Synthesis of site-specific deuterium substituted methyl 6-O-[(R)- and (S)-1-carboxyethyl]-α-D-galactopyranoside and conformational analysis thereof based on J couplings

Article abstract of DOI:10.1016/S0008-6215(98)00240-7

Methyl 6-(S)-²H-6-O-[(R)- and (S)-1-carboxyethyl]-α-D-galactopyranoside have been synthesized from 6-(S)-²H-1,6-anhydro-β-D-galactopyranoside. Conformational analysis of the exocyclic dihedral angles has been performed based on ³J(H,H) and ³J(C,H) coupling constants of the title compounds together with their non-deuterated counterparts. The ω dihedral angle (O-5-C-5-C-6-O-6) can be described by a conformational equilibrium where the order of conformers is gauche-trans> trans-gauche> gauche-gauche, i.e., the same as for galactose but with a different population distribution. The C-5-C-6-O-6-C-2' torsion populates only the gauche states, as deduced from J couplings. This is at variance with previous Langevin dynamics simulations in which the proportion of the trans state is significant. For the C-6-O-6-C-2'-H2' torsion an equal population of the three staggered conformers is compatible with experimental data, although previous simulations have indicated gauche conformers to be preferred. Since about half the number of the possible conformational states for the major degrees of freedom are significantly populated, as corroborated by experimental data, the flexibility of the substituting 1-carboxyethyl group is large. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(98)00240-7

Carbohydrate Research 312 (1998) 117±121
Synthesis of site-speci®c deuterium substituted
methyl 6-O-[(R)- and (S)-1-carboxyethyl]-ꢀ-d-
galactopyranoside and conformational
analysis thereof based on J couplings
Roland Stenutz a, Ake Pilotti a, Per-Erik Jansson b, Goran Widmalm a,*
a Department of Organic Chemistry, Arrhenius Laboratory, Stockholm University,
S-106 91 Stockholm, Sweden
b Clinical Research Centre, Analytical unit, Karolinska Institute, Huddinge Hospital, NOVUM,
S-141 86 Huddinge, Sweden
Received 28 May 1998; accepted 27 August 1998
Abstract
Methyl 6-(S)-2H-6-O-[(R)- and (S)-1-carboxyethyl]-ꢀ-d-galactopyranoside have been synthe-
sized from 6-(S)-2H-1,6-anhydro-ꢁ-d-galactopyranoside. Conformational analysis of the exocyclic
3
3
dihedral angles has been performed based on JH,H and JC,H coupling constants of the title
compounds together with their non-deuterated counterparts. The ! dihedral angle (O-5±C-5±C-6±
O-6) can be described by a conformational equilibrium where the order of conformers is gauche-
trans> trans-gauche> gauche-gauche, i.e., the same as for galactose but with a di€erent popula-
tion distribution. The C-5±C-6±O-6±C-20 torsion populates only the gauche states, as deduced
from J couplings. This is at variance with previous Langevin dynamics simulations in which the
proportion of the trans state is signi®cant. For the C-6±O-6±C-20±H20 torsion an equal population
of the three staggered conformers is compatible with experimental data, although previous simu-
lations have indicated gauche conformers to be preferred. Since about half the number of the
possible conformational states for the major degrees of freedom are signi®cantly populated, as
corroborated by experimental data, the ¯exibility of the substituting 1-carboxyethyl group is large.
# 1998 Elsevier Science Ltd. All rights reserved
Keywords: 1-Carboxyethyl; J Coupling; Deuterium; NMR
1. Introduction
and anomeric con®guration as well as by the sites
of substitution. In addition, modi®cation of the
sugars can occur, e.g., by substitution with O-
acetyl, phosphate or sulfate groups, and other
groups such as pyruvate acetals and 1-carboxyethyl
ethers. The latter, mostly found in bacterial
polysaccharides, occur both with the (R)- and
The three-dimensional structure of saccharides is
determined by the constituent sugars, their absolute
* Corresponding author. Tel.: +46-8-163742; fax: +46-
8-154908; e-mail: gw@admin.organ.su.se
0008-6215/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved
PII S0008-6215(98)00240-7

118
R. Stenutz et al./Carbohydrate Research 312 (1998) 117±121
(S)-con®guration. Di€erent substituents will alter
the shape and the electrostatics as well as the
hydrophobicity of the saccharides. These groups
may also change the conformation of the sacchar-
ide and may themselves exhibit internal ¯exibility.
