Stereoselective Anti Aldol Reactions of Erythrulose Derivatives. Functionalized Chiral d3 and d4 Synthons

Article abstract of DOI:10.1021/jo0356356

An improved procedure for the synthesis of anti aldols from protected erythrulose derivatives is reported. The preparation of functionalized d ³ and d⁴ synthons with various stereochemical arrays by means of this methodology is descr

Full text of DOI:10.1021/jo0356356

Ster eoselective An ti Ald ol Rea ction s of Er yth r u lose Der iva tives.
F u n ction a lized Ch ir a l d 3 a n d d 4 Syn th on s
J uan Murga,† Purificacio´n Ruiz,† Eva Falomir,† Miguel Carda,*,† Gabriel Peris,‡ and
J . Alberto Marco*,§
Departamento de Qu´ımica Inorga´nica y Orga´nica, and Servicios Centrales de Instrumentacio´n,
Universidad J aume I, E-12080 Castello´n, and Departamento de Qu´ımica Orga´nica,
Universidad de Valencia, E-46100 Burjassot, Valencia, Spain
alberto.marco@uv.es
Received November 7, 2003
An improved procedure for the synthesis of anti aldols from protected erythrulose derivatives is
reported. The preparation of functionalized d3 and d4 synthons with various stereochemical arrays
by means of this methodology is described and subsequently applied to a stereoselective formal
synthesis of the natural metabolite goniothalesdiol.
SCHEME 1. Ald ol Rea ction s of Er yth r u lose
Der iva tives 1 a n d 3 (TBS ) ter t-Bu tyld im eth ylsilyl;
d r ) Dia ster eom er ic Ra tio)
In tr od u ction
The aldol reaction1 has proven to be a powerful and
general method for the stereocontrolled construction of
carbon-carbon bonds and has relevant application in the
synthesis of natural polyoxygenated molecules such as
macrolide and polyether antibiotics.2 Our current interest
in the development of erythrulose3 as a useful chiral C4
building block for the stereocontrolled construction of
polyfunctionalized structures has prompted us to inves-
tigate the enolization of protected derivatives thereof and
the subsequent addition of the resulting enolates to
aldehydes. We have reported that L-erythrulose acetals
of the general formula 1 (Scheme 1, protecting group P
) triethylsilyl, TES; tert-butyldimethylsilyl, TBS; or tert-
butyldiphenylsilyl, TPS), readily prepared in two steps
from L-erythrulose,4 can be transformed into boron eno-
lates provided that chlorodicyclohexylborane (Chx2BCl)
is used as the enolization reagent.5 These boron enolates
were then allowed to react with a range of achiral
aldehydes to yield aldol adducts of the general formula
2 with a high degree of syn 1,2- and 1,3-induction (i.e.,
the 2,4-syn/4,5-syn relationship in 2).6
† Departamento de Qu´ımica Inorga´nica y Orga´nica, Universidad
J aume I.
‡ Servicios Centrales de Instrumentacio´n, Universidad J aume I.
§ Departamento de Qu´ımica Orga´nica, Universidad de Valencia.
(1) (a) Evans, D. A.; Nelson, J . V.; Taber, T. R. Top. Stereochem.
1982, 13, 1-115. (b) Mukaiyama, T. Org. React. 1982, 28, 203-331.
(c) Masamune, S.; Choy, W.; Petersen, J . S.; Sita, L. R. Angew. Chem.,
Int. Ed. Engl. 1985, 24, 1-30. (d) Heathcock, C. H. In Asymmetric
Synthesis; Morrison, J . D., Ed.; Academic Press: Orlando, 1984; Vol.
