On the direct 2,3-hydroxyl-group differentiation of tartaric acid esters

Article abstract of DOI:10.1016/S0040-4039(02)00662-7

Direct desymmetrization of tartaric acid esters with TIPS-triflate proceeds in up to 99% yield giving the useful intermediates 2a-b. Selective reduction of the ester groups provides access to fully differentiated threonolactone derivatives while reduction of both esters of 2a followed by periodate mediated diol cleavage allows access to 2-O-TIPS-protected L-glyceraldehyde.

Full text of DOI:10.1016/S0040-4039(02)00662-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 3857–3861
On the direct 2,3-hydroxyl-group differentiation of
tartaric acid esters†
James McNulty* and Justin Mao
Institute of Molecular Catalysis, Department of Chemistry, Brock University, St. Catharines, Ontario, Canada L2S 3A1
Received 1 April 2002; accepted 4 April 2002
Abstract—Direct desymmetrization of tartaric acid esters with TIPS-triflate proceeds in up to 99% yield giving the useful
intermediates 2a–b. Selective reduction of the ester groups provides access to fully differentiated threonolactone derivatives while
reduction of both esters of 2a followed by periodate mediated diol cleavage allows access to 2-O-TIPS-protected L-glyceraldehyde.
© 2002 Elsevier Science Ltd. All rights reserved.
The 1,2-diol subunit occurs in a large number of natu-
ral products of various classes including isoprenoids,
alkaloids, polyketides and carbohydrate derivatives.
Methods for the asymmetric synthesis of both syn- and
anti-1,2-diols have received much attention including
catalytic asymmetric routes to syn-1,2-diols via Sharp-
less asymmetric dihydroxylation1 and more recently
both syn- and anti-1,2-diols via catalytic asymmetric
aldol reactions.2 Tartaric acid 1 is a classic chiral-pool
alternative for the rapid asymmetric synthesis of syn-
1,2-diols. It is readily available in either enantiomeric
form and possesses an innate syn-2,3-diol subunit. Tar-
taric acid has been used extensively as a chiral precur-
sor in the asymmetric synthesis of natural and
non-natural products3a as well as C2-symmetrical
molecules including chiral ligands.4 One limitation of
the chiral-pool approach from tartaric acid is that often
several steps are necessary to desymmetrize (C2 to C1)
the molecule through either differentiation of the 1,4-
carbonyl groups or of the internal 2,3-diol. A similar
problem exists for differentiation of the 1,2-diol func-
tionality in non-symmetrical adducts produced through
catalytic asymmetric processes.2 We recently reported
on a very successful and general method for the differ-
entiation of the 1,4-carboxylate groups of tartaric acid
via mono addition of Grignard reagents to the bis-
Weinreb amide derivative.5 This method allows rapid
access to asymmetric extended linear fragments con-
taining the internal 1,2-diol functionality. We next
turned our attention to the problem of the 2,3-diol
differentiation.
A comprehensive literature survey
revealed that essentially four methods have been uti-
lized to achieve this transformation in the tartrate
series. The ‘statistical’ approach (vide infra), conducted
by employing 1 equiv. of reagent, is most simple and
direct although unfortunately usually low yielding. The
other general methods reported include conversion of
the diol to a cyclic O-stannylene acetal followed by
alkylation or acylation,6 desymmetrization through
reduction of a 2,3-diol derived benzilidene acetal,7 and
mono-acetylation of a cyclic orthoacetate derivative.8
The latter three methods all require at least two steps to
achieve the desired de-symmetrization.
Returning to the statistical approach, it is well known
that the mono-functionalization of similar or homo-
topic 1,2-diols can be a challenging problem due to the
similar reaction rates of the diol and intermediate mono
Scheme 1. TIPS protection of tartrate diesters.
* Corresponding author. Fax: 905 682 9020; e-mail: jmcnulty@chemiris.labs.brocku.ca
† Dedicated to Professor Ian W. J. Still on the occasion of his 65th birthday.
