Stereoselective titanium-mediated syn-aldol reaction from a lactate-derived chiral ethyl ketone

Article abstract of DOI:10.1016/j.tetlet.2004.05.077

Stereoselectivity of the titanium-mediated aldol process based on (S)-2-benzyloxy-3-pentanone, 1, is dramatically modified by the presence of a Lewis acid. Among the Lewis acids surveyed, TiCl₄ has given access to the corresponding 2,4-anti-4,5-syn aldol adducts with the highest diastereomeric ratios. The excellent stereocontrol exerted by the aforementioned ketone has been demonstrated in double asymmetric reactions involving chiral α-methyl-β-OTBDPS aldehydes.

Full text of DOI:10.1016/j.tetlet.2004.05.077

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 5379–5382
Stereoselective titanium-mediated syn-aldol reaction from
a lactate-derived chiral ethyl ketone
*
*
ꢀ
Joan G. Solsona, Pedro Romea and Felix Urpı
ꢁ
Departament de Quꢁımica Orgꢀanica, Universitat de Barcelona, Martꢁı i Franquꢁes 1-11, 08028 Barcelona, Catalonia, Spain
Received 26 March 2004;revised 14 May 2004;accepted 17 May 2004
Abstract—Stereoselectivity of the titanium-mediated aldol process based on (S)-2-benzyloxy-3-pentanone, 1, is dramatically
modified by the presence of a Lewis acid. Among the Lewis acids surveyed, TiCl4 has given access to the corresponding 2,4-anti-4,5-
syn aldol adducts with the highest diastereomeric ratios. The excellent stereocontrol exerted by the aforementioned ketone has been
demonstrated in double asymmetric reactions involving chiral a-methyl-b-OTBDPS aldehydes.
ꢀ 2004 Elsevier Ltd. All rights reserved.
Although the Lewis acid-mediated aldol-type addition
of a silyl enol ether to an aldehyde, the so-called
Mukaiyama reaction,1 has been thoroughly studied to
the point that it constitutes nowadays one of the most
powerful methodologies for the stereoselective con-
struction of carbon–carbon bonds,2 reactions of metal
enolates with aldehydes in the presence of Lewis acids
have received much less attention.3 Regardless of their
reaction mechanisms, which are not fully understood
yet, highly ordered transition-state models are com-
monly invoked to rationalize the stereochemical out-
come of such processes. Then, considering that we have
recently disclosed that the titanium-mediated aldol
reaction based on the lactate derived chiral ketone 14
affords the corresponding 2,4-syn-4,5-syn adducts (2 in
Scheme 1), presumably through a chelated cyclic tran-
sition state,5 we envisioned that the conformationally
rigid titanium enolate shown in Scheme 1 might provide
access to other adducts than those previously obtained if
the reaction proceeded through a different transition
state. We now report our studies on the Lewis acid-
mediated aldol reaction involving the titanium enolate
derived from 1.6–8
As expected, preliminary experiments showed that the
introduction of a Lewis acid produced dramatic changes
on the stereochemical outcome of the above-mentioned
process. In these early experiments, 1 equiv of several
Lewis acids and 1.5 equiv of isobutyraldehyde, a, were
subsequently added to a solution containing the tita-
nium enolate from ketone 1 (method A). After 2 h,
analysis of the reaction mixtures revealed the formation
of up to three diastereomers, 2a–4a (R ¼ i-Pr in Scheme
1).9 The results are summarized in Table 1.
Cln
O
OH
O
OH
O
OH
O
Ti
a
b
O
2
4
2
4
2
4
+
+
5
R
5 R
5 R
BnO
BnO
BnO
BnO
BnO
1
2,4-syn-4,5-syn
2,4-anti-4,5-syn
4,5-anti
2
3
4
Scheme 1. Reagents and conditions: (a) TiCl4, i-Pr2NEt, CH2Cl2, )78 ꢁC, 1.5 h;(b) Lewis acid, RCHO, )78 ꢁC.