3, pp 111-212. (e) Heathcock, C. H. Aldrichim. Acta 1990, 23, 99-
111. (f) Comprehensive Organic Synthesis; Trost, B. M., Fleming, I.,
Winterfeldt, E., Eds.; Pergamon Press: Oxford, 1993; Vol. 2. (g)
Mekelburger, H. B.; Wilcox, C. S. In ref 1f, pp 99-131. (h) Heathcock,
C. H. In ref 1f, pp 133-179 and 181-238. (i) Kim, B. M.; Williams, S.
F.; Masamune, S. In ref 1f, pp 239-275. (j) Rathke, M. W.; Weipert,
P. In ref 1f, pp 277-299. (k) Paterson, I. In ref 1f, pp 301-319. (l)
Franklin, A. S.; Paterson, I. Contemp. Org. Synth. 1994, 1, 317-338.
(m) Braun, M. In Houben-Weyl’s Methods of Organic Chemistry,
Stereoselective Synthesis; Helmchen, G., Hoffmann, R. W., Mulzer, J .,
Schaumann E., Eds.; Georg Thieme Verlag: Stuttgart, 1996; Vol. 3,
pp 1603-1666 and 1713-1735. (n) Mahrwald, R. Chem. Rev. 1999,
99, 1095-1120.
Prior to our research, only one case of a syn aldol
addition mediated by this reagent had been reported.
Paterson and co-workers described the use of an ethyl
ketone bearing an R-benzyloxy group, where a syn aldol
was formed with good diastereoselectivity.7 The authors
related this unanticipated syn bias to the formation of a
Z enolate instead of the expected E isomer, a feature
(4) (a) For an improved preparation of silylated L-erythrulose
acetonides 1 (P ) TES, TBS, TPS) from L-erythrulose hydrate, see:
Carda, M.; Rodr´ıguez, S.; Murga, J .; Falomir, E.; Marco, J . A.; Ro¨per,
H. Synth. Commun. 1999, 29, 2601-2610. (b) For the preparation of
protected D- and L-erythrulose derivatives using chiral precursors other
than erythrulose itself, see: Marco, J . A.; Carda, M.; Gonza´lez, F.;
Rodr´ıguez, S.; Murga, J . Liebigs Ann. Chem. 1996, 1801-1810.
(5) For a general review of boron aldol reactions, see: Cowden, C.
J .; Paterson, I. Org. React. 1997, 51, 1-200.
(6) Murga, J .; Falomir, E.; Gonza´lez, F.; Carda, M.; Marco, J . A.
Tetrahedron 2002, 58, 9697-9707. Aldol reactions of ketone 1 with
chiral aldehydes have also been investigated: Marco, J . A.; Carda, M.;
D´ıaz-Oltra, S.; Murga, J .; Falomir, E.; Roeper, H. J . Org. Chem. 2003,
68, 8577-8582.
(2) (a) Recent Progress in the Chemical Synthesis of Antibiotics;
Lukacs, G., Ohno, M., Eds.; Springer: Berlin, 1990. (b) Tatsuta, K. In
ref 2a, pp 1-38. (c) Blizzard, T., Fisher, M., Mrozik, H.; Shih, T. In ref
2a, pp 65-102. (d) Isobe, M. In ref 2a, pp 103-134. (e) Beau, J .-M. In
ref 2a, pp 135-182. (f) Yonemitsu, O.; Horita, K. In ref 2a, pp 447-
466. (g) Norcross, R. D.; Paterson, I. Chem. Rev. 1995, 95, 2041-2114.
(3) Marco, J . A.; Carda, M.; Gonza´lez, F.; Rodr´ıguez, S.; Castillo,
E.; Murga, M. J . Org. Chem. 1998, 63, 698-707 and references therein.
10.1021/jo0356356 CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/21/2004
J . Org. Chem. 2004, 69, 1987-1992
1987

Murga et al.
