Aldol Addition of Lithium and Boron Enolates of 1,3-Dioxan-5-ones to Aldehydes. A New Entry into Monosaccharide Derivatives

Article abstract of DOI:10.1021/jo0002238

Methods allowing control of stereoselectivity in aldol reactions of enolates derived from 1,3-dioxan-5-ones (4) are described. Boron enolates, generated in situ, react with benzaldehyde to give the corresponding anti aldol selectively (the anti:syn ratio of up to 96:4) in high yield. Lithium enolates give high anti selectivity only with aldehydes branched at the α-position. Enantioselective deprotonation of C_S symmetrical dioxanones (e.g., 4b) can be accomplished efficiently, with enantiomeric excess of up to 90%, with chiral lithium amide bases of general structure PhCH-(Me)N(Li)R (9, 10) if the R group is sufficiently bulky (e.g, R = adamantyl) or is fluorinated (e.g., R = CH₂CF₃). Dioxanone boron and lithium enolates react readily with glyceraldehyde derivatives (19), yielding protected ketohexoses (20 and 21).

Full text of DOI:10.1021/jo0002238

5152
J . Org. Chem. 2000, 65, 5152-5160
Ald ol Ad d ition of Lith iu m a n d Bor on En ola tes of 1,3-Dioxa n -5-on es
to Ald eh yd es. A New En tr y in to Mon osa cch a r id e Der iva tives
Marek Majewski* and Pawel Nowak
Department of Chemistry, University of Saskatchewan, 110 Science Place,
Saskatoon, SK, Canada S7N 5C9
majewski@sask.usask.ca
Received February 16, 2000
Methods allowing control of stereoselectivity in aldol reactions of enolates derived from 1,3-dioxan-
5-ones (4) are described. Boron enolates, generated in situ, react with benzaldehyde to give the
corresponding anti aldol selectively (the anti:syn ratio of up to 96:4) and in high yield. Lithium
enolates give high anti selectivity only with aldehydes branched at the R-position. Enantioselective
deprotonation of CS symmetrical dioxanones (e.g., 4b) can be accomplished efficiently, with
enantiomeric excess of up to 90%, with chiral lithium amide bases of general structure PhCH-
(Me)N(Li)R (9, 10) if the R group is sufficiently bulky (e.g, R ) adamantyl) or is fluorinated (e.g.,
R ) CH2CF3). Dioxanone boron and lithium enolates react readily with glyceraldehyde derivatives
(19), yielding protected ketohexoses (20 and 21).
Sch em e 1
Aldol addition of ketone enolates to aldehydes provides
one of the most useful synthetic tools for construction of
carbon-carbon bonds. The reaction has been studied
extensively during the past 3 decades, and a number of
its mechanistic features and some effective methods for
controlling diastereoselectivity have been elaborated.1
More recently, the scope of the aldol addition reaction
was expanded by chiral lithium amide bases which could
be used to generate chiral enolates stereoselectively, thus
affording another approach to synthesis of enantiomeri-
cally pure aldols.2 Aldol reaction is very general, and, over
the years, many diverse ketones were enolized and added
to aldehydes. Protected 1,3-dihydroxyacetone (1) could
be an interesting starting material for synthesis of
polyoxygenated natural products of general structure 3
via the sequential aldol approach, if a number of stereo-
selectivity challenges were solved (Scheme 1). Some
phenomena associated with the ketone structure affecting
the course of the aldol reaction, e.g., selectivity problems
due to the presence of a heteroatom at one of the R-carbon
atoms, were described before.3 When studying enolization
of several CS symmetrical ketones, we were intrigued by
structural features of 1,3-dioxan-5-ones (4) that are
potential synthetic equivalents of 1,3-dihydroxyacetone.
An efficient control of both the diastereoselectivity and,
ultimately, the enantioselectivity of the aldol reaction in
the dioxanone system would be necessary if dioxanones
were to become useful building blocks.4 We described
earlier a general method of synthesis of dioxanones and
preliminary studies on their enolization.5 Chemistry of
these compounds proved nontrivial due to side reactions
and low stability of the dioxanone system, especially
under acidic conditions. Below, we present results of
recent studies on aldol addition reactions of lithium and
boron enolates of 1,3-dioxanones to aldehydes and ap-
plication of these results to synthesis of monosaccharide
derivatives.6
Resu lts a n d Discu ssion
As reported before,5 early experiments with aldol
addition of dioxanone lithium enolates were not very
promising. The yields were rather low (55-77%) and,
although the anti aldols were produced predominantly,
in agreement with the Zimmerman-Traxler model,1c the
selectivity in reactions with typical aldehydes (e.g.,
benzaldehyde) was low and the syn:anti ratio oscillated
around 1:2. The aldol reaction could yield up to four
different diastereoisomeric products: syn-cis, syn-trans,
anti-cis, and anti-trans (Scheme 2; note that the trans
products, which are possible when R and R′ are different,
are not shown for brevity), and we hypothesized that the
diastereoselectivity should be influenced by the size of
the substituents R and R′ at the 2-position in the
dioxanone molecule and by the structure of the R1 group
in the aldehydessteric effects in the cyclic transition
(1) (a) Heathcock, C. H. In Comprehensive Organic Synthesis; Trost,
B. M., Ed.; Pergamon: Oxford, U.K., 1991; Vol. 2, p 181. (b) Kim, B.
M.; Williams, S. F.; Masamune, S. In Comprehensive Organic Synthe-
sis; Trost, B. M., Ed.; Pergamon: Oxford, U.K., 1991, Vol. 2, p 239. (c)
Braun, M. In Methods of Organic Chemistry (Houben-Weyl). Stereose-
lective Synthesis; Helmchen, G., Hoffmann, R. W., Mulzer, J ., Schau-
mann, E., Eds.; Thieme: New York, 1996; Vol. E-21, p 1603.
(2) Review: O’Brien, P. J . J . Chem. Soc., Perkin Trans. 1 1998, 1439.
(3) (a) Murray, D. H.; Albizati, K. F. Tetrahedron Lett. 1990, 31,
4109. (b) Martin, V. A.; Murray, D. H.; Pratt, N. E.; Zhao, Y.; Albizati,
K. F. J . Am. Chem. Soc. 1990, 112, 6965.
(4) For other pertinent studies of reactions and synthetic applica-
tions of dioxanones see: (a) Enders, D. In Stereoselective Synthesis;
Ottow, E., Schollkopf, K., Schulz, B.-G., Eds.; Springer: Berlin, 1994;
p 63. (b) Enders, D.; Prokopenko, O. F.; Raabe, G.; Runsink, J .
Synthesis 1996, 1095. (c) Hirama, M.; Noda, T.; Ito, S. J . Org. Chem.
1988, 53, 706.
(5) Majewski, M.; Gleave, D. M.; Nowak, P. Can. J . Chem. 1995,
73, 1616.
(6) Preliminary communication: Majewski, M.; Nowak, P. Synlett
1999, 1447.
10.1021/jo0002238 CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/29/2000

Aldol Reactions of Enolates of 1,3-Dioxan-5-ones
Ta ble 1. Ald ol Rea ction of Bor on En ola tes of 4a w ith P h CHOa
anti:syn
J . Org. Chem., Vol. 65, No. 17, 2000 5153
yieldb
(%)
entry
reagent/amine
workup
(6a a :7a a )
1
2
3
4
5
6
7
8
9
BBNOTf/DIPEA
Bu2BOTf/DIPEA
Cp2BOTf/DIPEA
Cy2BCl/Et3N
Cy2BCl/Et3N
Cy2BCl/Et3N
Cy2BCl/Et3N
Cy2BCl/Et3N
Cy2BCl/Et3N
Cy2BCl/Et3N
H2O2, pH 7
H2O2, pH 7
H2O2, pH 7
H2O2, pH 7
HOCH2CH2NH2 (1 equiv)
HOCH2CH2NH2 (2 equiv)
HOCH2CH2NH2 (3 equiv)
NaBO3
O3
DDO
53:47
73:27
84:16
96:4
51
46
47
64
60
64
47
31
83
81
92:8
70:30
71:29
94:6
95:5
96:4
10
a
b
Lithium enolate of dioxanone 4a gave 6a a and 7a a in a ratio of 65:35 and in 55% yield. Combined yield of purified syn and anti
products.
Sch em e 2
Sch em e 3
with the increasing size of the ligands on boron (Table
1, entries 1-4). Dicyclohexylboron enolate afforded the
aldol product 6a in reasonably high purity (entry 4).