Keywords: Aldol reactions;Asymmetric reactions;Titanium enolates;Lewis acids.
* Corresponding authors. Tel.: +34-93-402-1247;fax: +34-93-339-7878;e-mail addresses: pedro.romea@ub.edu; felix.urpi@ub.edu
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.05.077

5380
J. G. Solsona et al. / Tetrahedron Letters 45 (2004) 5379–5382
Table 1. Stereoselectivity of the Lewis acid-mediated aldol reaction of 1
again, 2,4-anti-4,5-syn aldol 3a was isolated in good
yield (76% yield) and excellent diastereomeric ratio (dr
2a:3a 1:99). In spite of these encouraging results, some
aspects deserve to be commented. Method B afforded 3a
but in a somewhat slightly diminished diastereomeric
ratio, which was dramatically eroded when other alde-
hydes were used.10 Otherwise, method C was more
suitable, but TiCl4-aldehyde precomplexation produced
a yellow suspension, which made its transfer more dif-
ficult and prevented its use on large scale preparations.
Therefore, the original experimental procedure (method
A) turned out to be the most suitable one and was used
henceforth.11
Entry
Lewis acid
dr (2a:3a:4a)a
1
2
3
4
5
6
7
8
––
83:17:––
7:86:7
BF3ÆOEt2
Et2AlCl
TiCl4
30:69:1
1:99:––
5:95:––
39:61:––
50:50:––
1:74:25
TiCl3(i-PrO)
TiCl2(i-PrO)2
Ti(i-PrO)4
SnCl4
a Diastereomeric ratios have been established by HPLC analysis of the
crude reaction mixtures.
Despite leading to the same diastereomer, the three re-
ported processes A–C do not necessarily follow identical
mechanistic pathways. The stereochemical outcome of
the reaction in the case of method C might be ruled by
an open transition state involving the antiperiplanar
approach of a TiCl4–aldehyde complex to the less hin-
dered Re face of the chelated enolate shown in Scheme
1.12a Alternatively, we speculate that the addition of a
second equivalent of TiCl4 (methods A and B) might
produce a new enolate-like species that could evolve
through a cyclic transition state.12b Mechanistic details
of these reactions are the subject of ongoing investiga-
tions.
It was clear indeed that addition of any Lewis acid
shifted the configuration of the major aldol adduct from
2,4-syn-4,5-syn, 2a, to 2,4-anti-4,5-syn, 3a (compare
entries 1 and 2–8 in Table 1). Most importantly, TiCl4
afforded 3a in an excellent diastereomeric ratio (see
entry 4 in Table 1).
Encouraged by these findings, we decided to study more
deeply the influence of TiCl4 on this transformation.
Given that the stereoselectivity can also be dependent on
the number of equivalents of Lewis acid and the order in
which the reactants are combined, both issues were
addressed.
Further studies revealed that the addition of the tita-
nium enolate derived from 1 to isobutyraldehyde was
very fast, and the starting material was almost con-
sumed in 5 min. Eventually, we were able to isolate pure
2,4-anti-4,5-syn aldol 3a in 85% yield after 30 min.
Moreover, no changes on the diastereomeric ratio were
observed from 5 min to 2 h, which suggests that the
process takes place under kinetic control and no equil-
ibration occurs over the reaction time.
Initially, we evaluated the influence of the supplemen-
tary equivalents of TiCl4 added to the reaction mixture
on the diastereoselectivity. The results are summarized
in Table 2. The stereochemical dependence of the Lewis
acid stoichiometry can be related to the existence of two
parallel mechanistic pathways. When no supplementary
TiCl4 is added to the reaction mixture, the process
probably goes through the chelated cyclic transition
state previously reported5 and the 2,4-syn-4,5-syn dia-
stereomer is mainly obtained (see Scheme 1). Then, as
the amount of extra Lewis acid increases a new pathway,
leading to the 2,4-anti-4,5-syn diastereomer, becomes
more important and finally overrides the former.
