Highly stereoselective TiCl4-mediated aldol reactions from (S)-2-benzyloxy-3-pentanone

Article abstract of DOI:10.1021/jo071048z

(Chemical Equation Presented) Stereoselectivity of TiCl₄- mediated aldol reactions from (S)-2-benzyloxy-3-pentanone is dramatically improved when the reaction is carried out in the presence of 1.1 equiv of tetrahydrofuran (THF) or 1,2-dimefhoxy

Full text of DOI:10.1021/jo071048z

Highly Stereoselective TiCl4-Mediated Aldol
Reactions from (S)-2-Benzyloxy-3-pentanone
5), presumably due to a chelated transition state resulting from
the abstraction of a chloride ion by the second equivalent of
Lewis acid. In the first case, sparteine can be replaced by a
3
combination of 1.1 equiv of i-Pr2NEt and 1.0 equiv of an achiral
additive as N-methyl-2-pyrrolidinone (NMP), which suggests
that the second equivalent of that base simply operates as a
Victor Rodr ´ı guez-Cisterna, Cristina Villar,
Pedro Romea,* and F e` lix Urp ´ı *
4,5
Departament de Qu ´ı mica Org a` nica, UniVersitat de Barcelona,
Mart ´ı i Franqu e´ s 1-11, 08028 Barcelona, Catalonia, Spain
ligand for titanium. Furthermore, Ghosh reported the dramatic
enhancement on the stereoselectivity of the titanium-mediated
chloroacetate aldol reactions from indan-derived chiral auxil-
iaries produced by additives as acetonitrile, NMP, or triph-
felix.urpi@ub.edu; pedro.romea@ub.edu
6
enylphosphine. Irrespective of the improvement on the stere-
oselectivity observed for these processes, the manner in which
such additives interact with the titanium center is still unclear
and, as a consequence, the intermediates or the transition states
affecting titanium enolates are mainly proposed to account for
the experimental results.
ReceiVed May 23, 2007
In this context, we have observed that the stereochemical
outcome of the titanium-mediated aldol reaction of chiral
R-benzyloxy ethyl ketone 1 (Scheme 1) relies on the Lewis acid
7
,8
engaged on the process. Indeed, the same experimental
procedure using TiCl4 or Ti(i-PrO)Cl3 leads to different syn-
aldol adducts, so that the substitution of a chloride by an
isopropoxide ligand on the Lewis acid determines the config-
uration of the resulting aldol. Unfortunately, the diastereose-
lectivity of both processes is rather different. The aldol reaction
based on Ti(i-PrO)Cl3 affords 2,4-anti-4,5-syn-aldols in high
diastereomeric ratios, but the TiCl4-mediated counterpart fur-
nishes 2,4-syn-4,5-syn-aldol adducts in an impractical manner
Stereoselectivity of TiCl
4
-mediated aldol reactions from
(S)-2-benzyloxy-3-pentanone is dramatically improved when
the reaction is carried out in the presence of 1.1 equiv of
tetrahydrofuran (THF) or 1,2-dimethoxyethane (DME). The
resultant 2,4-syn-4,5-syn adducts are then obtained in
diastereomeric ratios up to 97:3, which proves that the
appropriate choice of the Lewis acid (TiCl
4
-THF or DME
vs Ti(i-PrO)Cl ) engaged in the process permits access to
3
7a,9
(Scheme 1).
Thus, considering that it would be highly
both syn-aldol adducts.
advantageous to get a better stereocontrol on the latter reaction,
we envisaged that the presence in the reaction mixture of
additives able to act as Lewis bases and bind to metal center
might have an influence on the diastereoselectivity of the
process. Herein, we disclose our studies on the impact of achiral
nonionic additives on the TiCl4-mediated aldol reactions from
Titanium(IV) enolates have been increasingly employed in
C-C bond-forming processes since Evans reported that they
can be directly prepared from the corresponding carbonyl
1
precursors with a titanium(IV) Lewis acid and a tertiary amine.
