Siamyl Glyoxylate as an Efficient Enophile in the Lewis Acid Mediated Glyoxylate Ene-Reaction with Allylic Ethers. Application to the Synthesis of a Taxol A-Ring Subunit

Article abstract of DOI:10.1021/jo971082l

In contrast to other simple alkyl glyoxylates, siamyl glyoxylate (1a) proved to be an effective enophile in the glyoxylate ene-reaction with the allylic ether 2a, which we earlier used for the synthesis of the A-ring subunit of taxol.

Full text of DOI:10.1021/jo971082l

J . Org. Chem. 1997, 62, 8735-8740
8735
Sia m yl Glyoxyla te a s a n Efficien t En op h ile in th e Lew is Acid
Med ia ted Glyoxyla te En e-Rea ction w ith Allylic Eth er s.
Ap p lica tion to th e Syn th esis of a Ta xol A-Rin g Su bu n it
J ohan Wennerberg, Magnus Polla, and Torbjo¨rn Frejd*
Organic Chemistry 1, Department of Chemistry, Lund University, P.O. Box . 124, S-221 00 Lund, Sweden
Received J une 16, 1997X
In contrast to other simple alkyl glyoxylates, siamyl glyoxylate (1a ) proved to be an effective enophile
in the glyoxylate ene-reaction with the allylic ether 2a , which we earlier used for the synthesis of
the A-ring subunit of taxol.
Sch em e 1. P a r tia l Retr osyn th etic Sch em e for th e
Syn th esis of Ta xol
In tr od u ction
The thermal and Lewis acid induced glyoxylate ene
reaction was introduced more than 30 years ago by
Klimova and Arbuzov et al.1-6 and was further developed
by several research groups (Achmatowicz et al.,7 Snider
et al.,8,9 Whitesell et al.,10 and Mikami et al.11,12 ). In
particular, the asymmetric induction achieved by using
chiral glyoxylate esters7,10,13-15 and the catalysis with
BINOL-Lewis acid complexes developed by Mikami et
al.11,12 were major achievements. In our efforts to
synthesize the taxol A-ring building block 4 (Scheme 1),
we previously used the SnCl4-catalyzed asymmetric ene
reaction of 8-phenylmenthyl glyoxylate 1 (R1 ) (-)-8-
phenylmenthyl) and the ene-component 2a (R2 ) PMB).16
This reaction resulted in a high yield of the ene-product
3 with a high enantiomeric purity (>98% ee).
There were, however, several drawbacks (to be dis-
cussed below), and the need for larger amounts of
precursor 3 required a more suitable method. We
therefore abandoned the asymmetric ene reaction and
instead used the (iPrO)2TiCl2-catalyzed reaction of the
easily available 1d and 2a (Scheme 2) to produce racemic
3d . Ethyl glyoxylate 1d was used in 5 molar excess over
2a . Relatively good yields of 3d were obtained in 1-g
scales but the yields dropped when the scale was in-
Sch em e 2
creased. In one such large scale experiment “catalyzed”
by 2 equiv of (iPrO)2TiCl2 we obtained 40 g (40%) of 3d .
This amount of material lasted for some time. To our
dismay we could not repeat this reaction with a reason-
able accuracy at a later date. A large number of
experiments were run applying many different condi-
tions, but, statistically, less than one out of 10 experi-
ments were successful i.e., gave yields of up to 40%. In
the other nine experiments there was no conversion of
the allylic ether. We therefore decided to further inves-
tigate this reaction, and we here report the successful
solution to this problem by using siamyl glyoxylate.
X Abstract published in Advance ACS Abstracts, November 1, 1997.
(1) Klimova, E.; Arbuzov, Y. Dokl. Akad. Nauk SSSR 1965, 167,
419-421.
(2) Klimova, E.; Arbuzov, Y. Dokl. Akad. Nauk SSSR 1967, 173,
1332-1335.
(3) Klimova, E.; Arbuzov, Y. Dokl. Akad. Nauk SSSR 1966, 167,
1060-1062.
(4) Klimova, E.; Arbuzov, Y. Dokl. Akad. Nauk SSSR 1967, 173,
1332-1335.
(5) Klimova, E.; Treshchova, E.; Arbuzov, Y. Dokl. Akad. Nauk SSSR
1968, 180, 865-868.
(6) Klimova, E.; Arbuzov, Y. J . Org. Chem. USSR Engl. 1968, 4,
1726-1728.
(7) Achmatowicz, J ., O.; Szechner, B. J . Org. Chem. 1972, 37, 964-
967.
Resu lts a n d Discu ssion
(8) Snider, B. B.; Straten, J . W. v. J . Org. Chem. 1979, 44, 3567-
3571.
(9) Snider, B. B. Acc. Chem. Res 1980, 13, 426-432.
(10) Whitesell, J . K.; Bhattacharya, A.; Buchanan, C. M.; Chen, H.
H.; Deyo, D.; J ames, D.; Liu, C.-L.; Minton, M. A. Tetrahedron 1986,
42, 2993-3001.
(11) Mikami, K.; Shimizu, M. Chem. Rev. (Washington, D.C.) 1992,
92, 1021-1050.
(12) Mikami, K.; Terada, M.; Narisawa, S.; Nakai, T. Synlett 1991,
255-265.
(13) Whitesell, J . K.; Battacharya, A.; Anguilar, D. A.; Henke, K. J .
Chem. Soc., Chem. Commun. 1982, 989-990.
(14) Whitesell, J . K.; Chen, H.-H.; Lawrence, R. M. J . Org. Chem.