We have previously synthesized and studied 1-
carboxyethyl substituted sugars by NMR and CD
spectroscopy [1]. Two of these, methyl 6-O-[(R)-
and (S)-1-carboxyethyl]-ꢀ-d-galactopyranoside 1R
and 1S, were further studied by molecular mechan-
ics and Langevin dynamics (LD) simulations to
investigate the conformational ¯exibility and the
dynamics of the major degrees of freedom in the
molecules [2]. LD is a technique which di€ers from
molecular dynamics simulations in that the solvent
is modelled by random and frictional forces. This
facilitates longer simulations to be carried out
than with explicit solvent. In addition to the
simulation study, the (R)-isomer could be crystal-
lized and showed an extended conformation [3]. In
the present study we have synthesized methyl 6-
(S)-2H-6-O-[(R)- and (S)-1-carboxyethyl]-ꢀ-d-gal-
actopyranoside 2R and 2S and together with the
protiated compounds 1R, 1S studied them with
NMR spectroscopy. The JH,H and JC,H coupling
constants have been measured and used, via
Karplus relationships, to describe the conforma-
tional equilibria.
3
3
2. Results and discussion
Methyl 6-(S)-2H-6-O-[(R)- and (S)-1-carboxy-
ethyl]-ꢀ-d-galactopyranoside (2R and 2S) were
synthesized from 6-(S)-2H-1,6-anhydro-ꢁ-d-gal-
actopyranoside (3), which was prepared via
photobromination and reduction with tri-n-butyl-
tindeuteride as described previously [4]. The
incorporation of deuterium at C-6 was >95% as
judged by 1H NMR. Benzylation of 3 using
sodium hydride and benzyl bromide gave 4.
Acetolysis gave an anomeric mixture which upon
methanolysis a€orded 5, which was separated
from its ꢁ-anomer by silica gel chromatography.
Treatment of 5 with sodium hydride and (S)- or
(R)-2-chloropropanoic acid gave methyl 2,3,4,-tri-
O-benzyl-6-(S)-2H-6-O-[(R)- and (S)-1-carboxy-
ethyl]-ꢀ-d-galactopyranoside, respectively (6R and
6S). Hydrogenolysis and puri®cation by gel per-
meation chromatography gave 2R and 2S, which
were used in the sodium salt form for the NMR
studies.
Stereospeci®c NMR assignments could be per-
formed for the signals from the H-6 protons since
in the deuterated analogues H-6 pro-S has been
substituted with a deuterium atom. In 1R the pro-
ton signals at ꢂH 3.76 and 3.59 are assigned to pro-
R (H-6R) and pro-S (H-6S), respectively, whereas in
1S the signals appear at ꢂH 3.56 and 3.70, respec-
tively. Thus, the order of the chemical shifts are
reversed for the signals of the H-6 protons in the
two isomers. Such an e€ect is also observed when
substitution in disaccharides takes place at C-6 by
3
ꢀ- or ꢁ-linked sugar residues [5]. The JH,H cou-
plings between H-5 and H-6R/H-6S (Table 1) were
measured to be used in the subsequent conforma-
tional analysis of the ! dihedral angle de®ned by
O-5±C-5±C-6±O-6. In 1R and 1S two more dihedral
0
angles are important, namely de®ned by C-5±C-
6±O-6±C-20 and '0 de®ned by C-6±O-6±C-20±H-20

R. Stenutz et al./Carbohydrate Research 312 (1998) 117±121
119
Table 1
Coupling constants in Hz for 1R and 1S measured at 30 C
larger than for 1S both in the simulation and as
determined by experiment. Although the simula-
tions led to gauche states, the experimental values
are compatible with an equal population of the
three staggered '0 conformers.
