Magnesium Halide-Catalyzed Anti-Aldol Reactions of Chiral N-Acylthiazolidinethiones

Article abstract of DOI:10.1021/ol025553o

matrix presented Diastereoselective direct aldol reactions of chiral N-acylthiazolidinethiones occur in high yield with preference for the illustrated anti diastereomer. This reaction is catalyzed by 10% MgBr₂-OEt₂ in the presence of

Full text of DOI:10.1021/ol025553o

ORGANIC
LETTERS
2002
Vol. 4, No. 7
1127-1130
Magnesium Halide-Catalyzed Anti-Aldol
Reactions of Chiral
N-Acylthiazolidinethiones
David A. Evans,* C. Wade Downey, Jared T. Shaw, and Jason S. Tedrow
Department of Chemistry & Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
eVans@chemistry.harVard.edu
Received January 2, 2002
ABSTRACT
Diastereoselective direct aldol reactions of chiral N-acylthiazolidinethiones occur in high yield with preference for the illustrated anti diastereomer.
This reaction is catalyzed by 10% MgBr2‚OEt2 in the presence of triethylamine and chlorotrimethylsilane. Yields range from 56 to 93% with
diastereoselectivity up to 19:1 for a variety of N-acylthiazolidinethiones and unsaturated aldehydes.
Ongoing efforts to further refine the aldol process have
resulted in significant recent advances from a number of
research groups.1 In this context, we recently demonstrated
that substoichiometric amounts of magnesium halides, in the
presence of an amine base2 and chlorotrimethylsilane
(TMSCl), catalyze the direct aldol reaction of chiral N-
acyloxazolidinones with high anti diastereoselectivity (eq 1).3
This procedure substantially reduces the cost of the enoliza-
tion process while affording access to a new auxiliary-based
anti aldol process. In this Letter, we describe the extension
of this methodology to chiral N-acylthiazolidinethiones (eq
2).4 The realization of this goal, combined with our stoi-
chiometric titanium- and boron-mediated aldol processes (eq
3),5 now provides access to both syn and anti propionate
aldol products of both absolute configurations.
(1) (a) Yoshikawa, N.; Kumagai, N.; Matsunaga, S.; Moll, G.; Ohshima,
T.; Suzuki, T.; Shibasaki, M. J. Am. Chem. Soc. 2001, 123, 2466-2467
and references therein. (b) Trost, B. M.; Ito, H.; Silcoff, E. R. J. Am. Chem.
Soc. 2001, 123, 3367-3368 and references therein. (c) Trost, B. M.; Silcoff,
E. R.; Ito, H. Org. Lett. 2001, 3, 2497-2500.
Initial optimization studies focused on the Mg(II) source
(Table 1). Aldol reactions with thiazolidinethione 1a, cin-
namaldehyde, triethylamine, and chlorotrimethylsilane in
EtOAc identified MgCl2 and MgBr2‚OEt2 as likely metal
(2) For a related soft-enolization procedure, see: Tirpak, R. E.; Olsen,
R. S.; Rathke, M. W. J. Org. Chem. 1985, 50, 4877-4879 and references
therein.
(3) Evans, D. A.; Tedrow, J. S., Shaw, J. T.; Downey, C. W. J. Am.
Chem. Soc. 2002, 124, 392-393.
(4) (a) Crimmins, M. T.; King, B. W.; Tabet, E. A.; Chaudhary, K. J.
Org. Chem. 2001, 66, 894-902 and references therein. (b) Fujita, E.; Nagao,
Y. AdV. Heterocycl. Chem. 1989, 49, 1-36. (c) Nagao, Y.; Yamada, S.;
Kumagai, T.; Ochiai, M.; Fujita, E. J. Chem. Soc., Chem. Commun. 1985,
1418-1420.
(5) (a) Evans, D. A.; Bartroli, J.; Shih, T. L.; J. Am. Chem. Soc. 1981,
103, 2127-2129. (b) Evans, D. A., Rieger, D. L., Bilodeau, M. T., Urpi,
F. J. Am. Chem. Soc. 1991, 113, 1047-1049. (c) Evans, D. A., Nelson, J.
V.; Taber, T. Top. Stereochem. 1982, 13, 1-115. (d) Gage, J.; Evans, D.
A. Org. Synth. 1990, 68, 83-91.
10.1021/ol025553o CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/14/2002

