Stereoelectronic control of oxazolidine ring-opening: Structural and chemical evidences

Article abstract of DOI:10.1016/S0040-4020(02)01287-5

Ring opening of oxazolidines derived from tris(hydroxymethyl)aminomethane, L-serine and L-threonine was investigated. It was shown that n(N)→σ^*(C-O) electron delocalization (endo-anomeric effect) occurring in the five-membered ring plays a major role in the cleavage of the intracyclic C-O bond. The present work establishes that when the nitrogen lone pair is conjugated with a carbonyl group (n(N)→π(C=O) delocalization) as happens in N-acyloxazolidines, both hydrolysis and reductive ring-opening become much more difficult as a consequence of a concomitant decrease of oxygen basicity and of an increase of the intracyclic C-O bond strength.

Full text of DOI:10.1016/S0040-4020(02)01287-5

Tetrahedron 58 (2002) 9559–9566
Stereoelectronic control of oxazolidine ring-opening: structural
and chemical evidencesq
†
‡
,
a
b
a
Jimmy Selambarom, Sophie Monge,a Francis Carre, Jean Pierre Roque
,
´
´
a
,
*
´
and Andre A. Pavia
a
´ `
Laboratoire de Chimie Organique Physique, Universite de Montpellier II, CC020, Place Eugene Bataillon,
F-34095 Montpellier Cedex 05, France
Laboratoire de Chimie Moleculaire et Organisation du Solide, UMR-CNRS 5073, Universite de Montpellier II, CC07,
`
Place Eugene Bataillon, F-34095 Montpellier Cedex 05, France
b
´
´
Received 4 July 2002; revised 18 September 2002; accepted 3 October 2002
Abstract—Ring opening of oxazolidines derived from tris(hydroxymethyl)aminomethane, L-serine and L-threonine was investigated. It was
shown that n(N)!sp(C–O) electron delocalization (endo-anomeric effect) occurring in the five-membered ring plays a major role in the
cleavage of the intracyclic C–O bond. The present work establishes that when the nitrogen lone pair is conjugated with a carbonyl group
(n(N)!p(CvO) delocalization) as happens in N-acyloxazolidines, both hydrolysis and reductive ring-opening become much more difficult
as a consequence of a concomitant decrease of oxygen basicity and of an increase of the intracyclic C–O bond strength. q 2002 Elsevier
Science Ltd. All rights reserved.
¨
1. Introduction
Bronsted or a Lewis acid. For instance, nucleophilic
Grignard3 and dialkylzinc4 additions are achieved in the
presence of either a large excess of the organometallic
reagent (5 equiv.) or a Lewis acid, e.g. ZnCl2, TiCl4,
BF3–OEt2, MgX2.
Synthetic routes employing oxazolidines as cyclic protec-
tion of a-aminoalcohols have been widely used for the
asymmetric synthesis of chiral amines, aminoalcohols
and/or aminoacids.1 Moreover, such derivatives were
assumed to have potential as prodrug forms,1a,b as they
undergo conversion to the parent compound. Indeed,
oxazolidine derived from ephedrine has been proved to
have sympathomimetic activity in several animal models.2c
In both cases, the success of the approach relies on the
cleavage of the 1,3-N,O ring.
These observations indicate activation of the C–O bond
through coordination of the Lewis acid to the oxygen, a
situation which is commonly postulated for glycosides and
related acetals reactions5 and considered to be one of the
consequence of the anomeric effect.6,7
Anomeric effect is due to a stereoelectronic preference for
an antiperiplanar arrangement of a lone pair of electrons and
an electron-acceptor bond that permits a no bond–double
bond resonance effect.8 Among other consequences, it is
The regioselectivity of oxazolidine ring-opening is long
known to be governed by the higher stability of the acyclic
iminium intermediate resulting from the C–O bond
cleavage (see Scheme 1), compared with that of the
oxocarbenium produced on C–N bond fission. Never-
theless, oxazolidine cleavage requires the presence of a
q Part IV. For parts I–III, see Refs. 9–11.
Keywords: oxazolidines; crystal structure; ring-opening; stereoelectronic
control.
*
Corresponding author. Tel./Fax: þ33-467-143-841;
e-mail: aapavia@univ-montp2.fr
Present address: Laboratoire de Chimie des Substances Naturelles et des
†
´
´
Sciences des Aliments (LCSNSA), Universite de la Reunion, 15 avenue
Rene Cassin, BP 7151, 97715 Saint-Denis Messageries, Cedex 9, France.
´
‡
´
Present address: Laboratoire Organisation Moleculaire: Evolution et
Materiaux Fluores (OMEMF), UMR 5073, Universite Montpellier II,
´
´
´
Place E. Bataillon, 34095 Montpellier, Cedex 5, France.
Scheme 1.
0040–4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(02)01287-5

´
J. Selambarom et al. / Tetrahedron 58 (2002) 9559–9566
9560
Table 1. Procedures used to reduce compound 4
Runs
Reactant
NaBH4
NaBH4
Solvent T (8C)
Product
Yield (%)
a
b
c
d
e
f
g
h
i
THF
Ethanol
THF
THF
THF
THF
THF
20
80
20
20
20
20
20
20
20
20
20
No reaction No reaction
No reaction No reaction
No reaction No reaction
No reaction No reaction
6a
6a
6a
6a
6a
6b
7
NaBH3CN
NaBH4–AcOH
NaBH4–TFA
NaBH3CN–TFA
NaBH4–AlCl3
91
86
89
88
76
80
95
NaBH3CN–AlCl3 THF
BH3
LiAlH4–AlCl3
LiAlH4
THF
THF
THF
j
k
2. Results and discussion
2.1. Oxazolidines and bis(oxazolidines) derived from
tris(hydroxymethyl)aminomethane
Scheme 2.