SCHEME 2. Ald ol Rea ction s of Er yth r u lose
attributed in turn to the deprotonation step taking place
in a chelate involving the boron, carbonyl oxygen, and
R-oxygen atoms.8 They further reasoned that replacement
of the benzyl by the electron-attracting benzoyl group
would make the R-oxygen atom more electron-poor and
thus prevent the formation of such a chelate, with
subsequent reversal of the stereochemical course to the
expected anti aldol formation via an E enolate. Their
conclusions were then borne out by experimental re-
sults.7a,b In view of this, we first checked whether a Z
boron enolate was being formed when ketone 1 was
treated with Chx2BCl. That was found indeed to be the
case.6 Then, we checked whether the introduction of
electron-withdrawing protective groups in 1 would change
the stereochemical course of the reaction. As a matter of
fact, it was found that, in line with Paterson’s reasonings,
aldol reactions with ketone 3, which bears two benzoyl
protecting groups, give rise stereoselectively to anti aldols
4 (relative configuration at C4-C5), most likely through
the corresponding E boron enolate. Quantum-mechanical
studies provided a theoretical basis for these results.6
These results are interesting from a synthetic point of
view, as the resulting aldols can be used for the synthesis
of various polyoxygenated molecules, even very complex
ones.9,10 However, ketone 3 has one drawback: its
synthesis requires eight steps from L-ascorbic acid.6 It
would be much more practical if it could be prepared from
L-erythrulose itself in fewer steps. However, the two
primary hydroxyl groups of this ketotetrose cannot be
efficiently differentiated from one another, this dif-
ferentiation being necessary if an orthogonality of the
protecting groups of the aldols is subsequently required.11
Since the key to having anti aldols seems to rely on the
presence of an R-acyloxy group, we wondered whether a
ketone with general structure 5 (Scheme 2, R ) acyl
group) would behave in the same way as 3. Ketones 5,
available from L-erythrulose, have two identical tert-
butyldimethylsilyl (TBS) groups at the primary hydroxyl
functions12 but, since one of them becomes secondary in
the final aldol, selective deprotection should be feasible.
Furthermore, the secondary hydroxyl group is acylated
and therefore equally amenable to selective cleavage. The
preparation of ketones 5, the investigation of their
behavior in aldol reactions, and the synthesis of poly-
oxygenated d3 and d4 synthons therefrom constitute the
Der iva tives 5a -c
TABLE 1. Yield s a n d Dia ster eom er ic Ra tios in Ald ol
Rea ction s of Keton es 5a -c
ketone
aldehyde
yielda (%)
drb
5a
5b
5c
5a
PhCHO
PhCHO
PhCHO
Me2CHCHO
86
77
82
87
92:8
90:10
92:8
85:15
a
b
Overall yield of the diastereomeric mixture. dr determined
by means of 1H and 13C NMR.
main contents of the present manuscript. In addition, a
formal, stereoselective synthesis of the natural metabolite
goniothalesdiol is described as a practical application of
this methodology.
Resu lts a n d Discu ssion
Ketones 5a -c bearing three different acyl groups,
which basically differred in their size, were readily
prepared from L-erythrulose in only two steps (Scheme
2). The aldol reactions of these three ketones with
benzaldehyde were first investigated. Gratifyingly, anti
aldols 6a -c were formed with good stereoselectivity,
similar to that observed with ketone 3.13 As seen in Table
1, the size of the acyl group does not play a decisive role
in the outcome of the process. However, yields and
diastereoselectivities were slightly higher with the ben-
zoylated derivative 5a . Since the benzoyl group addition-
ally provided UV-fluorescence in TLC plates, only ketone
(7) (a) Paterson, I.; Wallace, D. J .; Vela´zquez, S. M. Tetrahedron
Lett. 1994, 35, 9083-9086. (b) Paterson, I.; Wallace, D. J .; Cowden, C.
J . Synthesis 1998, 639-652. After our results, a further example of
preparation of syn aldols using Chx2BCl appeared in the literature:
Galobardes, M.; Gasco´n, M.; Mena, M.; Romea, P.; Urp´ı, F.; Vilarrasa,
J . Org. Lett. 2000, 2, 2599-2602.
(8) Recent experimental findings of our group indicate that forma-
tion of syn aldols from R-oxygenated ketones may occur even in cases
where the formation of chelates is unlikely: Murga, J .; Falomir, E.;
Carda, M.; Gonza´lez, F.; Marco, J . A. Org. Lett. 2001, 3, 901-904.