ꢀ
Compound H-6R,
H-6S
H-5, H-5, H-6R, H-6S, C-6,
C20
H-6R H-6S
C-20
H-20
1R
1S
11.0
10.4
7.1
7.6
4.6
4.6
4.9
4.8
3.7
4.2
3.7
3.3
The conformational preference of the 0 dihedral
a
3
angle was calculated from the JC,H values in
Table 1 and eq (2). The results are given in Table 2.
a
From ref. [2].
3
For both isomers the JC,H coupling constants
in order to describe the conformational ¯exibility
of the 1-carboxyethyl group. The conformational
between C-20 and H-6R are of the same magnitude
(within experimental error) and quite large. The
LD simulation showed the trans conformer to be
most abundant, 68 and 60%, followed by the g
state with 18 and 38% for 1R and 1S, respectively.
0
preferences of and '0 can be investigated by
3
analysis of the heteronuclear JC,H coupling con-
stants. Carbon atoms C-6 and C-20 were site-selec-
3
tively excited and the long-range JC,H couplings
The di€erence in the JC-2,H-6 values between 1R and
S
[6,7] to H-20 and H-6R/H-6S, respectively0, could be
measured (Table 1). For the ! and dihedral
angles two three-bond coupling constants were
obtained whereas for '0 only one was acquired.
These J values were then used in the subsequent
conformational analysis.
The preferred dihedral angle can be studied
using Karplus equations. For the ! dihedral angle
the following equation [8] can be used for de®ning
the population distribution of staggered con-
formers from 3JH5,H6R and 3JH5,H6S values.
1S is larger than the experimental error showing that
conformational di€erences between the molecules
do exist. Analysis of the NMR data leads to a
somewhat higher population of the g+ state com-
pared to the g state, for both isomers. In contrast to
the simulations no signi®cant population of a trans
state can be corroborated by experimental data.
In galactopyranose the hydroxymethyl group
populates all three staggered conformers [10] as
opposed to glucopyranose in which only the gt and
gg conformers are signi®cantly populated. The
distribution of the ! dihedral angle in 1R and 1S
3
3JHH ˆ P1 cos2 ꢃ ‡ P2 cos ꢃ ‡ P3 ‡ ÆÁꢄi
ꢀ1†
(Table 2) was calculated from the JH,H coupling
È
É
constants in Table 1 and eq (1). The preferred
orientation of ! is similar for the two isomers of 1.
The gt conformer is most abundant, followed by
the tg and then the gg conformers, as in galactose,
but the relative proportions are changed. The
agreement between the LD simulation and experi-
ment was good for 1R with relative proportions
between gt:tg:gg of 68:25:7 in the simulation.
For 1S the gt conformer was populated 95% of
the time in the simulation. It is the major con-
former as determined by experiment, albeit not
that pronounced.
P4 ‡ P5 cos2ꢀꢅꢃ ‡ P6jÁꢄij†
In all analyses a restriction is also imposed, viz.,
that the sum of the fractional populations is equal
to unity. The two 0other major degrees of freedom
in the molecule, and '0, were analyzed by long-
3
range heteronuclear JC,H coupling constants. The
following Karplus type relationship was used [9].
3JHCOC ˆ 5:7 cos2 ꢃ 0:6 cos ꢃ ‡ 0:5
ꢀ2†
For the '0 dihedral angle the magnitude of the
3JC,H coupling constants in 1R and 1S are similar
and are consistent with equal populations in all
three rotameric states. In the LD simulations [2]
lasting 30 ns for each isomer, gauche states were
observed for 98 and 100% of the time for 1R and
Table 2
Preferred orientation a in 1R and 1S of ! and 0 dihedral angles
calculated from JH,H and JC,H values, respectively
3
3
+
Compound
Dihedral
angle
Pg
Pt
Pg
1R
1R
1S
1S
!
