Synthesis and stereochemistry of some 1,3-oxazolidine systems based on TRIS (α,α,α-trimethylolaminomethane) and related aminopolyols skeleton. Part 1: (Di)spiro-1,3-oxazolidines

Article abstract of DOI:10.1016/S0040-4020(00)00268-4

Synthesis of the title compounds is examined from the diastereoselectivity point of view; their stereochemical assignments are supported by high resolution NMR data, which are used to define their chirality and isomerisation aspects. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4020(00)00268-4

TETRAHEDRON
Pergamon
Tetrahedron 56 (2000) 3785±3798
Synthesis and Stereochemistry of Some 1,3-Oxazolidine Systems
Based on TRIS (a,a,a-Trimethylolaminomethane) and Related
Aminopolyols Skeleton. Part 1: (Di)spiro-1,3-oxazolidines
a,*
b
a
a
Mircea Darabantu, G e rard Pl e , Ion Silaghi-Dumitrescu, Carmen Maiereanu, Istvan Turos,
a
a
Ioan A. Silberg and Sorin Mager
a
a
Department of Organic Chemistry, Babes-Bolyai University, 11 Aranyi Janos Street, 3400 Cluj-Napoca, Romania
Facult e des Sciences, Universit e de Rouen, UPRES A 6014 CNRS, IRCOF, 76821 Mont Saint-Aignan Cedex, France
b
Received 7 October 1999; revised 6 March 2000; accepted 23 March 2000
AbstractÐSynthesis of the title compounds is examined from the diastereoselectivity point of view; their stereochemical assignments are
supported by high resolution NMR data, which are used to de®ne their chirality and isomerisation aspects. q 2000 Elsevier Science Ltd. All
rights reserved.
9
alcohol. A preparative study and patents called attention
8
Introduction
2
1
to the use of these compounds as pesticides. It is empha-
sised that no structural assignment accompanied these
papers (except for mass-spectra); moreover, even the
w
TRIS
(a,a,a-trimethylolaminomethane or 1,1,1-tris-
hydroxymethylaminomethane) is the trivial name of a
8
1
,2
8
very well known aminopolyol belonging to the large
class of b-aminoalcohols. A great number of heterocyclic
saturated compounds based on its skeleton, mainly as ®ve-
syntheses were the subject of some controversy. Therefore,
in this paper, our approach is primarily focused on prepara-
tive aspects.
3
±10
membered (substituted-1,3-oxazolidines)
membered structures(substituted-1-aza-3,7-dioxabicyclo[3.3.0]-
are mentioned in the literature. However, only
minor attention has been paid to spiro-1,3-oxazolidines
derived from TRIS (and of its analogous)
Scheme 1).
and fused ®ve-
On the other hand, the stereochemistry of the title structures
has been completely neglected so far. This includes the
fundamental aspects of conformational analysis, diastereo-
isomerism, central and conformational chirality.
3
±25
octanes),
4
,5,8±10,24
(
Thus, we also considered it of interest to enlarge the scope
of the present study, in connection with related and recent
work in the area.
3
Starting from the pioneering works of Bergmann, Pierce
4,5
1
0
26±32
and Senkus, the reaction of some simple cyclic ketones
with b-aminoalcohols was explored either aiming at investi-
gating the general reactivity of carbonyl compounds with
amine nucleophiles or for applicative purposes only.
Recently, azaoxaspiranes having an oxazolidine ring
(Scheme 1) have been considered prodrugs acting as
delivery systems because of their increased lipophilicity
compared with that of the corresponding free b-amino-
Results and Discussion
Synthesis
As depicted in Scheme 2, ten (di)spiro-1,3-oxazolidines
Scheme 1.
Keywords: diastereoselection; oxazolidines; spirocompounds; tautomerism.
*
Corresponding author. Fax: 140-64-190-819; e-mail: mircea.darabantu@ismra.fr
0040±4020/00/$ - see front matter q 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(00)00268-4