Monitoring of the reaction by 1H NMR indicated that the
formation of boron enolates was fast, clean, and quanti-
tative and the subsequent reaction with benzaldehyde
was also fast and clean. Loss of the product, visible in
low yields, occurred on workup, which turned out to be
the critical factor. Dioxanones are sensitive to acids and
bases, and some reagents commonly used for removal of
boron byproducts at the workup stage (e.g., ethanol-
amine, sodium perborate) caused decomposition. Entries
5-7 in Table 1 illustrate how increasing amounts of
ethanolamine adversely affected the reaction. Workup
with a buffered hydrogen peroxide solution proved some-
what capricious with high concentration of the sodium
phosphate buffer being an important variable (the more
concentrated the better). Finally, both ozone and dimeth-
yldioxirane (DDO) proved very good reagents for oxida-
tive workup (entries 9 and 10), with the latter being much
more convenient to use. To our knowledge, the use of
DDO for working up reactions involving boron reagents
is unprecedented and could offer a useful protocol in a
more general sense, when labile compounds are involved.
Conditions for highly diastereoselective reaction of diox-
anone boron enolates were thus successfully developed.
Next, we briefly explored the sequential bis-aldol reaction
(Scheme 3).
The presence of two C-H acidic sites in a ketone at
the R and R′ positions (one on each side of the carbonyl
group) does not necessarily mean that two sequential
aldol reactions would work efficiently, even if a single
aldol was facile.9 Diastereoselectivity is also an important
issue in bis-aldol reactions because up to eight different
isomers could be produced. We felt that a brief model
study was necessary before planning a synthesis along
the lines presented in Scheme 1. A “one pot” reaction of
dioxanone 4a with dicyclohexylboron chloride followed
by addition of benzaldehyde, then by another equivalent
of the boron reagent, and another equivalent of cyclo-
state should be maximized. In one promising experiment
cyclohexanecarboxaldehyde 5b had reacted with the
lithium enolate of 2-tert-butyl-2-methyl-1,3-dioxan-5-one
4b to give only one product, which had been identified
by NMR and X-ray crystallography as the anti-cis isomer
6bb.5
Further work on the aldol addition was clearly needed
before this reaction could be used in synthesis. In an
effort to optimize the reaction conditions we have exam-
ined the effect of lithium halide additives and the source
of lithium enolate on the aldol reaction of dimethyldi-
oxanone 4a . Addition of LiCl or LiBr (3 times molar
excess) to the solution of the lithium enolate did not affect
diastereoselectivity (a small decrease in de was ob-
served).7 To generate the dioxanone lithium enolate, we
used lithium diisopropylamide (LDA), LiTMP, or LiH-
MDS (HMDS ) 1,1,1,3,3,3-hexamethyldisilane) and also
amine-free conditions (the enolate was generated from
the corresponding tetramethylsilane (TMS) enol ether by
MeLi in Et2O). In all cases the subsequent addition to
benzaldehyde resulted in low diastereoselectivity (the
anti isomer accounted for 56-65% of the purified prod-
uct), but the yield was significantly higher under the
amine-free conditions (86%). The potential for control of
stereoselectivity in the dioxanone system seemed very
limited at this stage.
Boron enolates are well-known to undergo addition to
aldehydes with higher diastereoselectivity than the cor-
responding lithium enolates.1b,8 We synthesized several
boron enolates by treating dioxanone 4a with dialkylbo-
ron triflate or halide in the presence of either triethyl-
amine or N,N-diisopropylethanamine (DIPEA) and ex-
amined their reactions with benzaldehyde, the aldehyde
which had shown little selectivity in previous studies
mentioned above. The results are summarized in Table
1 and Scheme 2. Diastereoselectivity increased steadily
(7) LiBr and other additives greatly affect aldol diastereoselectivity
in some systems: (a) Majewski, M.; Gleave, D. M. Tetrahedron Lett.
1989, 30, 5681. (b) J uaristi, E.; Beck, A. K.; Hansen, J .; Matt, T.;
Mukhopadhyay, T.; Simson, M.; Seebach, D. Synthesis 1993, 1271.
(8) Heathcock, C. H. In Asymmetric Synthesis; Morrison, J . D., Ed.;
Academic Press: Toronto, 1984; Vol. 3, p 111.
(9) (a) McCarthy, P. A.; Kageyama, M. J . Org. Chem. 1987, 52, 4618.
(b) Luke, G. P.; Morris, J . J . Org. Chem. 1995, 60, 3013.

5154 J . Org. Chem., Vol. 65, No. 17, 2000
Majewski and Nowak
Ta ble 2. En a n tioselective Dep r oton a tion of 4b w ith
Ch ir a l Lith iu m Am id es 9-15 F ollow ed by Rea ction w ith
Cycloh exa n eca r boxa ld eh yd e (cf., Sch em e 2)
hexanecarboxaldehyde afforded a mixture of four com-
pounds in the ratio of 83:9:6:2 (Scheme 3; the numbers
were obtained by integration of the H NMR spectrum
1
entry
lithium amide
LiCla
eeb (%)
yieldc (%)
of the crude product). Column chromatography yielded
readily the pure major isomer 8a (anti-trans-anti as
shown in Scheme 3) in 64% yield. A similar experiment
with the second aldehyde being PhCHO gave the analo-
gous aldol 8b in 60% yield. These results signify that, in
general, the boron-mediated aldol and bis-aldol reactions
of dioxanones can be controlled and should provide an
efficient synthetic method.
E n a n t ioselect ive Dep r ot on a t ion . Treatment of
achiral, CS symmetrical ketones with chiral lithium
amides leads to the formation of nonracemic lithium
enolates.2 The base discriminates between two enan-
tiotopic protons HR and HS (cf., structure 4 in Scheme
2), and the resulting enolate could be trapped with
electrophiles yielding ultimately chiral products. Enan-
tioselective deprotonation has been applied successfully
to synthesis of several natural products.10 Although this
method is now well-established and quite general, there
is no “magic bullet”. The stereoselectivity of deprotona-
tion varies greatly with the structure of the starting
ketone and the structure of lithium amide, and each new
system has to be studied experimentally.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
9a
9b
9b
9c
9d
9e
9f
9f
10a
10a
10b
10c
10c
11
11
11
12
12
1
0
0.5
0.5
1
1
1
1
0
0.5
0.5
0
0.5
0
0.5
1
0
(+) 19
(+) 13
(+) 39
(+) 20
(+) 70
(+) 60
(+) 90
(+) 87d
(-) 10
(-) 4
(-) 16
(+) 15
(-) 32
(+) 18
(+) 60
(+) 59
(-) 50
(-) 60
(+) 80
(+) 90
(+) 20
63
32
56
51
60
76
61
86d
43
63
55
70
53
58
51
49
41
43
91
95
51
0.5
1
1
13
14
15
0.5
a
b
Mmoles of LiCl per 1 mmol fo lithium amide. The (+) and
(-) refer to the dextrorotatory and the levorotatory product 6bb
(R1 ) cyclohexyl), respectively. c Yields refer to chromatographi-
d
cally purified 6bb. Lithium amide was generated in situ from
A brief study of enantioselective deprotonation of 2-tert-
butyl-2-methyl-1,3-dioxan-5-one 4b had demonstrated
that the most popular chiral lithium amide reagents did
not work very well with dioxanones (vide infra). We also
observed that in all cases enantioselectivity increased
when LiCl was used as additive. To develop a selective
deprotonating reagent for dioxanones, we investigated a
number of chiral lithium amides (9-15). The aldol
the corresponding amine hydrochloride.
As mentioned above, chiral lithium amides 11 and 15
that are well-known for their ability to deprotonate cyclic
symmetrical ketones with high enantioselectivity2 did not
work very well with dioxanones. Addition of LiCl did
improve the selectivity in these cases, but the enantio-
meric excess (ee) was still too low to be useful practically
(Table 2, entries 14-16 and 21). In general, addition of
LiCl to the solution of lithium amide almost always
resulted in an increase of enantioselectivity (cf., entries
2 and 3, 9 and 10, 14 and 15, 17 and 18). In one
remarkable case the additive caused the reversal of the
absolute stereochemistry from the reaction favoring the
dextrorotatory isomer to the levorotatory one (entries 12
and 13). We reported before that changing the amount
of LiCl in this system caused a strong response and a
monotonic plot of ee vs molar ratio of LiCl/lithium amide
could be generated.11 In most cases, the enantioselectivity
did not increase further after 0.5 molar equiv of LiCl (per
amide) were added. Toward the end of these studies we
settled the lithium amide to LiCl-to-ketone ratio of 1:1:1
since we believed these conditions to be closest to the
optimum for diverse systems. The results in Table 2
indicate that increasing the size of the R group in chiral
lithium amides of general structure 9 (or 10) is beneficial
to selectivity. Entries 10, 11, 13, 3, 5, and 20 in Table 2
refer to the series of lithium amides where the size of
the R group increases from methyl through isopropyl,
neopentyl, and diphenylmethyl until, ultimately, the very
bulky dinaphthylmethyl group in compound 14 yields
excellent enantioselectivity. The electronic effects also
seem to matter, and R groups having electron-withdraw-
ing character result in more selective reagents (cf., entries
1 and 3, 4 and 13, 1 and 7). This is especially apparent
in the reaction with fluorinated chiral amide 9f (entries
7, 8) which, together with other fluorinated chiral lithium
amides, was developed by Koga and proved an effective
reaction of lithium enolate of 4b with cyclohexanecar-
boxaldehyde (5a ) was used as the model system (Scheme
2). This reaction gave only one diastereoisomer of the
aldol, the anti-cis isomer 6bb (out of four possible); thus,
the measurement of enantiomeric excess was greatly
simplified. Results of these enantioselective deprotona-
tion studies are summarized in Table 2.