1985, 50, 4663-4664.
(15) Whitesell, J . K.; Lawrence, R. M.; Chen, H.-H. J . Org. Chem.
1986, 51, 4779-4784.
(16) Polla, M.; Frejd, T. Tetrahedron 1993, 49, 2701-2710.
Lew is Acid s. A number of standard Lewis acids such
as TiCl4, SnCl4, Et2AlCl, BF3‚OEt2, BINOL‚TiCl2, and
BINOL‚AlCl3 were previously tried in the glyoxylate ene
reactions involving ethyl glyoxylate 1d and 2a without
success.16 As mentioned, TiCl2(OiPr)2 did not work well
with 1d , although in some small scale experiments yields
of up to 80% of 3d could be achieved, but most often no
product was formed at all. Application of two other Lewis
acids particularly recommended for the ene reaction,
FeCl3 and ZnCl2,17 gave only intractable reaction prod-
ucts (Table 1, entries 17 and 18). Curiously, the PMB-
grouping migrated to give 5 when the ene-reaction was
8
S0022-3263(97)01082-7 CCC: $14.00 © 1997 American Chemical Society

8736 J . Org. Chem., Vol. 62, No. 25, 1997
Ta ble 1. Lew is Acid Ca ta lyzed Glyoxyla te En e Rea ction s of 1a -d a n d 2a -c
Wennerberg et al.
product (yield %)
entry
R1(1)
R2(2)
catalyst
ratio 1:cat.:2
conda
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Sia (1a )
1a
1a
1a
1a
Et (1d )
1d
1a
1d
1d
1d
Cy (1b)
PMB (2a )
2a
2a
2a
2a
2a
2a
Bn (2b)
(iPro)2TiCl2
5:2:1
1:1:1
2:1:1
2:0.1:1
5:2:1
5:2:1
2:0.1:1
5:2:1
1:1:1
1:1:1
1:1:1
1:1:1
1:1:1
1:1:1
1:1:1
1:1:1
1:1:1
1:1:1
A
B
B
B
B
A
A
A
A
A
A
A
A
C
C
C
D
A
3a (97-99)
3a (30)
3a (47)
-(0)b
(+)-3a (95, 95% ee)
3d (0-40)c
3d (9)
3g impure
3e (0)
d
(iPrO)2TiCl2
(iPrO)2TiCl2
(iPrO)2TiCl2
(iPrO)2TiCl2/BINOL
(iPrO)2TiCl2
(iPrO)2TiCl2
(iPrO)2TiCl2
(iPrO)2TiCl2
(iPrO)2TiCl2
(iPrO)2TiCl2
(iPrO)2TiCl2
(iPrO)2TiCl2
SnCl4
2b
TBDMS (2c)
2a
2a
2a
2a
2a
2a
2a
2a
3d (0)e
3b (57)f
3c (65)f
3c (63)f
3a (30)
3b (58)f
g
2-MeCy (1c)
1c
1a
1b
1d
1d
SnCl4
SnCl4
FeCl3
ZnI2
g
a
b
Conditions; A: 48 h, -17 °C; B: 96 h, -17 °C; C: 5 min., -78 °C; D: 20 min., -25 °C. All of the starting material 2a was recovered.
c In most cases no reaction at all took place. The product deteroriated during chromatography on silica. e The reaction was performed
d
in toluene. f Yields after reesterfication to the corresponding ethyl esters. A large number of compounds were formed.
g
Sch em e 3
attempted with BF3‚OEt2 as Lewis acid or in the presence
of zeolite-â. This rearrangement is the subject of a
separate investigation.18
Thus, even if (iPrO)2TiCl2 only worked sporadically it
was still the only Lewis acid that gave at least some ene-
product. It was therefore used in the further testing of
other allylic ethers as ene components and glyoxylates
as enophiles.
P r otectin g Gr ou p of th e En e Com p on en t. We
initially chose the PMB-ether protection of the allylic
alcohol since the PMB-group may be conveniently re-
moved in high yields later in the reaction sequence.16
However, Lewis acid interactions with the PMB group
may have contributed to the problems related to the
unreproducibility of the ene reaction. Therefore we also
tested the reaction between 1d and allylic ethers carrying
the benzyl, ethyl, TIPS, and TBDMS groups. The TB-
DMS and benzyl protective groups in the ene-component
were reported to be compatible with the glyoxylate ene
reaction conditions (SnCl4, -78 °C).19 It was mentioned,
however, that the benzyl ether of 2-buten-1-ol did not give
any ene product, which was also true in our experiment
with 1d and benzyl ether 2b (entry 9). The TBDMS allyl
ether 2c20 gave a green spot and retention on TLC,
characteristic of the expected ene product, but the
product was too unstable to allow isolation in the pure
state (entry 10). On the other hand the TIPS allyl ether
2d gave a fair yield of ene product 6a (Scheme 3, eq 1).
Thus, the oxy-methylene hydrogens were more reactive
than the methine hydrogen. For comparison we also
tried the ethyl allyl ether 2e, which gave a 1:1 mixture
of the two ene products 3h and 6b (Scheme 3, eq 2). In
this case the reactivities of the different allylic hydrogens
were obviously identical. In other cases there may be a
certain degree of unselectivity; e.g. oxomalonate gave a
completely unselective reaction with substrate 8.21 How-
ever, for allyloxy systems such lack of regioselectivity has
previously only been reported for allylic acetates and
trifluoro acetates.11,19
It should be mentioned that we checked our technique
and reagents by repeating some relevant cases of Mikami
et al.,19 which could be perfectly reproduced.