3
786
M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
Scheme 2.
have been prepared starting from TRIS 1 and its alkyl
analogues 2 and 3 following the classical method of
diastereoselectivity as compared to 4c. In the case of
compound 4e, the DG (ˆ26.53 kJ/mol) value for the C-2
equatorial methyl group in the starting 2-methylcyclo-
hexanone (over 90% population of the 2-Me-equatorial
conformer) would appear to be responsible for the complete
diastereoselectivity found (Table 1). Anyhow, the explana-
tion is based on also considering the closer position of the
methyl group to the CvO double bond.3
4
,5
10
Pierce and Senkus. It is of note that our choice is,
from a stereochemical point of view, supported by the
different environment of the hydroxymethyl groups in the
starting materials: homotopic (1) and enantiotopic (2,3).
Monospiranes 4d±g and all dispiranes 5 are new
compounds. The representative results are depicted in
Table 1.
3±35
p
The RHF/3-21G molecular orbital calculations (with full
Monospiranes. Variation of diastereoselectivity, shown by
the synthesis of C-6-8 methylated monospiranes 4c±e
geometry optimisation) support the above observations;
Table 2 lists the relative differences between total energies
of the possible conformers and diastereomers in the series
4c±e.
(
Table 1, Scheme 3) could be explained by the preference
of the bulky nucleophile (±OH group) for one of the
diastereotopic faces of the imine double bond taking into
account the stereocontrol promoted by the methyl group and
Burgi±Dunitz trajectory requirements. All syntheses have
been performed under thermodynamic control (e.g. up to
Thus, the calculations reveal that from the four (E-1±4)
conformational or con®gurational equilibria only E-1 and
E-2 (Scheme 3) are considerable and shifted to NH-eq.±
Me-eq. diastereomers. The positive value DEˆ11.84 kJ/
mol for the compound 4e (E-2, Scheme 3) suggests an
axial position of the C-6 methyl group, in order to avoid
1
8 h in re¯uxing toluene). One can see the general prefer-
ence for the second axial nucleophilic attack and the good
3-alkylketone effect' (compound 4d) that gives a better
`
Table 1. Quantitative and qualitative results of the synthesis of (di)spiro-1,3-oxazolidines 4,5
1
R
2
R
Compound
n
Yield (%) /
mp (8C)
Diastereoselectivity (%) according to
NMR spectra performed
DMSO-d
6
CDCl
3
4
4
4
4
4
4
4
5
5
5
a
±CH
±CH
±CH
±CH
±CH
±CH
±CH
±CH
Me
2
OH
H
H
8-Me
7-Me
6-Me
H
0
1
1
1
1
2
1
±
±
±
75 / 118±120
86 / 118±119
80 / 95±99
50 / 83±86
20 / 93±96
90 / 91±92
80 / 178±180
81 / 216±217
75 / 144±145
75 / 134±135
±
±
±
±
b
2
2
2
2
2
2
2
OH
OH
OH
OH
OH
OH
OH
a
c
60c140t
85t115c
100c
±
100c
100t
100u-t
100u-t
61c139t
82t118c
90c110t
±
75c125t
±
a
d
a
e
f
a
b
b
t
g
a
8- Bu
±
±
±
b
65u-t135l-t
65u-t135l-t
b
c
Et
a
Isolated as mixtures (and racemic, if chiral); the dominant diastereomer will be labelled as follows: 4c(8-c), 4d(7-t), 4f(6-c) and 4g(8-c); descriptors: c (cis)
and t (trans) discriminate the diastereomer (O-1 and R group as references); e.g. in 4c(8-c) the O-1 and Me group (linked to C-8) are cis with respect to
2
3
8,39
cyclohexane ring
Isolated as racemic mixtures (if chiral); the diastereomers are labelled as u (unlike, meso form) or l (like, racemic) with respect to the con®gurations at C-3
(see Scheme 3).
b
3
8,39
and C-11 (see further discussion and Schemes 2 and 4);
cyclohexane ring.
descriptors t (trans) or c (cis) refer to the disposal of O-1 and O-9 ligands with respect to

M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
3787
Scheme 3.
two gauche interactions with the oxazolidine moiety, but
combined COSY and NOE-difference experiments have
proved its equatorial position (see Experimental and Stereo-
chemistry sections).
equatorial position). Thus, the result of the counterpoise
method demonstrates that the unsubstituted monospirane
3
3
4b is also a rigid structure in the cyclohexane part, in
DMSO-d , in agreement with RHF/3-21G molecular
p
6
orbital calculations (DEˆ23.64 kJ/mol, ±NH± group in
To evaluate more properly the steric preference of the
nucleophilic attack on
equatorial position). In turn, in CDCl , compound 4b should
3
a
(substituted)cyclohexanone
be a ¯ipping spirane and NMR spectra are consistent with a
single averaged structure. More evidence regarding this
followed by a ®ve-membered ring closure, compound 4g
was prepared. During the synthesis, ring inversion in the
cyclohexane part is now ruled out (Scheme 3) and the dia-
stereomeric ratio 4g(8-c) vs 4g(8-t) (the latter if formed)
also indicates the conformational preference of an alkyl-
amino group vs an alkoxy one when tied together in
3
6,37
solvent dependent behaviour
will be discussed later.
One can ®nally conclude that the synthesis of compounds
4 (Table 1 and 2) and their stereochemistry in the alicycle
part are closely related.
1,3-oxazolidine ring. The synthesis was completely dia-
stereoselective, as revealed by NMR spectra performed in
Dispiranes. The synthesis of the dispiranes 5 afforded
diastereomerically pure compounds only, as shown by the
DMSO-d , but in CDCl we found a mixture of the two
6
NMR spectra performed in DMSO-d with increased sensi-
6
3
possible diastereomers (Table 1, compound 4g); the corre-
sponding A-value is about ±2.72 kJ/mol (±NH± group in
tivity (300, 400 and 600 MHz) (Table 1). Therefore, the
present discussion is, for the moment, restricted to these
preliminary results. Thus, the parent compound 5a exhibited
a trans-(t) diequatorial con®guration of the 1,4-cyclohexane
unit. Then, unexpected complete diastereoselectivity
occurred in the synthesis of 5b,c isolated as unlike
p
Table 2. Results of the RHF/3-21G molecular orbital calculations for the
cis±trans diastereomers 4c±e (major diastereomer in bold) (see Scheme 3)
a
Equilibrium
DE (kJ/mol) (position of methyl group)
3
8,39
(
u)-trans (t) diastereomers
(meso forms, Scheme 4).
E-1
23.52 (C-8)
Ratio
23.77 (C-7)
Ratio
24.44 (C-6)
Ratio
b
b
b
Since in the starting materials 2 and 3, the two hydroxy-
methyl groups are enantiotopic, they have different pro-
chirality and any structure built from two identical units
(aminodiols 212 or 313) would yield a pair of dia-
stereomers [u (unlike) and l (like)]. To explain these results,
we believe that such long distance stereocontrol is the
consequence of the existence of a discrete intermediate in
which steric requirements should be crucial from both ends
of the molecule. The conformational preference of the
4
c(8-c):4c(8-t)
1.5:1
4d(7-t):4d(7-c)
5.7:1
4.6:1
24.23 (C-7)
27.45 (C-7)
215.44 (C-7)
4e(6-c):4e(6-t)
only 4e(6-c) found
9:1
11.84(C-6)
24.52 (C-6)
27.11 (C-6)
b
DMSO-d
6
CDCl
E-2
3
1.7:1
23.52 (C-8)
23.77 (C-8)
210.80 (C-8)
E-3
E-4
a
Negative values indicate the equilibria shifted as depicted in Scheme 3.
Diastereoselective ratios calculated based on H NMR and (or) QC
NMR spectra (see Table 1).
b
1