Compound 1b was first converted into 4 by a Williamson
addition12 according to a solid–liquid phase transfer
catalysis process.13 Compound 4 was then submitted to
various treatments summarized in Scheme 3 and Table 1
whose data call for the followings remarks:
assumed that the heteroatom antiperiplanar to a doubly-
occupied sp3 orbital is made more basic because the
resonance effect increases its electron density and conse-
quently its affinity for a Lewis acid.
1. Open-chain derivative 5 was readily obtained by smooth
acid hydrolysis.
We recently obtained structural evidences for the manifes-
tation of a strong anomeric effect occurring in the N–C–O
fragment of 1,3-oxazolidines derived from tris(hydroxy-
methyl)aminomethane (TRISw),9 L-serine and L-threonine
2. Using routine conditions, NaBH4, NaBH3CN and
LiAlH4 proved unable to achieve ring-opening (runs
a–d, k), even when NaBH4 was previously treated with
acetic acid14,15 (run d). However, as expected, ethyl ester
was easily reduced by LiAlH4 (run k).
1
methyl ester.10,11 Both H and 13C NMR spectroscopy as
well as X-ray diffraction analyses proved unambiguously
the existence of a n(N)!sp(C–O) electron delocalization
characteristic of the anomeric effect. Such phenomenon was
shown to control the conformational properties of the above
derivatives both in the solid state and in solution.10,11
3. In contrast, the ester group was not affected whereas
ring-opening occurred on treatment with an excess of
both NaBH4 and NaBH3CN previously treated with TFA
(runs e, f) or used in the presence of AlCl3 (runs g, h). A
similar result was obtained by reacting 4 with BH3 in
THF (run i).
The work reported herein aims at establishing that the
anomeric effect occurring in the oxazolidine derivatives
reported in Scheme 2 governs also ring-opening.
4. Concomitant ring-opening and ester reduction occurred
when 4 was reacted with LiAlH4 in the presence of AlCl3
(run j).
It is worthwhile to note that reductive ring-opening is
observed as soon as the reducing reagent is used in the
presence of a Lewis acid (AlCl3) or if the reducing reagent
displays Lewis acid properties (BH3). Most probably, the
Lewis acid acts by coordinating the oxygen atom, which in
turn weakens the C–O bond and therefore increases the rate
of reduction. Boron and aluminum are known to have high
affinity for oxygen as evident from the bond strengths in
several molecules (metal-oxygen B–O¼808 kJ/mol and
Al–O¼511 kJ/mol).16 These values should be compared
with those related to nitrogen atom: B–N¼380 kJ/mol and
Al–N¼296 kJ/mol.
As mentioned above, ring-opening occurred also on treat-
ment with NaBH4 or NaBH3CN previously reacted with
TFA. One can postulate that in such conditions, Lewis acid
species are formed in situ according to one of the following
equations. It is known that acyloxyborohydride derivatives
are generated on reaction of metal hydrides with TFA:17
NaBH4 þ CF3COOH ! CF3COONa þ H2 þ BH3
ð1Þ
Scheme 3.

´
J. Selambarom et al. / Tetrahedron 58 (2002) 9559–9566
9561
Scheme 4.
oxazolidine allowed antiperiplanar configuration of the
nitrogen lone pair and of the C-2–O-3 acceptor bond (see
the ORTEP representation reported in Fig. 1(a)); such
arrangement is suitable to the manifestation of an endo-
anomeric effect along the N-1–C-2–O-3 bond sequence.
This hypothesis was unambiguously confirmed by X-ray
diffraction analysis of oxazolidines 8a and 8b.
NaðCF3COOÞxBH42x þ CF3COOH ! CF3COONa
þ H2 þ ðCF3COOÞxBH32x
ð2Þ
The fact that part of the reducing agent is consumed to
generate Lewis acid species explains the necessity to use a
large excess of metal hydride (2.5 molar equiv.) whereas
reduction is completed with nearly stoechiometric amount
of BH3 in THF (1.1 molar equiv.).
The three-dimensional structure of compound 8a (Fig. 1(a))
clearly shows the OE envelope conformation adopted by the
oxazolidine ring. Careful examination of crystal data
In contrast to TFA, acetic acid is probably not strong enough
to promote equilibria represented in Eqs. (1) and (2) to
a sufficient extent. As a consequence, NaBH4/CH3COOH
(run d) is unable to achieve ring-opening. The present
results corroborate previous observations3,4 indicating that
effective oxazolidine ring-opening requires the presence of
an oxophilic catalyst (proton or Lewis acid).
In compound 1b, crystal data revealed the manifestation of a
cooperative anomeric effect resulting from delocalization of
electrons over the bond sequence O-3–C-2–N-1–C-8–O-
7.9 The conformation of ring A allowed n(N)!sp(C8–O7)
electron interaction characteristic of a classical endo-
anomeric effect (Scheme 4) whereas that of ring B allowed
antiperiplanar interaction between the pseudo-equatorial sp3
lone-pair on oxygen-3 with the C-2–N-1 antibonding
orbital sp.9 The reverse motion of electron in ring B is the
result of N-1 becoming a much better electron acceptor than
expected due to the endo-anomeric effect taking place in
ring A. Electron delocalization mentioned above should
influence the basicity of the oxygen atoms, as already
observed for acetals,5 thence their relative ease of protona-
tion and/or coordination with a Lewis acid catalyst.
Consequently, the ability of the C-8–O-7 bond to cleave
should be enhanced whereas that of C-2–O-3 should be
reduced.
Hydrolysis of compound 1b was performed using various
reagents, stoichiometries, temperatures and reaction times.