Apparently, our ene-substrate has two more or less
equally reactive hydrogen sites, which of course may be
a complicating factor, even if this is reportedly not the
case with other closely related allyl ether substrates.19
We could verify an earlier observation9 that methine
hydrogens are less reactive than methylene hydrogens
in some carbonyl ene-reactions (Scheme 3, eq 3).
It seems reasonable to take into account the coordina-
tion of the Lewis acid to the allylic ether oxygen. In the
TIPS ether it is likely that the nucleophilicity of the
silicon-carrying oxygen is lowered both due to steric
hindrance and to the low availability of the oxygen
(17) Audran, G.; Monti, H.; Le´andri. G.; Monti, J .-P. Tetrahedron
Lett. 1993, 34, 3417-3418.
(18) Wennerberg, J .; Eklund, L.; Polla, M.; Frejd, T. Chem. Commun.
1997, 445.
(19) Mikami, K.; Shimizu, M.; Nakai, T. J . Org. Chem. 1991, 56,
2952-2953.
(20) Ikura, K.; Ryu, I.; Kambe, N.; Sonoda, N. J . Am. Chem. Soc.
1992, 114, 1520-1521.
(21) Salomon, M. F.; Pardo, S. N.; Salomon, R. G. J . Am. Chem. Soc.
1984, 106, 3797-3802.

Lewis Acid Mediated Glyoxylate Ene-Reaction
J . Org. Chem., Vol. 62, No. 25, 1997 8737
n-electrons for complexation.22-24 Thus, the effective bulk
at the oxymethylene hydrogens would be smaller in the
TIPS ether 2d as compared with the corresponding
benzyl (2b), PMB (2a ), and ethyl (2e) ethers. As a
consequence the oxymethylene hydrogens would be more
available in 2d . In the less sterically demanding 2e the
bulk of the Lewis acid complex is probably intermediate
at the region of the oxymethylene hydrogens, making the
methine and the methylene hydrogens equally reactive.
If this reasoning is correct, the coordination of the
Lewis acid at the allylic ether oxygen of the PMB ether
2a builds up too much steric hindrance for the oxymeth-
ylene hydrogens to be available for the reaction. Appar-
ently, allylic ethers of the general stucture 2 are par-
ticularly difficult ene components, which may give regio-
unselective reactions as well as the 1,4-rearrangement
mentioned.
shown that aldehyde trimers revert to monomer in the
presence of protic acids25 little is known about glyoxylates
in this respect. One may speculate that, e.g., trimeric
cyclic glyoxylates, having many electron rich oxygens,
may trap the Lewis acid by forming less reactive com-
plexes. We therefore argued that the use of glyoxylates
of sufficient bulk to increase the chances that the
monomer is present under the reaction conditions would
be desirable.
Despite that its drawbacks were essentially as de-
scribed for 8-phenylmenthyl glyoxylate 1c still gave the
reproducibly highest yields (entries 13 and 14). We
therefore reasoned that a structurally similar ester unit,
generated by removing two of the methylene groups of
the cyclohexane ring of 1c to create the siamyl group,
would be worth testing. The siamyl group has the
branching points as in the 2-methylcyclohexyl group but
is of course more flexible.
Glyoxyla tes. As is seen in Table 1 (entries 6, 7, 11,
17, 18) ethyl glyoxylate (1d ) was inferior in the Lewis
acid induced ene-reactions with 2a as were the n-butyl,
isopropyl, and methyl glyoxylates (not shown). It should
also be mentioned that the thermal noncatalyzed reaction
did not give an ene product at all. The more bulky
glyoxylates such as 1 (R1 ) 8-phenylmenthyl, 2-phenyl-
cyclohexyl,16 2-methylcyclohexyl (entries 13, 14) or cy-
clohexyl (entry 12, 16) gave on the other hand repeatedly
moderate to good yields of the corresponding ene prod-
ucts. However, there were serious drawbacks by using
these bulky glyoxylates. First, taking the 8-phenylmen-
thyl case as an example, the yields dropped drastically
when the scale was increased to more than ca. 1 gram
of substrate; second, the bulk of the ester function
prevented further reaction at the carbonyl carbon of the
ene product 3 by the enolate of ethyl acetate, and
therefore the ester function had to be replaced by a
smaller one; third, the asymmetric auxiliary by necessity
must be present in stoichiometric amounts, and fourth,
the resulting ene products 3, carrying bulky ester groups,
were not stable enough to make the method reliable. As
we earlier observed, these compounds readily decomposed
on standing or on contact with silica gel used for their
purification,16 while the ene products carrying smaller
ester groups (Me and Et) were more stable. Details of
this decomposition have not been investigated. Unfor-
tunately, only traces of ene product could be detected in
the (S)-BINOL-TiCl2 catalyzed reaction between 2a and
ethyl glyoxylate 1d .16 Thus, it was necessary to find a
more suitable glyoxylate.