3
788
M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
Scheme 4.
1,4-cyclohexandione (as a twist-boat conformer) is, in our
opinion, not relevant in this case, but a double imine inter-
mediate 6 is (Scheme 5, see also Scheme 3).
Stereochemistry
Monospiranes. In this section, attention is focused on
the in¯uence of the two types of chirality encountered
in this series: central and axial. Thus, only the last
one is expected to reveal the conformational behaviour
of the oxazolidine ring; concerning the alicycle ring,
it is obviously rigid (compounds 4c±e, 4g) or
unpredictable, at least from our data (compounds
4a,f). If an axis is the single stereogenic element
(Scheme 6), monospiranes without any chiral centre
should be a pair of conformational enantiomers, aR
and aS.
Its suggested anti con®guration has, as a positive conse-
quence, the same steric compression against methylene
protons linked to C-2, -3, -5, -6 in cyclohexane moiety.
Moreover, 6 is itself a meso form, having two opposite
sense chirality axes (aR, aS). Even if a chair±chair
homomeric inversion of 6 is considered, the u-t-spiro-
diastereomers (meso forms) 5b,c are formed. If the chirality
axes are identical, e.g. 6 (aR,aR) or 6 (aS,aS) (not depicted),
the double imine derivative 6 is an l-diastereomer, yielding
the racemic mixtures, l-t-5b,c. Another stereochemical
requirement (revealed by manipulation of Dreiding models)
is a double axial attack on the .CvN bonds; this again
The envelope±envelope ring inversion of the oxazolidine
fragment, as proposed in Scheme 6 (¯exibility around
C-2), is supported only by careful inspection of Dreiding
models (¯exibility around N-4 can be also considered).
Hereafter, all evidence regarding the stereochemistry of
monospiranes will be discussed based on NMR data
(Tables 3 and 4, Scheme 2 and 6).
involves opposite prochirality of the ±CH OH nucleophiles
2
at both ends of the molecule: (±CH OH) at C-1 and
2
Si
1
(
substituent, is placed near the equatorial cyclohexanyl
±CH OH) at C-4. Thus, the R group, the less bulky
2 Re
protons to afford t-5b,c spirodiastereomer (meso form).
Scheme 5.

M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
3789
Scheme 6.
Monospiranes with no chiral centerÐthe enantiomeric
inversion: As a general remark, the conformational beha-
viour of the oxazolidine ring, depicted in Scheme 6, is
con®rmed (compounds 4a±c,f,g). Thus, the C-2 geminal
two distinct AX ±CH OH systems should be observed and
2
two different ±CH OH signals since the substitution test
2
would yield 4 (2 respectively) diastereomers. If ¯ipping
(aR±aS), protons having different prochirality but belong-
ing to different hydroxymethyl groups become enantio-
topic and a single AX system is obtained (Fig. 1,
compound 4b).
heterocyclic protons (H , ) diastereotopic in a rigid
b1 b2
structure, have been found enantiotopic (isochronous)
because the structure is ¯ipping. The same is valid,
mutatis±mutandis, for the C-3 geminal hydroxymethyl
groups. However, the substitution test reveals, for each
hydroxymethyl group, the same geminal anisochronism
Next, data listed in Table 3 illustrate that the diastereo-
topicity DdˆH 2H maintains the same range (0.06±
a1
a2
(
ment). If the entire molecule is rigid (aR or aS) (Scheme 6),
protons H , are diastereotopic in each aR or aS environ-
0.08 ppm) that depends neither on the shape of the alicycle
fragment, nor on its alkyl substitution at C-8. The
a1 a2
1
Table 3. Relevant chemical shifts d (ppm) and coupling constants J (Hz) in H NMR spectra of the monospiranes 4 (major diastereomer is labelled in bold)
2
R
Compound and solvent
%
n
d Ha1
d Ha2
Dd Ha1, a2
J
gem
H
a1
,
a2
d Hb1
d Hb2
J
gem
H
b1,b2
4
a
H
0
a
DMSO-d
CDCl
6
6
6
±
±
3.34(s)
3.65(d)
3.34(s)
3.59(d)
0.00
0.06
0.00
10.9
3.51(s)
3.62(s)
3.51(s)
3.62(s)
±
±
b
3
4
b
H
1
1
±
±
±
DMSO-d
CDCl
±
±
3.62(d)
3.61(d)
3.58(d)
3.53(d)
0.04
0.08
10.9
10.8
3.85(s)
3.68(s)
3.85(s)
3.68(s)
3
4
c
DMSO-d
8-Me
Unresolved multiplet
in the 3.30±3.28ppm
region
(
(
8-c)
8-t)
60
40
±
±
±
±
3.55(s)
3.57(s)
3.55(s)
3.57(s)
±
±
CDCl
3
(
(
8-c)
8-t)
61
39
3.62(d)
3.59(d)
3.54(d)
3.52(d)
0.08
0.07
11.0
11.2
3.67(s)
3.69(s)
3.67(s)
3.69(s)
±
±
4d
DMSO-d
7-Me
1
6
(
7-t)1(7-c)
CDCl
7-t)
±
3.32(s)
3.60(d)
3.29(d)
0.03
11.1
3.56(d)
3.64(s)
3.68(s)
3.54(d)
3.63(s)
3.68(s)
8.6
±
3
(
82
18
3.52(d)
3.52(d)
3.53(d)
3.53(d)
0.08
0.07
0.07
0.06
10.8
10.8
10.9
10.9
3
.59(d)
3.60(d)
.59(d)
(
7-c)
±
3
4
e
6-Me
1
DMSO-d
6-c)
CDCl
6
(
100
3.42(d)
3.35(d)
0.07
11.7
3.58(d)
3.54(d)
8.4
3
(
(
6-c)
6-t)
90
10
3.58(d)
3.58(d)
3.53(d)
3.51(d)
0.05
0.07
10.8
10.8
3.64(d)
3.64(d)
3.61(d)
3.61(d)
6.9
6.9
4
f
H
2
1
DMSO-d
CDCl
6
3.33(d)
3.59(s)
3.29(d)
3.53(d)
0.04
0.06
10.8
10.8
3.53(s)
3.64(s)
3.53(s)
3.64(s)
±
3
t
8- Bu
4
g
DMSO-d
8-c)
CDCl
6
(
100
3.31(s)
3.31(s)
0.00
0.00
3.35(s)
3.35(s)
±
3
(
(
8-c)
8-t)
75
25
3.68(d)
3.61(d)
3.57(d)
3.53(d)
0.11
0.08
11.3
11.0
3.72(s)
3.70(s)
3.72(s)
3.70(s)
±
±
a
Singlet.
Doublet.
b