57
63
62
61
26
18
40
12
50
3
1S, respectively. The value of the JC,H coupling
0
3
26
11
constants between H-20 and C-6 were calculated to
be 3.1 and 2.6 Hz, to be compared to those mea-
sured from NMR which were 3.7 and 3.3 Hz for 1R
!
0
a
0
! refers to the dihedral angle O-5±C-5±C-6±O-6 and to
3
the dihedral angle C-5±C-6±O-6±C-20.
and 1S, respectively. The JC,H value for 1R was

120
R. Stenutz et al./Carbohydrate Research 312 (1998) 117±121
If instead the experimental data are compared to
the low energy conformers identi®ed previously by
energy minimizations [2], gauche states of are
a gradient enhanced version for 13C site selective
excitation as devised by Nishida et al. [7] For C-6
the selective excitation used a Gaussian shaped
pulse of 200 ms duration and a spectral width of
4000 Hz was sampled with 16,384 data points
(accuracy ‹0.3 Hz) using 23,000 transients. Band-
selective decoupling of the methyl group (ꢂH 1.33)
was applied during the acquisition period. For C-20
the selective excitation used a half-Gaussian
shaped pulse of 50 ms duration and a spectral
width of 1200 Hz was sampled with 8192 data
points using 19,000 and 33,000 transients for 1R
and 1S. The accuracy is estimated to be ‹0.2 Hz,
except for C-20 to H-6 pro-R in 1S where it is
‹0.5 Hz. The FIDs were processed using Varian
VNMR software. An exponential weighting func-
tion with a line broadening of 0.6 Hz was applied
0
observed for the lowest energy conformers of both
isomers, although these states were not the most
populated in the LD simulation. The outcome of
the simulation is sensitive to small changes in the
potential energy surface as the energy di€erence
between conformational states is ꢁ1 kcal/mol or
less, with low barriers for interconversion. Hence,
the poor agreement between simulation and
experiment.
In summary, the 1-carboxyethyl substituted
galactose residue in this study is characterized by
large conformational ¯exibility of its exocyclic
dihedral angles. For the ! angle all staggered con-
0
formers are signi®cantly populated. The angle
3
has both gauche conformers equally populated and
the '0 angle has at least two staggered conformers
populated to a large extent.
prior to Fourier transformation. The JC,H value
on excitation of C-6 was measured directly from
the separation of the anti-phase components in the
H20 signal. Otherwise the J-doubling procedure
was used with zero-®lling eight times of the FID
and eight delta functions in the frequency domain
[11,12].
3. Experimental
General.ÐSodium hydride (55±60%, suspension
in oil) was washed with hexane prior to use. Thin
layer chromatography (TLC) was performed on
pre-coated plates (Merck Silica Gel 60 F254) which
were developed by charring with aqueous 8%
H2SO4. Column chromatography of synthetic
intermediates was performed on Matrex silica gel
60 (35±70 ꢆm, Amicon). Gel permeation chroma-
tography of deprotected products was carried out
on polyacrylamide gel (Bio-Gel P-2, 2.6Â90 cm) in
0.1 M pyridinium acetate bu€er. Solutions were
concentrated under reduced pressure at tempera-
tures not exceeding 50 ꢀC. Atoms in the 1-carboxy-
ethyl group are numbered from the carboxyl group
(10). A g+ state (gt for !) denotes a dihedral angle
of +60ꢀ, a g state (gg) has its dihedral around
60ꢀ and a t state (tg) has its dihedral around
180ꢀ.