We were very satisfied to find that 1a (5 equiv) gave
reliably very high yields of 3a using 2 equiv of (iPrO)2TiCl2
as a catalyst and 1 equiv of the allylic PMB-ether 2a
(entry 1). The reaction gave consistently 97-99% yields
of 3a in a 1 mmol scale. Other stoichiometries gave lower
yields, and in the presence of only 0.1 equiv of the Lewis
acid there was essentially no reaction with 1a (entry 4),
and a very inefficient reaction with 1d (entry 7). As seen
in Table 1 the 1:1:1 ratio was not optimal but was used
for screening purposes since it gave yields high enough
to be of diagnostic value. In large scale experiments (250
mmol of 2a ) we used 3 equiv of the glyoxylate, which gave
80-85% yields of 3a , repeatedly. The excess of the
glyoxylate 1a was recovered by distillation at reduced
pressure and could be reused several times. Even the
previously unreactive benzyl protected ether 2b gave a
product with 1a (entry 8) using the 5:2:1 conditions that
by TLC analysis indicated it to be the expected one
(characteristic green spot by anisaldehyde spray). How-
ever, it could not be obtained sufficiently pure by silica
gel chromatography to give an interpretable NMR spec-
trum.
A reasonable explanation for the “siamyl effect” on this
ene reaction has not yet been found. We have hitherto
only been able to dismiss some possible mechanistic
details, such as ester exchange between the glyoxylate
and the titanium isopropoxides, product catalysis, and
induction period. Nor has any NMR spectroscopic evi-
dence been obtained of a complex between the Lewis acid
and the glyoxylates.
NMR spectra of a series of glyoxylates in CDCl3 showed
that the free aldehyde content was negligible for the
lower alkyl glyoxylates but increased with larger ester
groups. For example, methyl and ethyl glyoxylate had
a free aldehyde content of less than 1%, while butylgly-
oxylate had 21%, cyclohexyl 22%, 2-methylcyclohexyl
41%, and 8-phenylmentyl glyoxylate >98%. A plausible
explanation for this is that the lower alkyl glyoxylates
tend to form trimers and hydrates to a larger extent than
the more bulky ones. Mechanistically, the ene-reaction
requires that the free aldehyde is the reacting species,
and consequently the larger glyoxylates were expected
to be better than the smaller ones. Even if it has been
Initially, we believed that the quality and/or the mode
of preparation of the glyoxylates were critical, but this
turned out to be not that important for 1a . Our preferred
synthesis of glyoxylates is the periodate cleavage of of
tartrates followed by distillation from P2O5.26 Siamyl
glyoxylate can be stored at 22 °C for several weeks
without loosing its efficiency in the ene reaction. Its
NMR spectrum revealed the presence of substantial
amounts (37%) of the free aldehyde.
Ap p lica tion to th e Syn th esis of a Ta xol A-Rin g
Su bu n it. We earlier found that the 4S,6R and the 4S,6S
isomers of epoxyallylsilane 12 (Scheme 4) underwent
different reactions when treated with BF3‚Et2O. The
4S,6R isomer cyclized to give compound (+)-13 while the
4S,6S isomer rearranged to ketone (-)-14.27,28 It would
(22) Kahn, S. D.; Keck, G. E.; Hehre, W. J . Tetrahedron Lett. 1987,
28, 279-280.
(23) Keck, G. E.; Castellino, S. Tetrahedron Lett. 1987, 28, 281-
284.
(25) Bo¨rjesson, L. Dissertation, Uppsala University, 1994.
(26) Kelly, R. K.; Schmidt, T. E.; Haggerty, J . G. Synthesis 1972,
544-545.
(24) Shambayati, S.; Blake, J . F.; Wierschke, S. G.; J orgensen, W.
L.; Schreiber, S. L. J . Am. Chem. Soc. 1990, 112, 697-703.

8738 J . Org. Chem., Vol. 62, No. 25, 1997
Wennerberg et al.
Sch em e 4a
Due to the relatively high cost of optically active BINOL,
large scale synthesis of the taxol A-ring subunit (+)-13
was not as attractive as we first imagined. As already
mentioned, the Ti-BINOL catalyst did not produce any
ene product (essentially no reaction took place) together
with 2a and ethyl glyoxylate 1d despite several varia-
tions of the reaction conditions.16
In conclusion, we solved the problem with the erratic
glyoxylate reaction of 2a by using siamyl glyoxylate 1a
together with (iPrO)2TiCl2 as a Lewis acid. The now
reliable and high yielding synthesis of the ene-product
3a made it practical to synthesize the optically active
taxol A-ring derivative (+)-13 in only eight steps from
PMB ether 2a (Scheme 4). Optically active (+)-3a could
be obtained by the use of 2 equivalents of BINOL‚Ti(OiPr)2
as promoter in the siamyl glyoxylate ene reaction. We
previously reported the syntesis of a seco taxane deriva-
tive starting from (+)-13,31 and the ample supply of this
compound allows us to resume our work toward the
synthesis of taxanes.
a
Reactions conditions: (a) TBDMSCl, imidazole, 95%; (b)
Exp er im en ta l Section
(TMS)2NLi, TMEDA, EtOAc, -78 °C to 0 °C, 95%; (c) KOtBu,
ClPO3Et2, 60%; (d) TMSCH2MgCl, Ni(acac)2, 65%; (e) DDQ, 85%;
(f) D-(-)-DIPT, Ti(OiPr)4, TBHP, -40 °C, 87%; (g) BF3‚Et2O, 0 °C,
36%.
Gen er a l. GC analyses were performed with a DBwax
(J &W Scientific) capillary column (30 m, 0.25 mm i.d., 0.25
µm stationary phase) and with a Beta-DEX 120 chiral column
for determination of enantiomeric compositions. NMR spectra
were recorded at 300 MHz using CDCl3 (CHCl3 δ 7.26 (1H)
and 77.0 (13C)) as solvent. Chromatographic separations were
performed on Matrex Amicon normal phase silica gel 60
(0.035-0.070 mm), and for thin-layer chromatography we used
Merck precoated TLC-plates Silica gel 60 F-254, 0.25 mm.