(9) (a) Falomir, E. Ph.D. Thesis, University of Castello´n, Spain, 1998.
(b) Carda, M.; Murga, J .; Falomir, E.; Gonza´lez, F.; Marco, J . A.
Tetrahedron 2000, 56, 677-683.
(10) Forsyth and co-workers have recently used ketone 3 in their
synthetic approach to azaspiracid: Forsyth, C. J .; Hao, J .; Aiguade´, J .
Angew. Chem., Int. Ed. 2001, 40, 3663-3667.
(11) For this reason, tri-O-benzoyl-L-erythrulose was not considered
for the purposes described here.
(12) Ketones 5 with triethylsilyl groups were not obtained in good
yield, due to the formation of appreciable amounts of trisilylated
erythrulose in the silylation step. Furthermore, ketones 5 bearing tert-
butyldiphenylsilyl groups were found less reactive and less diastereo-
selective as well. Thus, only ketones 5 with tert-butyldimethylsilyl
groups are discussed here.
(13) The configurations of the major aldols and of its reduction
products were established with the aid of chemical correlations
described in the Supporting Information. In the case of aldol 6a , further
comformation was provided by the conversion into goniothalesdiol
precursor 23 (see Scheme 6). The configuration of aldol 7 was secured
by means of an X-ray diffraction analysis. Crystallographic data of this
compound (excluding structure factors) have been deposited at the
Cambridge Crystallographic Data Centre as Supporting Information
with reference CCDC-223372. Copies of the data can be obtained, free
of charge, on application to CCDC, 12 Union Road, Cambridge, CB2
1EZ, UK [fax: +44(0)-1223-336033 or e-mail: deposit@ccdc.cam.ac.uk].
The minor aldols were found to be also anti according to NMR
examination of the aldol mixture (diagnostic value of J 4,5 ∼ 5-7 Hz
for anti aldols but e 2 Hz for syn aldols). The isolation of these minor
aldols, however, was not attempted.
1988 J . Org. Chem., Vol. 69, No. 6, 2004

Anti Aldol Reactions of Erythrulose Derivatives
SCHEME 3. F u n ction a lized d 3 a n d d 4 Ch ir a l
Syn th on s Gen er a ted fr om Syn Ald ols of
Er yth r u lose Der iva tives
SCHEME 4. Syn th esis of F u n ction a lized d 3 a n d d 4
Ch ir a l Syn th on s fr om An ti Ald ols of Er yth r u lose
Der iva tivesa
5a was used in subsequent work. Aldol reaction of the
latter with isobutyraldehyde gave aldol 7 with dr 85:15.13
While the aldol process was in this case somewhat less
diastereoselective than with ketone 3, this was amply
compensated by the much shorter route to 5a .
After having established the feasibility of anti aldol
reactions with ketone 5a , we set out to develop practical
applications for this synthetic methodology. In a previous
communication of our group, we showed that polyfunc-
tionalized d3 and d4 synthons (Scheme 3) could be easily
prepared from syn aldols generated from ketones 1.14 The
preparation of chiral synthons of the d3 class involve
oxidative cleavage of one C-C bond of the initial eryth-
rulose moiety. They thus preserve three of the four
carbons of the sugar as well as the two newly generated
stereogenic carbon atoms. Chiral synthons of the d4 class,
generated from aldols without C-C oxidative cleavage,
are still more interesting from the point of view of atom
economy15 as they preserve all four erythrulose carbons
together with the initially resident chirality. As shown
in Scheme 3, syn,syn-stereotriads and syn,syn,syn-ste-
reotetrads can be created in this way.