NMR spectroscopy.ÐNMR spectra were recor-
ded at 30 ꢀC on JEOL GSX-270 MHz, Varian
Unity 500 MHz or Varian Inova 600 MHz NMR
spectrometers. For solutions in CDCl3 or D2O tet-
ramethylsilane (ꢂH/ꢂC 0.00 ppm) or sodium 3-tri-
methylsilylpropanoate-2H4 (ꢂH 0.00 ppm) were
used as internal references, respectively. Measure-
ment of long-range 13C,1H coupling constants were
performed for D2O solutions of the sodium salt of
1R and 1S (pD 8, 40 and 60 mM, respectively) using
Synthesis.Ð6-(S)-2H-1,6-anhydro-2,3,4-tri-O-ben-
zyl-ꢁ-d-galactopyranoside (4). A solution of 6-
(S)-2H-1,6-anhydro-ꢁ-d-galactopyranoside [4] (3,
780 mg, 4.8 mmol, >95% 2H according to 1H
NMR) in DMF (10 mL) was added dropwise to a
suspension of NaH (1.0 g) in DMF (10 mL), fol-
lowed by benzyl bromide (1.9 mL, 16.0 mmol) in
DMF (10 mL). After 3 h water was added dropwise
and the reaction mixture was extracted with
CHCl3. The organic phases were dried with anhy-
drous MgSO4 and evaporated to yield (2.01 g,
97%) crude 4 which was used without further
puri®cation. 13C NMR data were in good agree-
ment with published values for the non-deuterated
derivative [13].
Methyl 6-(S)-2H-2,3,4-tri-O-benzyl-a-d-galacto-
pyranoside (5). 6-(S)-2H-1,6-anhydro-2,3,4-tri-O-
benzyl-ꢁ-d-galactopyranoside (4, 340 mg, 0.79 mmol)
was dissolved in warm acetic anhydride (1 mL).
ꢀ
The solution was cooled to 0 C and 0.05 mL of a
freshly prepared mixture of concentrated H2SO4
(0.1 mL) in acetic anhydride (5 mL) was added.
After completion of the reaction as indicated by
TLC (toluene±EtOAc, 4:1), the mixture was
poured into ice water (6 mL). The water was de-
canted after 1 h from the oily product and replaced.
After 20 h the water was decanted again and the
product dissolved in CHCl3. The organic layer was

R. Stenutz et al./Carbohydrate Research 312 (1998) 117±121
121
washed with aqueous 10% sodium bicarbonate,
water, dried (MgSO4) and evaporated to dryness.
Column chromatography (toluene±EtOAc, 9:1)
yielded an anomeric mixture of 6-(S)-2H-1,6-di-
O-acetyl-2,3,4-tri-O-benzyl-d-galactopyranoside
(308 mg, 73%, ꢀ/ꢁ&3:1). 13C NMR (CDCl3)
showed, inter alia, signals at 94.2 (C-1-ꢁ), 90.8 (C-
1-ꢀ) and 63.0 (triplet, C6).
The anomeric mixture of 6-(S)-2H-1,6-di-O-
acetyl-2,3,4-tri-O-benzyl-d-galactopyranoside (2.9 g,
5.4 mmol) was dissolved in dry methanol (80 mL)
and treated with 2M HCl (80 mL) in MeOH. After
10 h the reaction mixture was concentrated to yield
an anomeric mixture of methyl glycosides (94%
2.4 g, ꢀ/ꢁ&1:1). Column chromatography (CHCl3±
EtOAc, 5:1) a€orded the pure ꢀ-anomer 5 (0.99 g,
41%). 13C NMR (CDCl3): 139.0±127.4 (Bn), 98.6
(C-1), 78.8, 76.2, 74.8, 73.4, 69.8 (C-2±C-5 and
3ÂCH2, overlapping signals), 55.1 (OMe). The ꢁ-
anomer was obtained as a by-product (0.74 g,
31%).
(300 mg, 2.76 mmol) and NaH (1.15 g), using the
same experimental conditions as in the synthesis
of 6R, methyl 6-(S)-2H-6-O-[(R)-1-carboxyethyl]-
2,3,4-tri-O-benzyl-ꢀ-d-galactopyranoside 6S was
obtained (108 mg, 82%). 13C NMR data were in
agreement with those of the non-deuterated com-
pound [1].
Compound 6S (92 mg, 0.18 mmol) in MeOH was
subjected to hydrogenolysis for 25 h at atmospheric
pressure using Pd±C as catalyst. Filtration and
concentration gave 2S (42 mg, 98%). 13C NMR
data were in agreement with those of the non-
deuterated compound [1]. Prior to NMR studies
compound 2S was treated with Chelex-100 resin,
®ltered, puri®ed by gel permeation chromato-
graphy and converted to its sodium salt by ion-
exchange (Dowex-50, Na+).