(10) For earlier examples cf.: (a) Cox, P. J .; Simpkins, N. S.
Tetrahedron: Asymmetry, 1991, 2, 1. Examples of recent syntheses:
(b) Honda, T.; Kimura, N. Sato, S.; Kato, D.; Tominaga, H. J . Chem.
Soc., Perkin Trans. 1 1994, 1043. (c) Honda, T.; Kimura, N.; Tsubuki,
M. Tetrahedron: Asymmetry 1993, 4, 1475. (d) Muraoka, O.; Okumura,
K.; Maeda, T.; Tanabe, G.; Momose, T. Tetrahedron: Asymmetry 1994,
5, 317. (e) Majewski, M.; Lazny, R.; Ulaczyk, A. Can. J . Chem. 1997,
75, 754. (f) Majewski, M.; Lazny, R. J . Org. Chem. 1995, 60, 5825.
(11) Majewski, M.; Lazny, R.; Nowak, P. Tetrahedron Lett. 1995,
36, 5465.

Aldol Reactions of Enolates of 1,3-Dioxan-5-ones
J . Org. Chem., Vol. 65, No. 17, 2000 5155
Sch em e 4
Sch em e 5
tiomerically pure compounds, while the latter suffered
from lack of diastereoselectivity in a broad sense, only
aldehydes branched at the R-carbon worked well. While
dioxanone boron enolates do not provide obvious means
to build chiral targets by manipulating the structure of
the enolate, one could attempt to construct a system in
which stereoselectivity rests in the chiral aldehyde
reagent.14 To explore this possibility, addition of the
dicyclohexylboron enolate of 4a to protected R-glyceral-
dehydes 19a ,b (Scheme 5) was tested. There are four
diasteroisomeric aldol products which could form in this
reaction corresponding to derivatives of four ketohex-
oses: D-tagatose, D-psicose, D-fructose, and D-sorbose.
Isopropylidene glyceraldehyde 19a afforded only two of
deprotonating reagent in several different systems.12
Overall, we were able to find two very selective chiral
lithium amide reagents, compounds 9e and 14, which
deprotonated dioxanones with synthetically useful selec-
tivity.
Absolute stereochemistry of dioxanone deprotonation
deserves mention. To establish which chiral lithium
amides abstracted the HR proton and which one preferred
the HS one, we needed to correlate one of the aldol
products with a compound of known absolute stereo-
chemistry. We were able to accomplish that, somewhat
indirectly, by degrading the aldol product 6bb to a
glyceraldehyde derivative (Scheme 4). Reduction of 6bb
with diisobutylaluminum hydride (DIBAH) afforded two
diastereoisomeric diols in a ratio of 92:8. Diol 16 was the
major product (isolated in 67% yield). The minor product,
having the newly created OH group at the R face, was
crystalline and provided us with a crystal structure, thus
furnishing proof of relative stereochemistry.5 Diol 16 was
then subjected to transacetalization, which afforded
compound 17 as the major product. Compound 17 was
cleaved readily with lead tetraacetate to give S-isopro-
pylideneglyceraldehyde 18. Thus absolute configuration
of the dextrorotatory isomer of compound 6bb, originat-
ing from deprotonation of dioxanone 4b with chiral
lithium amide 9d , was established to be as drawn in
Scheme 4.
Syn th esis of Ca r boh yd r a te Der iva tives. Potential
polyoxygenated synthetic targets were identified in
Scheme 1 by structures 2 and 3. The dioxanone-based
methodology is readily amenable to synthesis of carbo-
hydrates, and it should be noted that although the main
themes in modern carbohydrate synthesis involve ma-
nipulation of readily available monosaccharides and
synthesis of oligosaccharides, stereoselective total syn-
thesis of “rare” sugars is a challenging and practically
important undertaking.13 To use dioxanones as building
blocks for carbohydrate synthesis, a control of relative
and absolute stereochemistry of aldol addition of diox-
anone enolates is necessary. In the foregoing sections we
have described the development of conditions for high
diastereoselectivity using boron enolates and high enan-
tioselectivity using lithium enolates. The former method
did not offer an obvious potential for synthesizing enan-
these products in a ratio of 85:15 and in a modest yield
of 59% (the DDO workup, described above, was used).
The major product was subsequently identified as the
D-tagatose derivative 20a a and the minor isomer as the
D-psicose derivative 21a a . Isopropylidene glyceraldehyde
is known to have a tendency to polymerize which could
adversely affect the reactions involving this reagent.
Trying to circumvent this potential problem, we tried
another glyceraldehyde derivative 19b having the diol
functionality protected with a different group.15 This
reagent afforded a mixture of compounds 20a a and 21a a
in 83% yield, but the stereoselectivity was somewhat
diminished and the ratio of the products was 80:20.
A combination of a chiral dioxanone enolate with a
chiral aldehyde could offer increased control of stereose-
lectivity due to double stereodifferentiation.16 To inves-
tigate this possibility, we have conducted experiments
with two enantiomeric lithium enolates of dioxanone 4b,
generated separately in high enantiomeric excess by
using chiral lithium amides 9f or 10d (Scheme 5, Table
3). Three R-glyceraldehyde derivatives were used (19a -
c), and it was noted that the isopropylidene-protected
glyceraldehyde 19a was the most selective reagent (but
the yields were higher when 19b was used). Treatment
(12) Aoki, K.; Koga, K. Tetrahedron Lett. 1997, 38, 2505, and
references cited therein.
(13) (a) J urczak, J .; Zamojski, A. In Preparative Carbohydrate
Chemistry; Hanessian, S., Ed.; Dekker: New York, 1997; p 593. (b)
Collins, P.; Ferrier, R. Monosaccharides: Their Chemistry and Their
Roles in Natural Products; Wiley: New York, 1995; p 53.
(14) C.f.: ref 1c, p 1713.
(15) Review: J urczak, J .; Pikul, S.; Bauer, T. Tetrahedron 1986, 42,
447.
(16) (a) Masamune, S.; Choy, W.; Petersen, J . S.; Sita, L. R. Angew.
Chem., Int. Ed. Engl. 1985, 24, 1. (b) Heathcock, C. H.; White, C. T. J .
Am. Chem. Soc. 1979, 101, 7076.

5156 J . Org. Chem., Vol. 65, No. 17, 2000
Majewski and Nowak
Ta ble 3. Dou ble Ster eoselection In volvin g Ch ir a l
En ola tes of Dioxa n on e 4b a n d P r otected
R-Glycer a ld eh yd es 19 (cf., Sch em e 5)
entry lithium amide aldehyde 20a a :21a a :(?)a yieldb (%)
1
2
3
4
5
6
10d
10d
10d
9f
9f
9f
19a
19b
19c
19a
19b
19c
82:10: (8)
71:15: (14)
60:16: (14:10)
3:97
16:74: (10)
16:78: (6)
77
96
82
81
97
80
a
The question mark symbolizes additional (not identified)
b
byproduct(s). Combined yields of all products after purification.
F igu r e 1. Relevant elements of stereocontrol. Comparison of
axial (ax) and equatorial (eq) approach of an electrophile to
dioxanone enolate and of the Felkin-Ahn (F) and Cram
chelate (C) models applied to glyceraldehyde.
of 4b with the chiral lithium amide 10d , followed by
addition of the protected glyceraldehyde 19a yielded
compound 20a a as the major product (Table 3, entry 1),
which was easily purified by column chromatography.
The structure of 20a a was assigned to be a derivative of
D-tagatose by chemical correlation. A sample of 20a a was
reduced with NaBH4 followed by removal of both ac-
etonide groups to yield a mixture of two corresponding
hexitols (22, Scheme 5), the spectra of which were
compared with these obtained from a commercially
available sample (cf., Experimental Section). The enan-
tiomeric enolate, produced using base 9f, yielded the
corresponding D-psicose derivative 21a a (entry 4). Ste-
reoselective synthesis of ketohexose derivatives was thus
successfully realized.