10
ketone 1 that eventually permits direct access to aldol adducts
Particularly important in the aldol arena, titanium enolates
participate in such a wide scope of highly stereoselective
reactions that they constitute nowadays an appealing alternative
2
and 3 in a highly stereocontrolled manner.
2
(3) (a) Crimmins, M. T.; King, B. W.; Tabet, E. A. J. Am. Chem. Soc.
to the more classic boron and lithium counterparts. Unfortu-
1
997, 119, 7883-7884. (b) Crimmins, M. T.; King, B. W.; Tabet, E. A.;
nately, there is a lack of information concerning the aggregation
state and the distribution of ligands around the metal, which
hampers the development of new methodologies.
Regarding these elusive issues, Evans speculated that the
titanium enolates from chiral N-propanoyl-1,3-oxazolidin-2-ones
Chaudhary, K. J. Org. Chem. 2001, 66, 894-902.
(4) Crimmins, M. T.; She, J. Synlett 2004, 1371-1374.
(5) Toru has also reported the effect of bases and phosphorous additives
on the titanium mediated aldol reactions of R-seleno ketones and esters:
Nakamura, S.; Hayakawa, T.; Nishi, T.; Watanabe, Y.; Toru, T. Tetrahedron
2001, 57, 6703-6711.
(6) (a) Ghosh, A. K.; Kim, J.-H. Org. Lett. 2004, 6, 2725-2728 and
references therein. (b) For the influence of triphenylphosphine on titanium-
mediated Mukaiyama-like aldol reactions, see: Palazzi, C.; Colombo, L.;
Gennari, C. Tetrahedron Lett. 1986, 27, 1735-1738.
1a
might be considered as six coordinated Z-chelated ate complexes.
Crimmins took advantage of this model and established that
the stereochemical outcome of the titanium-mediated aldol
reactions from related 1,3-oxazolidine-2-thione auxiliaries de-
pends on the base and the equivalents of TiCl4 involved: Evans
syn-aldol adducts are obtained with an excellent diastereose-
lectivity (dr >97:3) when 2.5 equiv of sparteine are used through
a putative nonchelated transition state, while reactions with 2
equiv of TiCl4 give non-Evans syn-aldol diastereomers (dr >95:
(7) (a) Solsona, J. G.; Romea, P.; Urp ´ı , F.; Vilarrasa, J. Org. Lett. 2003,
5
4
2
, 519-522. (b) Solsona, J. G.; Romea, P.; Urp ´ı , F. Tetrahedron Lett. 2004,
5, 5379-5382. (c) Solsona, J. G.; Nebot, J.; Romea, P.; Urp ´ı , F. Synlett
004, 2127-2130.
(8) For parallel studies on â-hydroxy ketones, see: (a) Solsona, J. G.;
Nebot, J.; Romea, P.; Urp ´ı , F. J. Org. Chem. 2005, 70, 6533-6536. (b)
Ward, D. E.; Beye, G. E.; Sales, M.; Alarcon, I. Q.; Gillis, H. M.; Jheengut,
V. J. Org. Chem. 2007, 72, 1667-1674.
(9) Moreover, the experimental procedure must be followed very carefully
(
1) (a) Evans, D. A.; Urp ´ı , F.; Somers, T. C.; Clark, J. S.; Bilodeau, M.
because small changes in the amounts of the reagents can modify the
diastereoselectivity of the process.
(10) (a) Mart ´ı n, R.; Romea, P.; Tey, C.; Urp ´ı , F.; Vilarrasa, J. Synlett
1997, 1414-1416. (b) Ferrer o´ , M.; Galobardes, M.; Mart ´ı n, R.; Montes,
T.; Romea, P.; Rovira, R.; Urp ´ı , F.; Vilarrasa, J. Synthesis 2000, 1608-
1614.
T. J. Am. Chem. Soc. 1990, 112, 8215-8216. (b) Evans, D. A.; Rieger, D.
L.; Bilodeau, M. T.; Urp ´ı , F. J. Am. Chem. Soc. 1991, 113, 1047-1049.