3
790
M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
1
Figure 1. H NMR spectrum of monospirane 4b (solvent DMSO-d
1
6
, from down®eld to up®eld): ±OH, Hb1,b2, Ha1,a2, ±NH±, cyclohexyl; detail: H NMR
3
spectrum in CDCl : Hb1,b2, Ha1,a2.
diastereotopicity due only to axial chirality is better
exhibited in CDCl than in DMSO-d .
spiranes 4d,e, the same ring inversion is now diastereomeric
(two stereogenic elements in each cis and trans diastereo-
mer, Scheme 7).
3
6
Monospiranes possessing one chiral center in the alicycle
partÐthe diastereomeric inversion: In the case of mono-
1
Therefore, we would have expected to detect four H AX
1
3
systems (H , ) and four C distinct signals for ±CH OH
a1 a2
2
units; unfortunately, complete discrimination of peaks has
been impossible because of overlapping (Tables 3 and 4).
The same is true for the diastereotopic C-2 methylene.
Finally, we point out two remarks:
1
3
Table 4. Relevant chemical shifts d (ppm) in C NMR spectra of the
monospiranes 4 (major diastereomer labelled in bold)
Compound
±CH OH
2
C-2
C-3
C-5
±CH
3
4
a
²
in the case of compound 4e, the ¯ipping oxazolidine
fragment should be hindered by the steric hindrance
promoted by the C-6 methyl group because NMR calcu-
lations are consistent with its equatorial position (see
Experimental).
DMSO-d
CDCl
6
6
6
63.2
65.5
68.3
69.3
66.4
77.3
104.4
±
±
±
3
4
b
DMSO-d
CDCl
63.5
65.3
67.8
68.9
66.2
66.3
94.8
96.1
±
±
3
²
spectra performed in CDCl allowed a better separation
3
4
DMSO-d
c
but not enough for the COSY experiments. Conse-
quently, only intensity of peaks and magnitude of
J-values (Table 3) support assignments of signals to
each diastereomer.
(
(
8-c)
8-t)
63.5
63.4
67.9
67.7
65.9
66.6
94.2
95.3
21.9
21.6
CDCl
3
(
(
8-c)
8-t)
65.3
64.6
68.8
68.9
65.9
66.7
95.6
96.6
21.9
21.5
To conclude, in the monospirane series 4, data from Tables
3 and 4 reveal about the same diastereotopicity if just one or
both stereogenic elements are present in the molecule
(single stereogenic axis or together with a stereogenic
center).
4
DMSO-d
d
6
(
7-t)
63.5
66.7
68
68
65.8
67.6
94.8
95.7
22.3
22.3
(7-c)
CDCl
3
(
(
7-t)
7-c)
65.3
64.63 and 64.57
69.1
68.7
65.9
66.8
96.2
97
22.4
22.4
Dispiranes. As their synthesis, conformational analysis of
dispiranes is supported by the NMR evidence restricted to
4
e
DMSO-d
6-c)
CDCl
6
DMSO-d used as solvent. Preliminary observation is that
6
(
63.0 and 63.5
68.7
65.6
96.5
14.8
t-5a has no chiral centre; meanwhile t-5b, c possess two
chiral centers, C-3, -11 [t (trans) indicates the trans disposal
of the two ±NH± groups in cyclohexane ring]. The H NMR
3
(
(
6-c)
6-t)
65.1 and 65.3
64.9 and 65.7
69.5
69.5
65.7
65.7
97.9
97.2
14.6
14.9
1
spectrum of dispirane t-5a is quite similar to 4b (Fig. 2,
Scheme 8).
4
f
DMSO-d
CDCl
6
63.2
65
67.5
68.8
66.4
66.3
98.2
100.2
±
±
3
Since the eight cyclohexanyl protons present an overlapped
AX system (hence, a rigid alicyclic conformation and a
trans con®guration regarding the substituents linked to
C-5 and C-8), based also on the stereochemical assignments
4
g
DMSO-d
8-c)
6
(
63.5
68
65.9
94.2
±

M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
3791
Scheme 7.
for monospirane 4b, one has now to consider two chirality
axes and the conformational equilibria depicted in Scheme
exhibiting a single averaged structure, is consistent with
these equilibria.
8. From all four stereoisomers, only three I, II (identical
with IV) and III are distinct structures: conformational
enantiomers I±III and conformational diastereomers I±II
and II±III. The already discussed H NMR spectrum,
Very dif®cult structural problems are raised by the
dispiranes t-5b,c in order to clarify their stereochemistry
(con®guration l or u at C-3 and C-11) according to the
1
1
Figure 2. H NMR spectrum of dispirane 5a (solvent DMSO-d
6
, from down®eld to up®eld): ±OH (4.58 ppm, t, Jˆ5.4 Hz), Hb1,b2 (3.57 ppm, s), Ha1,a2 (3.33 and
3
.29 ppm, dd, Jˆ11.1 Hz), ±NH± (2.41 ppm, br.s), cyclohexyl (1.59 and 1.56 ppm, dd, Jˆ12.4 Hz).
Scheme 8.