Once the reaction had been optimized, representative
aldehydes were surveyed in order to gain insight into the
synthetic potentiality of the process. The results are
summarized in Table 3. As shown, aliphatic aldehydes
a–c reacted with uniformly excellent diastereoselectivi-
ties and a single diastereomeric aldol adduct 3 (see
Scheme 1) was observed (dr 2:3 1:99), even in the case of
propanal. In the case of aromatic and a,b-unsaturated
aldehydes, d–g, the corresponding aldol adducts 3 were
isolated in equally good yields but with a slightly lower
diastereoselectivity (dr 2:3 6:94).9
Next, two new experimental protocols were examined.
Simultaneous addition of 2 equiv of TiCl4 in the enoli-
zation step (method B), keeping constant all the other
reaction parameters, also gave access to 3a in good yield
(77% yield) and very good diastereomeric ratio (dr 2a:3a
3:97). Alternatively, the isobutyraldehyde was added
precomplexed with 1 equiv of TiCl4 (method C);once
Table 2. Influence of the additional TiCl4 on the stereoselectivity of
the titanium-mediated aldol reaction of 1
Table 3. TiCl4 Mediated aldol reaction of 1
Entry
Aldehyde
R
dr (2:3)a Yieldb (%)
Entry
Supplementary TiCl4
equivalents
dr (2a:3a)a
1
2
3
4
5
6
7
a
b
c
d
e
f
i-Pr
i-Bu
1:99
1:99
1:99
6:94
6:94
6:94
6:94
85
82
81
88
87
65
70
1
2
3
4
5
6
––
83:17
62:38
38:62
12:88
1:99
Et
Ph
0.25
0.50
0.75
1.0
4-ClPh
(E)CH3CH@CH
H2C@C(CH3)
g
2.0
1:99
a Diastereomeric ratios have been established by HPLC analysis of the
crude reaction mixtures.
b Overall isolated yield.
a Diastereomeric ratios have been established by HPLC analysis of the
crude reaction mixtures.

J. G. Solsona et al. / Tetrahedron Letters 45 (2004) 5379–5382
5381
O
OTBDPS
H
OH
O
OH OTBDPS
5
OH
a
a
O
H
Cln
TiCl4
BnO
BnO
6
8
H
BnO
BnO
H
Ti
O
77% dr 99:1
93%
O
OTBDPS
BnO
H
H
O
OH OTBDPS
ent-5
H
O
OH
TiCl4
H
H
7
9
HO
79% dr 98:2
96%
Scheme 2. Reagents and conditions: (a) TBAFÆ3H2O, AcOH, THF, rt, 2 h.
7883–7884;(f) Crimmins, M. T.;King, B. W.;Tabet, E.
A.;Chaudhary, K. J. Org. Chem. 2001, 66, 894–902;(g)
Ghosh, A. K.;Kim, J.-H. Tetrahedron Lett. 2002, 43,
5621–5624;(h) Crimmins, M. T.;McDougall, P. J. Org.
Lett. 2003, 5, 591–594.
Finally, the stereocontrol exerted by 1 under the afore-
mentioned experimental procedure was challenged in
double stereodifferentiating reactions13 involving chiral
aldehydes 5 and ent-5.14 As represented in Scheme 2,
both processes were highly stereoselective and virtually
one diastereomer was obtained irrespective of the con-
figuration of the aldehyde. The configuration of adducts
ꢁ
4. Ferrero, M.;Galobardes, M.;Mart ın, R.;Montes, T.;
Romea, P.;Rovira, R.;Urp ı, F.;Vilarrasa, J. Synthesis
ꢁ
ꢁ
2000, 1608–1614.
5. Solsona, J. G.;Romea, P.;Urp ı, F.;Vilarrasa, J. Org.
Lett. 2003, 5, 519–522.
