation of the cyclopropane (Table 1, entry 6). The
formation of 5a was also performed using [Cp2ZrCl2], but the
efficiency of the reaction was reduced (Table 1, entry 7). Only
direct alkylation was observed when the reaction was run
without the titanium catalyst (Table 1, entry 8).
The scope of the reaction was explored next (Table 2).
Cyanoesters prepared from either donor- or acceptor-sub-
stituted aromatic acids are equally good substrates (Table 2,
entries 1–4) and 1-furyl-substituted cyanoester 3 f was
smoothly converted into 5 f (Table 2, entry 5). Use of aliphatic
cyanoesters afforded the corresponding diketones in compa-
rable yields (Table 2, entries 6–8), except for 3j (Table 2,
entry 9). In this latter case, the cyclopropane 4 was also
obtained in a significant yield (12%). Diketones bearing a
conjugated alkene moiety can also be obtained (Table 2,
entries 10–11). Cyanoesters derived from substituted cyano-
hydrins are well tolerated, thus allowing the construction of
1,5-disubstituted 5-hydroxy-1,4-diketones (Table 2, entry 12).
Finally, the reaction is not limited to EtMgBr, as shown by the
8
9
3j
10
11
12
13[c]
3k
3l
3m
3a
14[d]
3a
5o 30
[a] Unless otherwise mentioned, reactions were carried out with 3 (1 mmol),
Ti(OiPr)4 (1.1 equiv), EtMgBr (2m in Et2O, 2.1 mmol) in Et2O (5 mL).
[b] Yield of isolated product. [c] Used iPrMgBr instead of EtMgBr. [d] Used
=
H2C CH-(CH2)2MgBr instead of EtMgBr.
Although the yields are lower, the products were obtained as
a single regioisomer (Table 2, entries 13–14).[14,15]
In light of the experimental results, a tentative rationale is
proposed (Scheme 4). The first step would imply the insertion
of the nitrile moiety of 3 into the titanacyclopropane
intermediate A (obtained from Ti(OiPr)4 and 2 equiv of
EtMgBr) to provide the five-membered titanacycle C. The
diketone 5 might be directly obtained from an intramolecular
=
direct insertion of the ester C O bond into the remaining
=
use of both iPrMgCl and H2C CH-(CH2)2MgBr, thus giving
titanium–carbon bond, however this possibility is unlikely
because of obvious geometric constrictions. A more plausible
access to diketones displaying other substitution patterns.
8692 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 8691 –8694

Angewandte
Chemie
then EtOAc (10 mL) and 1m HCl were added to obtain two clear
phases. The aqueous phase was extracted with EtOAc (3 ꢀ 20 mL).
The combined organic phases were washed with saturated aqueous
NaHCO3 and dried (MgSO4). After evaporation of the solvents, the
residue was purified by flash chromatography on silica gel to afford
the desired product (hexanes/EtOAc then EtOAc).
Received: June 28, 2010
Published online: September 30, 2010
Keywords: carbonyl compound · grignard reagents ·
.
metallacycles · synthetic methods · titanium
[1] For reviews, see: a) S. E. Gibson, S. E. Lewis, N. Mainolfi, J.
Organomet. Chem. 2004, 689, 3873 – 3890; b) G. Piancatelli, M.
DꢁAuria, F. DꢁOnofrio, Synthesis 1994, 867 – 889.
[2] For reviews, see: a) S. F. Kirsch, Org. Biomol. Chem. 2006, 4,
2076 – 2080; b) F. Bellina, R. Rossi, Tetrahedron 2006, 62, 7213 –
7256; c) T. Eicher, S. Hauptmann, A. Speicher, The Chemistry of
Heterocycles, Wiley-VCH, Weinheim, 2003; d) B. H. Lipshutz,
Chem. Rev. 1986, 86, 795 – 819.
Scheme 4. Proposed mechanism for the formation of 1,4-dicarbonyl
compounds.
[3] For recent preparations of 1,4-diketones: a) A. Voituriez, L. E.