(2) (a) Ghosh, A. K.; Shevlin, M. In Modern Aldol Reactions; Mahrwald,
R., Ed.; Wiley-VCH: Weinheim, 2004; Vol. 1, pp 63-125. (b) Schetter,
B.; Mahrwald, R. Angew. Chem., Int. Ed. 2006, 45, 7506-7525.
1
0.1021/jo071048z CCC: $37.00 © 2007 American Chemical Society
Published on Web 07/18/2007
J. Org. Chem. 2007, 72, 6631-6633
6631

SCHEME 1. Stereoselective Titanium-Mediated Aldol
Reactions from Ketone 1
TABLE 1. Preliminary Studies on the TiCl4-Mediated Aldol
Reaction of 1 and Isobutyraldehyde
a
b
entry
additive, L
equiv
t (min)
dr 2a/3a
yield (%)
Since TiCl4-mediated aldol reactions based on ketone 1
1c
83:17
91:9
82:18
93:7
92:8
75:25
84:16
74
55
60
70
70
60
75
provided particularly low stereocontrol in the case of aliphatic
2
3
4
5
6
7
Ph3P
Ph3P
THF
THF
Ph3P
THF
1.1
1.1
1.1
1.1
2
10
60
10
60
10
10
aldehydes,7a we chose isobutyraldehyde to test the influence of
additives on such processes. Then, we focused our attention on
the changes that two Lewis bases as triphenylphosphine (Ph3P)
and tetrahydrofuran (THF) produced on the TiCl4-mediated aldol
reaction of 1 and isobutyraldehyde. Keeping in mind that our
main concern dealt with stereoselectivity, we decided to apply
the previously optimized experimental procedure with deliber-
ately extended reaction times. Importantly, enolization was
carried out in the absence of those additives to avoid potentially
undesired interactions with TiCl4. The results are summarized
in Table 1.11
As expected, diastereoselectivity was affected by the presence
in the reaction mixture of either of the two Lewis bases.
Moreover, diastereomeric ratios were dependent on the time
elapsed between the addition of the Lewis base and the aldehyde
to the enolate (compare entries 2 and 3 in Table 1) and, more
importantly, on the amounts of Ph3P or THF used (compare
respectively entries 2 and 6, and 4 and 7 in Table 1). This
behavior suggests that both additives might act as ligands that
affect to the aggregation state of the enolate and the coordination
sphere around the metal.
2
a
Determined by HPLC. b Overall isolated yield. c See ref 7a.
TABLE 2. Influence of Additives (1.1 equiv) on the
TiCl4-Mediated Aldol Reaction of 1 and Isobutyraldehyde
a
b
entry
additive, L
dr 2a/3a
yield (%)
1c
2
3
4
5
6
7
8
9
83:17
81:19
84:16
90:10
92:8
91:9
81:19
93:7
74
25
89
67
42
55
72
70
60
72
TMEDA
CH3CN
NMP
Ph3PdO
Ph3P
Et2O
THF
DME
94:6
81:19
1
0
1,4-dioxane
a
Determined by HPLC. b Overall isolated yield. c See ref 7a.
Having established the key issues of the modified experi-
mental procedure, we next surveyed the influence of 1.1 equiv
of other common Lewis bases that could potentially bind to
the titanium. The results gathered in Table 2 reveal that they
have a variable influence on the diastereoselectivity. The
rationale for these results is at present elusive. Surprisingly,
TMEDA afforded the worst diastereomeric ratio and yield (see
enolization should be carried out for 30 min, followed by
dropwise addition of a solution containing 1.1 equiv of additive
(see the General Procedure in the Experimental Section). After
1
0 min, 1.5 equiv of freshly distilled aldehyde was added, and
the resultant reaction mixture was stirred at -78 °C for 1.5 h.
Unfortunately, the addition of DME slows down the aldol
reaction and unreacted starting ketone 1 is usually observed in
the crude reaction mixtures. Longer reaction times do not
substantially improve the yields, and higher temperatures have
a deleterious effect on the diastereoselectivity. Once optimized,
the experimental procedure was next applied to a representative
set of aliphatic, aromatic, and R,â-unsaturated aldehydes. Results
are summarized in Table 3.