We reported previously that acetal dioxanones (R )
H, R′ ) alkyl) were much more difficult to work with then
ketal dioxanones (R and R′ ) alkyl groups).5 Searching
for a better starting material, we briefly investigated two
dioxanones having the phenyl group at C-2 (compounds
4c and 4d ). Compound 4c did not offer significant
advantages over compound 4a (although the yields of
aldol addition experiments were often higher with 4c by
up to 20%, as in the case of lithium enolate added to
PhCHO). Compound 4d displayed a significantly lower
enantioselectivity in deprotonation experiments with
chiral lithium amides, which could be attributed to the
A value of the phenyl group being smaller than the A
value of tert-butyl,17 resulting in lower conformational
stabilization of the starting material.18
Ster eoch em ica l Con sid er a tion s. Reactions of diox-
anones, described above, show some unusual features.
We reported previously on striking electrophilicity of
dioxanones; their high reactivity toward nucleophiles
approaches the reactivity of aldehydes rather then ke-
tones, and these compounds are very prone to reduction
by LDA.5 These features are related to inductive effects
of the two oxygen atoms in the dioxanone ring. These
two oxygen atoms also precipitate stereoelectronic effects
which were used to rationalize the preference of 1,3-
dioxane-5-carboxylate esters for equatorial alkylation.19
Stereoselectivity of reactions involving dioxanone eno-
lates also has some unusual aspects: aldehydes add to
dioxanone enolates via equatorial attack (as exemplified
by the transformation of 4b to 6bb), the reactions with
protected glyceraldehydes follow the Cram chelate model,20
which results in the anti-syn isomers (20) being formed
faster than the anti-anti isomers (21), and not the
Felkin-Anh model (nucleophilic addition to glyceralde-
hyde usually favors the latter15), and addition of chiral
lithium enolates of 4b to glyceraldehyde derivatives is
highly stereoselective and shows reversal of the sense of
diastereoselectivity as the enolate is changed from the
R isomer to the S isomer (cf., Scheme 5 and Table 3). We
cannot fully rationalize all of these phenomena at this
time; however, some comments can be made. Reactions
of cyclic enolates with electrophiles were discussed in
detail by Evans,21 and it was pointed out that the axial
attack of the electrophile, albeit favorable on stereoelec-
tronic grounds, could be disadvantaged by steric effects.
This is likely the case in our system, where the axial
attack suffers an unfavorable 1,3-steric interaction from
the pseudoaxial Me group (Figure 1).
Nucleophilic attack on R-isopropylidene glyceraldehyde
usually favors the si face, in agreement with the Felkin-
Anh model, and results in the relative stereochemistry
of the two newly created stereogenic centers being
anti.15,22 Exceptions are known, however, and syn-selec-
tive addition was observed in several systems.15,23 Diox-
anones appear to be one of these exceptions; and the
addition of achiral dioxanone enolates to 19a proceeds
predominantly to the re face of the aldehyde to afford
tagatose derivative 20 as the major product (Scheme 5).
Chiral enolates of dioxanones apparently obey the rules
of a reagent-controlled system;24 chirality of the enolate
dominates reactivity. Examples of reagent-control sys-
tems of this type, where diastereotopic face selectivity of
the aldehyde reagent was reversed, have been reported
before.25
In summary, 1,3-dioxan-5-ones are promising synthetic
building blocks. Their boron enolates were generated
efficiently and provided good control of relative stereo-
chemistry in the aldol addition reaction. Reagents for
efficient enantioselective deprotonation of dioxanones in
up to 90% ee were identified. Enantioselective depro-
tonation combined with the reaction of the resulting
(20) Eliel, L. E. In Asymmetric Synthesis; Morrison, J . D., Ed.;
Academic Press: Toronto, 1983; Vol. 2, p 125.
(17) Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of
Organic Compounds; Wiley-Interscience: Toronto, 1994; p 695.
(18) The importance of “conformational anchoring” in cyclic ketones,
related to the belief that, due to stereoelectronic effects, axial protons
are abstracted by bases faster than equatorial protons was pointed
out before. In some cases, however, the conformational bias seems to
matter little; cf.: Majewski, M.; MacKinnon, J . Can. J . Chem. 1994,
72, 1699.
(21) Evans, D. A. In Asymmetric Synthesis; Morrison, J . D., Ed.;
Academic Press: Toronto, 1984; Vol. 3, p 1.
(22) Reetz, M. T. Angew. Chem., Int. Ed. Engl. 1984, 23, 556.
(23) (a) Mulzer, J .; Chucholowski, A. Angew. Chem., Int. Ed. Engl.
1982, 21, 777. (b) Ohdan, S.; Okamoto, T.; Maeda, S.; Ichikawa, T.;
Araki, Y.; Ishido, Y. Bull. Chem. Soc. J pn. 1973, 46, 981.
(24) Ref 17, p 969.
(25) (a) Grauert, M.; Schollkopf, U. Liebigs Ann. Chem. 1985, 1817.
(b) Roush, W.; Halterman, R. L. J . Am. Chem. Soc. 1986, 108, 294.
(19) Ndibwani, A.; Deslongchamps, P. Can. J . Chem. 1986, 64, 1788.

Aldol Reactions of Enolates of 1,3-Dioxan-5-ones
J . Org. Chem., Vol. 65, No. 17, 2000 5157
purified by DFC (hexane, then 4:1 hexane:ethyl acetate). The
yields reported refer to the chromatographically pure material.
P r oced u r e A2. En a n tioselective Dep r oton a tion in th e
P r esen ce of Lith iu m Ch lor id e (Ad d ed a s Solid ) F ollow ed
by Ald ol Rea ction . One of the chiral amines 9-15 (1.00
mmol) and LiCl (1.00 mmol, 42 mg) were dissolved in THF
(10 mL). The solution was cooled to 0 °C, and n-BuLi was
added (0.44 mL, 1.10 mmol, 2.5 M solution in hexanes). After
1 h, the solution was cooled to -78 °C, and dioxanone 4b (172
mg, 1.00 mmol in 0.5 mL of THF) was added over 5 min. After
stirring for 0.5 h, cyclohexanecarboxaldehyde 5b (0.13 mL, 1.10
mmol) was added and the reaction mixture was stirred for 5
min at -78 °C. The reaction was then quenched with a
concentrated pH 7 buffer (20 mL), and the mixture was
extracted with Et2O (3 × 50 mL). The extracts were washed
with brine (20 mL) and dried with MgSO4. The solvents were
evaporated, and the enantioselectivity of the reaction was
determined by 1H NMR using a chiral shift reagent, (+)-
Eu(hfc)3. The crude product was purified by DFC (hexane
followed by hexane:ethyl acetate, 4:1). The yields reported refer
to the chromatographically pure material.
P r ocedu r e B: Bor on -Mediated Aldol Reaction of 2-Su b-
stitu ted 1,3-Dioxa n -5-on e. This procedure was adapted from
the literature.30 A tertiary amine (either DIPEA or NEt3; 2.00
mmol) was dissolved in CH2Cl2 (10 mL), and the solution was
cooled to 0 °C. The dialkylboron halide or triflate (c.f, Table 1,
1.00 mmol) was added with stirring, followed by addition of
the dioxanone (1.00 mmol), and the resulting mixture was
stirred for 15 min. The aldehyde (1.10 mmol) was then added
at 0 °C (procedure B1), or, alternatively, the reaction mixture
was cooled to -78 °C and the aldehyde was added at this
temperature (procedure B2). After an additional 15 min of
stirring, the reaction was quenched with concentrated pH 7
buffer (20 mL) and extracted with Et2O (3 × 50 mL). The
combined extracts were washed with brine (2 × 10 mL) and
dried with MgSO4. The solvents were evaporated, and the
residue was worked up in one of the following ways:
lithium enolates with protected glyceraldehyde provided
a quick synthetic entry into chiral polyoxygenated natu-
ral products. The utility of the method was demonstrated
by synthesis of monosaccharide derivatives having all the
OH groups protected differently, a feature which might
be useful for further synthetic use of these sugars.
Exp er im en ta l Section
Gen er a l Meth od s. All air-sensitive reactions were carried
out under nitrogen. Tetrahydrofuran (THF) and Et2O were
distilled under nitrogen from sodium and benzophenone.
Dichloromethane and diisopropylamine were distilled from
CaH2. LiCl was dried at 130-150 °C under vacuum overnight,
and it was used either as a solid or as a solution in THF.
Flash column chromatography (FCC)26 and dry flash chro-
matography (DFC)27 were carried out using Merck silica gel
60 (230-400 mesh) and Sigma silica gel type H (10-40 µm),
respectively. Thin-layer chromatography (TLC) was performed
on precoated glass plates (Merck, silica gel 60, F254). The spots
were detected using UV light (254 nm) or with a developing
solution prepared by dissolving concentrated H2SO4 (50 g),
cerium(IV) sulfate (10 g), and phosphomolybdic acid hydrate
(40 g) in water (1 L) followed by charring on a hot plate.