Zimmer, A. B. Charette, J. Org. Chem. 2010, 75, 1244 – 1250;
b) M. Sauthier, N. Lamotte, J. Dheur, Y. Castanet, Y. Mortreux,
New J. Chem. 2009, 33, 969 – 971; c) M. D. Clifts, C. N. Taylor,
R. J. Thomson, Org. Lett. 2007, 9, 4667 – 4669; d) S. Xue, L.-Z.
Li, Y.-K. Liu, Q.-X. Guo, J. Org. Chem. 2006, 71, 215 – 218; e) M.
Yuguchi, M. Tokuda, K. Orito, J. Org. Chem. 2004, 69, 908 – 914.
[4] See for instance: a) M. Christmann, Angew. Chem. 2005, 117,
2688 – 2690; Angew. Chem. Int. Ed. 2005, 44, 2632 – 2634;
b) A. E. Mattson, A. R. Bharadwaj, K. A. Scheidt, J. Am.
Chem. Soc. 2004, 126, 2314 – 2315; c) Y. Hanzawa, N. Tabuchi,
K. Narita, A. Kakuuchi, M. Yabe, T. Taguchi, Tetrahedron 2002,
58, 7559 – 7571; d) B. H. Lipshutz, T. R. Elworthy, Tetrahedron
Lett. 1990, 31, 477 – 480; e) D. Seyferth, R. C. Hui, J. Am. Chem.
Soc. 1985, 107, 4551 – 4553; f) D. Seyferth, R. C. Hui, Tetrahe-
dron Lett. 1986, 27, 1473 – 1476.
[5] a) S. Aoki, T. Fujimura, E. Nakamura, I. Kuwajima, Tetrahedron
Lett. 1989, 30, 6541 – 6544; b) I. Ryu, M. Ikebe, N. Sonoda, S.
Yamato, G. Yamamura, M. Komatsu, Tetrahedron Lett. 2002, 43,
1257 – 1259.
[6] a) S. A. Cohen, P. R. Auburn, J. E. Bercaw, J. Am. Chem. Soc.
1983, 105, 1136 – 1143; b) M. L. Steigerwald, W. A. Goddard III,
J. Am. Chem. Soc. 1985, 107, 5027 – 5035.
[7] For reviews dealing with synthetic uses of titanacyclopropanes,
see: a) A. Wolan, Y. Six, Tetrahedron 2010, 66, 15 – 61; b) A.
Wolan, Y. Six, Tetrahedron 2010, 66, 3097 – 3133; c) O. G.
Kulinkovich, A. de Meijere, Chem. Rev. 2000, 100, 2789 – 2834;
d) F. Sato, H. Urabe, S. Okamoto, Chem. Rev. 2000, 100, 2835 –
2886.
[8] a) O. G. Kulinkovich, S. V. Sviridov, D. A. Vasilevsky, Synthesis
1991, 234; b) O. G. Kulinkovich, Eur. J. Org. Chem. 2004, 4517 –
4529.
[9] a) V. Chaplinski, A. de Meijere, Angew. Chem. 1996, 108, 491 –
492; Angew. Chem. Int. Ed. Engl. 1996, 35, 413 – 414; b) A.
de Meijere, S. I. Kozhushkov, A. I. Savchenko, J. Organomet.
Chem. 2004, 689, 2033 – 2055.
pathway would involve a C,N-bis(metallated) intermediate
such as D,[16] which results from the ring-opening of C by
magnesium alkoxides or Grignard reagents. Several related
metallacycle ring-openings have already been proposed in the
literature.[17–19] For instance, the addition of Grignard reagents
to five-membered oxatitanacyclopentanes has been proposed
by Kulinkovich et al. to explain the need for excess Grignard
in the synthesis of cyclopropanols.[17] Recently, the azatitana-
cyclopent-2-ene ring-opening has been postulated by Cha
et al. to explain the nonselective cyclopropanation of
nitriles.[18] The next step is the cyclization of the open
bimetallic intermediate D. The amide 4 (Scheme 4, path a)
could be obtained after acyl transfer of the metallated
cyclopropylamine (not shown). The addition could alterna-
tively occur on the ester moiety (Scheme 4, path b) to afford
E. This last intermediate can be seen as a protected diketone,
since it is not affected by an excess of Grignard reagent (see
Table 1, entry 4). Finally, the diketone 5 can be obtained by
hydrolysis of E. This indicates the divergent synthesis of 4 and
5 from the common intermediate D. Nevertheless, the
astonishing difference in reactivity between cyanoesters 1
and 3 remains unexplained at the moment.