None of these experiments indicated that hydrolysis
proceeded in two steps. No intermediate oxazolidine was
detected in the reaction mixture. Whatever the experimental
conditions, the latter was shown to contain only the final
product 9 or a mixture of compounds 1b and 9. This
observation seemed to indicate that either there was no
difference of reactivity between rings A and B or, because of
nitrogen inversion, the conformation of the intermediate
Figure 1. ORTEP views of compounds 8a (a) and 8b (b).

´
J. Selambarom et al. / Tetrahedron 58 (2002) 9559–9566
9562
Scheme 5.
˚
reveals that O-3 is 0.584 A out of the plane C-2, N-1, C-5,
C-4 (P1) whereas the latter makes a 4088 dihedral angle with
the envelope tip. More importantly, the pseudo-equatorial
lone pair of electron on N-1 is almost antiperiplanar to the
C-2–O-3 bond allowing n(N)!sp(C2–O3) electron delo-
calization. Indeed, the manifestation of the anomeric effect
is corroborated by the contraction of the C2–N1 bond
and acetone), these molecules existed as an unequal mixture
of cis (syn)- and trans (anti)-rotational isomers (see
Scheme 6), as commonly reported in similar situations18–20
(iv) signals coalescence was observed when the temperature
of the deuteriochloroform solution of 12 for instance, was
increased to 428C. The above observations unquestionably
reflect delocalization of the nitrogen lone pair along the
N-1–CO-Ph (CH3) bond sequence.
˚
˚
(1478 A versus 1492 A for the N1–C5 reference bond).
Consequently, once ring A is opened, the basicity of O-3
becomes similar to that of O-8 as does the reactivity of
rings A and B. A similar situation is observed for the
p-methoxyphenyl analog 8b (Fig. 1(b)).
2.2.2. Crystal structure of compound 13, stereoelectronic
effects. Crystal data for compound 13 are presented in
Section 4. X-Ray structure and atoms labeling are shown in
Fig. 2. Estimated standard deviations (e.s.d.s.) for bond
˚
2.2. N-Acyloxazolidines derived from L-threonine
lengths and bond angles are #0.003 A and #0.28,
respectively.
One may expect electron-withdrawing and electron-donat-
ing groups attached to the nitrogen atom to affect the
reactivity of the five-membered ring. In order to assess the
possible influence of electronic and/or steroelectronic
effects on the behavior of oxazolidines, various N-alkyl
and N-acyloxazolidines derived from L-threonine were
prepared from cycloadducts 2a,b (3a,b)18 (see Scheme 5).
The main features of the three-dimensional structure of 13
may be summarized as follows:
1. The five-membered ring displays an OE envelope
˚
conformation in which O-3 is 0.624 A out of P1 plane
defined by atoms C-2, N-1, C-5, C-4 (rms deviation of
˚
fitted atoms, 0.031 A). The dihedral angle between P1
2.2.1. Synthesis. Compounds 11–13 were readily obtained
by treatment of cycloadducts 2b with the appropriate acyl
chlorides or anhydrides derivatives at room temperature in a
1:1 water/acetonitrile mixture.
and the envelope tip (plane C-2, O-3, C-4) is equal to
4387.
2. Atoms C-2, N-1, C-5, C-10 can be considered as coplanar
˚
(average deviation from mean plane: 0.004 A). That
nitrogen atom has gained a marked sp2 character is
corroborated by the sum of bond angles around N-1
The following spectroscopic features are worth to mention:
(i) in the mass spectra, the molecular ion peak is always
accompanied by that of the acylium ion as well as by the
residual fragment resulting from the cleavage of the amide
bond (ii) all proton NMR signals were found significantly
downfield compared with the equivalent signals in N-alkyl
1
derivatives 2a,b (3a,b) (iii) H and 13C NMR signals were
duplicated indicating that, in CDCl3 (or deuterio-benzene
Scheme 6.
Figure 2. ORTEP views of compound 13.

´
J. Selambarom et al. / Tetrahedron 58 (2002) 9559–9566
9563
(35885) as well as by a substantial decrease of the N-1–
˚
4. Experimental
˚
C-10 bond length: 1.348 A compared with 1.469 A for
N-1–C-2.
4.1. General
3. In terms of stereoelectronic effect, these data confirm that
electron delocalization of the lone pair of electron on
nitrogen occurs essentially towards the carbonyl acceptor
bond (n(N)!p(CvO) conjugation) to the detriment of
the endo-anomeric effect (n(N)!sp(C–O)) previously
observed in NH and N-alkyloxazolidines.9–11
The progress of reactions and the homogeneity of
compounds were monitored by thin layer chromatography
(TLC Merck 254). Detection was achieved by exposure to
UV light (254 nm) or by spraying a 1% potassium
permanganate aqueous solution or 5% ninhydrin ethanolic
solution and heating at 1508C. Column chromatography was
performed on silica gel Si 60 (40–63 mm, Merck). Melting
points were measured on a Bu¨chi 530 apparatus and are
4. N-1 becoming more electropositive, electron delocali-
zation of the sp3 orbital on the ring oxygen into the sp
antibonding orbital of the C-2–N-1 acceptor bond takes
place as illustrated in Scheme 6. The anomeric effect
occurs now in the opposite direction as a result of the
polar C–NCOR receiving electrons from the oxygen
atom. Confirmation of such behavior arises from the
1
reported uncorrected. The H and 13C-NMR spectra were
recorded at 200, 250 or 400 MHz on a Bruker apparatus;
atoms numbering is reported in Schemes 2 and 3, and Figs. 1
and 2. Chemicals shifts are given in ppm relative to
tetramethylsilane using the deuterium signal of the solvent
as a heteronuclear reference for 1H and 13C. Coupling
constants are given in Hz. For rotational isomers, only
characteristic peaks are listed. Elemental analysis were
performed by the Service Central de Microanalyse du
CNRS at Vernaison (France).