Concentrated phosphate buffer, used to quench reactions, was
prepared by dissolving Na2HPO4 (47 g) and NaH2PO4 (32 g)
in H2O (0.5 L). Optical rotations were measured on a Perkin-
Elmer 241 polarimeter (1 dm, 1 mL cell); all concentrations
are given in grams per 100 mL. Parent chiral amines of
reagents 9-15 were synthesized by following the well-
established literature procedures,28 and the dioxanone sub-
strates (4) were synthesized as described before.5 Derivatives
of glyceraldehyde (19) were synthesized according to literature
procedures.29
Gen er a l P r oced u r es. P r oced u r e A1. En oliza tion of
Dioxa n on e a n d Ald ol Ad d ition . Diisopropylamine (0.17 mL,
1.20 mmol) was dissolved in THF (10 mL). The solution was
cooled to 0 °C, and n-BuLi was added (0.44 mL, 1.10 mmol,
2.5 M solution in hexanes). After 0.5 h, the solution of the
lithium amide was cooled to -78 °C and dioxanone 4 (1.00
mmol), dissolved in THF (0.5 mL), was added over 5 min. After
stirring for an additional 5 min, the aldehyde (1.10 mmol) was
added. The mixture was stirred for 5 min at -78 °C and then
was quenched with a concentrated pH 7 buffer (20 mL) and
extracted with Et2O (3 × 50 mL). The combined extracts were
washed with brine (20 mL) and dried with MgSO4. The
solvents were evaporated, and diastereoselectivity of the
(a) The residue was dissolved in MeOH (9 mL) and cooled
to 0 °C. Concentrated pH 7 buffer (3 mL) and H2O2 (3 mL,
30%) were next added. The solution was stirred at 0 °C for 3
h. Next, Et2O (150 mL) was added, and the solution was
washed with saturated NaHCO3 (2 × 15 mL) and brine (15
mL) and was dried with MgSO4. The solvents were then
removed affording the crude product.
(b) The residue was dissolved in a cold (0 °C) acetone
solution of dimethyldioxirane (30 mL, 0.1 M). After 1 h, acetone
was removed and the mixture was dissolved in Et2O (100 mL);
the solution was washed with saturated NaHCO3 (2 × 15 mL)
and brine (15 mL) and was dried with MgSO4. The solvents
were then removed affording the crude product.
1
reaction was determined by H NMR. The crude product was
(c) The residue was dissolved in CH2Cl2 (50 mL). The
solution was cooled to -78 °C and was saturated with ozone.
The excess of ozone was removed with a stream of argon. Next,
the solution was washed with saturated NaHCO3 (2 × 15 mL)
and was then dried with MgSO4. The solvents were then
removed, affording the crude product.
(26) Still, W. C.; Khan, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923.
(27) Harwood: L. M. Aldrichim. Acta 1985, 18, 25.
(28) (a) Rossiter, B. E.; Eguchi, M.; Miao, G.; Swingle, N. M.;
Hernandez, A. E.; Vickers. D.; Fluckiger, E.; Patterson, R. G.; Deddy,
K. V. Tetrahedron 1993, 49, 965. (b) Cain, C. M.; Cousins, R. P. C.;
Coumbarides, G.; Simpkins, N. S. Tetrahedron 1990, 46, 523. (c)
Shiray, R.; Aoki, K.; Sato, D.; Kim, H.-D.; Murakata, M.; Yasukata,
T.; Koga, K. Chem. Pharm. Bull. 1994, 42, 690. (d) J uaristi, E.; Murer,
P.; Seebach, D. Synthesis 1993, 1243. (e) Barch, R. F.; Bernstein, M.
D., Durst, H. D. J . Am. Chem. Soc. 1971, 93, 2897. (f) Wittig, G.; Thiele,
U. Liebigs Ann. Chem. 1969, 726, 1. (g) Overman, L. E.; Okazaki, M.
E.; Mishra, P. Tetrahedron Lett. 1986, 27, 4391. (h) Overberger, C. G.;
Marullo, N. P.; Hiskey, R. G. J . Am. Chem. Soc. 1961, 83, 1374. (i)
Fraser, R. R.; Boussard, G.; Postescu, I. D.; Whiting, J . J .; Wigfield,
Y. Y. Can. J . Chem. 1972, 51, 1109. (j) Guthrie, R. D.; J aeger, D. A.;
Meister, W.; Cram, D. J . J . Am. Chem. Soc. 1971, 93, 5137. (k) Kamer,
P. C. J .; Cleij, M. C.; Nolte, R. J . M.; Harada, T.; Hezemans, A. M. F.;
Drenth, W. J . Am. Chem. Soc. 1988, 110, 1581. (l) Paquette, L. A.;
Freeman, J . P. J . Am. Chem. Soc. 1969, 91, 7548. (m) Bordwell, F. G.;
Lynch, T.-Y. J . Am. Chem. Soc. 1989, 111, 7558. (n) Duhamel, L.;
Ravard, A.; Plaquevent, J .-C.; Davoust, D. Bull. Soc. Chim. Fr. 1990,
127, 787. (o) Higashiyama, K.; Inoue, H.; Yamauchi, T.; Takahashi,
H. J . Chem Soc., Perkin Trans. I 1995, 111.
Diastereoselectivity of the reaction was determined by 1H
NMR on the crude product. The crude product was then
purified by DFC (hexane, then 1:1 hexane:AcOEt). The yields
reported refer to the combined yields of all isomers.
a n ti- a n d syn -4-(Hyd r oxyp h en ylm eth yl)-2,2-d im eth yl-
1,3-d ioxa n -5-on e (6a a , 7a a ).5 Yields and selectivity of the
boron-mediated aldol addition of dioxanone 4a to PhCHO are
summarized in Table 1. E.g., procedure A1 afforded the
mixture of 6a a and 7a a as a colorless oil in 55% yield. The
diastereoselectivity of the reaction was 65:35. The isomers
were separated by DFC (hexane-hexane:ethyl acetate (4:1))
to provide the minor isomer 7a a in 16% yield and the major
isomer 6a a in 27% yield. Procedure B2b afforded a mixture of
6a a and 7a a in 81% yield and in a ratio of 96:4. Spectral
properties of 6a and 7a were in agreement with data published
before.5
(29) (a) Schmid, C. R.; Bryant, J . D. Org. Synth. 1995, 72, 6. (b)
Ohgo, Y.; Yoshimura, J .; Kono, M.; Sato, T. Bull. Chem. Soc. J pn. 1969,
42, 2957. (c) Sugiyama, T.; Sugawara, H.; Watanabe, M.; Yamashita,
K. Agric. Biol. Chem. 1984, 48, 1841.
(30) Evans, D. A.; Nelson, J . V.; Vogel, E.; Taber, T. R. J . Am. Chem.
Soc. 1981, 103, 3099.

5158 J . Org. Chem., Vol. 65, No. 17, 2000
Majewski and Nowak
4-(Cycloh exylh yd r oxym eth yl)-6-(h yd r oxyp h en ylm eth -
yl)-2,2-d im eth yl-1,3-d ioxa n -5-on e (8a ). Triethylamine (0.28
mL, 2.00 mmol) was dissolved in CH2Cl2 (10 mL), and the
solution was cooled to 0 °C. Dicyclohexylboron chloride (2.00
mL, 0.5 M solution in hexane, 1.00 mmol) was added, and the
solution was stirred for 5 min. Next, dioxanone 4a (0.130 g,
1.00 mmol) was added, and, after stirring for 15 min, PhCHO
(0.10 mL, 1.00 mmol) was added. After stirring for another 15
min, NEt3 (0.28 mL, 2.00 mmol) was added followed by
addition of dicyclohexylboron chloride (2.00 mL, 0.5 M solution
in hexane, 1.00 mmol) and, after 15 min, aldehyde 5b (0.17
mL, 1.50 mmol). After stirring for 15 min, the reaction was
quenched with concentrated pH 7 buffer (20 mL) and extracted
with diethyl ether (3 × 50 mL). The extracts were washed with
brine (2 × 10 mL) and dried with MgSO4. The solvents were
evaporated, and the residue was dissolved in methanol (18 mL)
and cooled to 0 °C. Concentrated pH 7 buffer (6 mL) and
hydrogen peroxide (6 mL, 30%) were next added. The solution
was stirred at 0 °C for 3 h, Et2O (150 mL) was added, and the
solution was washed with saturated NaHCO3 (2 × 15 mL) and
brine (15 mL) and was dried with MgSO4. The solvents were
removed, and the mixture was analyzed by 1H NMR. Four
isomers were detected in the ratio of 83:09:6:2. The major
product was isolated by DFC (9:1 hexane:CH2Cl2 followed by
9:1:2 hexane:CH2Cl2:AcOEt), which gave a colorless liquid (224
mg, 64%). Only the major product 8a was characterized.