1
6 (anti Felkin) and 7 (Felkin) was established by H
ꢁ
NMR through analysis of coupling constants 3J(H4–H5)
and NOE studies on hemiacetals derivatives 8 and 9
generated by deprotection of the silyl ether (diagnostic
NOE interactions are indicated in Scheme 2).15;16
6. The synthetic utility of lactate-derived ketones in stereo-
selective syn and anti boron-mediated aldol reactions was
firmly established by Paterson. See: (a) Paterson, I.;
ꢁ
Wallace, D. J.;Vel azquez, S. M. Tetrahedron Lett. 1994,
35, 9083–9086;(b) Paterson, I.;Wallace, D. J. Tetrahedron
Lett. 1994, 35, 9087–9090;(c) Paterson, I.;Wallace, D. J.;
Cowden, C. J. Synthesis 1998, 639–652.
In summary, it has been documented the dramatic
influence of Lewis acids on the titanium-mediated aldol
reaction based on ketone 1, which gives access to
enantiomerically pure syn-aldol adducts in high yields.
Furthermore, considering how easily the aldols gener-
ated from lactate-derived ketones can be manipu-
lated,6b;c;17 the ready availability of 1 from lactate esters
and the use of inexpensive TiCl4, the reported method-
ology represents an appealing and practical entry to the
synthesis of polypropionate-like natural products.
7. For previous boron- and titanium-mediated aldol reac-
tions based on a-silyloxy ketones derived from lactic acid,
ꢁ
ꢁ
see: (a) Figueras, S.;Mart ın, R.;Romea, P.;Urp ı, F.;
Vilarrasa, J. Tetrahedron Lett. 1997, 38, 1637–1640;(b)
ꢁ
Galobardes, M.;Gasc on, M.;Mena, M.;Romea, P.;Urp ı,
F.;Vilarrasa, J. Org. Lett. 2000, 2, 2599–2602.
ꢁ
8. For a catalytic aldol process based on the trichlorosilyl
enolates from a lactate-derived ketone, see Denmark, S.
E.;Pham, S. M. Org. Lett. 2001, 3, 2201–2204.
9. The relative and absolute configuration of syn-aldols 2 and
3 have been established by comparison of 1H and 13C
NMR spectra as well as specific rotations with those
previously reported by us. See Ref. 5. The absolute
configuration of anti-aldols 4 has not been established.
10. Method B afforded a 15:85 mixture of 2d:3d in the case of
benzaldehyde.
Acknowledgements
We thank Professor Ian Paterson for helpful sugges-
tions. Financial support from the Ministerio de Ciencia
y Tecnologıa and Fondos FEDER (grant BQU2002-
01514), and from the Generalitat de Catalunya
(2001SGR00051), as well as a doctorate studentship
(Universitat de Barcelona) to J.G.S. are acknowledged.
ꢁ
11. Typical experimental procedure: TiCl4 (0.12 mL,
1.1 mmol) was added dropwise to
a solution of 1
(198 mg, 1.0 mmol) in CH2Cl2 (5 mL) at )78 ꢁC under
N2, and, 2 min later, i-Pr2NEt (0.19 mL, 1.1 mmol). The
resulting dark red enolate solution was stirred for 1.5 h at
)78 ꢁC. Then, TiCl4 (0.11 mL, 1.0 mmol) was added
dropwise, the solution was allowed to stir for 10 min,
and 1.5 equiv of aldehyde was added. Stirring was contin-
ued for 30 min at )78 ꢁC. The reaction was quenched by
the addition of saturated NH4Cl (5 mL) and vigorously
stirred at rt. The mixture was diluted with Et2O, washed
with saturated NaHCO3, and brine. The aqueous phases
were extracted with Et2O and the combined organic
extracts were dried (MgSO4) and concentrated. The
resulting oil was analyzed by HPLC and purified by flash
chromatography on silica gel (hexanes/EtOAc).
References and notes
1. Mukaiyama, T. Org. React. 1982, 28, 203–331.
2. Mahrwald, R. Chem. Rev. 1999, 99, 1095–1120.
3. (a) Walker, M. A.;Heathcock, C. H. J. Org. Chem. 1991,
56, 5747–5750;(b) Yan, T.-H.;Lee, H.-C.;Tan, C.-W.