Table 1. Magnesium Sources for the Enolization Process
Table 2. Addition of 1a to Aldehydes (eq 2)
entry
MgX2
MgCl2
MgCl2
MgCl2
MgBr2‚OEt2
MgBr2‚OEt2
MgBr2‚OEt2
Mg(NTf2)2
Mg(OTf)2
Mg(ClO4)2
concn (1a )
convn (%)b
dr (%)b,c
entry
R
adduct
dr (%)b,c
yield (%)d
1
2
3
4
5
6
7
8
9
0.2 M
0.4 M
0.2 M
0.2 M
0.4 M
1.0 M
0.2 M
0.2 M
0.2 M
94
94
92
91
99
94
87
22
64
9:1
10:1
10:1
8:1
10:1
13:1
5:1
1
2
3
4
5
6
7
Ph
2a e
2bf
2cf
2d e
2ee
2fe
19:1
19:1
19:1
10:1
19:1
7:1
85
92
91
87
90
84
56
4-MeC6H4
4-MeOC6H4
CHdCHPh
C(CH3)dCHPh
2-naphthyl
C(CH3)dCH2
2gf
7:1
12:1
4:1
a (1) MgBr2‚OEt2 (10 mol %), RCHO (1.1 equiv), Et3N (2 equiv), TMSCl
(1.5 equiv), 0.4 M in EtOAc, 24 h; (2) 5:1 THF/1.0 N HCl. b Determined
by 500 MHz 1H NMR of the unpurified, silylated reaction mixture. c dr )
diastereomeric ratio ) (desired anti):(Σ other isomers). d Isolated yield of
major diastereomer. e Absolute configuration assigned by chemical cor-
relation to known material. f Absolute configuration assigned by X-ray
crystallography.
a (1) MgX2 (10 mol %), RCHO (1.1 equiv), Et3N (2 equiv), TMSCl
(1.1 equiv), EtOAc, 24 h; (2) 5:1 THF/1.0 N HCl. b Determined by 500
MHz 1H NMR of the unpurified, silylated reaction mixture. c dr )
diastereomeric ratio ) (desired anti):(Σ other isomers).
halide candidates. Other magnesium salts afforded aldol
adducts in lower conversion or selectivity (diastereomer ratio
(dr) ) (desired anti):(Σ other isomers)). Upon varying the
concentration of 1a, MgBr2‚OEt2 emerged as the preferred
catalyst.6
Examination of a number of trialkylamines clearly indi-
cated triethylamine as the base of choice. While Hunig’s base
effected a selective reaction (7:1 dr), conversion declined
(37%).7 An increase in the amount of TMSCl (1.5 equiv)
led to high conversion with no loss of diastereoselectivity.8
We speculate that TMSCl may also be a scavenger of trace
amounts of water in the solvent,9 which was used without
purification.
(entries 4, 5, and 7). The major diastereomers were easily
purified by chromatography. Recrystallization was possible
for adducts 2b, 2c, and 2g, allowing proof of absolute
stereochemistry by X-ray crystallography. Each adduct, when
cleaved from the auxiliary, was established to be antipodal
to its corresponding oxazolidinone-derived adduct (eq 1).3
We next extended the reaction scope to thiazolidinethiones
1b-e, synthesized in analogy to literature precedent.4a,11 Each
N-acylthiazolidinethione reacted smoothly with benzaldehyde
in high yield and selectivity (Table 3). Some cases required
higher catalyst loading to effect full conversion (entries 2
and 5).
In the N-acyloxazolidinone system (eq 1), addition of
substoichiometric amounts of NaSbF6 sometimes afforded
aldol adducts in higher conversion and selectivity.3 Interest-
ingly, this halide scavenger had little positive effect on the
thiazolidinethione reaction; indeed, an increase in additive
loading led to decreased selectivity and conversion.10
Addition of 1a to various aldehydes was accomplished
under the optimized reaction conditions (Table 2). Aromatic
aldehydes (entries 1-3 and 6) afforded products with
diastereoselectivity as high as 19:1 in excellent yield. Other
unsaturated aldehydes also serve as viable substrates, al-
though with diminished yield in the case of methacrolein
Table 3. Addition of Representative N-Acylthiones
entry
R
adduct
dr (%)b,c
yield (%)d
1
2
3
4
5
CH3
CH2CH3
CH2CCH2CH2
CH2Ph
CH2CH(CH3)2
2a e
3bf
3cf
3d h
3eh
19:1
10:1
13:1
8:1
85
88g
91
84
93i
(6) Low conversion at high concentration may be due to poor solubility
of some intermediates. White precipitate forms upon addition of amine base
and persists throughout the reaction.
19:1
a (1) MgBr2‚OEt2 (10 mol %), PhCHO (1.1 equiv), Et3N (2 equiv),
TMSCl (1.5 equiv), 0.4 M in EtOAc, 24 h; (2) 5:1 THF/1.0 N HCl.
b Determined by 500 MHz 1H NMR of the unpurified, silylated reaction
mixture. c dr ) diastereomeric ratio ) (desired anti):(Σ other isomers).
d Isolated yield of major diastereomer. e Absolute configuration assigned
by chemical correlation to known material. f Absolute configuration assigned
by X-ray crystallography. g 20 mol % of MgBr2‚OEt2 was necessary to
induce full conversion. h Absolute configuration assigned by analogy. i 15
mol % of MgBr2‚OEt2 was necessary to induce full conversion.
(7) Other bases (2,6-lutidine, N-methylmorpholine, imidazole) were
inferior in terms of conversion or selectivity.
(8) Other silylating agents were less reactive (TESCl, conv ) 0%) or
selective (TMSBr, anti/syn ) 4:1).
(9) Ethyl acetate, isopropyl acetate, dichloromethane, and acetonitrile
were effective solvents in terms of conversion and selectivity. Ethyl acetate
was selected for further studies.
(10) Addition of 13 mol % of NaSbF6 to the reaction of 1a and
cinnamaldehyde: 90% y, dr ) 90%. 25 mol % of NaSbF6: 80% conv, dr
) 82%. 46 mol % of NaSbF6: 57% conv, dr ) 83%.
1128
Org. Lett., Vol. 4, No. 7, 2002