We thus wanted to expand this methodology to the anti
aldols described here. With this purpose in mind, the
previously used aldolization/in situ reduction sequence14
was applied again. Thus, the E boron enolate generated
from 5a was added to isobutyraldehyde and benzalde-
hyde. Prior to the standard workup, LiBH4 was added
to the reactions mixture with the aim of performing an
in situ stereoselective reduction to the corresponding syn
1,3-diol.16 However, the anti 1,3-diols 8a /8b were the sole
compounds obtained (Scheme 4).13,17 The two free hy-
droxyl groups in compounds 8a /8b were then protected
as the corresponding acetonides 9a /9b. Selective cleavage
of the primary silyl group to yield alcohols 10a /10b took
place under mild conditions using the hydrogen fluoride-
pyridine complex. After mesylation of the alcohol func-
a
Reagents and conditions: (a) Chx2BCl, Et3N, 0 °C, 1 h, then
RCHO, -78 °C, 5 h, then LiBH4, -78 °C, 2 h, then diastereomer
separation (8a , 76%; 8b, 74%); (b) acetone, 2,2-dimethoxypropane,
CSA, 3 Å MS, rt, 12 h (9a , 87%; 9b, 80%); (c) HF-pyridine, THF,
rt, 6-12 h (10a , 79%; 10b, 72%); (d) (1) MsCl, Et3N, CH2Cl2, rt, 1
h, (2) KOH/MeOH, rt, 1 h (11a , 72%; 11b, 69%, overall yields for
the two steps); (e) K2CO3, MeOH, rt, 1 h (12a , 83%; 12b, 82%); (f)
(1) PivCl, CH2Cl2-py, 0 °C, 30 min, (2) MsCl, Et3N, CH2Cl2, rt, 1
h, (3) KOH/MeOH, rt, 24 h (13a , 52%; 13b, 48%, overall yields for
the three steps); (g) NaIO4, aq THF, rt, 1 h (14a , 93%; 14b, 95%).
Abbreviations: TBS, tert-butyldimethylsilyl; Chx, cyclohexyl; CSA,
camphor-10-sulfonic acid; MsCl, methanesulfonyl chloride; Piv,
pivaloyl.
tion, the benzoate group was hydrolyzed by alkaline
treatment. The intermediate hydroxy mesylates cyclized
spontaneously to yield epoxides 11a /11b. These are
functionalized d4 synthons with a syn/syn/anti relative
arrangement of the oxygen atoms, thus diastereomeric
to those depicted in Scheme 3.14
Functionalized d4 synthons 13a /13b displaying the
anti/syn/anti stereotetrad were obtained with the same
methodology used in our previous paper.14 Thus, after
alkaline cleavage of the benzoate function in 10a /10b to
yield diols 12a /12b, the primary hydroxyl group was
(14) Murga, J .; Falomir, E.; Carda, M.; Marco, J . A. Tetrahedron:
Asymmetry 2002, 13, 2317-2327.
(15) (a) Trost, B. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 259-
281. (b) Trost, B. M. Acc. Chem. Res. 2002, 35, 695-705.
(16) Paterson, I.; Channon, J . A. Tetrahedron Lett. 1992, 33, 797-
800.
J . Org. Chem, Vol. 69, No. 6, 2004 1989

Murga et al.
SCHEME 5. Syn th esis of F u r th er F u n ction a lized
d 3 a n d d 4 Ch ir a l Syn th on s w ith Alter n a tive
Ster eoch em ica l Ar r a n gem en tsa
selectively acylated as its pivalate ester. Mesylation of
the secondary alcohol furnished the intermediate piv-
alate-mesylate (depicted in Scheme 4) which, without
purification, was subjected to alkaline treatment. This
caused hydrolysis of the pivaloyl moiety, followed by
rapid closure to epoxides 13a /13b via intramolecular
displacement of the mesylate group with configurational
inversion. Diols 12a /12b served also as precursors of the
functionalized d3 synthons 14a /14b (syn/anti stereotriad)
via oxidative diol cleavage with sodium metaperiodate.