After elution, the TLC-plates were sprayed with a solution of
p-methoxybensaldehyde (26 mL), glacial acetic acid (11 mL),
concentrated sulfuric acid (35 mL), and 95% ethanol (960 mL),
and the compounds were visualized upon heating. Unless
otherwise stated, reagent grade chemicals were used, and the
reactions were perfomed in septum-capped, oven-dried flasks
under an atmospheric pressure of argon. THF was distilled
under N2 from sodium-benzophenone ketyl, and CH2Cl2 was
distilled from P2O5 prior to use. Molecular sieves were
activated at 400 °C for 3 h and all organic extracts were dried
using MgSO4. The cyclohexyl32 and ethyl26 glyoxylates, 1b and
1d , respectively, were prepared according to literature proce-
dures as was 2-isopropyl-2-propenol.33
3-Meth yl-2-bu tyl Glyoxyla te (1a ). The procedure of Kelly
et al.26 was adopted. Paraperiodic acid (11.4 g, 50.0 mmol)
was added in portions during 1 h to an ice-cooled solution of
bis(3-methyl-2-butyl) tartrate (14.5 g, 50.0 mmol) in ether (300
mL). The mixture was stirred for another 1 h and was then
decanted from the solid residue. The ether solution was dried
with 4A molecular sieves and filtered. Evaporation of the
solvent at reduced pressure gave a pale yellow oil (13.5 g)
which was distilled to give 1a (10.8 g, 75%) as a colorless oil:
bp 85-88 °C/20 mbar; IR: (film) 3430, 1735 cm-1; 1H NMR δ
9.41 (s, <1H), 5.30 (m, <1H), 4.86 (m, 1H), 3.60 (broad s, exch.
with D2O), 1.87 (m, 1H), 1.25 (d, J ) 6.5 Hz, 3H), 0.91 (d, J )
6.9 Hz, 6H); 13C NMR δ 184.3, 156.4, 87.2, 32.7, 17.9, 16.8.
Anal. HRMS Calcd for C7H16O3N [M + NH4]: 162.1130.
Found: 162.1129.
therefore be possible to use racemic 11 and then intro-
duce the epoxide stereoselectively by the use of the
Sharpless asymmetric epoxidation reaction for the syn-
thesis of optically active 12.29 Thus, racemic 9 was
treated with the enolate of ethyl acetate, followed by
diethyl chlorophosphate to give phosphoenolate 10. We
were delighted to find that the enolate reaction resulted
in a very high yield (95%) of the chain extended â-keto
carboxylate.
Subsequent Ni-catalyzed coupling with TMSCH2MgCl
and deprotection with DDQ)16 gave racemic 11. Sharp-
less epoxidation29 of 11 using D-(-)-diisopropyl tartrate
(D-(-)-DIPT) gave a diastereomeric mixture of 12, which
was treated with BF3‚Et2O to give the cyclized derivative
(+)-13 and the rearranged derivative (-)-14. These
compounds were easily separated by silica gel chroma-
tography. The absolute configuration of the major enan-
tiomer of (+)-13 was confirmed by comparison with an
authentic sample.27 Based on the optical rotation and
on NMR analysis of the Mosher ester of (+)-13, the ee
was determined to be 80-85%. Although this eight-step
route to (+)-13 gave material with lower enantiomeric
purity than was previously obtained via the more tedious
routes,16,27 we believe that it is advantageous especially
on a larger scale.
Optically active 3a would be the ideal starting material
for the construction of 13 and should be obtainable by
the use of the asymmetric catalytic carbonyl-ene reaction
as described by Mikami et al. who used 5 mol % of
optically active BINOL‚TiCl2 as a catalyst.11,30 Satisfac-
torily, application of this complex to 1a and 2a gave (+)-
3a of 95% ee in 95% yield (Table 1, entry 5). However,
it was necessary to use 2 equiv of the Ti-BINOL
complex; a catalytic amount (5 mol %) gave no conversion.
Asym m etr ic En e Rea ction Ca ta lyzed by (S)-BINOL‚
TiCl2. P r ep a r a tion of 3′-Meth yl-2′-bu tyl (2S)-2-Hyd r oxy-
4-[[(4-m e t h oxyb e n zyl)oxy]m e t h yl]-5-m e t h yl-4-h e xe n -
oa te ((+)-3a ). A solution of (i-PrO)2TiCl2 (237 mg, 1.00 mmol)
in CH2Cl2 (2 mL) was added to a mixture of (S)-BINOL (286
mg, 1.00 mmol) and pulverized activated 4A molecular sieves
(4 g) in CH2Cl2 (25 mL) at 0 °C under an argon atmosphere.
(27) Pettersson, L.; Frejd, T.; Magnusson, G. Acta Chem. Scand.
1993, 47, 196-207.
(28) Pettersson, L.; Frejd, T.; Magnusson, G. Tetrahedron Lett. 1987,
(31) Pettersson, L.; Frejd, T. J . Chem. Soc., Chem. Commun. 1993,
28, 2753-2756.
1823-1825.
(29) Hanson, R. M.; Sharpless, K. B. J . Org. Chem. 1986, 51, 1922-
(32) Ferna´ndez, F.; Garcia, G.; Rodriguez, J . E. Synth. Commun.
1925.
1990, 20, 2837-2847.
(30) Mikami, K.; Terada, M.; Nakai, T. J . Am. Chem. Soc. 1990, 112,
3949-3954.
(33) Green, M. B.; Hickinbottom, W. J . J . Chem. Soc. 1957, 3262-
3270.