˚
observation that C-2–O-3 bond (1.412 A) is noticeably
˚
shorter (<0.03 A i.e. 10£e.s.d.) than the O-3–C-4
˚
reference bond (1.441 A). Therefore, O-3–C-2 bond is
strengthened while the basicity of the intracyclic oxygen
is reduced.
5. For all these reasons, we predicted ring opening to be
much more difficult in N-acyloxazolidines than in their
N–H and N-alkyl counterparts.
4.1.1. 5-Hydroxymethyl-1-aza-3,7-dioxabicyclo[3.3.0]-
octane (1a). Tris(hydroxymethyl)aminomethane (TRIS)
(2.00 g, 16.5 mmol) and paraformaldehyde (1.50 g,
50 mmol) were stirred at 658C in anhyd. toluene (50 mL)
until the reaction mixture becomes clear. The solvent was
evaporated to dryness and the crude residue dissolved in
diethylether (50 mL). The etheral solution was washed with
a saturated Na2CO3 solution (10 mL). The aqueous phase
was extracted with diethylether (3£50 mL) and the
combined organic phase was dried over anhyd. Na2SO4
then evaporated to a white powder. The latter was column
chromatographed (ethyl acetate) to afford colorless crystals
Indeed, acid treatment of derivatives 11–13 according to
the experimental conditions used for compounds 2 (3)
proved unable to achieve ring-opening. Even the mixture
TFMSA–10%/TFA reported in the literature21 as being an
efficient mean to hydrolyse oxaproline moieties of pseudo-
peptide sequences was unsuccessful.
When experimental conditions used in run g (Table 1) were
applied to compound 13 during 5 days, no ring-opening
occurred. Treatment of the reaction mixture followed by
column chromatography allowed to recover the unchanged
starting material. Similarly, when 13 was treated according
to the procedure employed in run i (24 h, 308C, 2.4-fold
excess of BH3 in THF), no open-chain derivative was
obtained.
1
(1.10 g, 45%, mp 608C). H NMR (200 MHz, CDCl3) d:
2.35 (1H, t, OH, J¼6.2 Hz); 3.60 (2H, d, 2H-9, J¼6.2 Hz);
3.78 and 3.84 (4H, AB syst., 2H-4, 2H-6, Jgem¼8.8 Hz);
4.46 and 4.54 (4H, AB syst., 2H-2, 2H-8, Jgem¼5.5 Hz); 13
C
NMR (100 MHz, CDCl3) d: 64.44 (C-9); 73.45 (C-4, C-6);
74.10 (C-5); 88.44 (C-2, C-8); FABþ MS (GT) m/z: 146
[MþH]þ; 114 [MþH2CH3OH]þ; Anal. calcd for
C6H11NO3: C, 49.65; H, 7.58; N, 9.65; O, 34.20; Found:
C, 49.08; H, 7.64; N, 9.51; O, 34.29.
All these observations supported the hypothesis that
reactivity of the oxazolidine ring is dramatically reduced
following amidification.
4.2. Hydrolysis—general procedure
A suspension of compound 1 or 4 (1 mmol) in a 0.1N
methanolic solution of hydrochloric acid (20 mL) was
heated for 1 h. After cooling, 20 mL of toluene was added to
the reaction mixture which was then concentrated. The
aqueous phase was extracted with diethylether (3£5 mL)
and subjected to freeze drying.
3. Conclusion
We have unambiguously established the crucial contri-
bution of the anomeric effect to both the structure and
reactivity of 1,3-oxazolidines. Indeed, C–O bond cleavage
is greatly facilitated when n(N)!sp(C–O) electron
delocalization (endo-anomeric effect) is possible, i.e.
when the lone pair of electron on the nitrogen atom is
antiperiplanar to the C–O acceptor bond. The intra-
cyclic oxygen becoming more basic, its interaction
with the Lewis acid is greatly enhanced and ring-opening
made easy. However, whenever the lone pair of electron
of the donor is not available, anomeric effect does not
occur and ring-opening becomes considerably more
difficult.
4.2.1. 2-Amino-2-(40-ethyloxycarbonyl-20-oxabutyl)pro-
pane-1,3-diol hydrochloride (5). The title compound was
1
obtained as an oily material (90%). H NMR (250 MHz,
D2O) d: 1.35 (3H, t, CH3CH2O, J¼7.1 Hz); 2.77 (2H, t,
2H-40, J¼5.9 Hz); 3.73 (2H,0 s, 2H-10); 3.81 (4H, s, 2H-1,
2H-3); 3.89 (2H, t, 2H-3 , J¼5.9 Hz); 4.28 (2H, q,
CH3CH2O, J¼7.1 Hz); NMR 13C (100 MHz, D2O) d:
13.64 (CH3CH2O); 34.94 (C-2); 61.07 (CH3–CH2O);
59.68 and 59.90 (C-1, C-3); 62.27, 67.24 and 68.34 (C-4,

´
J. Selambarom et al. / Tetrahedron 58 (2002) 9559–9566
9564
C-10, C-30); 175.01 (CvO); FABþ MS (NBA) m/z: 222
[MþH]þ; 244 [MþNa]þ.