3
792
M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
than DMSO-d , the unexpected role of `mobile' protons has
6
been surprisingly put in evidence. In DMSO-d , the
6
hydroxyl protons have been found as clear triplets (in fact
overlapped doublets of doublets) (Figs. 2 and 3), triplets
near coalescence, or broad singlets but located in about
the same region (4.65±4.70 ppm). One can assume their
®xed location in hydroxymethyl group, chelate developed
with the solvent or a rapid change with the solvent.
d-Values of amine protons (broad singlet) ranged about
2
.5 ppm.
In CDCl , the NMR spectra have a completely different
3
appearance. Thus, all `mobile' protons have been found as
a unique broad singlet, shifted up®eld (in the region
2±3 ppm), suggesting an intramolecular rapid change of
their environment. This hypothesis has been also supported
by the very poor solubility of all spiranes in CDCl (5±10
3
times less than in DMSO-d ) and allowed us to consider the
6
effect of intramolecular hydrogen bonds.
1
Figure 3. H NMR spectrum of dispirane 5b (600 MHz, solvent DMSO-d
from down®eld to up®eld): ±OH, Hb1,b2, Ha1,a2, ±NH±, cyclohexyl, ±CH
6
3
.
On the other hand, the results listed in Table 1 indicate
different compositions of diastereomers, depending on the
NMR solvent used. Then, it is pertinent to rule out that these
differences should be the result of a different sensitivity of
detection (CDCl vs DMSO-d ). We believe that, as soon as
reaction model proposed (Scheme 4 and 5). A serious draw-
back of our study has been the fact that they have shown
good solubility only in DMSO-d . Increased sensitivity of
6
1
H NMR detection (300, 400, 600 MHz instruments) did not
3
6
separate the peaks, in the cyclohexane part, crucial for de®-
nite assignments. Thus, just a single type of stereoisomer
has been detected in the spectrum of each compound [e.g.
one singlet for methyl groups in t-5b, one (overlapped)
doublet of doublets for methyl groups in t-5c etc.] (see
Fig. 3, Scheme 9 and further Table 6).
intramolecular hydrogen bonds are effective (CDCl ), an
3
isomerisation occurs and hereafter we will call it ring±
ring tautomerism. It is favoured by the intramolecular
hydrogen bonds that increase ±OH group nucleophilicity
as well as non-ionic transition states (S 2-like mechanism).
N
The equilibria seem to be fast because spectra re-recorded
after several days indicated the same ratio between
`tautomers'.
To discriminate, by means of NMR, the two possible l- or
u-diastereomers (Scheme 9, just a single enantiomer is
depicted) one must observe that in u-t-5b,c diastereomers,
cyclohexanyl protons labelled with the same letter are
enantiotopic whereas, in l-t-5b,c they are homotopic.
Two types of tautomerism should be taken into account.
² Homomeric ring±ring tautomerism (compounds 4a, b, f):
1
0
0
2
Unfortunately, no H AA XX system (4 signals, 32
peaks, u-form) and no (two) distinct H AX systems (4
as illustrated in Scheme 10 (R ˆH), in this type of tauto-
1
merism the two tautomers could not be revealed by NMR
because of the rapid conformational equilibrium between
conformers V±VII. A typical example is monospirane 4b
(Fig. 1). In this trivial case, the two tautomers are, in fact,
conformational diastereomers.
signals, 8 peaks, l-form) were revealed (Fig. 3, compound
5b). Obviously, no useful data have been provided by the
C NMR spectra because in each of the u- or l-environment
1
3
only two types of distinct methylene carbons can be
exhibited (see nuclei labelled as X and Y, Scheme 9 and
Table 6). That is, we have assigned a u-con®guration for
both t-5b,c compounds according to the chemistry depicted
in Scheme 5.
² Diastereomeric ring±ring tautomerism (compounds 4c,
d, e, g): in this case, the two tautomers are distinct as cis,
trans diastereomers (O-1 and R as reference). The most
2
2
t
relevant example is monospirane 4g (R ˆ Bu, Fig. 4).
The ring±ring tautomerism
The ratio cis:trans (about 3:1, in CDCl ) allows us to esti-
3
mate best the incidence of equatorial attack vs. axial attack
on the diastereotopic faces of the hypothetical imine
intermediate VI (Scheme 10), that is now an authentic
ananchomeric structure.
The failure of any attempt to complete structural assign-
ments for dispiranes has required a re-examination of the
NMR investigation. That is, by using NMR solvents other
Scheme 9.