12
entry 2 in Table 2), whereas acetonitrile had apparently no
13
influence on the process (see entry 3 in Table 2). Otherwise,
most of the additives improved the stereoselectivity, with ethers
as THF or 1,2-dimethoxyethane (DME) the most suitable among
them (see entries 8 and 9 in Table 2).14
Hence, our interest was centered on both additives and
experimental conditions were evaluated in depth. After a
considerable experimental optimization, it was established that
Analysis of the diastereomeric ratios obtained for such
aldehydes indicates that both additives have a dramatic influence
on the stereochemical outcome of the process. Indeed, addition
(
11) Configurations of the resulting aldol diastereomers have been
previously established. See ref 7a.
12) TMEDA was successfully used by Crimmins on aldol reactions
(
(14) In striking contrast, diethyl ether and 1,4-dioxane afford low
diastereomeric ratios. In regard of the influence of ethers, it is worth recalling
the large solvent effects (THF, Et2O, i-Pr2O) observed in aldol reactions
based on N-acyl-1,3-oxazolidinones involving titanium enolates prepared
from their lithium counterparts; see: Nerz-Stormes, M.; Thornton, E. R. J.
Org. Chem. 1991, 56, 2489-2498.
based on N-acyl oxazolidinethiones and thiazolidinethiones. See refs 3
and 4.
(13) Ghosh reported that acetonitrile produces a dramatic enhancement
of diastereoselectivity on titanium-mediated chloroacetate aldol reactions.
See ref 6a.
6
632 J. Org. Chem., Vol. 72, No. 17, 2007

TABLE 3. Influence of THF and DME (1.1 equiv) on the
TiCl4-Mediated Aldol Reaction of 1
In summary, we have documented the influence of neutral
Lewis bases on the TiCl4-mediated aldol reactions based on
chiral ketone 1, with tetrahydrofuran and 1,2-dimethoxyethane
the most suitable additives to achieve high diastereomeric ratios.
Hence, it is possible to gain access to both 2,4-syn-4,5-syn (2)
and 2,4-anti-4,5-syn (3) adducts through the appropriate choice
of the Lewis acid used (TiCl4-THF or DME vs Ti(i-PrO)Cl3
respectively) (Scheme 2).
Although the mechanistic details of the reaction have been
not disclosed, the aforementioned results lend support to the
idea that the coordination sphere of the metal plays a crucial
role on the stereochemical outcome of the process. Eventually,
comparison of these results with previously reported studies on
the effect of additives on the stereoselectivity of titanium-
mediated aldol reactions suggests that the influence of ligands
bound to the metal closely relies on the structure of the enolate,
and in consequence, the space around the titanium must be
carefully crafted to achieve highly stereoselective processes.
a
b
entry additive, L aldehyde
R
dr 2/3 yield (%)
1c
2
a
a
a
b
b
b
c
c
c
d
d
d
e
e
e
f
CH(CH3)2
CH(CH3)2
CH(CH3)2
CH2CH(CH3)2
CH2CH(CH3)2
CH2CH(CH3)2
CH2CH3
83:17
93:7
97:3
82:18
88:12
90:10
69:31
88:12
94:6
93:7
95:5
97:3
91:9
94:6
94:6
92:8
94:6
96:4
74
79
64
95
85
57
92
84
45
91
92
56
96
91
64
83
82
41
THF
DME
3
4
c
5
6
7
THF
DME
c
8
9
0
THF
DME
CH2CH3
CH2CH3
c
Experimental Section
1
1
1
1
1
1
1
1
1
Ph
Ph
Ph
1
2
3
THF
DME
General Procedure. Neat TiCl
slowly to a solution of 1 (196 mg, 1 mmol) in CH
-78 °C under N . The resulting yellow mixture was stirred for 3
min, and i-Pr NEt (0.19 mL, 1.1 mmol) was added dropwise. The
resulting dark red solution was stirred for 30 min at -78 °C, and
a 0.55 M solution of THF or DME in CH Cl (2 mL, 1.1 mmol)
4
(0.12 mL, 1.1 mmol) was added
c
4-ClPh
4-ClPh
4-ClPh
C(CH3)dCH2
C(CH3)dCH2
C(CH3)dCH2
2
2
Cl (3 mL) at
4
5
6
7
8
THF
DME
2
2
THF
DME
f
f
2
2
was slowly added. The mixture was stirred for 10 min at -78 °C,
and freshly distilled aldehyde was added (1.5 mmol). The reaction
a
Determined by HPLC. b Overall isolated yield. c See ref 7a.