4.3.2. 5-(50-Hydroxy-20-oxapentyl)-cis-2,8-diphenyl-1-
aza-3,7-dioxabicylo[3.3.0]octane (7). A solution of com-
pound 4 (0.20 g, 0.5 mmol) in anhyd. THF (5 mL) was
added dropwise to a cooled suspension of lithium–
aluminium hydride (0.04 g, 1.0 mmol) in anhyd. THF
(5 mL). After stirring 1 h at 08C, the mixture was cooled
and decomposed cautiously by 10% sodium hydroxide
aqueous solution. The mixture was then diluted with
diethylether (10 mL) and the solution washed with brine
(3£5 mL). The aqueous layer was extracted with diethy-
lether (3£10 mL). The combined organic layer was dried
over anhyd. Na2SO4 and evaporated to dryness to afford a
white solid (0.17 g, 95%, mp 66–688C). 1H NMR
(250 MHz, CDCl3) d: 1.74 (2H, m, 2H-40); 1.86 (1H, s,
OH); 3.40 (2H, s, 2H-10); 3.52 (2H, t, 2H-30, J¼5.8 Hz);
3.67 (2H, t, 2H-50, J¼5.8 Hz); 3.86 and 4.03 (4H, AB, 2H-4,
2H-6, Jgem¼8.9 Hz); 5.51 (2H, s, 1H-2, 1H-8); 7.30–7.60
(10H, m, H-ar); 13C NMR (100 MHz, CDCl3) d: 32.51 (C-
40); 62.12 and 71.40 (C-10, C-30); 73.88 (C-4, C-6); 76.12 (C-
5); 97.50 (C-2, C-8); 127.56, 128.60, 128.88, 139.67 (C-ar);
FABþ MS (NBA) m/z: 356 [MþH]þ; 378 [MþNa]þ; 266
4.2.2. Tris(hydroxymethyl)aminomethane hydrochlo-
ride (9). The title compound was obtained as an oily
material (80%). 1H NMR (250 MHz, CDCl3) d: 3.81 (6H, s,
2H-2, 2H-3, 2H-4); 4.92 (1H, s, OH); FABþ MS (GT) m/z:
122 [MþH]þ; 144 [MþNa]þ; 105 [MþH2(H2O)]þ.
4.3. Reductive opening—2-(N,N0-dibenzyl)amino-2-(40-
ethyloxycarbonyl-20-oxabutyl)propane-1,3-diol (6a)
Runs e–h (see Table 1). Trifluoroacetic acid or aluminium
trichloride (5 mmol) was added dropwise or portionwise
respectively to a suspension of the metal hydride (5 mmol)
in anhyd. THF (5 mL) at 08C. A solution of compound 4
(1.0 mmol) in anhyd. THF (5 mL) was added dropwise to
the cooled mixture. After stirring 1 h at room temperature,
the mixture was cooled and decomposed cautiously by 10%
sodium hydroxide aqueous solution. The mixture was then
concentrated and extracted with ethyl acetate (3£25 mL).
The combined organic layer was dried over anhyd. Na2SO4
and evaporated to dryness. The oily residue was column
chromatographed (petroleum ether/ethyl acetate 30:70, v/v)
to afford a white solid (82–91%, mp 498C). 1H NMR
(200 MHz, CDCl3) d: 1.28 (3H, t, CH3CH2O, J¼7.1 Hz);
2.60 (2H, t, 2H-40, J¼5.8 Hz); 2.84 (2H, s, 2OH); 3.65 (2H,
s, 2H-10); 3.66 (2H, t, 2H-30, J¼5.8 Hz); 3.64 and 3.79 (4H,
AB syst., 2H-1, 2H-3, Jgem¼11.5 Hz); 3.96 (4H, s,
2£CH2Ph); 4.20 (2H, q, 2H-CH3CH2O, J¼7.1 Hz); 7.10–
7.30 (10H, m, H-ar); FABþ MS (NBA) m/z: 402 [MþH]þ;
[MþH2CH3OCH2CH2CH2OH]þ;
Anal.
calcd
for
C21H25NO4: C, 70.98, H, 7.04, N, 3.94; Found: C, 69.27;
H, 7.04; N, 3.68.
4.4. Synthesis of compounds 8a,b—general procedure
TRIS (82.5 mmol) and benzaldehyde (or p-anisaldehyde)
(82.5 mmol) were poured into a 250 mL round flask
containing 150 mL of hot (80–1008C) toluene, fitted with
a Dean and Stark apparatus. The mixture was refluxed until
the expected amount of water was recovered in the
graduated tube (4–5 h). The solvent was removed under
reduced pressure, the residue dissolved in ethyl acetate
(200 mL) followed by filtration to remove the unreacted
material (Tris). The filtrate was evaporated to dryness and the
solid residue purified by recrystallization from ethyl acetate.
Pure title compounds were obtained as colorless crystals.
424
[MþNa]þ;
370
[MþH2CH3OH]þ;
270
[MþH2CH3O–CH2–CH2CO2Et]þ; 91 [PhCH2]þ; Anal.
calcd for C23H31NO5: C, 68.82; H, 7.73; N, 3.49; O, 19.95;
Found: C, 68.48; H, 7.64; N, 3.57; O, 20.17.
Run i (BH3–THF). A solution of compound 4 (1 mmol) in
anhyd. THF (5 mL) was added dropwise to a cooled
commercial solution of 1 M BH3 in THF (1.1 molar equiv.).
The reaction mixture was treated as reported above to afford
the title compound (76%, mp 498C).
4.4.1. 5,5-Dihydroxymethyl-2-phenyl-1,3-oxazolidine
1
(8a). 86%, Mp 64–658C). H NMR (400 MHz, CDCl3) d:
3.10 (3H, br., NH, 2£OH); 3.20 and 3.65 (4H, AB syst., 2H-
6, Jgem¼11.2 Hz); 3.58 and 3.74 (4H, AB syst., 2H-60,
Jgem¼10.8 Hz); 3.55 and 3.82 (2H, AB syst., 2H-4,
Jgem¼8.6 Hz); 5.45 (1H, s, H-2); 7.30–7.50 (5H, m, H-
ar);013C NMR (100 MHz, CDCl3) d: 64.70 and 65.23 (C-6,
C-6 ); 67.64 (C-4); 70.77 (C-5); 92.49 (C-2); 126.49,
129.03, 129.44, 139.06 (C-ar); FABþ MS (NBA) m/z: 210
[MþH]þ; 232 [MþNa]þ; 178 [MþH2(CH3OH)]þ.