Lewis Acid Mediated Glyoxylate Ene-Reaction
J . Org. Chem., Vol. 62, No. 25, 1997 8739
After 15 min, the temperature was raised to 20 °C, and the
stirring was continued for 1 h. The reaction mixture was then
cooled to -78 °C, and 2a (110 mg, 0.50 mmol) in CH2Cl2 (1
mL) was added followed by glyoxylate 1a (362 mg, 2.50 mmol)
in CH2Cl2 (1 mL). After 48 h at -17 °C, the mixture was
diluted with ether and washed with a saturated aqueous
NaHCO3 and brine. The organic extract was dried, and then
the solvent was evaporated at reduced pressure. Chromatog-
raphy of the residue (heptane-EtOAc, 3:1) gave (+)-3a (173
Gen er a l P r oced u r e for En e Rea ction s Ca ta lyzed by
Sn Cl4. P r ep a r a t ion of Cycloh exyl 2-H yd r oxy-4-[[(4-
m eth oxyben zyl)oxy]m eth yl]-5-m eth yl-4-h exen oa te (3b)
a n d Der iva t iza t ion t o It s Cor r esp on d in g E t h yl E st er .
A solution of 2a (220 mg, 1.00 mmol) in CH2Cl2 (5 mL) was
added to a solution of 1b (156 mg, 1.00 mmol) in CH2Cl2 (5
mL) at -78 °C under an argon atmosphere, followed by
dropwise addition of SnCl4 (118 µL, 1.00 mmol). The dark red
reaction mixture was stirred for 5 min and was then diluted
with ether and washed with saturated aqueous NaHCO3 and
brine. The organic extract was dried, and then the solvent
was evaporated at reduced pressure which gave 280 mg of
opalescent oil. The product decomposed on silica and could
therefore not be properly purified for spectroscopic analysis.
Instead it was transformed to its corresponding ethyl ester in
the following manner. Crude 3b (280 mg, 0.74 mmol) was
hydrolyzed by refluxing for 8 h in a mixture of THF (5 mL),
MeOH (10 mL), and 1 M NaOH (5 mL). The mixture was
extracted with heptane whereafter the aqueous phase was
acidified with 1 M HCl (to pH ) 2) and extracted with ether.
The collected ethereal extract was washed with brine (15 mL)
and dried, and the solvent was evaporated at reduced pressure
to give the crude hydroxy acid, which was dissolved in benzene
(5 mL) containing DBU (0.37 mL, 5.0 mmol) and EtBr (0.40
mL, 10 mmol). This mixture was refluxed for 5 h, diluted with
mg, 95%): [R]21 +0.2° (c 2, CDCl3). The ee was determined
D
to be 95% by GC and 1H NMR analysis of the corresponding
Mosher ester.34
Gen er a l P r oced u r e for En e Rea ction s Ca ta lyzed by
(i-P r O)2TiCl2. P r ep a r a tion of 3′-Meth yl-2′-Bu tyl-2-h y-
d r oxy-4-[[(4-m eth oxyben zyl)oxy]m eth yl]-5-m eth yl-4-h ex-
en oa te ((()-3a ). The reaction conditions were varied accord-
ing to Table 1. Glyoxylate 1a (1.81 g, 12.5 mmol) in CH2Cl2
(2.5 mL) was added to a solution of (i-PrO)2TiCl2 (1.18 g, 5.00
mmol) in CH2Cl2 (10 mL) at -78 °C and under Ar followed by
addition of PMB-ether 2a (0.550 g, 2.50 mmol) in CH2Cl2 (2.5
mL). The reaction mixture was kept at -17 °C for 48 h and
was then diluted with ether and washed with saturated
aqueous NaHCO3 and brine. The organic extract was dried,
and then the solvent was evaporated at reduced pressure.
Chromatography of the residue (heptane-EtOAc, 3:1) gave 3a
(0.90 g, 99%) as a mixture of diastereomers: IR: (film) 3450,
1735 cm-1; 1H NMR δ 7.28 (d, 2H, J ) 8.7 Hz), 6.88 (d, 2H, J
) 8.7 Hz), 4.80 (m, 1H), 4.45 (s, 2H), 4.25-4.05 (m, 2H), 3.98-
3.89 (m, 1H), 3.80-3.68 (m, 1H), 3.79 (s, 3H), 2.73 (m 1H),
2.45 (m, 1H), 1.91-1.71 (m, 1H), 1.75 (s, 3H), 1.72 (s, 3H), 1.18
(dd, J ) 6.4 Hz, J ) 3.1 Hz, 3H), 0.90 (dd, J ) 4.8 Hz, J ) 2.0
Hz, 6H); 13C NMR δ 174.6, 159.3, 142.7, 135.1, 130.0, 130.0,
129.5, 125.2, 125.2, 113.8, 76.3, 76.3, 72.2, 70.8, 70.7, 69.6, 69.6,
55.3, 37.1, 36.9, 32.6, 32.6, 21.0, 20.7, 18.1, 18.0, 16.7, 16.6.
Anal. HRMS Calcd for C21H33O5 [M + H] 365.2328. Found:
365.2330.
ether and washed with 1 M HCl, saturated NaHCO3
solution,
and brine. The ethereal extract was dried, and the solvent
was then evaporated at reduced pressure. Chromatography
of the residue (heptane-EtOAc, 1:1) gave 3b (187 mg, 58%).