Tetrahedron Lett. 1993, 34, 3559–3562;(c) Wang, Y.-C.;
Hung, A.-W.;Chang, C.-S.;Yan, T.-H. J. Org. Chem.
1996, 61, 2038–2043;(d) Ghosh, A. K.;Onishi, M. J. Am.
Chem. Soc. 1996, 118, 2527–2528;(e) Crimmins, M. T.;
King, B. W.;Tabet, E. A. J. Am. Chem. Soc. 1997, 119,
12. (a) The stereochemical outcome of the aldol reaction of
boron enolates derived from oxazolidinones in the pres-
ence of Lewis acids has been rationalized using this model.

5382
J. G. Solsona et al. / Tetrahedron Letters 45 (2004) 5379–5382
See Ref. 3a;(b) It has been suggested that additional TiCl 4
equivalents can produce the abstraction of a chloride
anion or the formation of chloro bridged species. See Ref.
3e.
3.07 (1H, br s, OH), 1.69–1.61 (1H, m, CHCH2), 1.35 (3H,
d, J ¼ 6:8 Hz, CH3 CHOBn), 1.18 (3H, d, J ¼ 6:8 Hz,
COCHCH3), 1.08 (9H, s, C(CH3)3), 1.00 (3H, d,
J ¼ 7:0 Hz, CH3CHCH2); 13C NMR (100 MHz, CDCl3)
d 215.5, 137.5, 135.6, 135.5, 133.0, 132.9, 129.8, 129.7,
128.4, 127.8, 127.7 (·3), 79.3, 74.4, 71.6, 68.2, 44.0, 37.4,
26.8, 19.1, 16.5, 12.6, 11.3.
13. Masamune, S.;Choy, W.;Petersen, J. S.;Sita, L. R.
Angew. Chem., Int. Ed. Engl. 1985, 24, 1–30.
14. Roush, W. R.;Palkowitz, A. D.;Ando, K. J. Am. Chem.
Soc. 1990, 112, 6348–6359.
15. Selected physical and spectroscopic data.
6: colourless oil; Rf (hexanes/EtOAc 90:10) ¼ 0.1; ½aꢀ )7.7
16. The deprotection of the aldol adduct 6 affords a thermo-
dynamical mixture of two hemiacetals, the major compo-
nent being 8 both in CDCl3 (dr 80:20) and C6D6 (dr
65:35). In the case of adduct 7, a single hemiacetal 9 was
observed. Selected NMR data:
D
(c 1.42, CHCl3);IR (film): 3496, 3072, 2933, 2860, 1711,
1455, 1428, 1113 cmꢁ1 1H NMR (400 MHz, CDCl3) d
;
1
7.69–7.67 (4H, m, ArH), 7.46–7.28 (11H, m, ArH), 4.58
(1H, d, J ¼ 11:7 Hz, PhCHxHyO), 4.52 (1H, d,
J ¼ 11:7 Hz, PhCHxHyO), 4.11 (1H, q, J ¼ 6:9 Hz,
CHOBn), 3.99 (1H, dt, J ¼ 8:7 Hz, J ¼ 2:9 Hz, CHOH),
3.78 (1H, dd, J ¼ 10:1 Hz, J ¼ 4:4 Hz, CHxHyOSi), 3.67
(1H, dd, J ¼ 10:1 Hz, J ¼ 6:1 Hz, CHxHyOSi), 3.53 (1H,
d, J ¼ 2:9 Hz, OH), 3.12 (1H, qd, J ¼ 7:0 Hz, J ¼ 2:9 Hz,
COCHCH3), 1.84–1.71 (1H, m, CHCH2), 1.38 (3H, d,
J ¼ 6:9 Hz, CH3CHOBn), 1.10 (3H, d, J ¼ 7:0 Hz,
COCHCH3), 1.04 (9H, s, C(CH3)3), 0.85 (3H, d,
J ¼ 7:0 Hz, CH2CHCH3); 13C NMR (100 MHz, CDCl3)
d 215.2, 137.7, 135.6, 133.0, 132.9, 129.8, 128.4, 127.8,
127.7, 79.3, 74.4, 71.7, 68.0, 44.2, 37.6, 26.8, 19.2, 17.4,
13.6, 8.7.