was quenched after 1.5 h by the addition of satd NH
The mixture was diluted with Et O and washed with H
NaHCO and brine. The aqueous layers were extracted with Et
and the combined organic extracts were dried (MgSO ) and
concentrated. The resulting oil was analyzed by H NMR and HPLC
and purified by flash chromatography.
4
Cl (5 mL).
O, satd
O,
2
2
SCHEME 2. Divergent Stereochemical Pathways of the
Titanium-Mediated Aldol Reaction of 1
3
2
4
1
(
2S,4R,5S)-2-Benzyloxy-5-hydroxy-4,6-dimethyl-3-heptan-
one (2a): colorless oil. R (hexanes/EtOAc 90:10) ) 0.10. HPLC
hexanes/i-PrOH 99:1) t ) 10.3 min. [R] ) -19.7 (c ) 1.0,
CHCl ). IR (film): ν 3510 (br), 2962, 2873, 1713, 1455, 1118 cm .
H NMR (400 MHz, CDCl ): δ 7.40-7.30 (5H, m), 4.57 (1H,
f
(
R
D
-1
3
1
3
AB system, J ) 11.7), 4.52 (1H, AB system, J ) 11.7), 4.06 (1H,
q, J ) 6.8), 3.46 (1H, dd, J ) 8.5, J ) 2.6), 3.18 (1H, qd, J ) 7.2,
J ) 2.6), 1.72-1.62 (1H, m), 1.39 (3H, d, J ) 6.8), 1.11 (3H, d,
1
3
J ) 7.2), 1.00 (3H, d, J ) 6.5), 0.83 (3H, d, J ) 6.7). C NMR
75.4 MHz, CDCl
(
7
3
): δ 217.4, 137.5, 128.5, 128.0, 127.8, 79.4,
6.2, 71.8, 42.4, 30.6, 19.1, 18.9, 17.3. 9.5. HRMS (+ESI): m/z
of 1.1 equiv of DME produces a remarkable improvement on
the diastereoselectivity for all the aldehydes, to the point that
most of them afford 2,4-syn-4,5-syn-aldols 2 in dr (2/3) higher
than 94:6 (compare entries 1, 4, 7, 10, 13, and 16 to 3, 6, 9, 12,
+
24 3
calcd for C16H O Na [M + Na] 287.1618, found 287.1614.
Acknowledgment. We acknowledge financial support from
the Spanish Ministerio de Ciencia y Tecnolog ´ı a (CTQ2006-
3249/BQU), the Generalitat de Catalunya (2005SGR00584),
and the Universitat de Barcelona (ACES-UB2006).
1
5, and 18, respectively). Similarly, THF improves the results
1
previously obtained and also provides highly stereoselective
aldol reactions (compare entries 1, 4, 7, 10, 13, and 16 to 2, 5,
8
, 11, 14, and 17, respectively). Although the diastereoselectivity
Supporting Information Available: Spectroscopic data and
copies of H and C NMR spectra of aldols 2. This material is
available free of charge via the Internet at http://pubs.acs.org.
of these reactions is slightly poorer than that achieved with
DME, the yields are consistently higher, which confer to THF
a particular interest for aromatic and R,â-unstaurated aldehydes
1
13
(see entries 11, 14, and 17).
JO071048Z
J. Org. Chem, Vol. 72, No. 17, 2007 6633