NMR data were identical with those earlier reported.16
3′-Met h yl-2′-b u t yl 2-[(ter t-b u t yld im et h ylsilyl)oxy]-4-
[[(4-m eth oxyben zyl)oxy]m eth yl]-5-m eth yl-4-h exen oate (9).
A solution of 3a (1.00 g, 2.70 mmol), imidazole (0.43 g, 6.30
mmol), and TBDMSCl (0.475 g 3.15 mmol) in DMF (7 mL) was
stirred overnight under Ar. The reaction mixture was then
diluted with ether and washed with 1 M HCl, saturated
aqueous NaHCO3, and brine. The organic extract was dried,
and the solvent was evaporated at reduced pressure. Chro-
matography of the residue (heptane-EtOAc, 3:1) gave 9 (1.28
This procedure was used for the preparation of the following
compounds in a 1 mmol scale.
Eth yl 2-Hydr oxy-4-(2-pr opyl)-5-[(tr iisopr opylsilyl)oxy]-
4-p en ten oa te (6a ). Yield: 241 mg (67%) starting from 2d .
IR: (film) 1740 cm-1; 1H NMR δ 4.95 (s, <1H), 4.84 (s, <1H),
4.21 (m, 4H), 2.98 (m, 1H), 2.83 (m, 1H), 1.78 (m, 1H), 1.41 (s,
3H), 1.32 (s, 3H), 1.29 (t, J ) 7.1 Hz, 3H), 1.11-1.05 (m, 3H),
1.05 (s, 18H); 13C NMR δ 172.4, 153.7, 142.7, 104.6, 83.5, 74.1,
61.1, 36.5, 28.1, 17.7, 14.2, 12.3. Anal. Calcd for C19H38O4Si:
358.2539. Found: 358.2533
g, 92%) as a mixture of diastereoisomers: IR: (film) 1740 cm-1
;
1H NMR δ 7.24 (d, J ) 8.7 Hz, 2H), 6.85 (d, J ) 8.7 Hz, 2H),
4.80 (m, 1H), 4.39 (s, 2H), 4.26 (m, 1H), 4.00 (s, 2H), 3.79 (s,
3H), 2.57 (m, 1H), 2.34 (s, 1H), 1.95-1.75 (m, 1H), 1.74, 1.69
(2s, 6H), 1.18 (dd, J ) 6.4 Hz, J ) 1.4 Hz, 3H), 0.90 (m, 6H),
0.85 (s, 9H), 0.09, -0.02 (2s, 6H); 13C NMR δ 173.8, 129.3,
113.9, 75.8, 72.1, 72.0, 69.5, 55.1, 36.7, 36.6, 32.7, 32.6, 25.7,
21.2, 20.6, 18.3, 18.1, 18.0, 17.9, 16.7, 16.6. Anal. Calcd for
C27H46O5Si: C, 67.7; H 9.7. Found: C, 67.4; H, 9.3. HRMS
Calcd for C27H47O5Si [M + H]: 479.3192. Found: 479.3200.
Eth yl (2Z)-4-[(ter t-Bu tyldim eth ylsilyl)oxy]-3-[(dieth oxy-
p h osp h or yl)oxy]-6-[[(4-m et h oxyb en zyl)oxy]m et h yl]-7-
m eth yl-2,6-octa d ien oa te (10). n-BuLi (1.6 M solution in
hexane, 2.2 mL, 3.5 mmol) was added to a solution of
hexamethyl-disilazane (0.80 mL, 3.8 mmol) in THF (3 mL) at
0 °C under an argon atmosphere. The mixture was cooled at
-78 °C, and EtOAc (0.17 mL, 1.8 mmol) was then added
dropwise. After stirring this mixture for 10 min a solution of
9 (0.57 g, 1.2 mmol) and TMEDA (0.52 mL, 3.5 mmol) in THF
(0.75 mL) was added. The cooling bath was removed, and the
mixture was stirred for 30 min at rt. Ether (15 mL) was added,
and the resulting mixture was washed with 1 M HCl and
brine. The organic layer was dried, and the solvent was
evaporated at reduced pressure. The crude â-keto ester was
dissolved in THF (3 mL) followed by addition of t-BuOK (136
mg, 1.2 mmol). The resulting mixture was stirred for 3 min
followed by addition of diethyl chlorophosphate (0.21 mL, 1.7
mmol). Stirring was continued for another 15 min, and then
ether (15 mL) was added followed by washing with saturated
NH4Cl solution and water. The etheral extract was dried, and
the solvent was evaporated under reduced pressure. Chro-
matography of the residue (heptane-EtOAc, 1:1) gave 10 (0.42
g, 58%): 1H NMR and 13C NMR are identical with those earlier
reported.16
E t h yl 5-E t h oxy-2-h yd r oxy-4-(2-p r op yl)-4-p en t en oa t e
(6b Z:E 1:1). Yield 84 mg (36%) starting from 2e. Rf ) 0.21;
1
IR (film) 1735 cm-1; H NMR δ 5.87 (s, <1H), 5.41 (s, <1H),
4.31 (q, J ) 7.1 Hz, 2H), 4.27-4.07 (m, 1H), 3.92-3.71 (m,
1H), 3.51 (q, J ) 7.0 Hz, 2H), 2.85-2.68 (m, 1H), 2.48-2.32
(m, 1H), 1.81 (m, 1H), 1.33 (t, J ) 7.1 Hz, 3H), 1.24 (t, J ) 7.0
Hz, 3H), 0.92 (d, J ) 6.8 Hz, 6H); 13C NMR δ 172.5, 153.7,
142.7, 70.6, 66.0, 61.1, 37.3, 20.9, 20.7, 15.0, 14.2. Calcd for
C12H22O4: 230.1518. Found: 230.1520.