Properties: Rf 0.35 (2:1 hexane:ethyl acetate); IR 3517, 1733,
90% ee, was obtained in 61% yield. The (2R,4S)-6bb enanti-
omer was obtained in an analogous way using the (R)-N-(2,2,2-
trifluoroethyl)-1-phenylethylamine in 90% ee and 70% yield.
Specific rotation: the 2S,4R-isomer of 6bb, [R]25D +56.3 (c 0.8,
MeOH, 90% ee); the 2R,4S-isomer had [R]25 -56.1 (c 1.0,
D
MeOH, 90% ee).
cis-2-(Cycloh exylh yd r oxym eth yl)-2-m eth yl-2-p h en yl-
1,3-d ioxa n -5-on e (6d b). Procedure A1 afforded the racemic
compound 6d b as a white solid in 55% yield. Only one
diastereisomer was detected. The optically active (2S,4R)-6d b
was obtained by procedure A2: (S)-N-(2,2,2-trifluoroethyl)-1-
phenylethylamine was used instead of diisopropylamine to
generate the lithium amide. The reaction was performed in
the presence of 1 equiv of LiCl. Pure compound 6d b, having
54% ee, was obtained in 64% yield. Properties: Rf 0.31 (4:1
hexane:ethyl acetate); [R]25 +35.9 (c 1.0 methanol, 54% ee,
D
2S,4R-6d b); IR 3523, 1736, 1169 cm-1; 1H NMR 7.50-7.58 (m,
2H), 7.31-7.45 (m, 3H), 4.42 (dd, J 1 ) 17.0, J 2 ) 1.5 Hz, 1H),
4.18 (d, J ) 17.0 Hz, 1H), 3.92 (d, J ) 7.5 Hz, 1H), 3.65-3.75
(m, 1H), 3.09 (d, J ) 2.0 Hz, 1H), 1.48-1.81 (m, 5H), 1.69 (s,
3H), 0.91-1.38 (m, 6H); 13C NMR 211.3, 140.9, 128.5, 128.3,
125.9, 101.6, 74.7, 74.5, 67.1, 38.3, 29.7, 26.9, 26.5, 26.3, 26.0,
25.7; MS (CI-NH3) 305 (M + 1, 31), 193 (100), 121 (18), 105
(27); HRMS (CI-N3) 305.1753 (M + 1), calcd for C18H25O4
305.1753.
Red u ction of Com p ou n d 6bb. Reduction of the ketone
group in 6bb was undertaken during structural assignment
studies. It was established, that this compound could be
reduced selectively to the cis-cis diol 16 by using DIBAH or to
the cis-trans epimer of 16, by using Raney nickel. The latter
compound yielded well-defined crystals allowing crystallo-
graphic determination of the relative stereochemistry.5
Com p ou n d 16. Aldol 6bb (0.860 g, 3.02 mmol) was dis-
solved in THF (25 mL) and cooled to -78 °C. DIBAH (2.00
mL, 1.5 M solution in toluene, 3.00 mmol) was added, and the
reaction was stirred for 1 h. The reaction was quenched with
MeOH (10 mL), and the mixture was warmed to room
temperature. Saturated NaHCO3 solution (5 mL) was added,
followed by Celite (10 g), and Et2O (100 mL). The slurry was
stirred for 15 min and filtered. The filter cake was washed
with Et2O (2 × 50 mL). The resulting organic solution was
washed with brine (2 × 20 mL) and dried with MgSO4, and
the solvents were evaporated. Two diastereoisomers were
detected in the crude product in a ratio of 92:8 (1H NMR). The
major product was isolated by DFC (hexane followed by 1:1
hexane:AcOEt) which gave a white solid (0.579 g, 67%). The
optically active 2S,4R,5S-16 was prepared in the same way
using the 2S,4R-6bb aldol having 70% ee as the starting
material. Properties: Rf 0.38 (2:1 hexane:ethyl acetate); mp
1
1221 cm-1; H NMR 7.18-7.43 (m, 5H), 4.90 (d, J ) 7.0 Hz,
1H), 4.33 (d, J ) 7.0 Hz, 1H), 4.07 (d, J ) 7.0 Hz, 1H), 3.64
(dd, J 1 ) 7.0, J 2 ) 2.5 Hz, 1H), 3.10 (br s, 2H), 1.42-1.89 (m,
6H), 1.37 (s, 3H), 1.25 (s, 3H), 1.02-1.40 (m, 5H); 13C NMR
213.9, 139.1, 127.9, 127.1, 101.4, 76.3, 74.3, 73.7, 72.6, 38.5,
29.4, 26.4, 26.3, 26.0, 23.6, 23.4; MS (CI-NH3) 366 (M + 18,
90), 260 (25), 254 (61), 243 (33), 240 (100), 219 (21), 184 (15),
165 (15), 161 (16); HRMS (CI-NH3) 366.2281 (M + 18), calcd
for C20H32NO5 366.2280.
4,6-Bis-(h ydr oxyph en ylm eth yl)-2,2-dim eth yl-1,3-dioxan -
5-on e (8b). The procedure described above for compound 8a
was followed on the same scale. Five isomers were detected
in the crude product in the ratio of 80:10:04:03:03. The major
product was isolated by DFC (9:1 hexane:CH2Cl2 followed by
9:1:2 hexane:CH2Cl2:AcOEt) to provide a colorless liquid (206
mg, 60%). Only the major product 8b was characterized.
Properties: Rf 0.32 (2:1 hexane:ethyl acetate); IR 3530, 1738,
1
1224, 1027 cm-1; H NMR 7.20-7.39 (m, 10H), 4.88 (d, J )
6.5 Hz, 2H), 4.22 (d, J ) 6.5 Hz, 2H), 3.39 (br s, 2H), 1.18 (s,
6H); 13C NMR 212.1, 139.0, 128.0, 127.0, 101.7, 76.4, 72.6, 23.3;
MS (CI-NH3) 360 (M + 18, 21), 254 (57), 240 (16), 237 (16),
220 (15), 219 (100), 178 (31), 165 (25), 161 (35), 131 (91), 124
(32), 113 (17), 105 (44); HRMS (CI-NH3) 360.1806 (M + 18),
calcd for C20H26NO5 360.1811.
96-99 °C; [R]25 +20.9 (c 0.7, chloroform, 70% ee, 2S,4R,5S-
D
isomer); IR 3379, 1167, 1066 cm-1; H NMR 4.00-4.13 (br s,
1
cis-2-ter t-Bu tyl-4-(cycloh exylh yd r oxym eth yl)-2-m eth -
yl-1,3-d ioxa n -5-on e (6b b ). Procedure A1 afforded racemic
compound 6bb as a white solid in 61% yield. Only one
diastereoisomer was detected. The spectral data were in
agreement with values published before.5 Compound 6bb was
also synthesized from the trimethylsilyl enol ether of 2-tert-
butyl-2-methyl-1,3-dioxan-6-one as follows: TMS enol ether
of 4b (synthesized according to a general procedure published
previously,5 122 mg, 0.50 mmol) was dissolved in THF (10 mL).
The solution was cooled to -78 °C and n-BuLi was added (0.22
mL, 0.55 mmol, 2.5 M solution in hexanes). After 1 h,
cyclohexanecarboxaldehyde (67 µL, 0.55 mmol) was added and,
after an additional 5 min the reaction was quenched with
concentrated pH 7 buffer (20 mL) and extracted with Et2O
(3 × 50 mL). The extracts were washed with brine (20 mL)
and dried with MgSO4. The solvents were evaporated, and the
crude product was purified by SCC (4:1 hexane:ethyl acetate)
to provide the pure aldol 6bb (136 mg, 96%). Only one isomer
was detected (1H NMR).
1H), 3.75-3.90 (m, 1H), 3.50-3.70 (m, 4H), 2.94-3.03 (br s,
1H), 1.50-1.83 (m, 6H), 1.36 (s, 3H), 0.87-1.30 (m, 5H), 0.93
(s, 9H); 13C NMR 102.2, 80.1, 71.2, 67.7, 63.7, 39.3, 39.1, 29.2,
26.5 (2×), 26.2, 25.6, 24.6, 12.1; MS (CI-N3) 288 (M + 2, 26),
287 (M + 1, 100), 173 (16), 169 (11); Anal. Calcd for C16H30O4:
C, 67.10; H, 10.55. Found: C, 67.89; H, 10.68.
Com p ou n d 17. Compound 16 (579 mg, 2.02 mmol, 70% ee)
was dissolved in MeOH (25 mL). Water (1 mL) and concen-
trated HCl (0.5 mL) were added, and the resulting solution
was refluxed for 15 min. Next, the solvent was evaporated,
and the residual water was removed by evaporation with
ethanol (2 × 50 mL) to provide a white solid (Rf 0.04, 1:1
hexane:ethyl acetate) in quantitative yield. The crude tetraol
(413 mg, 2.02 mmol) was dissolved in THF (10 mL) and treated
with 2,2-dimethoxypropane (1 mL), acetone (1 mL), and
concentrated HCl (0.1 mL). After 0.5 h, when the TLC analysis
indicated that all the starting material was consumed, the
reaction was quenched with concentrated NH4OH (1 mL). The
solvent was evaporated, and the residue was dissolved in Et2O
(100 mL). The resulting solution was washed with brine (3 ×
10 mL), dried with MgSO4, and concentrated. The GC analysis
indicated the presence of three products in the ratio of 20:54:
26. DFC (hexane followed by 1:1 hexane:ethyl acetate) pro-
vided the three products in pure form as colorless oils.