M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
3793
1
the case of compound 5b (R ˆMe), the hyper®ne structure
1
of H NMR spectrum allows signi®cant assignments
(
Fig. 5a±d).
The C-2, -10 methylenes H ,
6,37
(four doublets of
Jgemˆ8.7 Hz) are placed in the
b1 b2
doublets, typical value3
region (A, B) and best separated to allow estimation of the
diastereomeric composition. In the region (C) there are 4
pairs of geminal protons H ,
3
6,37
(typical value
Jgemˆ10.4±10.5 Hz) that is consistent with four environ-
a1 a2
ments (in IX, two distinct AX systems H , are expected
a1 a2
instead of four because of ¯ipping of the molecule as a
homomeric inversion). In the zone (D) one can identify
four types of methyl groups and the most intense peak
belongs to VIII (u-t-5b), then to X (l-t-5b) and the smallest
two ones, but having the same intensity, are reasonably
assigned to diastereomers IX (l- and u-c-5b) since they
are located on different ¯ipping structures (homomeric
inversion).
Scheme 10.
Ring±ring tautomerism in the case of dispiranes.
Obviously, the most complex isomerisation has been exhib-
ited by the dispiranes 5. Unfortunately, the parent
compound t-5a was quite insoluble in both CDCl and
3
C D thus preventing us from recording NMR spectra.
6
Similar results have been derived from the spectra of dis-
pirane 5c (Table 6) but intermediates of type IX have been
detected only in small traces that makes their assignments
hazardous.
6
Consequently, no information was available about the pos-
sible changing of the cis:trans diastereoisomerism in the
alicycle fragment (trans-1,4-di-eq.±NH± against cis-1,4-
ax.±eq.±NH±). In turn, compounds u-t-5b,c provided
evidence for modi®cation of spectra (Table 5 and 6, Scheme
Extension of these phenomena has been observed by
choosing another pair of solvents: D O and C D (Fig. 6,
2 6 6
11).
compound 5b).
Although the complete description of all terms VIII±X
involved in the above equilibria is quite impossible, the
number of typical coupling patterns can be estimated. In
As expected, in D
whereas in C both u-t- and l-t-5b are detected in a
O just one diastereomer is present,
2
D
6
6
1
Figure 4. H NMR spectrum of monospirane 4g (solvent DMSO-d
t
, from down®eld to up®eld): ±OH, Hb1,b2, Ha1,a2, ±NH±, cyclohexyl, Bu; detail: H NMR
1
6
3
spectrum in CDCl : Hb1,b2, Ha1,a2 (coupling patterns as shown).
Table 5. Possible tautomers formed by isomerisation of dispiranes 5b, c (see also Scheme 11)
Starting tautomer (DMSO-d
6
)
Mixture of tautomers observed (CDCl
3
)
Disposal of ±NH±groups
Con®guration at C-3 and C-11
Symbol
Conformation
u-t-5b,c VIII
trans (t)
cis (c)
cis (c)
like (R1R or S1S)
like (R1R or S1S)
unlike (R1S)
l-t-5b,c
l-c-5b,c
u-c-5b,c
Rigid X
Flipping IX
Flipping IX

3
794
M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
d
d

M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
3795
1
Figure 5. H NMR of the dispirane 5b (solvent CDCl
3
, from down®eld to up®eld): (a) Hb1(IX), Hb1(X), Hb1(VIII), Hb2(IX); (b): Hb1(IX), Hb2(X), Hb2(IX),
(X), ±CH (IX), ±CH (VIII).
H
b2(VIII); (c) Ha1,a2(VIII±X); and (d) ±CH (IX), ±CH
3
3
3
3

3
796
M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
Scheme 11.
1
Figure 6. H NMR spectrum of dispirane 5b (solvent D
1
2
O, from down®eld to up®eld): ±OH and ±NH±, Hb1,b2, Ha1,a2, cyclohexyl, ±CH
3
); detail H NMR
spectrum in C
6
D
6
: Hb1(X), Hb1(VIII), Hb2(VIII), Ha1(X), Hb2(X), Ha1(X), Ha2(X), Ha2(VIII).
different ratio than in CDCl . Moreover, the tautomers l-c-
3
Experimental
Melting points (Boetius) are not corrected. All reagents
were purchased from Aldrich . H NMR and C NMR
were performed on VARIAN-Gemini 300 and Brucker
AM 400 spectrometers for all compounds. For dispiranes
and u-c-5b,c are not detectable in C D .
6
6
w
1
13
Conclusion
5
instrument. QC NMR spectra were performed on
VARIAN-Gemini 300 operating at 75 MHz for C, by
using 80±100 mg/sample in standard tubes and a delay
b,c, spectra were also recorded on a Brucker AM 600
Mono- and dispirooxazolidines derived from TRIS or some
of its analogues are available in satisfactory yields and
distereoselectivity; their stereochemical behaviour is
versatile but suitable to NMR monitoring and dependent
on the solvent used. Chirality is better explained by a
stereogenic axis than a stereogenic center. Isomerisations
may occur and they are promoted by the intramolecular
mobility (hydrogen bonds included) of amine and hydroxyl
protons. We have de®ned this phenomenon as ring±ring
tautomerism.
1
3
time D1ˆ10. No SiMe was added; chemical shifts were
4
measured against the solvent peak. Samples were prepared
by using commercially available DMSO-d and measured
after complete dissolution in standard tubes (about 30 mg
6
1
13
for H and 60±70 mg for C NMR spectra). For spectra
recorded in CDCl3 5±10 mg/sample in standard tubes
were used. Hyper®ne structure of H NMR was obtained
1