Further functionalized d4 synthons with alternative
stereochemical arrangements were obtained through a
modification of the aldol reduction. This was not per-
formed, however, in the expected manner. During our
initial attempts to reduce the aldol in the anti mode, we
subjected the isolated aldols to reduction with tetra-
methylammonium triacetoxyborohydride (TABH). This
reaction, however, did not take the expected stereochem-
ical course. Instead of the anticipated anti 1,3-diol,18
treatment of aldols 6a and 7 with TABH under the
prescribed conditions gave rise to syn 1,3-diols 15a /15b
as the main compounds (dr ∼ 9:1) (Scheme 5).13,19
Hereafter, we proceeded in the same manner described
in Scheme 4. In this way, the functionalized d4 synthons
18a /18b and 20a /20b displaying, respectively, the anti/
anti/anti and the syn/anti/anti stereotetrads were ob-
tained. Similarly, oxidative cleavage of diols 19a /19b
gave the functionalized d3 synthons 21a /21b with an anti/
anti stereotriad.
We now wish to illustrate the synthetic usefulness of
the aforementioned methodology in the case of a chiral
natural product. (+)-Goniothalesdiol 22 is a 2,3,4,5-
tetrasubstituted tetrahydrofuran derivative isolated in
1998 from the bark of the Malaysian tree Goniothalamus
(17) This unexpected stereochemical outcome was previously ob-
served in the in situ LiBH4 reduction of aldol mixtures generated from
ketone 3: Carda, M.; Falomir, E.; Murga, J .; Castillo, E.; Gonza´lez,
F.; Marco, J . A. Tetrahedron Lett. 1999, 40, 6845-6848. A plausible
explanation of this unexpected stereochemical course is shown in the
following transition states (TS). The syn boron aldolates formed from
ketone 1 adopt a half-chair conformation where the bulky R group
tends to be situated in an equatorial-like orientation (D ) dioxolane
fragment), thus avoiding a steric repulsion with the voluminous boron
ligands. Attack of the external hydride donor (LiBH4) takes place at
the carbonyl Si face along a sterically uncrowded, and stereoelectroni-
cally favored, Felkin-Anh approach (anti to the OTBS group), with
formation of the observed all-syn stereoisomer (syn-1,3-diol moiety).
By the same line of reasoning, anti-aldolates formed from 3 or 5 also
adopt a half-chair conformation with an equatorial-like R group (D′ )
benzoyl-containing fragment). Once again, hydride attack at the
carbonyl Si face becomes favored from both steric and stereoelectronic
factors. However, this yields here an anti-1,3-diol moiety.
a
Reagents and conditions: (a) TABH, AcOH/MeCN, -30 °C,
12 h (15a , 94%; 15b, 84%, yields of the diastereomeric mixture);
(b) acetone, 2,2-dimethoxypropane, CSA, 3 Å MS, rt, 12 h, then
diastereomer separation (16a , 84%; 16b, 83%); (c) HF-pyridine,
THF, rt, 6-12 h (17a , 79%; 17b, 73%); (d) (1) MsCl, Et3N, CH2Cl2,
rt, 1 h, (2) KOH/MeOH, rt, 1 h (18a , 71%; 18b, 69%, overall yields
for the two steps); (e) K2CO3, MeOH, rt, 1 h (19a , 84%; 19b, 84%);
(f) (1) PivCl, CH2Cl2-py, 0 °C, 30 min, (2) MsCl, Et3N, CH2Cl2, rt,
1 h, (3) KOH/MeOH, rt, 24 h (20a , 48%; 20b, 48%, overall yields
for the three steps); (g) NaIO4, aq THF, rt, 1 h (21a , 93%; 21b,
94%). Abbreviations: TABH, tetramethylammonium triacetoxy-
borohydride (see also Scheme 4).
borneensis (Annonaceae)20 and found to exhibit a mea-
surable cytotoxic activity against P388 mouse leukemia
cells. Two syntheses of this bioactive metabolite have
been published so far.21 One of them used D-glucurono-
lactone as the starting material and finally provided (-)-
goniothalesdiol, the enantiomer of the natural compound,
through a lengthy 16-step reaction sequence.21a The other
synthesis started from D-mannitol and required nine
steps, one of them nonstereoselective (∼1:1 mixture of
stereoisomers), to yield lactone 23, subsequently con-
verted into 22 by means of a few additional transforma-
(18) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J . Am. Chem.