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)00662-7

3858
J. McNulty, J. Mao / Tetrahedron Letters 43 (2002) 3857–3861
functionalized derivative with the alkylating or
acylating agent.9 Even when a stoichiometric amount of
such an agent is employed, mixtures of diol, mono- and
bis-protected product are formed. In the case of tartaric
acid diesters, mono-functionalization of the internal
2,3-diol (acetylation, tosylation, benzylation etc.) gener-
ally proceeds in 30–60% yield6 using the ‘statistical’
method. Additives such as silver oxide10 and lanthanide
chlorides9 have been reported to assist mono-function-
alization in some cases. A direct, high yielding mono-
functionalization of tartaric acid diester did not appear
to be a very rewarding venture. Notwithstanding this
prognosis, we have observed (Scheme 1) that silyl-pro-
tection of dimethyltartrate 1a with triisopropylsilyl
(TIPS) triflate in either dichloromethane or dimethyl-
formamide proceeds rapidly to the desymmetrized
mono-TIPS 2a derivative but very slowly to the bis-
TIPS derivative 3. We now report an optimized proce-
dure for this simple and direct de-symmetrization
proceeding routinely in 88–99% yield as well as some
synthetic transformations of the useful desymmetrized
product monoalcohols 2a and 2b.
tant in these rigid cyclic threo-imides due to the lesser
degree of rotational freedom allowing bis TIPS-silyla-
tion to proceed. The use of the bulky TIPS group then
was key to the high selectivity observed in this kineti-
cally controlled differentiation of the homotopic
hydroxyl groups in the acyclic tartrate esters.
We have investigated several useful synthetic elabora-
tions of the desymmetrized tartaric diesters 2a and 2b.
Threonic acid 5 and threonolactone 6 derivatives are
conveniently prepared by partial reduction and cycliza-
tion of 2,3-protected tartrates, sometimes via a cyclic
anhydride 4 (Scheme 2) and have been utilized in
several valuable synthetic transformations.3a The value
of these intermediates is limited however as in almost
all cases the 2,3-diol unit is not differentially pro-
tected.3b The mono-TIPS protected esters 2a and 2b
potentially offer a rapid entry to fully differentiated
threonolactones provided selective ester reductions can
be carried out. The selective reduction of an a-hydroxy
ester in the presence of a second ester fortunately is a
well know transform typically achieved by the use of
borane/sodium borohydride combinations.8 In the
present application, the desymmetrized dimethyl tar-
trate 2a reacted with borane–dimethyl sulfide and
sodium borohydride (Scheme 3) in THF to give a
mixture of the two readily separable isomeric diols 7a
(58%) and 8a (38%), 96% total isolated yield. The diol
7a slowly cyclized of its own accord and readily so
upon treatment with mild acid (pyridinium-4-toluene
sulfonate (PPTS), toluene, 21°C, 4 h) to give lactone 9
quantitatively. The minor diol 8a obtained from the
reduction underwent cyclization under slightly more
forcing conditions (PPTS, toluene, 70°C, 3 h) to give
the isomeric lactone 10, also in quantitative yield. Thus,
in addition to achieving the selective reduction of the
a-hydroxy ester, lactone formation proved to be a very
convenient method to discriminate between the primary
and secondary alcohol functions in 7a and 8a. Reduc-
tion of the 2,3-diol differentiated isopropyl ester 2b
under the same conditions proved to be highly regiose-
lective, although more sluggish in comparison to 2a.