Scheme 1. Proposed Catalytic Cycle
Table 4. Isomerization Experiments
A proposed catalytic cycle is outlined in Scheme 1.
Thiazolidinethione-magnesium complex A reacts with tri-
ethylamine, yielding magnesium enolate B, which adds
reversibly3 to the aldehyde, forming the magnesium aldolate
C. Chlorotrimethylsilane then irreversibly traps this aldolate,
which is subsequently displaced from the metal center by
another molecule of N-acylthiazolidinethione.12 Because
enolate silylation is not obserVed under the reaction condi-
tions, a Mukaiyama-type pathway can be reasonably ex-
cluded.
a (1) MgBr2‚OEt2 (10 mol %), Et3N (2 equiv), TMSCl (1.5 equiv), 0.4
M in EtOAc, 24 h. b Determined by 500 MHz 1H NMR of the unpurified,
silylated reaction mixture. dr ) diastereomeric ratio ) (desired anti):(Σ
other isomers). c Anti/syn ) 77:23.
aldol cleavage might be competitive with this aldol dia-
stereomer. In conclusion, while these experiments document
the reversibility of some of the steps, they by no means
establish that reaction diastereoselection is operating under
thermodynamic control.
To test the reversibility of the carbon-carbon bond-
forming step (B f C), a number of experiments were
performed (Table 4). Diastereomerically pure (dr > 100:1)
desilylated aldol adduct 2e was subjected to the optimized
The mechanistic intricacies of these reactions are note-
worthy. This reaction presents two stereochemical issues:
enolate diastereoface selectivity and anti aldol diastereo-
selection. On the basis of the weight of circumstantial
evidence, all enolization procedures to date to form boron,
titanium, lithium, or sodium enolates with this family of
imides implicate the intervention of (Z) metal enolates. Given
the assumption that this is the geometry of the intervening
enolate, the enolate face selectivity observed for the N-
acyloxazolidinone-derived magnesium enolate is fully con-
sistent with a chelate-controlled process (eq 4) as is observed
in analogous metal enolate alkylation and Michael addition
reactions.14 In contrast, the N-acylthiazolidinethione-deriVed
magnesium enolates exhibit the opposite face selection in
these reactions. This observation negates the possibility that
the thione CdS moiety is coordinating to the Mg center
during the aldol transition state. Rather, enolate face selection
derived from the minimization of CdOTCdS dipole
interactions provides the stereocontrol on the enolate reaction
partner (eq 5) as is observed for the analogous boron aldol
reactions.5a
1
reaction conditions (entry 1). After 24 h, H NMR analysis
showed that the minor anti diastereomer constituted 9% of
the silylated product mixture. This result is consistent with
a rapid retro-aldol cleavage of magnesium aldolate C,
regenerating the aldehyde and magnesium enolate B. Selec-
tive reformation of C, followed by silylation, would produce
a mixture of aldol adducts as observed. Accordingly, the rate
of retro-aldol cleaVage appears to be faster than magnesium
aldolate silylation.
This conclusion is supported by the analogous equilibration
experiment carried out with the minor anti aldol diastereomer
5. A diasteromerically pure (dr > 100:1) sample of 5 was
isomerized to a 7:1 mixture of 2e:5 (entry 2). Even the
independently synthesized4a,b syn adduct 6 could be converted
to 4 in good yield (entry 3).13 Finally, syn aldol adduct 7
did not interconvert cleanly to the principal anti adduct 4
(entry 4), suggesting that the rates of silylation and retro-
(11) Delaunay, D.; Toupet, L.; Le Corre, M. J. Org. Chem. 1995, 60,
6604-6607.
(12) Silylated adducts are configurationally stable under the reaction
conditions.
(13) Further experiments to eliminate epimerization as a source of
isomerization have been conducted and will be reported in due course.
(14) (a) Evans, D. A.; Ennis, M. D.; Mathre, D. J. J. Am. Chem. Soc.
1982, 104, 1737-1739. (b) Evans, D. A.; Nelson, S. G. J. Am. Chem. Soc.
1997, 119, 6452-6453.
Org. Lett., Vol. 4, No. 7, 2002
1129