8: H NMR (400 MHz, C6D6) d 7.30–7.00 (5H, m, ArH),
4.44 (1H, d, J ¼ 11:3 Hz, OCHxHyPh), 4.06 (1H, d,
J ¼ 11:3 Hz, OCHxHyPh), 3.88 (1H, t, J ¼ 11:8 Hz,
CHxHyO), 3.46 (1H, q, J ¼ 6:2 Hz, CHOBn), 3.34 (1H,
dd, J ¼ 11:8 Hz, J ¼ 5:3 Hz, CHxHyO), 3.05 (1H, br s,
CHOH), 2.23 (1H, qd, J ¼ 7:1 Hz, J ¼ 2:8 Hz,
CH3CHC(OH)O), 1.75–1.65 (1H, m, CH3CHCH2O),
1.38 (3H, d, J ¼ 6:2 Hz, CH3CHOBn), 0.78 (3H, d,
J ¼ 7:1 Hz, CH3CHC(OH)O), 0.64 (3H, d, J ¼ 6:9 Hz,
CH3CHCH2O).
1
9: H NMR (500 MHz, C6D6) d 7.30–7.00 (5H, m, ArH),
4.42 (1H, d, J ¼ 11:7 Hz, OCHxHyPh), 4.16 (1H, d,
J ¼ 11:7 Hz, OCHxHyPh), 3.58 (1H, t, J ¼ 11:2 Hz,
CHxHyO), 3.53 (1H, dd, J ¼ 11:2 Hz, J ¼ 5:2 Hz,
CHxHyO), 3.46 (1H, q, J ¼ 6:3 Hz, CHOBn), 3.17 (1H,
t, J ¼ 10:0 Hz, CHOH), 1.92 (1H, dq, J ¼ 10:0 Hz,
J ¼ 6:6 Hz, CH3CHC(OH)O), 1.56–1.46 (1H, m,
7: colourless oil; Rf (hexanes/EtOAc 90:10) ¼ 0.1; ½aꢀ )9.7
D
(c 1.18, CHCl3);IR (film): 3505, 3072, 2933, 2860, 1711,
1457, 1428, 1113 cmꢁ1 1H NMR (400 MHz, CDCl3) d
;
7.69–7.66 (4H, m, ArH), 7.47–7.29 (11H, m, ArH), 4.57
(1H, d, J ¼ 11:7 Hz, PhCHxHyO), 4.54 (1H, d,
J ¼ 11:7 Hz, PhCHxHyO), 4.20–4.00 (2H, m, CHOBn
and CHOH), 3.73 (1H, dd, J ¼ 10:2 Hz, J ¼ 3:9 Hz,
CHxHyOSi), 3.60 (1H, dd, J ¼ 10:2 Hz, J ¼ 4:7 Hz,
CHxHyOSi), 3.27 (1H, quintet, J ¼ 6:8 Hz, COCHCH3),
CH3CHCH2O),
1.18
(3H,
d,
J ¼ 6:6 Hz,
CH3CHC(OH)O), 1.14 (3H, d, J ¼ 6:3 Hz, CH3CHOBn),
0.74 (3H, d, J ¼ 6:5 Hz, CH3CHCH2O).
ꢁ
17. (a) Esteve, C.;Ferrer o, M.;Romea, P.;Urp ı, F.;Vilarrasa,
J. Tetrahedron Lett. 1999, 40, 5083–5086;(b) Solsona, J.
ꢁ
ꢁ
G.;Romea, P.;Urp ı, F. Org. Lett. 2003, 5, 4681–4684.