Acknowledgements
This work was supported by grants from the
Swedish Natural Science Research Council. The
Swedish NMR Centre is thanked for access to the
Varian Unity 500 NMR spectrometer.
Methyl 6-(S)-2H-6-O-[(R)-1-carboxyethyl]-a-d-
galactopyranoside (2R). A solution of methyl 6-
(S)-2H-2,3,4-tri-O-benzyl-ꢀ-d-galactopyranoside
(5, 106 mg, 0.23 mmol) and (S)-chloropropanoic
acid (50 mg, 0.46 mmol) in 1,4-dioxane (20 mL) was
ꢀ
stirred with NaH (0.6 g) at 50 C for 16 h, where-
after the mixture was cooled and water (20 mL)
was added. Extraction with petroleum ether (bp
References
[1] M. Andersson, L. Kenne, R. Stenutz, and
G. Widmalm, Carbohydr. Res., 254 (1994) 35±41.
[2] R. Stenutz and G. Widmalm, Glycoconjugate J., 14
(1997) 973±981.
[3] L. Eriksson, A. Pilotti, R. Stenutz, and G. Widmalm,
Acta. Cryst., C52 (1996) 2285±2287.
[4] H. Ohrui, Y. Nishida, and H. Meguro, Agric. Biol,
Chem., 48 (1984) 1049±1053.
[5] P.-E. Jansson, L. Kenne, and I. Kolare, Carbohydr.
Res., 257 (1994) 163±174.
[6] T. Nishida, G. Widmalm, and P. Sandor, Magn.
Reson. Chem., 33 (1995) 596±599.
[7] T. Nishida, G. Widmalm, and P. Sandor, Magn.
Reson. Chem., 34 (1996) 377±382.
[8] C.A.G. Haasnoot, F.A.A.M. de Leeuw, and
C. Altona, Tetrahedron, 36 (1980) 2783±2792.
[9] I. Tvaroska, M. Hricovõni, and E. Petrakova, Car-
bohydr. Res., 189 (1989) 359±362.
[10] Y. Nishida, H. Ohrui, and H. Meguro, Tetrahedr.
Lett., 25 (1984) 1575±1578.
ꢀ
60±70 C, 2Â25 mL) was followed by acidi®cation
of the water phase (pH&3, AcOH). The water
phase was extracted with CHCl3 (3Â25 mL) and
the combined CHCl3 phases were dried (MgSO4),
®ltered, and concentrated to give methyl 6-(S)-2H-
6-O-[(R)-1-carboxyethyl-2,3,4-tri-O-benzyl-ꢀ-d-gal-
actopyranoside 6R (90 mg, 74%) as a syrup. 13C
NMR data were in agreement with those of the
non-deuterated compound [1].
Compound 6R (88 mg, 0.17 mmol) in MeOH was
subjected to hydrogenolysis for 25 h at atmospheric
pressure using Pd±C as catalyst. Filtration and
concentration gave 2R (40 mg, 98%). 13C NMR
data were in agreement with those of the non-
deuterated compound [1]. Prior to NMR studies
compound 2R was treated with Chelex-100 cation
exchange resin, ®ltered, puri®ed by gel permeation
chromatography and converted to its sodium salt
by ion-exchange (Dowex-50, Na+).
Methyl 6-(S)-2H-6-O-[(S)-1-carboxyethyl]-a-d-
galactopyranoside (2S). Starting from methyl 6-
(S)-2H-2,3,4-tri-O-benzyl-ꢀ-d-galactopyranoside (5,
116 mg, 0.24 mmol), (R)-chloropropanoic acid
[11] L. McIntyre and R. Freeman, J. Magn. Reson., 96
(1992) 425±431.
[12] F. del Rio-Portilla, V. Blechta, and R. Freeman, J.
Magn. Reson., A111 (1994) 132±135.
[13] P.-M. Aberg and B. Ernst, Acta. Chem. Scand., 48
(1994) 228±233.