The optically active (2S,4R)-6bb enantiomer was obtained
by procedure A2: (S)-N-(2,2,2-trifluoroethyl)-1-phenylethyl-
amine was used instead of diisopropylamine to generate the
lithium amide 9f. The reaction was performed in the presence
of 1 equiv of LiCl. The pure compound (2S,4R)-6bb, having

Aldol Reactions of Enolates of 1,3-Dioxan-5-ones
J . Org. Chem., Vol. 65, No. 17, 2000 5159
Compound 17 (the 3,4-acetonide, side chain numbered from
the cyclohexane ring) was the major product (eluted second)
and was obtained in 50% yield. The other products were
identified as the 1,2-acetonide (20%) and the bis-acetonide
product consisted of three isomers in a ratio of 74:16:10 (97%
yield). Only the major isomer 21bb was characterized. Proper-
ties: Rf 0.39 (2:1 hexane:ethyl acetate); [R]25 +62.9 (c 1.1,
D
1
chloroform); H NMR 4.34-4.46 (m, 1H), 4.18-4.30 (m, 1H),
(13%). Properties: Rf 0.39 (1:1 hexane:ethyl acetate); [R]25
3.85-4.15 (m, 5H), 2.65 (br s, 1H), 1.30-1.64 (m, 10H), 1.35
(s, 3H), 1.01 (s, 9H); 13C NMR 206.0, 110.3, 103.3, 78.6, 74.3,
74.1, 69.4, 66.7, 40.2, 35.8, 34.5, 25.1, 25.0, 23.9, 23.7, 15.9;
IR 3464, 1739, 1162, 1096, 1052 cm-1; MS 342 (M+, 20), 299
(32), 199 (21), 141 (70), 127 (100), 101 (26), 99 (25), 83 (24), 81
(20), 57 (23), 55 (29); HRMS (EI) 342.2041 (M), calcd for
D
-6.0 (c 2.4, chloroform, sample having 70% ee); 1H NMR 3.78-
4.22 (m, 4H), 3.33-3.45 (m, 1H), 2.75 (br s, 2H), 0.91-2.00
(m, 11H), 1.38 (s, 3H), 1.33 (s, 3H); 13C NMR 108.8, 108.3, 77.4,
70.8, 68.0, 65.6, 39.8, 27.7, 28.1, 26.7, 26.5, 26.2, 25.7.
Cleavage of compound 17 (0.27 mmol sample having 70%
optical purity) with lead tetraacetate (0.30 mmol) in CDCl3 (2
mL, room temperature, 30 min) yielded a sample of 18 having
C
18H30O6 342.2042.
(-)-(2R,4S)-2-(ter t-Bu tyl)-4-[(2S)-(2R)-1,4-d ioxa sp ir o-
[R]25 -13.5 (c 3.3, CHCl3).
D
[4.5]d e c-2-yl(h yd r oxy)m e t h yl]-2-m e t h yl-1,3-d ioxa n -5-
on e (20bb). Procedure A2 with base10d gave a colorless oil
in 96% yield. The crude product consisted of three isomers in
a ratio of 71:15:14. Only the major isomer 20bb was charac-
(-)-1,3:5,6-Di-O-isop r op ylid en e-D-ta ga tose (20a a ) a n d
(+)-1,3:5,6-Di-O-isop r op ylid en e-D-p sicose (21a a ). Proce-
dure A1 gave the mixture of aldols in 70% yield. Three
diastereoisomers were formed in the ratio of 55:33:12. Only
the two major isomers were isolated and characterized.
Procedure B2a gave a mixture of aldols in 39% yield containing
two diastereoisomers in the ratio of 85:15. Procedure B2b gave
a mixture of aldols in 59% yield containing two diastereoiso-
mers in the ratio of 87:13. Properties: Compound 20a a (major
product). Mp 103-105 °C (white solid); Rf 0.23 (2:1 hexane:
ethyl acetate); [R]25D -167.2 (c 1.1, chloroform); 1H NMR 4.26-
4.32 (m, 3H), 3.98-4.10 (m, 2H), 3.83-3.91 (m, 2H), 3.17 (d,
terized. Properties: Rf 0.38 (2:1 hexane:ethyl acetate); [R]25
D
1
-45.0 (c 1.0, chloroform); H NMR 4.10-4.48 (m, 3H), 3.91-
4.08 (m, 2H), 3.58-3.89 (m, 2H), 1.26-1.63 (m, 11H), 1.36 (s,
3H), 0.97 (s, 9H); 13C NMR 209.0, 109.8, 104.2, 76.1, 74.5, 70.9,
69.6, 65.3, 40.2, 35.8, 34.8, 25.1, 25.0, 23.9, 23.7, 16.2; IR 3469,
1739, 1164, 1103, 1042 cm-1; MS 342 (M, 14), 299 (30), 199
(23), 171 (22), 141 (60), 127 (100), 126 (49), 101 (37), 99 (37),
55 (20); HRMS (EI) 342.2042 (M), calcd for C18H30O6 342.2042.
(-)-(2S,4R)-4-[(1R,2R)-2,3-Bis(b en zyloxy)-1-h yd r oxy-
p r op yl]-2-(ter t-bu tyl)-2-m eth yl-1,3-d ioxa n -5-on e (21bc). A
modified procedure A2 with base 9f (the aldehyde was used
neat) gave a colorless oil in 80% yield. The crude product
consisted of three isomers in a ratio of 78:16:6. Only the major
isomer 21bc was characterized. Properties: Rf 0.35 (2:1
hexane:ethyl acetate); [R]25D -33.0 (c 1.4, chloroform); 1H NMR
7.18-7.40 (m, 10H), 4.8--4.81 (m, 1H), 3.88-4.65 (m, 8H),
3.65-3.82 (m, 2H), 3.30 (br s, 1H), 1.30 (s, 3H), 0.99 (s, 9H);
13C NMR 210.3, 138.3, 137.9, 128.2, 128.1, 127.5 (2×), 127.3,
127.2, 103.7, 76.2, 74.5, 73.3, 73.0, 71.4, 70.0, 69.2, 40.0, 25.1,
15.9; IR 3425, 1728, 1103 cm-1; MS (CI-NH3) 460 (M + 18,
44), 443 (M + 1, 16), 442 (M, 13), 425 (26), 335 (27), 317 (47),
288 (73), 181 (48), 173 (23), 108 (33), 91 (100); HRMS (EI)
442.2356 (M), calcd for C26H34O6 442.2355.
(+)-(2R,4S)-4-[(1S,2R)-2,3-Bis(b en zyloxy)-1-h yd r oxy-
p r op yl]-2-(ter t-bu tyl)-2-m eth yl-1,3-d ioxa n -5-on e (20bc). A
modified procedure A2 with base 10d (the aldehyde was used
neat) gave a colorless oil in 82% yield. The crude product
consisted of four isomers in a ratio of 60:16:14:10. Only the
major isomer 20bc was characterized. Properties: Rf 0.34 (2:1
hexane:ethyl acetate); [R]25D +14.6 (c 1.2, chloroform); 1H NMR
7.15-7.40 (m, 10H), 3.88-4.95 (m, 9H), 2.58-3.84 (m, 2H),
2.15 (s, 1H), 1.35 (s, 3H), 1.05 (s, 9H); 13C NMR 205.7, 138.1,
137.8, 128.2, 128.1, 127.5 (2×), 127.3, 127.1, 102.8, 77.9, 77.2,
73.2, 72.9, 71.8, 69.8, 68.8, 39.7, 25.1, 15.0; IR 3427, 1732, 1105
cm-1; MS (CI-NH3) 460 (M + 18, 46), 443 (M + 1, 13), 442 (M,
11), 425 (29), 335 (20), 317 (30), 288 (60), 181 (26), 108 (32),
91 (100); HRMS (EI) 442.2357 (M), calcd for C26H34O6 442.2355.