M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
3797
on diluted samples only by using RESOLV macro on
VARIAN-Gemini 300, after 128 transients (special atten-
tion was paid to neglect negative peaks). All reactions were
monitored by TLC on MERCK silica gel, by using benzene:
trans-4,12-Diaza-3,11-di(bishydroxymethyl)-1,9-dioxa-
dispiro[4.2.4.2]tetradecane, 5a. 81; 216±217 (DMF);
white crystalline powder. Anal. Calcd for: C H N O
1
4
26
2
6;
found: C 53.00%, H 8.15%, N 8.75%; found: C 52.81%,
H 8.23%, N 8.80%. dC (100 MHz; DMSO-d ) 94.2 (2C,
methanol 3:1 v/v as eluent (visualisation in I bath). The
2
6
p
RHF/3-21G molecular orbital calculations were performed
by using Spartan 5.0 package of programs; Spartan version
C-5, -8), 67.9 (2C, C-3, -11), 66.1 (2C, C-2, -10), 34.5
(4C, C-6, -7, -13, -14), 63.4 (4C, 4£-CH OH).
2
5
3
.0, Wavefunction Inc., 18401 Von Karman Avenue, Suite
70, Irvine, CA 92612, USA.
p
(3R ,11S )-trans-4,12-Diaza-3,11-bishydroxymethyl-3,11-
p
p
dimethyl-1,9-dioxadispiro[4.2.4.2]tetradecane, u-t-5b; (3R ,
p
11R )-trans-4,12-diaza-3,11-bishydroxymethyl-3,11-di-
methyl-1,9-dioxadispiro[4.2.4.2]tetradecane, l-t-5b;
Preparation of mono- and dispiranes (typical procedure)
p
p
TRIS as 0.30±0.50 M suspension in toluene (50 mL) [for
aminodiols 2, 3 in benzene (50 mL)] and cyclic ketone
(3R ,11S )-cis-4,12-diaza-3,11-bishydroxymethyl-3,11-
dimethyl-1,9-dioxadispiro[4.2.4.2]tetradecane, u-c-5b;
p
p
(
molar ratio ketone/TRISˆ1.5±3.5: 1; for synthesis of
(3R ,11R )-cis-4,12-diaza-3,11-bishydroxymethyl-3,11-
dimethyl-1,9-dioxadispiro[4.2.4.2]tetradecane, l-c-5b. 75;
144±145 (Acetone); white crystalline powder. Anal. Calcd
dispiranes stoichiometric ratio) were stirred under re¯ux
in a Dean±Stark trap until TLC monitoring showed no
more evolution of the reaction (18±21 h). In the case of
monospiranes 4 the crude reaction mixture was neutralised
with solid Na CO and ®ltered hot. The organic solution was
for C14H
N
2
O
4
: C, 58.50%, H, 9.25%, N, 10.00%; found:
26
C, 58.72%, H, 9.15%, N, 9.78%.
2
3
p
(3R ,11S )-trans-4,12-diaza-3,11-diethyl-3,11-bishydroxy-
p
evaporated in vacuo and the residue was crystallised from
an appropriate solvent. Dispiranes 5 were preliminary
isolated as crude products by ®ltering the reaction mixture
at room temperature then recrystallised from an appropriate
solvent. Monospiranes 4d±g and all dispiranes 5 have not
been previously reported. Compounds 4a±c were prepared
p
methyl-1,9-dioxadispiro[4.2.4.2]tetradecane u-t-5c; (3R ,
p
11R )-trans-4,12-diaza-3,11-diethyl-3,11-bishydroxymethyl-
1,9-dioxadispiro[4.2.4.2]tetradecane, l-t-5c. 75; 134±135
(Acetone); grey crystalline powder. Anal. Calcd for
C
61.12%, H, 9.61%, N, 8.91%.
H
16
N
2
O
4
: C, 61.00%, H, 9.71%, N, 9.10%; found: C,
30
8
according to literature.
Yields (%), mp (8C, solvent), elemental analysis and NMR
data (if not collected in Table 3, 4 and 6) are listed below.
References
4
-Aza-3,3-bishydroxymethyl-t-7-methyl-r-1-oxaspiro[4.5]-
1. Aldrichimica Acta 1983, 16, 35.
2. Eachus, A. C. Chim. Oggi 1994, 12, 24±31.
decane and 4-aza-3,3-bishydroxymethyl-c-7-methyl-r-1-
oxaspiro[4.5]decane, 4d. 50; 83±86 (n-heptane); white
crystalline powder. Anal. Calcd for C H NO : C,
3. Bergmann, D. E. Chem. Rev. 1953, 53, 309±352.
4. Pierce, J. S.; Lunsford, C. D.; Raiford Jr., R. W.; Rush, J. L.;
Riley, D. W. J. Am. Chem Soc. 1951, 73, 2595±2596.
5. Pierce, J. S.; Lunsford, C. D. J. Am. Chem. Soc. 1951, 73, 2596±
1
1
21
3
6
9
1.50%, H, 10.05%, N, 6.60%; found: C, 61.37%, H,
.83%, N, 6.50%.
2
598.
4-Aza-3,3-bishydroxymethyl-c-6-methyl-r-1-oxaspiro[4.5]-
decane and 4-aza-3,3-bishydroxymethyl-t-6-methyl-r-1-
6. Cabb, R. U.S. Patent 3,660,488; Chem. Abstr. 1972, 77,
P61258b.
oxaspiro[4.5]decane, 4e. 20; 93±96 (Et O1n-heptane);
2
7. Susi, P. U.S. Patent 4,125,518; Chem. Abstr. 1979, 90,
P72897d.
white crystalline powder. Anal. Calcd for C H NO : C,
1
1
21
3
6
9
1.15%, H, 9.59%, N, 6.45%; found: C, 61.37%, H,
.83%, N, 6.50%. dH (400 MHz; DMSO-d ). Assignment
8. Sokolov, V. B.; Krykov, L. N.; Koloniets; Sololskii, G. A. Zh.
Org. Khim. 1979, 15, 1777±1779.
6
of equatorial position of C-6 methyl group: NOE on doublet
at 0.85 ppm (Me group, d, Jˆ5.7 Hz) showed the enhance-
ment of multiplet located in 1.56±1.49 ppm region,
assigned to H , H7a,7e; COSY experiment showed that
9. Buur, A.; Bundgaard, H. Arch. Pharm. Chem. Sci. Ed. 1987, 15,
76±86.
10. Senkus, M. J. Am. Chem. Soc. 1945, 67, 1515±1519.
11. American Cyanamid Co. Br. Patent 564,506; Chem. Abstr.
6
a
3
H10e 1.75 ppm (ddd, Jˆ4.5, 4.6 and 13.0 Hz) is coupled
1946, 40, P4084.
12. Hedge, B. E. U.S. Patent 2,433,609; Chem. Abstr. 1948, 42,
P3430h.
with H , H which were located in a different region
1
0a
9a,e
than above, 1.45±1.36 ppm.
13. Cabb, R. U.S. Patent 3,843,726; Chem. Abstr. 1975, 82,
P86193t.
4
4
-Aza-3,3-bishydroxymethyl-1-oxaspiro[4.6]undecane,
f. 90; 91±92 (Et O); white crystalline powder. Anal. Calcd
14. Eastman Kodak. Fr. Patent 1,504,886; Chem. Abstr. 1969, 70,
P57863h.
2
for C H NO : C, 61.50%, H, 9.88%, N, 6.77%; found: C,
21
1
1
3
61.36%, H, 9.83%, N, 6.50%.
15. Schnegelberger, H.; Bellinger, H. German Patent 2,218,348;
Chem. Abstr. 1974, 80, P52285t.
4
-Aza-3,3-bishydroxymethyl-c-8-tertbutyl-r-1-oxaspiro-
4.5]decane and 4-aza-3,3-bishydroxymethyl-t-8-tert-
16. Sidi, H.; Johnson, H. R. Fr. Patent 2,279,828; Chem. Abstr.
1976, 85, P162040v.
[
butyl-r-1-oxaspiro[4.5]decane, 4g. 80; 178±180 (Et O);
white crystalline powder. Anal. Calcd. for C H NO : C,
4
17. Cunningham, A. L., Mathai, J. U.S. Patent 4,101,527; Chem.
Abstr. 1979, 90, P24900e.
2
1
27
3
6
1
5.60%, H, 10.40%, N, 5.50%; found: C, 65.33%, H,
0.57%, N, 5.44%.
18. Tilford, C. H.; Van Campen Jr., M. G.; Shelton, R. S. J. Am.
Chem. Soc. 1947, 69, 2902±2906.