Starting diester still remained after 24 h, however the
only reduction product was the diol 7b (60% isolated
yield from 2b) with no trace of isomeric 8b being
observed. In contrast to methyl esters 7a and 8a, the
isopropyl ester 7b was more stable presumably due to
the increased steric hindrance around the isopropoxy
substituted acyl group. Threonolactone derivatives 9
and 10 are densely functionalized, fully differentiated
chirons in contrast to the literature derivatives outlined
in Scheme 2 (6a–d) and offer considerable promise for
a variety of transformations towards linear syn-1,2-diol
Reaction of dimethyltartrate 1a in dichloromethane
with TIPS-OTf (1.2 equiv.) at 21°C in the presence of
2,6-lutidine (1.2 equiv.) provided alcohol 2a in 88–99%
yield.11 Diisopropyl tartrate 1b reacted somewhat
slower under the same conditions to give the monopro-
tected non symmetrical alcohol 2b in 80% yield. Under
these conditions the TLC profile indicates only a trace
of the bis-silylated product 3. We were somewhat sur-
prised at the high degree of selectivity shown in this
transformation and wondered initially if this was not
the result of some unexpected complexation (or precipi-
tation) of an intermediate from which 2a/b was
obtained by hydrolysis. However, even when 2a was
treated with an excess TIPS-OTf and 2,6-lutidine (4.0
equiv. each) the second silylation was very slow. The
bis-TIPS derivative 3 could only be obtained in good
yield when this reaction was subjected to sonication
(Branson 5510, 22°C, 4 h) giving 3 in 80% yield. The
selectivity observed appears to be simply due to steric
hindrance imposing a high activation barrier for the
second silylation step. The reaction with t-
butyldimethylsilyl (TBDMS) triflate was less selective
under these conditions but still provided the mono-
TBDMS derivative from 1a in a useful 78% yield as
well as the bis-TBDMS derivative corresponding to 3
(14.5%). Cyclic 2,3-bis-TBDMS protected tartrate
derivatives have been previously reported in good yields
through normal silylation protocols,12 and a bis-TIPS
protected tartaric acid derived cyclic imide derivative
has been reported.13 Steric factors are likely less impor-
Scheme 2. Literature syntheses of 1,2-diol non-differentiated threonic acid and threonolactone derivatives.3a

J. McNulty, J. Mao / Tetrahedron Letters 43 (2002) 3857–3861
3859
Scheme 3. Synthesis of differentiated threonic acid and threonolactone derivatives.
containing fragments. The stability of 7b offers an
can lead to subsequent chemoselective incompatibilities
also. For example Wittig or Horner–Wadsworth–
Emmons olefination of 12b gives an O-benzyl contain-
ing olefin14 in which hydrogenolysis or hydrogenation
could not be effected selectively. We felt that the 2-O-
alternate route to open chain threonic ester derivatives.
Optically active, protected derivatives of
D- and L-gly-
ceraldehyde, such as the acetonide 12a and the 2-O-ben-
zyl derivative 12b (Scheme 4), are valuable C3-building
blocks and have been utilized extensively in organic
synthesis.14 While these two compounds are the most
important members of protected chiral glyceraldehydes
available, both suffer inherent limitations. For example
in the acetonide 12a, the primary and secondary diols
are not differentiated and would require selective
manipulations later on in a synthesis subsequent to
acetonide hydrolysis. In derivative 12b, the benzyl ether
TIPS- -glyceraldehyde 12c would be an ideal C3 build-
L
ing block not subject to either of the limitations stated
above. Reduction of both esters of the mono-O-TIPS
protected tartaric acid 2a was readily achieved using
sodium borohydride/lithium chloride in ethanol to give
the differentiated 2-O-TIPS protected L-threitol deriva-
tive 11 in 82% yield. Periodate cleavage of the 1,2-diol
in 11 was carried out under standard conditions14 giv-
ing the desired 2-O-TIPS protected
L-glyceraldehyde
Scheme 4. Synthesis of differentiated allylic alcohols from 2-O-TIPS-L-glyceraldehyde.