In conclusion, the 1,2-dianionic reactivity of titanacyclo-
propanes was applied for the first time to the synthesis of
1,4-dicarbonyl compounds by proceeding through a double
nucleophilic addition. From a synthetic point of view, this
method represents efficient access to g-diketones from
carboxylic acids, via the formation of acyl cyanohydrins.
Additional work is underway to better understand the
mechanism and expand the scope of this reaction.
[10] a) P. Bertus, J. Szymoniak, Chem. Commun. 2001, 1792 – 1793;
b) P. Bertus, J. Szymoniak, Synlett 2007, 1346 – 1356.
[11] J. Lee, Y. G. Kim, J. G. Bae, J. K. Cha, J. Org. Chem. 1996, 61,
4878 – 4879.
[12] A sequential addition of two different electrophiles has been
performed with imides, acylpyrroles, and CO2 as first electro-
philes, but only a very few electrophiles (H2O, NBS, I2, O2) react
Experimental Section
EtMgBr (2.1 mmol, 3m in Et2O) was added dropwise to a solution of
the cyanoester 3 (1 mmol) and Ti(OiPr)4 (0.33 mL, 1.1 mmol) in Et2O
(5 mL) that was maintained at 08C under argon. After the addition of
the Grignard reagent, the mixture was warmed to RT and stirred for
1 h. The turbid yellow mixture was quenched with water (1 mL), and
À
with the remaining Ti C bond; see: a) J. Lee, J. D. Ha, J. K. Cha,
J. Am. Chem. Soc. 1997, 119, 8127 – 8128; b) O. G. Epstein, J. M.
Angew. Chem. Int. Ed. 2010, 49, 8691 –8694
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 8693

Communications
Seo, N. Masalov, J. K. Cha, Org. Lett. 2005, 7, 2105 – 2108; c) F.
Cadoret, P. Retailleau, Y. Six, Tetrahedron Lett. 2006, 47, 7749 –
7753.
[16] The nature of the metals linked to the intermediates D and E
(Scheme 4) is unknown. The representation given here is only a
possibility.
[13] a) P. Bertus, J. Szymoniak, Synlett 2003, 265 – 267; b) C. Laroche,
D. Harakat, P. Bertus, J. Szymoniak, Org. Biomol. Chem. 2005, 3,
3482 – 3487.
[17] a) O. G. Kulinkovich, Pure Appl. Chem. 2000, 72, 1715 – 1719;
b) D. G. Kananovich, O. G. Kulinkovich, Tetrahedron 2008, 64,
1536 – 1547.
[14] The relative position of the substituent was determined by an
HMBC NMR experiment. See the Supporting Information.
[15] For a review dealing with the regioselectivity on titanium-
mediated coupling reactions, see: H. A. Reichard, M. McLaugh-
lin, M. Z. Chen, G. C. Micalizio, Eur. J. Org. Chem. 2010, 391 –
409.
[18] D. Astashko, H. G. Lee, D. N. Bobrov, J. K. Cha, J. Org. Chem.
2009, 74, 5528 – 5532.
[19] A related EtMgBr-mediated ring-opening of azazirconacycles
has been postulated in the [Cp2ZrCl2]-catalyzed ethylmagnesa-
tion of imines, see: V. Gandon, P. Bertus, J. Szymoniak, Eur. J.
Org. Chem. 2001, 3677 – 3681.
8694 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 8691 –8694