Synthesis of and asymmetric cycloaddition with chiral diiron acyl complexes

Full text of DOI:10.1021/ja964195f

J. Am. Chem. Soc. 1997, 119, 3399-3400
3399
Synthesis of and Asymmetric Cycloaddition with
Chiral Diiron Acyl Complexes
Scott R. Gilbertson* and Omar D. Lopez
Department of Chemistry
Washington UniVersity
St. Louis, Missouri 63130
ReceiVed December 5, 1996
The use of stoichiometric transition metal reagents in organic
synthesis has seen a considerable amount of growth in the
last 20 years. These reagents bring a number of useful and
unique characteristics to organic synthesis.1,2 Such reagents
often allow access to reaction manifolds that are not available
to standard organic reagents. There are also examples of
transition metal compounds exhibiting increased reactivity due
to their electronic nature. In other cases, stoichiometric orga-
nometallic reagents bring structures to organic synthesis that
have potential in directing the stereochemical outcome of reac-
tions. Of these unique structure types, there are few asymmetric
transition metal complexes where the chirality resides at the
metal.3-5 The synthesis of these complexes is often compli-
cated, involving resolution of diastereomers. This paper reports
the first synthesis of asymmetric diiron acyl complexes6-9 in
an optically active form. These complexes exhibit enhanced
reactivity and unique selectivity in the dipolar cycloaddition of
nitrones, giving optically active 4-substituted isoxazolidenes.
The cycloaddition of nitrones with R,â-unsaturated ester
analogs,10-15 and other carbonyl compounds,16-20 in high
diastereoselectivity has been an ongoing problem. The chiral
diiron reagents reported in this paper perform this reaction with
high selectivity providing access to optically active 4-isox-
azolidenes of either enantiomer.
Figure 1.
hope was that the chirality in the thiol would induce selective
addition of the acid chloride to one of the two diastereotopic
iron atoms in the anionic diiron intermediate (2, Figure 1). Given
the distance the thiol-based chiral center was from the two iron
atoms it must discriminate between, we felt this was an
ambitious goal. The acid chloride could also add to the complex
from the side away from the bridging thiol. It was difficult to
see how such an addition could be controlled by a chiral thiol
attached to the opposite side of the complex. Despite these
potential problems, this approach was taken. If reaction of the
acid chloride and the diiron anion failed to be selective, we
would still have diastereomeric complexes and the potential to
separate them. The first thiol used was the thiol from N-Boc-
protected prolinol (1).21 This thiol was chosen because of the
rigid structure of the five-membered ring and the large steric
volume of the Boc group attached to the nitrogen. Reaction of
triiron dodecacarbonyl with this thiol and triethylamine in THF
gave the characteristic yellow solution (2). Addition of crotonyl
or acryloyl chloride to this solution, followed by overnight
stirring, gave the desired acyl complexes; crotonyl in 65% yield
and acrylate in 45% yield.
Our inital approach to the synthesis of asymmetric diiron
complexes was to take a thiol containing a chiral center and
carry out the standard procedure developed by Seyferth.6-9 The
Because of hindered rotation of the Boc-protected proline,
the NMR line shape of the product complexes was broad,
making it difficult to assess the ratio of diastereomers. On the
basis of analysis of the 13C NMR resonances for the tert-butyl
of the Boc group, the ratio of the major product to the sum of
the two minor products was approximately 10:1. Along with
the potential for different complexes due to rotational isomers,
there are two other types of diastereomers possible with these
diiron complexes. The acid chloride could have added to either
of the two iron atoms, giving diastereomers that differ in the
sense of chirality between the iron atoms and the chiral carbon
from the thiol. This was the selectivity we set out to control.
The other type of diastereomeric relationship possible is between
the sulfur atom and the other chiral centers in the molecule.
We have seen this type of isomer with these complexes before
and have shown that this relationship does not have a profound
effect on the reactivity of these complexes.15,22
(1) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles
and Applications of Organotransition Metal Chemistry; University Science
Books: Mill Valley, CA, 1987.