3860
J. McNulty, J. Mao / Tetrahedron Letters 43 (2002) 3857–3861
12c as a colorless viscous oil. Reaction of the aldehyde
with the Horner–Emmons reagent provided the (E)-
olefin 13 in 72% yield from 11. Olefin 13 is a C5 chiron
incorporating five-differentially functionalized carbons
and is available in only four steps from dimethyl tar-
trate. The ready access and presence of useful function-
ality, including a free primary alcohol, a protected
allylic alcohol as well as an unsaturated ester, offers
Ito, H.; Silcoff, E. R. J. Am. Chem. Soc. 2001, 123,
3367–3368.
3. (a) For a comprehensive review of synthetic applications
using tartaric acid, see: Gawronski, J.; Gawronska, K.
Tartaric and Malic Acids in Synthesis; John Wiley: New
York, 1999; (b) Partially-differentiated threonic acid/
threonolactone derivatives can be prepared from L-ascor-
bic acid, see: Wei, C. C.; Bernardo, S. D.; Tengi, J. P.;
Borgese, J.; Weigele, M. J. Org. Chem. 1985, 50, 3462–
3467.
much potential for 13 as
a
useful synthetic
intermediate.
4. For the synthesis of C2-symmetrical tartrate derived chi-
ral ligands, see: (a) Seebach, D.; Beck, A. K.; Heckel, A.
Angew. Chem., Int. Ed. 2001, 40, 92–138; (b) Beck, A. K.;
Gysi, P.; La Veccia, L.; Seebach, D. Org. Synth. 1998, 76,
12–22; (c) Dieguez, M.; Orejon, A.; Masdeu-Bulto, A.
M.; Echarri, R.; Castillon, S.; Claver, C.; Ruiz, A. J.
Chem. Soc., Dalton Trans. 1997, 4611–4618; (d) McNulty,
J.; Mo, R.; Capretta, A.; Frampton, C. S. Chem. Com-
mun. 2001, 2384–5
5. McNulty, J.; Grunner, V.; Mao, J. Tetrahedron Lett.
2001, 42, 5609–5612.
6. Nagashima, N.; Ohno, M. Chem. Lett. 1987, 141–144.
See also: Barton, D. H. R.; Cleophax, J.; Gateau-Olesker,
A.; Gero, S. D.; Tachdjian, C. Tetrahedron 1993, 49,
8381–8396.
Finally we made an interesting observation during the
Horner–Wadsworth–Emmons reaction. If slight
a
excess of LiHMDS was employed during this reaction,
a minor isomeric product could be isolated from the
reaction. This compound was subsequently identified as
the product of silyl migration to the less hindered
primary position, 13 to 14. The isomerization could be
completely avoided employing a 1:1 ratio of phospho-
nate to base during the olefination reaction. In addition
when a purified sample of 13 was treated independently
with LiHMDS in dry THF, isomerization occurred
readily giving 14 as the major product (82% yield).
Compound 14 contains a free allylic alcohol as well as
an unsaturated ester and further increases the versatility
of this methodology from the differentiated glyceralde-
hyde 12c. 1,2-Silyl migrations from one hydroxyl to
another are well documented with less hindered silanes
such as TBDMS15a and appear to be rare in the case of
the TIPS group.15b
7. (a) Adam, G.; Seebach, D. Synthesis 1988, 373–375; (b)
Guindon, Y.; Girard, Y.; Berthiaume, S.; Gorys, V.;
Lemieux, R.; Yoakim, C. Can. J. Chem. 1990, 68, 897–
902.
8. Saito, S.; Ishikawa, T.; Kuroda, A.; Koga, K.; Moriwake,
T. Tetrahedron 1992, 48, 4067–4086.
In conclusion, we report a simple one-step method for
the direct 2,3-diol differentiation in tartaric diester
derivatives. Selective reduction of the ester residues has
been demonstrated leading to the potentially valuable
fully differentiated threonolactone derivatives 9 and 10.
Complete reduction and periodate cleavage of the
desymmetrized tartrate diester opens a route to a valu-
able differentiated glyceraldehyde derivative 12c. This
compound has been elaborated to the two highly func-
tionalized linear C5 chirons 13 and 14. Application of
this methodology towards the preparation of biologi-
cally active compounds containing extended linear frag-
ments is currently in progress in our laboratories.