4.3.1. 2-(N,N0-Dibenzyl)amino-2-(50-hydroxy-20-oxapen-
tyl)propane-1,3-diol (6b). Aluminum trichloride (0.84 g,
6.3 mmol) was added portionwise to a suspension of
lithium–aluminium hydride (0.25 g, 6.3 mmol) in anhyd.
THF (5 mL). A solution of compound 4 (0.50 g, 1.3 mmol)
in anhyd. THF (5 mL) was added dropwise to the cooled
supension. After stirring 1 h at room temperature, the
mixture was cooled and decomposed cautiously by 10%
sodium hydroxide aqueous solution. The mixture was
concentrated and extracted with ethyl acetate (3£10 mL).
The combined organic layer was dried over anhyd. Na2SO4
and evaporated to dryness to give a colorless oil (0.36 0g,
80%). 1H NMR (200 MHz, CDCl3) d: 1.90 (2H, m, 2H-4 );
2.51 (1H, br., C-50 –OH); 3.10 (2H, br., C-1–OH, C-3–
OH); 3.56 (2H, t, 2H-50, J¼5.8 Hz); 3.62 (2H, s, 2H-10);
3.80 (6H, m, 2H-1, 2H-3, 2H-30); 3.96 (4H, s, 2£CH2Ph);
7.10–7.20 (10H, m, H-ar); FABþ MS (NBA) m/z: 360
[MþH]þ; 382 [MþNa]þ; 328 [MþH2CH3OH]þ; 91
[PhCH2]þ. Anal. calcd for C21H29NO4: C, 70.19; H, 8.07;
N, 3.90; O, 17.82; Found: C, 70.48; H, 8.37; N, 3.77; O,
18.02.
4.4.2. 5,5-Dihydroxymethyl-2-(4-methoxy)phenyl-1,3-
oxazolidine (8b). 60%, Mp 114–1158C. 1H NMR
(200 MHz, CDCl3) d: 2.10 (3H, br., NH, 2£OH); 3.60 and
3.75 (4H, AB syst., 2H-6, Jgem¼1.2 Hz); 3.69 and 3.85 (4H,
AB syst., 2H-60, Jgem¼10.7 Hz); 3.62 and 3.90 (2H, AB
syst., 2H-4, Jgem¼8.6 Hz); 3.84 (3H, s, OCH3); 5.45 (1H, s,
H-2); 7.10–7.40 (5H, m, H-ar); 13C NMR (100 MHz,
CDCl3) d: 55.75 (OCH3); 64.98 and 65.69 (C-6, C-60); 67.56
(C-4); 70.72 (C-5); 92.30 (C-2); 114.37, 127.81, 131.28 (C-
ar–H); 160.04 (C-ar–O); FABþ MS (NBA) m/z: 240
[MþH]þ; 208 [MþH2CH3OH]þ; Anal. calcd for
C12H17NO4: C, 60.25; H, 7.11; N, 5.85; O, 26.72; Found:
C, 60.48; H, 7.27; N, 5.77; O,26.98.

´
J. Selambarom et al. / Tetrahedron 58 (2002) 9559–9566
9565
4.5. Reduction of compounds 2a,b—general procedure
OCH3); 4.10–4.40 (2H, m, H-4, H-5); 4.85 and 5.45, 5.09
and 5.27 (2H, 2AB syst., 2H, 2H-2, Jgem¼5.0 Hz and
Jgem¼3.7 Hz); 5.79 (1H, dd, CH2v, Jgem¼1.8 Hz, Jcis¼10.2
Hz); 6.22 (1H, dd, CH2v, Jgem¼1.8 Hz, Jtr¼16.8 Hz); 6.43
(1H, dd, CHv, Jcis¼10.2 Hz, Jtr¼16.8 Hz); 13C NMR
(100 MHz, CDCl3) d: (syn and anti rotational isomers)
18.84, 19.10 (CH3); 53.08, 53.40 (CH3O); 63.60, 64,.0
(C-4); 78.70, 79.12 (C-5); 79.60, 80,15 (C-2); 127.85,
128.01 (CH2v); 129.90 (CHv); 163.10 (NCvO); 170.29,
170.85 (OCvO); FABþ MS (NBA) m/z: 200 [MþH]þ; 198
[M2H]þ; 222 [MþNa]þ; 55 [CH2vCHCO]þ; 399
[2MþH]þ; Anal. calcd for C9H13NO4: C, 54.27; H, 6.53;
N, 7.03; O,32.16; Found: C, 54.68; H, 6.39; N, 7.17; O,
32.07.
A solution of compound 2a or 2b (1.0 mmol) in anhyd. THF
(5 mL) was added dropwise to a cooled suspension of
sodium borohydride (0.38 g, 10.0 mmol) in anhyd. THF
(5 mL). After stirring 5 h at room temperature, the mixture
was cooled and cautiously decomposed with 10% NaOH
aqueous solution. The mixture was concentrated and
extracted with dichloromethane (3£20 mL). The combined
organic layer was dried over anhyd. Na2SO4 and evaporated
to dryness. The crude product was column chromatographed
(diethylether/acetonirile/ ammonia 25%, 10:1:0.1, v/v) to
afford colorless oils.
4.5.1. N-methyl-L-serine methyl ester (10a). 30%; 1H
NMR (200 MHz, CDCl3) d: 2.43 (2H, br., NH, OH); 2.45
(3H, s, NCH3); 3.40 (1H, m, H-2); 3.72 (2H, m, 2H-3); 3.75
(3H, s, OCH3); FABþ MS (NBA) m/z: 134 [MþH]þ; 156
[MþNa]þ.