Scheme 2. PM3 Calculations Suggest the Invervention of a Boat Transition State
Since these aldol addition reactions afford the unexpected
adducts are complementary to adducts derived from N-
acyloxazolidinones. Investigations continue into the mech-
anism, diastereoselectivity, and potential enantioselectivity
of this reaction.
anti aldol product from the (Z) metal enolate, the intervention
of a chair Zimmerman-Traxler transition state is precluded.
On this basis we wish to suggest that these reactions are
proceeding via either five- or six-coordinate Mg coordination
geometries where the boat-metal transition structures such
as E-H (Scheme 2) are reasonable.15 Semiempirical calcula-
tions (PM3) support the premise that a twist-boat conforma-
tion should be favored for the magnesium aldol transition
state when the length of the forming carbon-carbon bond
is constrained to 2.0 Å.16 Chair F is 2.5 kcal/mol higher in
energy than boat E; chair H is 2.8 kcal/mol higher in energy
than boat G.17 Thus, judicious auxiliary choice controls the
enolate face-selectivity of the magnesium aldol, allowing
access to either anti diastereomer.
Acknowledgment. Support has been provided by the
NSF. The NIH (GM-33327-16) is acknowledged for post-
doctoral fellowship support (J.T.S). The NIH Shared Instru-
mentation Grant Program (1-S10-RR04870) and the NSF
(CHE 88-14019) are acknowledged for providing NMR
facilities.
Supporting Information Available: Experimental pro-
cedures for all new compounds as well as calculation
procedures and X-ray crystallographic data for 2b, 2c, 2g,
3b, and 3c. This material is available free of charge via the
Internet at http://pubs.acs.org.
In conclusion, MgBr2‚OEt2 selectively catalyzes the direct
aldol of N-acylthiazolidinethiones. The anti propionate
OL025553O
(15) For the sake of space, we have not presented the analogous five-
coordinate Mg transition states. We speculate that the O-Mg-OdC bond
angle in these complexes will also be ∼90° with the OdC (RCHO) ligand
oriented in the equatorial plane and the O-enolate ligand in axial position
to optimize the electronic effects conferred on the reaction partners by the
metal center.
(17) Transition-structure optimizations were performed at the PM3 level
using the Spartan Semiempirical Program 5.0 (Wavefunction Inc., Irvine,
CA 92715) on a Silicon Graphics Impact 10000 (195 MHz, 128 M RAM)
running IRIX 6.2. Calculations were performed without solvent using these
parameters: optcycle ) 2000, maxcycle ) 2000, charge ) 0, multiplicity
) 0. Calculations converged (energy difference between cycles <0.0005
kcal/mol) in e 4 h CPU time.
(16) Li, Y.; Paddon-Row: M. N.; Houk, K. N. J. Org. Chem. 1990, 55,
481-493.
1130
Org. Lett., Vol. 4, No. 7, 2002