9. Clarke, P. A.; Holton, R. A.; Kayaleh, N. E. Tetrahedron
Lett. 2000, 41, 2687–2690.
10. Bouzide, A.; Sauve, G. Tetrahedron Lett. 1997, 38, 5945–
5948.
11. Dimethyl-L-tartrate 1a (50.0 mg, 0.281 mmol) was dis-
solved in dry CH2Cl2 (1.0 mL, dist. CaH2) by stirring at
rt for 15 min at which time triisopropylsilyltri-
fluoromethanesulfonate (0.337 mmol, 1.2 equiv.) was
added slowly over 5 min. When the addition was com-
pleted, the mixture was stirred at rt for 10 min. before a
solution of 2,6-lutidine (0.337 mmol, 1.2 equiv.) was
added dropwise. The entire mixture was stirred at rt until
starting material had disappeared; approx.
3
h
(TLC:hexane:EtOAc, 50:50; compound (Rf): 1a (0.12), 2a
(0.78), 3 (0.91)). The reaction was quenched with satu-
rated ammonium chloride (0.5 mL), diluted with CH2Cl2,
dried over Na2SO4 and evaporated under reduced pres-
sure to give 2a as a colourless oil. In most cases the
material was pure enough for use in the next reaction. If
desired, purification by flash chromatography, 1:4 ethyl
acetate:hexane gave 2a as a colourless oil. The yield
ranged from 88 to 99% over 10 independent runs and
may be scaled up to the gram level without incident
(88–90% yield). When scaling up, it is important to
maintain the same concentration as above. More concen-
trated solutions provided slightly higher yield of the
bis-sililated product under otherwise identical conditions.
Acknowledgements
We thank the Natural Sciences and Engineering
Research Council of Canada and Research Corpora-
tion (Cottrell Scholar Science Award to J.McN.) for
financial support.
References
1. Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B.
Chem. Rev. 1994, 94, 2483–2547.
2. (a) Notz, W.; List, B. J. Am. Chem. Soc. 2000, 122,
7386–7387; (b) Yoshikawa, N.; Kumagai, N.; Matsunaga,
S.; Moll, G.; Ohshima, T.; Suzuki, T.; Shibasaki, M. J.
Am. Chem. Soc. 2001, 123, 2466–2467; (c) Trost, B. M.;
1
2a had [h]2D2 +23.0° (c 2.87, MeOH); H NMR 300 MHz
(CDCl3): l 1.04–1.10 (m, 21H), 3.80 (s, 3H), 3.81 (s, 3H),
4.55 (br s, 1H), 4.83 (m, 1H); 13C NMR (CDCl3): l 172.2,
171.4, 74.50, 73.83, 52.93, 52.67, 18.21, 12.95; IR (KBr):

J. McNulty, J. Mao / Tetrahedron Letters 43 (2002) 3857–3861
1991, 32, 3401–3404.
3861
3559, 2868–2947, 1764, 1464, 1438, 1262, 1157, 1104, 883,
683; EIMS m/z (%) 291 (M+−isopropyl, 100), 231 (44.2),
203 (77.8), 145 (75.5); HREIMS: calcd (C12H23SiO6)
291.1264, found 291.1261.
14. Steuer, B.; Wehner, V.; Lieberknecht, A.; Jager, V. Org.
Synth. 1996, 74, 1–11.
15. (a) Greene, T. W.; Wuts, P. G. M. Protective Groups in
Organic Synthesis, 3rd ed.; Wiley-Interscience: New
York, 1999; pp. 114–116; (b) Seela, F.; Frohlich, T. Helv.
Chim. Acta 1994, 77, 399–408.
12. Yoda, H.; Shirakawa, K.; Takabe, K. Chem. Lett. 1991,
489–490.
13. Yoda, H.; Shirakawa, K.; Takabe, K. Tetrahedron Lett.