3
798
M. Darabantu et al. / Tetrahedron 56 (2000) 3785±3798
30. Grosu, I.; Mager, S.; Pl e , G. J. Chem. Soc., Chem. Commun.
1
1
2
9. Laurent, A. P.; Riehl, M.; Frazao, S. C. Bull. Soc. Chim. Fr.
967, 10, 3868±3872.
0. Brois, D.; Ryer, J.; Winans, E. U.S. Patent 4,199,463; Chem.
1995, 167±168.
31. Grosu, I.; Mager, S.; Pl e , G. J. Chem. Soc., Perkin Trans. 2
1995, 1351±1356.
Abstr. 1980, 93, P117076r.
1. For example: Barbulescu, N.; Moga, S. Gh.; Sintamarian, A.;
Cuza, O.; Vasilescu, V.; Rom. Patent 83,939; Chem. Abstr. 1985,
2
32. Grosu, I.; Mager, S.; Batiu, I. Rev. Rom. Chim. 1995, 40,
1165±1171.
1
02, P149252r and related patents cited therein.
2. Zayed, S. E. Pak. J. Sci. Ind. Res. 1987, 30, 432±438; Chem.
33. Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic
Compounds; Wiley: New York, 1994; pp. 732; for the counterpoise
2
Abstr. 1988, 108, 94446y.
3. Nouguier, R.; Crozet, M.; Vanelle, P.; Maldonado, J.
Tetrahedron Lett. 1985, 26, 5523±5524.
4. Vanelle, P.; De Meo, M. P.; Maldonado, J.; Nouguier, R.;
Crozet, M. P.; Laget, M.; Dumenil, G.; Eur J. Med. Chem. 1990,
method see: pp. 699, 700, 704±706, op. cit.
2
3
1
3
1
3
4. Brown, H. C.; Borkowsky, M. J. Am. Chem. Soc. 1952, 74,
894±1902.
2
5. Brown, H. C.; Brewster, J. H.; Shechter, H. J. Am. Chem. Soc.
954, 76, 467±474.
6. Darabantu, M.; Pl e , G.; Mager, S.; Cotora, E.; Gaina, L.;
25, 241±250.
25. Mattson, A.; Norin, T. Synth. Commun. 1994, 24, 1489±1491.
26. Gange, D.; Magnus, P.; Bass, L.; Arnold, E.; Cardy, J. J. Am.
Costas, L.; Mates, A. Tetrahedron 1997, 53, 1873±1890.
7. Darabantu, M.; Pl e , G.; Mager, S.; Gaina, L.; Cotora, E.;
Mates, A.; Costas, L. Tetrahedron 1997, 53, 1891±1908.
3
Chem. Soc. 1980, 102, 2134±2136.
2
2
1
2
7. Dodziuk, H. J. Chem. Soc., Perkin Trans. 2 1986, 249±253.
8. Dodziuk, H.; Mirowicz, M. Tetrahedron Asymmetry 1990, 1,
71±172.
38. Pure and Applied Chemistry 1996, 12, 2195±2222.
3
9. Gawley, R. E.; Aub e , J. Principles of Asymmetric Synthesis,
Tetrahedron Organic Chemistry Series, Pergamon: Oxford, 1996;
Vol. 14, pp 15±40.
9. Dodziuk, H.; Sitkovski, J.; Stefanian, I.; Mursakulov, I. G.;
Guseinov, M. M.; Kurbanova, E. A. Struct. Chem. 1992, 3, 269±
71.
2