4.5.2. N-Methyl-L-threonine methyl ester (10b). 31%; 1H
NMR (200 MHz, CDCl3) d: 1.22 (3H, d, NCH3, J¼6.4 Hz);
2.40 (3H, s, NCH3); 2.60 (2H, br., NH, OH); 2.85 (1H, m,
H-2); 3.70 (2H, m, 2H-3); 3.75 (3H, s, OCH3); FABþ MS
(NBA) m/z: 148 [MþH]þ; 116 [MþH2CH3OH]þ.
4.6.3. N-Benzoyl-2(S)-ethyloxycarbonyl-3(R)-methyl-
oxazolidine (13). Column chromatography (petroleum
ether/diethylether 50:50, v/v) gave a solid material (60%)
which recrystallized from diethylether to afford colorless
crystals (mp 97–988C). 1H NMR (200 MHz, CDCl3) d: (syn
and anti rotational isomers) 1.54 (3H, d, CH3); 3.83 (3H, s,
CH3O); 4.25 (1H, m, H-4); 4.42 and 5.03 (2H, br., 2H-2);
5.11 (1H, d, H-5, J¼4.4 Hz); 7.40–7.60 (5H, m, Har); 13C
NMR (100 MHz, CDCl3) d: (syn and anti rotational
isomers) 15.65 and 18.89 (CH3); 53.03 (CH3O); 63.74 and
66.23 (C-3); 78.57 (C-2); 81.12 (C-5); 127.87, 128.94,
131.73, 135.21 (C-ar); 169.02 (NCvO); 170.42 (OCvO);
FABþ MS (NBA) m/z: 250 [MþH]þ; 105 [Ph–CvO]þ;
272 [MþNa]þ; Anal. calcd for C9H15NO4: C, 62.65; H,
6.02; N, 5.62; O,25.70; Found: C, 62.98; H, 6.09; N, 5.47;
O, 25.97.
4.6. General procedure for the N-acylation of
compounds 2a,b
A solution of acyl chloride or acid anhydride (3 mmol) was
added dropwise to a solution of compound 2a or 2b
(1 mmol) in a 1:1 mixture of acetonitrile/saturated aqueous
solution of Na2CO3 (10 mL). The pH was maintained above
7 by addition of solid Na2CO3. After stirring 16 h at room
temperature, the mixture was concentrated, washed with a
saturated NaCl aqueous solution and extracted with
diethylether (3£30 mL). The combined organic layer was
dried over anhyd. Na2SO4 and evaporated to dryness. The
residue was column-chromatographed to afford N-acyl
derivatives (11–13).
4.7. Supplementary material
Full crystallographic data (tables of crystallographic details,
non-hydrogen coordinates, bond distances and bond angles,
anisotropic thermal parameters, hydrogen coordinates and
isotropic thermal parameters) have been deposited with
the Cambridge Crystallographic Data Centre (CCDC
no.188949, 188950 and 188951 for compounds 8a, 8b and
13, respectively). This material is available on request from
The Director, CCDC, 12 Union Road, Cambridge CB21EZ,
UK (tel.: þ44-1223-336408; fax: þ44-1223-336033; e-mail
deposit@ccdc.cam.ac.uk or http://www/ccdc.cam.ac.uk).
4.6.1. N-Acetyl-2(S)-ethyloxycarbonyl-3(R)-methyl-
oxazolidine (11). Column chromatography (ethyl acetate)
1
gave a colorless oil (70%). H NMR (200 MHz, CDCl3) d
(syn and anti rotational isomers): 1.46 and 1.48 (3H, 2d,
CH3C, J¼6.1 Hz); 3.77 and 3.82 (3H, s, CH3O); 4.00 and
4.11 (2H, 2d, H-5, J¼6.5 Hz); 4.20 and 4.35 (1H, 2m, H-4);
4.87 and 5.38, 5.01 and 5.13 (2H, 2AB syst., 2H-2, Jgem¼5.0
and 3.5 Hz); 13C NMR (100 MHz, CDCl3) d: (syn and anti
rotational isomers) 18.93 and 19.13 (CH3); 22.55 and 25.59
(CH3CvO); 53.02 and 53.37 (CH3O); 63.38 and 64.74
(C-4); 79.01 and 79.45 (C-2); 79.50 and 80.11 (C-2); 167.34
and 168.85 (NCvO); 170.45 and 170.58 (OCvO); FABþ
MS (NBA) m/z: 188 [MþH]þ; 210 [MþNa]þ; 146
[MþH2CH3CO]þ; 43 [CH3CO]þ; 375 [2MþH]þ; Anal.
calcd for C8H13NO4: C, 51.33; H, 6.95; N, 7.48; O,34.22;
Found: C, 50.68; H, 7.27; N, 7.77; O, 34.07.
Acknowledgements
Authors are indebted to R. Astier (Laboratoire Commun de
´
Mesures Physiques, Universite de Montpellier II) for X-ray
data collection.
References
1. (a) Wagner, B.; Gonzalez, G. I.; Tran Hun Dau, M. E.; Zhu, J.
Bioorg. Med. Chem. 1999, 7, 737–747. (b) Guerrier, L.;
Royer, J.; Grierson, D. S.; Husson, H. P. J. Am. Chem. Soc.
1983, 105, 7754–7755. (c) Berrien, J. F.; Royer, J.; Husson,
H. P. J. Org. Chem. 1994, 59, 3769–3774. (d) Herdeis, C.;
4.6.2. N-Acryloyl-2(S)-ethyloxycarbonyl-3(R)-methyl-
oxazolidine (12). Column chromatography (petroleum
ether/diethylether 80:20, v/v) gave a colorless oil (63%).