(2) AdVances in Metal-Organic Chemistry; Liebeskind, L. S., Ed.; JAI
Press Inc.: Greenwich, CT, 1989-96; Vol. 1-5.
(3) Davies, S. G. Chem. Brit. 1989, 268.
(4) Davies, S. G. Aldrichim. Acta 1990, 31.
(5) Davies, S. G.; Ichihara, O. J. Chem. Soc., Chem. Commun. 1990,
1554.
(6) Seyferth, D.; Womack, G. B.; Dewan, J. C. Organometallics 1985,
4, 398.
(7) Seyferth, D.; Womack, G. B.; Archer, C. M.; Dewan, J. C.
Organometallics 1989, 8, 430.
(8) Seyferth, D.; Archer, C. M.; Ruschke, D. P. Organometallics 1991,
10, 3363.
(9) Seyferth, D.; Brewer, K. S.; Wood, T. G. Organometallics 1992, 11,
2570.
(10) Gothelf, K. V.; Jørgensen, K. A. J. Org. Chem. 1995, 60, 6847.
(11) Gothelf, K. V.; Hazell, R. G.; Jørgensen, K. A. J. Org. Chem. 1996,
61, 346.
(12) Gothelf, K. V.; Thomsen, I.; Jørgensen, K. A. J. Am. Chem. Soc.
1996, 118, 59.
With all of these diastereomers possible, we were both
surprised and pleased with the results obtained upon reaction
with a selection of nitrones (Table 1).23-26 We found that
reaction of the mixture of diastereomeric complexes with
nitrones gave the expected cycloadducts with high selectivity.
After cycloaddition, the cycloadduct was liberated from the iron
(13) Murahashi, S.-I.; Imada, Y.; Kohno, M.; Kawakami, T. Synlett 1993,
395.
(14) Olsson, T.; Stern, K.; Westman, G.; Sundell, S. Tetrahedron 1990,
46, 2473.
(15) Gilbertson, S. R.; Dawson, D. P.; Lopez, O. D.; Marshall, K. L. J.
Am. Chem. Soc. 1995, 117, 4431.
(16) Rispens, M. T.; Keller, E.; de Lange, B.; Zijlstra, R. W. J.; Feringa,
B. L. Tetrahedron: Asymmetry 1994, 5, 607.
(17) Langlois, N.; Van Bac, N.; Dahuron, N.; Delcroix, J.-M.; Deyine,
A.; Griffart-Brunet, D.; Chiaroni, A.; Riche, C. Tetrahedron 1995, 51,
3571.
(21) Volante, R. P. Tetrahedron Lett. 1981, 22, 3119.
(22) Gilbertson, S. R.; Zhao, X.; Dawson, D. P.; Marshall, K. L. J. Am.
Chem. Soc. 1993, 115, 8517.
(23) Huisgen, R.; Seidl, H.; Bruning, I. Chem. Ber. 1969, 102, 1102.
(24) Black, D. S.; Crozier, R. F.; Davis, V. C. Synthesis 1975, 205.
(25) Tufariello, J. J. Acc. Chem. Res. 1979, 12, 396.
(26) Keirs, D.; Moffat, D.; Overton, K.; Tomanek, R. J. Chem. Soc.,
Perkin Trans. 1 1991, 1041.
(18) Panfil, I.; Belzecki, C.; Urbanczyk-Lipkowska, Z.; Chmielewski,
M. Tetrahedron 1991, 47, 10087.
(19) Saito, S.; Ishikawa, T.; Moriwake, T. Synlett 1994, 279.
(20) Kanemasa, S.; Uemura, T.; Wada, E. Tetrahedron Lett. 1992, 33,
7889.