Str u ctu r e Deter m in a tion of Ketoh exose Der iva tives
20 a n d 21. Samples of commercially available D-fructose,
L-sorbose, D-psicose, and D-tagatose were converted to the
corresponding alcohols by the following procedure (representa-
tive example):
J ) 3.5 Hz, 1H), 1.48 (s, 3H), 1.45 (s, 3H), 1.24 (s, 3H), 1.07 (s,
1
3H); H NMR (benzene-d6) 4.25-4.32 (m, 2H), 3.98 (dd, J 1
)
8.0, J 2 ) 7.5 Hz, 1H), 3.74-3.81 (m, 2H), 3.65 (m, 1H), 3.59
(d, J ) 17.5 Hz, 1H), 3.16 (d, J ) 3.0 Hz, 1H), 1.48 (s, 3H),
1.35 (s, 3H), 1.24 (s, 3H), 1.07 (s, 3H); 13C NMR 210.4, 109.1,
101.3, 75.2, 73.4, 70.1, 66.7, 65.6, 26.3, 25.6, 23.6, 23.5; IR 3475,
1744, 1374, 1222, 1068 cm-1; MS: (CI-NH3) 278 (M + 18, 27),
261 (M + 1, 44), 243 (100), 220 (81), 203 (51), 131 (25), 101
(46); Anal. Calcd for C12H20O6: C, 55.37; H, 7.74. Found: C,
55.57; H, 7.89.
Com p ou n d 21a a (Min or P r od u ct). Colorless oil; Rf 0.16
(2:1 hexane:ethyl acetate); [R]25 +144.7 (c 3.0, chloroform);
D
1H NMR 4.46 (dd, J 1 ) 3.5, J 2 ) 1.5 Hz, 1H), 4.26-4.36 (m,
2H), 3.99-4.09 (m, 4H), 2.73 (d, J ) 6.0 Hz, 1H), 1.45 (s, 6H),
1.35 (s, 3H), 1.32 (s, 3H); 13C NMR 207.3, 109.2, 100.6, 75.9,
74.9, 71.9, 67.0, 66.1, 26.2, 25.2, 24.4, 23.2; IR 3469, 1747, 1374,
1223, 1069; MS (CI-NH3) 278 (M + 18, 56), 261 (M + 1, 100),
243 (88), 220 (63), 203 (51), 101 (47); HRMS (FAB) 261.1333
(M + 1), calcd for C12H21O6 261.1338.
(+)-(2S,4R)-2-(ter t-Bu t yl)-4-[(2R)-[(4R)-2,2-d im et h yl-
1,3-d ioxola n -4-yl](h yd r oxy)m eth yl]-2-m eth yl-1,3-d ioxa n -
5-on e (21ba ). Procedure A2, with 9f, gave a colorless oil in
81% yield. The crude product consisted of two isomers in a
ratio of 93:7. Only the major isomer was characterized.
Properties: Rf 0.30 (2:1 hexane:ethyl acetate); [R]25 +59.6 (c
D
1
1.3, chloroform); H NMR 4.24-4.51 (m, 3H), 3.93-4.21 (m,
4H), 2.55 (br s, 1H), 1.40 (s, 3H), 1.34 (s, 3H), 1.32 (s, 3H),
1.06 (s, 9H); 13C NMR 206.6, 109.6, 103.4, 78.3, 74.6, 74.1, 69.5,
68.5, 66.9, 40.1, 26.1, 25.1, 15.9; IR 3488, 1739, 1151, 1058
cm-1; MS (CI-NH3) 303 (M + 1, 100), 287 (20), 285 (20), 245
(71), 203 (30), 173 (21), 131 (24), 115 (18), 101 (57); HRMS
(CI-NH3) 303.181 (M + 1), calcd for C15H27O6 302.181.
(-)-(2R,4S)-2-(ter t-Bu tyl)-4-[(2S)-[(4R)-2,2-d im eth yl-1,3-
d ioxola n -4-yl](h yd r oxy)m et h yl]-2-m et h yl-1,3-d ioxa n -5-
on e (20a a ). Procedure A2, with base 10d , gave a colorless oil
in 77% yield. The crude product consisted of three isomers in
a ratio of 82:10:8. Only the major isomer was characterized.
D-Glu citol a n d D-Ma n n itol.31 D-Fructose (0.500 g, 2.78
mmol) was dissolved in EtOH (25 mL), RaNi (0.500 g) was
added, and the resulting slurry was refluxed overnight in
hydrogen atmosphere. Next, the nickel catalyst was filtered
off on a Celite pad, and the solvent was evaporated. The oily
product (0.490 g, 98%) was dissolved in D2O (1.0 mL), and the
13C NMR spectrum was recorded. The methyl signal of the
residual ethyl alcohol (17.4 ppm) was used as the internal
standard. The product consisted of a mixture of two isomers,
D-glucitol and D-mannitol, in the ratio of 85:15.
The 13C NMR data (in D2O), for different hexahexaols
(hexitiols) thus obtained, were as follows: 2S,3R,4R,5R-isomer
(D-glucitol): 73.4, 71.6, 71.5, 70.1, 63.4, 62.9; 2R,3R,4R,5R-
isomer (D-mannitol): 71.3, 69.7, 63.8; 2S,3R,4R,5S-isomer (L-
Properties: Rf 0.29 (2:1 hexane:ethyl acetate); [R]25 -45.8 (c
D
1
1.2, chloroform); H NMR 4.12-4.50 (m, 4H), 4.06 (dd, J 1
)
8.0, J 2 ) 6.5 Hz, 1H), 3.93 (dd, J 1 ) 6.5, J 2 ) 5.0 Hz, 1H), 3.84
(dd, J 1 ) 7.5, J 2 ) 0.5 Hz, 1H), 3.10 (br s, 1H), 1.46 (s, 3H),
1.42 (s, 3H), 1.37 (s, 3H), 1.04 (s, 9H); 13C NMR 209.1, 109.3,
104.2, 76.1, 75.1, 71.2, 69.6, 65.8, 40.2, 26.3, 25.4, 25.1, 16.2;
IR 3485, 1738, 1158, 1058 cm-1; MS (CI-NH3) 303 (M + 1, 78),
245 (42), 220 (31), 203 (26), 173 (100), 148 (28), 131 (51), 115
(63), 101 (60); HRMS (CI-NH3) 303.181 (M + 1), calcd for
C
15H27O6 302.181.
(+)-(2S,4R)-2-(ter t-Bu tyl)-4-[(2R)-1,4-dioxaspir o[4.5]dec-
2-yl(h yd r oxy)m eth yl]-2-m eth yl-1,3-d ioxa n -5-on e (21bb).
Procedure A2 with base 9f gave a colorless oil. The crude
(31) Angyal, S. J .; Le Fur, R. Carbohydr. Res. 1980, 84, 201.

5160 J . Org. Chem., Vol. 65, No. 17, 2000
Majewski and Nowak
iditol): 72.2, 71.6, 63.2; 2R,3R,4S,5S-isomer (23b, D-allitol):
72.9, 72.8, 63.0; 2R,3R,4S,5R-isomer (23a ) 22b, D-talitol):
73.2, 72.1, 71.3, 71.0, 63.6, 62.6; 2S,3R,4S,5R-isomer (22a ,
D-galactitol): 70.8, 70.0, 63.8.
Hexitols fr om Com p ou n d s 20 a n d 21. The following two
step procedure was used:
methyl signal of residual ethyl alcohol (17.4 ppm) was used
as the internal standard.
Compound 20a a yielded initially (step 1) a mixture of two
isomers having Rf’s of 0.24 and 0.26 (1:1 hexane:ethyl acetate)
in 86% yield. Step 2 provided a mixture of 22b and 22a in a
ratio of 67:33 and in 99% yield.
Step 1. A sample of 20a a or 21a a (0.40 mmol) was dissolved
in methanol (5 mL). The solution was cooled to 0 °C, and
sodium borohydride (15 mg, 0.40 mmol) was added. After
strirring for 45 min, the reaction was quenched with a
concentrated pH 7 buffer (2 mL), extracted with Et2O (100
mL), washed with brine (2 × 15 mL), and dried with MgSO4.
The solvents were evaporated, and the residue was purified
by SCC (3:1 hexane:dichloromethane f 3:1:3 hexane:dichlo-
romethane:ethyl acetate).
Compound 21a a yielded a mixture of two isomers having
Rf’s of 0.22 and 0.12 (1:1 hexane:ethyl acetate) in 78% in step
1. Step 2 provided a mixture of 23b and 23a in a ratio of 64:
36 and in 92% yield.
Compounds 20ba , 20bb, 20bc and 21ba , 21bb, 21bc were
converted to mixtures of the corresponding alcohols in the
same way. In each case compound 20 yielded a mixture of 22a
and 22b, and compound 21 afforded 23a and 23b.
Step 2. The resulting diol was next dissolved in MeOH.
Concentrated hydrochloric acid (0.5 mL) was added, and the
solution was refluxed for 1 h. Next, EtOH (50 mL) was added,
and the solvents were evaporated. The residue was dissolved
in D2O (0.5 mL), and 13C NMR spectrum was recorded. The
Ack n ow led gm en t. The authors are grateful to the
Natural Sciences and Engineering Research Council of
Canada for financial support.
J O0002238