1H NMR (250 MHz, CDCl3) d: (syn and anti rotational
isomers) 1.48 (3H, d, CH3, JCH3,H-3¼6.1 Hz); 3.79 (3H, s,
¨
Aschenbrenner, A.; Kirfel, A.; Schwabenlander, F. Tetrahe-
dron: Asymmetry 1997, 8(14), 2421–2432.

´
J. Selambarom et al. / Tetrahedron 58 (2002) 9559–9566
9566
¨
2. (a) Lazar, L.; Fu¨lop, L. Acta Pharm. Hung. 1999, 69(4),
202–209. (b) Johansen, M.; Bundgaard, H. J. Pharm. Sci.
1983, 72(11), 1294–1297. (c) Walker, R. B.; Huang, M.-J.;
Leszczynski, J. J. Mol. Struct. 2001, 549, 137–146, and
references cited.
Chemical Society: Washington DC, 1993. (d) Juaristi, E.;
Cuevas, G. The Anomeric Effect. CRC: Boca Raton, FL, 1995.
8. Romers, C.; Altona, C.; Buys, H. R.; Havinga, E. Top.
Stereochem. 1969, 4, 39–97.
´ ´
9. Monge, S.; Selambarom, J.; Carre, F.; Verducci, J.; Roque,
J. P.; Pavia, A. A. Carbohydr. Res. 2000, 328, 123–127.
´ ´
10. Selambarom, J.; Monge, S.; Carre, F.; Fruchier, A.; Roque,
J. P.; Pavia, A. A. Carbohydr. Res. 2001, 330, 43–51.
´ ´
11. Selambarom, J.; Fruchier, A.; Carre, F.; Roque, J. P.; Pavia,
A. A. Tetrahedron 2002, 58, 4439–4444.
3. (a) Wu, M. J.; Pridgen, L. N. J. Org. Chem. 1991, 56,
1340–1344. (b) Pridgen, L. N.; Mokhallalati, M. K.; Wu, M. J.
J. Org. Chem. 1992, 57, 1237–1241. (c) Mehmandoust, M.;
Marazano, C.; Das, B. C. J. Chem. Soc., Chem. Commun.
1989, 1185–1187.
´
12. Monge, S.; Selambarom, J.; Roque, J. P.; Pavia, A. A.
Tetrahedron 2001, 57, 9979–9987.
4. (a) Andres, C.; Gonzalez, A.; Pedrosa, R.; Perez-Encabo, A.;
Garcia-Granda, S.; Salvado, M. A.; Gomez-Beltran, F.
Tetrahedron Lett. 1992, 33(33), 4743–4746. (b) Nishiyama,
T.; Kishi, H.; Kitano, K.; Yamada, F. Bull. Chem. Soc. Jpn
1994, 67, 1765–1768.
13. Starks, C. M.; Liotta, C. Phase Transfer Catalysis, Principles
and Techniques. Academic: New York, 1978; pp 126–140.
14. Umino, U.; Iwakuma, T.; Itoh, N. Tetrahedron Lett. 1976, 10,
763–766.
5. Deslongchamps, P. Stereoelectronic Effetcs in Organic
Chemistry. Pergamon: New York, 1983.
15. Gribble, G. W.; Lord, P. D.; Skotnicki, J.; Dietz, S. E.; Eaton,
J. T.; Johnson, J. L. J. Am. Chem. Soc. 1988, 53, 223–224.
16. Lide, D. R. Handbook of Chemistry and Physics. 78th ed.
CRC: New York, 1997–1998.
6. (a) Lemieux, R. U.; Chu¨, N. J. Abstract of Paper, 133rd
Meeting, American Chemical Society, 1958; p. 31N. See
(b) Lemieux, R. U. Rearrangements and Isomerisations in
Carbohydrate Chemistry. In Molecular Rearrangements. de
Mayo, P., Ed.; Intersciences: New York, 1964; p 709.
(b) Lemieux, R. U. Explorations with Sugars. How Sweet It
Was. In Profiles, Pathways and Dreams. Autobiographies of
Eminent Chemists. Seeman, J. I., Ed.; American Chemical
Society: Washington, DC, 1990; pp 75–102 185 pp. (c)
Lemieux, R. U.; Koto, S. Tetrahedron 1974, 30, 1933–1944.
7. (a) Jeffrey, G. A. In Anomeric Effect Origin and Consequences
Szareck. Szareck, W. A., Horton, D., Eds.; American
Chemical Society: Washington DC, 1979; pp 54–59.
(b) Kirby, A. J. The Anomeric Effect and Related Stereo-
electronic Effects at Oxygen. Springer: New York, 1983.
(c) Thatcher, G. R. J. The Anomeric Effect and Associated
Stereoelectronic Effect. ACS Symposium Series 593, American
17. (a) Brown, H. C.; Subba Rao, B. C. J. Am. Chem. Soc. 1960,
681–686. (b) Gribble, G. W.; Ferguson, D. C. J. Chem. Soc.,
Chem. Commun. 1975, 535–536.
18. In fact, the major compound obtained on condensation of
L-serine methyl ester with formaldehyde was assigned the
bicyclo[4.4.1]undecane structure (3a). The latter reacts
similarly to its isomeric N,N0-methylenebis(oxazolidine)
counter-part (see Ref. 10).
19. Taylor, W. G.; Hall, T. S.; Schreck, C. E. Can. J. Chem. 1992,
70, 165–172.
20. Taylor, W. G.; Schreck, C. E. J. Pharm. Sci. 1989, 78(2),
109–113.
¨
21. Wohr, T.; Wahl, F.; Nefzi, A.; Rohweder, B.; Sato, T.; Sun, X.;
Mutter, M. J. Am. Chem. Soc. 1996, 118, 8227–9218.