S0002-7863(96)04195-9 CCC: $14.00 © 1997 American Chemical Society

3400 J. Am. Chem. Soc., Vol. 119, No. 14, 1997
Communications to the Editor
Table 1. Nitrone Additions to Chiral Diiron Acyl Complexes
Figure 2.
ratio may have been incorrectly determined.27 For this reason
we chose to introduce the chiral center of an asymmetric oxa-
zolidinone. This places a well-defined chiral group closer to
the other chiral centers in the molecule and gives us further
evidence that the reported product ratios are correct and not
due to an inability to resolve diastereomeric compounds in the
NMR.
This correlation now makes it possible to know the correct
enantiomer of proline to use to obtain the desired configuration
of the three chiral centers formed in this reaction. We are
currently attempting to determine the absolute configuration of
the starting diiron complex. This knowledge will allow us to
determine whether the diiron complexes react in an s-cis or an
s-trans configuration. While this information will increase our
understanding of these reagents, it is not critical for the use of
this reaction, since we know which absolute configuration of
the starting proline provides a given absolute configuration of
the product. As a final proof, we synthesized the diiron complex
using the thiol from D-proline. Reaction with N-methyl-C-
phenyl nitrone gave, after oxidaton, the opposite enantiomer in
good yield and high selectivity.
a The R,â-unsaturated complexes exist as two isomers at sulfur. The
cycloaddition reactions are run in the mixture of isomers. b The endo/
exo ratios were determined by NMR after conversion to the thioester.
by oxidation with Ce(IV).15,22 Oxidative removal of the metal
not only converts the organometallic species to the synthetically
useful thioester moiety but also removes one of the potential
centers of chirality. This reduces the number of possible
diastereomers to 2 and allows for accurate determination of the
product ratio.
In each case, the cycloaddition of nitrones was run on the
mixture of complexes (Table 1). In every case, oxidative
removal of the transition metal showed the cycloaddition to
proceed with high selectivity and in good yield. Reaction with
either the crotonyl (3) or acrylate (4) derived complex gave
useful results, with the crotonyl-derived complexes reacting with
higher selectivity. A variety of nitrones were examined. The
reaction proceeded in good yield and high selectivity with
nitrones that have N-aryl or -alkyl groups. The reaction of
C-carbethoxy-N-benzyl nitrone with the crotonyl complex also
proceeded in good yield and with high selectivity.
We are currently working on the development of this chem-
istry as a route to â-amino acids and â-lactams. Along with
that, work is being done to determine the structure of the diiron
anion as well as the diiron acyl complexes. Through this
information, it may be possible to determine the origin of both
the selectivity observed in the synthesis of the acyl complexes
and their selectivity in 1,3-dipolar cycloadditions.
While it was quickly clear that the cycloadditions of nitrones
with these complexes proceeded with good selectivity, the
absolute sense of this selectivity remained to be determined.
This was accomplished by correlation of the products of two
different cycloadditions with known compounds (Figure 2).
Thioesters 5 and 6 were converted to know isoxazolidines 9
and 10, respectively.10-12 Hydrolysis of thioesters 5 and 6 gave
the respective acids (7 and 8) which were then converted to the
isoxazolidines by standard chemistry. These molecules were
found to have the same optical rotations as those reported for
9 and 10.11 The transformation of 6 to 10 was also used as a
check that our determination of the ratio of diastereomers formed
in the reaction sequence was correct. The ratios in Table 1
Acknowledgment. This work was supported by the National
Science Foundation (Grant No. CHE-9316821) and Washington
University. We also gratefully acknowledge the Washington University
High-Resolution NMR Facility, partially supported by NIH RR02004,
RR05018, and RR07155, and the Washington University Mass
Spectrometry Resource Center, partially supported by NIHRR00954,
for their assistance.
Supporting Information Available: Experimental details and
spectroscopic data (9 pages). See any current masthead page for
ordering and Internet access instructions.
JA964195F
1
were determined by analysis of H and 13C NMR spectra for
diastereomeric thioesters. This approach left the possibility that
if two different diastereomers had similar NMR spectra their
(27) The ratio of greater than 25:1 represents none of the minor isomer
observed in the proton NMR of the crude reaction mixture.