Soc. 1988, 110, 3560-3578.
1990 J . Org. Chem., Vol. 69, No. 6, 2004

Anti Aldol Reactions of Erythrulose Derivatives
tions.21b Compound 23 became here our synthetic target.
The starting material was aldehyde 14b, one of the d3
syntons depicted in Scheme 4. Wittig-Horner-Emmons
olefination of 14b under mild conditions22 furnished trans
ester 24. Hydrolytic treatment of the latter not only
caused cleavage of the acetonide moiety and the TBS
group but also in situ intramolecular Michael addition
to yield tetrahydrofurane ester 25. Acid treatment of the
latter provided a lactone 26 which was different in its
spectral properties from 23. We concluded therefore that
the intramolecular Michael addition took place with the
undesired stereochemical course (Scheme 6).23 Most
likely, the molecule of the precursor adopts conformation
A (R ) H or TBS) for the Michael addition, where both
SCHEME 6. F or m a l Syn th esis of
(+)-Gon ioth a lesd iol 22a
torsional
inter-
actions between the chain substituents and allylic 1,3-
strain are minimized.
We then reasoned that the intramolecular Michael
addition would be compelled to take the required stereo-
(19) As we reported in ref 14, TABH reduction of the syn aldols
generated from ketone 1 takes place mainly in the expected anti
manner, even though with
a
moderate stereoselectivity (∼7:3). A
possible explanation of the diverging stereochemical outcome with
aldols from ketones 3 or 5a is proposed below. The internal hydride
transfer through a six-membered chairlike TS with the R group in
equatorial orientation (depicted below as TS-1 for aldols from 1, D )
dioxolane moiety), as proposed by Evans and co-workers (ref 18), faces
important pentane gauche repulsive interactions in the case of aldols
from 3 or 5a (see the chairlike TS-2, which leads to the minor anti-
1,3-diol). Note that a C-C bond rotation in TS-2 may eliminate this
gauche interaction but with simultaneous increase of the dipolar
repulsion between the CdO and C-OBz bonds. As an alternative
explanation, we propose the boatlike TS-3, where these unfavorable
interactions are relieved in part, and which predicts the formation of
the syn-1,3-diol. In any case, it is worth remembering here that TABH
reductions of densely oxygenated R,R’,â-chiral â-hydroxy ketones, such
as those discussed here, are not extensively documented. In fact, cases
of low diastereoselectivity in TABH reductions of R-alkyl-â-hydrox-
yketones are not unprecedented: J eong, E. J .; Kang, E. J .; Sung, L.
T.; Hong, S. K.; Lee, E. J . Am. Chem. Soc. 2002, 124, 14655-14662.
a
Reagents and conditions: (a) (EtO)2POCH2COOEt, LiCl, DI-
PEA, MeCN, rt, 18 h (85%); (b) PPTS, MeOH, 50 °C, 18 h (62%);
(c) CSA, tol, 80 °C, 12 h (70%); (d) (CF3CH2O)POCH2COOMe,
KHMDS, 18-crown-6, -78 °C, 30 min (86%); (e) PPTS, MeOH, 50
°C, 12 h (68%); (f) TBAF, THF, rt, 1 h (78%). Abbreviations:
DIPEA, ethyl N,N-diisopropylamine; PPTS, pyridinium p-tolu-
enesulfonate; tol, toluene; KHDMS, potassium hexamethyldisilyl-
amide; TBAF, tetra-n-butylammonium fluoride hydrate (see also
Schemes 4 and 5).
chemical pathway provided that the process occurred in
a precursor having already the lactone ring in place.