In this experiment the following was also formed:
E t h yl 2-H yd r oxy-4-(et h oxym et h yl)-5-m et h yl-4-h ex-
en oa te (3h ). Yield: 86 mg (37%); Rf ) 0.43; IR: (film) cm-1
:
1
1740; H NMR δ 4.20 (q, J ) 7.2 Hz, 2H), 4.12 (m, 1H), 4.07
(s, 2H), 3.86 (m, 1H), 3.51 (q, J ) 7.0 Hz, 2H), 2.76 (m, 1H),
2.42 (m, 1H), 1.76 (s, 3H), 1.74 (s, 3H), 1.30 (t, J ) 7.2 Hz,
3H), 1.24 (t, J ) 7.0 Hz, 3H); 13C NMR δ 174.1, 135.2, 124.9,
70.8, 70.6, 65.9, 61.0, 37.1, 20.8, 20.4, 14.8, 14.4. Anal. Calcd
for C12H22O4 230.1518. Found: 230.1511.
Eth yl 2-Hyd r oxy-3-(6-m eth yl-1-cycloh exen yl)p r op a n -
oa te (7, m ixtu r e of d ia ster eom er s). Yield: 176 mg (83%)
1
starting from 8. IR: (film) 3470, 1730 cm-1; H NMR δ 5.51
(s, 1H), 4.24 (q, J ) 7.1 Hz, 2H), 4.29-4.20 (m, 1H), 2.68-
2.60 (d, J ) 13.0 Hz, 1H), 2.52-2.30 (m, 1H), 2.23-2.07 (m,
2H), 2.03-1.92 (m, 2H), 1.77-1.27 (m, 4H), 1.25 (t, J ) 7.1
Hz, 3H), 1.07 (d, J ) 7.0 Hz, 3H); 13C NMR δ 174.9, 137.3,
125.6, 69.1, 61.5, 40.3, 31.3, 30.8, 25.7, 19.6, 19.5, 14.2. Anal.
Calcd for C12H21O3 [M + H]: 213.1491. Found: 213.1498
Eth yl (4RS,6R)-(2Z)-4-[(ter t-Bu tyld im eth ylsilyl)oxy]-
6,7-ep oxy-6-(h yd r oxym eth yl)-7-m eth yl-3-[(tr im eth ylsilyl-
)m eth yl]-2-octen oa te (12). The same two-step procedure as
(34) Ward, D. E.; Rhee, C. K. Tetrahedron Lett. 1991, 32, 7165-
7166.

8740 J . Org. Chem., Vol. 62, No. 25, 1997
Wennerberg et al.
described earlier16,27 was used for the synthesis of 12 from 10
including nickel catalyzed coupling of 10 and [(trimethylsilyl-
)methyl]magnesium chloride to give 11, followed by Sharpless
asymmetric epoxidation using D-(-)-diisopropyl tartrate. After
chromatography (heptane:EtOAc, gradient 10:1 f 5:1) 12 (11.0
g, 87%) was isolated as a mixture of diastereomers (Rf
(heptane:EtOAc 3:1) 0.30) as indicated by the 1H NMR
spectrum, which showed two sets of signals as earlier described
for the separate compounds.27
reported. The Mosher ester gave 1H signals at 5.0 and 5.3
ppm suitable for determination of the ee, which with this
method was found to be 85%. Compound (+)-13 was eluted
before (-)-14. A small sample of pure (-)-14 was obtained;
[R]20 ) -21 (c 1.10, CDCl3), having 1H NMR data as desci-
D
bed.27
Ack n ow led gm en t. We thank the Swedish Natural
Science Research Council, the Knut and Alice Wallen-
berg Foundation, and the Crafoord Foundation for
financial support. Wacker AG is acknowledged for a
generous gift of TBDMSCl.
Eth yl (1R,3S,5S)-5-[(ter t-Bu tyld im eth ylsilyl)oxy]-3-h y-
d r oxy-3-(h yd r oxym et h yl)-2,2-d im et h yl-6-m et h ylen ecy-
cloh exa n eca r boxyla te (13). A solution of BF3‚OEt2 (3.4 mL,
27.6 mmol) in CH2Cl2 (12 mL) was added to a solution of 12
(11.0 g, 26.3 mmol) in dry CH2Cl2 (1300 mL) under an argon
atmosphere at 0 °C. The reaction mixture was then stirred
for 10 min followed by addition of aqueous NaHCO3 (sat., 220
mL) and ether (300 mL). The organic phase was washed with
brine and dried. Evaporation of the solvent at reduced
pressure followed by column chromatography (heptane: EtOAc,
gradient 10:1 f 3:1) of the crude product gave (+)-13 (3.0 g,
36%) as a clear oil. [R]21D ) +64 (c 2.39, CDCl3) corresponding
Su p p or tin g In for m a tion Ava ila ble: 13C NMR spectra for
all compounds with high-resolution mass spectra and experi-
mental details for bis(1,2-dimethylpropyl)tartrate, bis(2-me-
thylcyclohexyl)tartrate, 1c, 2b-e, and large scale preparation
of 3a (15 pages). This material is contained in libraries on
microfiche, immediately follows this article in the microfilm
version of the journal, and can be ordered from the ACS; see
any current masthead page for ordering information.
to 80% ee based on the literature value27 [R]21 +81 (c 1.10,
D
CDCl3). Its NMR data were identical with those earlier
J O971082L