Thus, Z-selective Still-Gennari olefination24 of aldehyde
14b furnished uneventfully ester 27. Acid-catalyzed
methanolysis of the acetonide ring was accompanied by
spontaneous lactonization to 28. Due to the steric con-
straints imposed by the lactone ring, the intramolecular
Michael addition of the free hydroxyl to the conjugated
(23) Had the Michael addition taken the desired course, the result-
ing tetrahydrofurane ester would necessarily have cyclized to lactone
23, as no other alternative is available. Since lactone 26 was syntheti-
cally useless, no efforts were invested in the unambiguous establish-
ment of its structure via exhaustive NMR measurements. Theoreti-
cally, ester 25 might also have cyclized to bicyclic lactone i. However,
the strained trans-fused dioxabicyclo[3.3.0]octane system of the latter
makes this a less likely alternative. In fact, the predicted values for
the 1H NMR vicinal coupling constants in 26 (3J ) 0-3 Hz, according
to modelization with PCMODEL) were very similar to the experimental
values (<2 Hz), in contrast with lactone i, where calculated values for
3J were always >7 Hz.
(20) Cao, S. G.; Wu, X. H.; Sim, K. Y.; Tan, B. K. T.; Pereira, J . T.;
Goh, S. H. Tetrahedron 1998, 54, 2143-2148.
(21) (a) Yoda, H.; Nakaseko, Y.; Takabe, K. Synlett 2002, 1532-
1534. (b) Babjak, M.; Kapita´n, P.; Gracza, T. Tetrahedron Lett. 2002,
43, 6983-6985. For the synthesis of the nonnatural 2-epimer, see ref
21b and: Yoda, H.; Shimojo, T.; Takabe, K. Synlett 1999, 1969-1971.
(22) Blanchette, M. A.; Choy, W.; Davis, J . T.; Essenfeld, A. P.;
Masamune, S.; Roush, W. R.; Sakai, T. Tetrahedron Lett. 1984, 25,
2183-2186.
(24) Still, W. C.; Gennari, C. Tetrahedron Lett. 1983, 24, 4405-4408.
J . Org. Chem, Vol. 69, No. 6, 2004 1991

Murga et al.
CdC bond was anticipated to take necessarily the desired
stereochemical course and yield an unstrained cis-fused
dioxabicyclo[3.3.0]octane system. Indeed, treatment of 28
with tetra-n-butylammonium fluoride not only caused the
expected desilylation but also induced an in situ cycliza-
tion to yield lactone 23, identical in its spectral properties
with the described compound.21b Since 23 has already
been converted into (+)-goniothalesdiol 22, this consti-
tutes a formal synthesis of this natural compound
(Scheme 6). The synthesis is not only comparatively short
but also highly stereoselective in each of its steps.
ated molecules and have demonstrated it with the
example of the bioactive metabolite (+)-goniothalesdiol.
Further investigations related to synthetic uses of the
aforementioned chiral synthons are underway and will
be reported in due course.
Ack n ow led gm en t . The financial support for this
project was provided through the Spanish Ministry of
Science and Technology (Project No. BQU2002-00468)
and by BANCAJ A (Project No. PI-1B2002-06). One of
the authors (J .M.) thanks the Spanish Ministry of
Science and Technology for a Ramo´n y Cajal fellowship.
We further thank Dr. H. Roeper, Cargill TDC Food
Europe, Cerestar Vilvoorde R&D Centre, Belgium, for
a very generous supply of L-erythrulose.
Con clu sion s
We have shown that anti aldols can be efficiently
obtained from a protected erythrulose derivative, in turn
readily available from L-erythrulose in only two steps.
Furthermore, various chiral, polyoxygenated d3 and d4
synthons can be easily prepared in few steps from the
aforementioned anti aldols. These synthons display the
same functional arrays as those prepared from the
previously reported syn aldols14 but show different ster-
eochemical features. We have also shown that such
synthons are very useful in the synthesis of polyoxygen-
Su p p or tin g In for m a tion Ava ila ble: Detailed experi-
mental procedures; complete analytical and spectral data of
ketones 5a -c and aldols 6a -c and 7; description of correlation
procedures and spectral data of compounds 8-21 and 23-28.
This material is available free of charge via the Internet at
http://pubs.acs.org.
J O0356356
1992 J . Org. Chem., Vol. 69, No. 6, 2004