Structural determination of a neutral exopolysaccharide produced by Lactobacillus delbrueckii ssp. bulgaricus LBB.B332

Article abstract of DOI:10.1016/j.carres.2007.09.005

The neutral exopolysaccharide produced by Lactobacillus delbrueckii ssp. bulgaricus LBB.B332 in skimmed milk was found to be composed of d-glucose, d-galactose, and l-rhamnose in a molar ratio of 1:2:2. Linkage analysis and 1D/2D NMR (¹H and ¹³C) studies carried out on the native polysaccharide as well as on an oligosaccharide generated by a periodate oxidation protocol, showed the polysaccharide to consist of linear pentasaccharide repeating units with the following structure:{A figure is presented}.

Full text of DOI:10.1016/j.carres.2007.09.005

Available online at www.sciencedirect.com
Carbohydrate Research 342 (2007) 2735–2744
Structural determination of a neutral exopolysaccharide produced
by Lactobacillus delbrueckii ssp. bulgaricus LBB.B332
a
a
b
Inmaculada S a´ nchez-Medina, Gerrit J. Gerwig, Zoltan L. Urshev and
a,
*
Johannis P. Kamerling
a
Bijvoet Center, Department of Bio-Organic Chemistry, Utrecht University, Padualaan 8, NL-3584 CH Utrecht, The Netherlands
b
LB Bulgaricum Plc., R&D Center, 12A Malashevska Str., Sofia 1202, Bulgaria
Received 15 August 2007; accepted 14 September 2007
Available online 22 September 2007
Abstract—The neutral exopolysaccharide produced by Lactobacillus delbrueckii ssp. bulgaricus LBB.B332 in skimmed milk was
found to be composed of D-glucose, D-galactose, and L-rhamnose in a molar ratio of 1:2:2. Linkage analysis and 1D/2D NMR
1
13
(
H and C) studies carried out on the native polysaccharide as well as on an oligosaccharide generated by a periodate oxidation
protocol, showed the polysaccharide to consist of linear pentasaccharide repeating units with the following structure:
ꢀ
2007 Elsevier Ltd. All rights reserved.
Keywords: Lactobacillus delbrueckii ssp. bulgaricus; Exopolysaccharide; Lactic acid bacteria; Structural analysis; NMR spectroscopy; Mass
spectrometry
1
. Introduction
strains as commercial yoghurt starters. Several EPSs
produced by Lb. delbrueckii ssp. bulgaricus have been
characterized, being mainly composed of Glc and
Microbial exopolysaccharides (EPSs) produced by lactic
acid bacteria are of increasing interest to the food indus-
try. Their particular physical and rheological properties,
which make them suitable as viscosifying, stabilizing,
3
–6
7–10
Gal or of Glc, Gal, and Rha.
Here, we report on the structure of the neutral EPS
produced by Lb. delbrueckii ssp. bulgaricus LBB.B332
in skimmed milk.
1
gelling, or emulsifying agents, in combination with
the GRAS (generally recognized as safe) status of
EPS-producing lactic acid bacteria, make EPSs promis-
ing as a new generation of food thickeners. To unravel
the relationship between their structure and physical
properties, structural studies have been performed on
EPSs produced by different species of lactic acid bacte-
ria, such as Lactobacillus, Lactococcus, and Streptococ-
2
. Results and discussion
2.1. Isolation, purification and composition of the
exopolysaccharide
2
cus genera.
Lactobacillus delbrueckii ssp. bulgaricus strains are
used in combination with Streptococcus thermophilus
The neutral EPS produced by Lb. delbrueckii ssp.
bulgaricus LBB.B332 was isolated via absolute ethanol
precipitation of the trichloroacetic acid-treated culture
medium, and further purified by anion-exchange chro-
matography on DEAE-Trisacryl Plus M. Its average
*
Corresponding author. Tel.: +31 30 253 34 79; fax: +31 30 254 09
0; e-mail: j.p.kamerling@uu.nl
8
0
008-6215/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2007.09.005

2736
I. S a´ nchez-Medina et al. / Carbohydrate Research 342 (2007) 2735–2744
Table 1. Methylation analysis data of Lactobacillus delbrueckii ssp. bulgaricus LBB.B332 neutral EPS and oligosaccharide TRI
a
b
b
Partially methylated alditol acetate
t
R
Structural feature
Molar ratio EPS
Molar ratio TRI
3
2
2
2
,4-Di-O-methyl-1,2,5-tri-O-acetylrhamnitol-1-d
,4-Di-O-methyl-1,3,5-tri-O-acetylrhamnitol-1-d
,3,4,6-Tetra-O-methyl-1,5-di-O-acetylglucitol-1-d
,4,6-Tri-O-methyl-1,3,5-tri-O-acetylglucitol-1-d
0.92
0.94
1.00
1.18
o
!2)-Rhap-(1!
!3)-Rhap-(1!
Glcp-(1!
1.00
0.83
—
—
0.7
1.0
—
!3)-Glcp-(1!
0.81
o
2
3
,4,6-Tri-O-methyl-1,3,5-tri-O-acetylgalactitol-1-d
,4,6-Tri-O-methyl-1,2,5-tri-O-acetylgalactitol-1-d
!3)-Galp-(1!
1.1
—
1
:23
1:83
!2)-Galp-(1!
a
GLC retention times relative to 2,3,4,6-tetra-O-methyl-1,5-di-O-acetylglucitol-1-d on EC-1.
b
Calculated from peak areas, not corrected by response factors.
6
molecular mass of 1.3 · 10 Da was determined by
gel-filtration chromatography on Sephacryl S-400 HR.
Quantitative monosaccharide analysis, including abso-
lute configuration determination, of the EPS gave D-Glc,
D-Gal, and L-Rha in a molar ratio of 1.0:2.0:2.1. Methyl-
ation analysis (Table 1) showed the presence of 3-substi-
tuted Glcp, 3-substituted Galp, 2-substituted Galp
a-configurations only. The spectrum presents one high-
field signal at d 1.307, arising from the methyl groups
of two Rha residues.
2.2. Periodate oxidation and analysis of the trisaccharide
TRI
1
1
(
coeluting with the 3-substituted Galp in the gas chro-
The material obtained after periodate oxidation of the
EPS and subsequent dialysis was subjected to monosac-
charide analysis, showing D-Glc, D-Gal, and L-Rha in a
molar ratio of 1.0:1.0:1.1, which indicated the loss of
one Gal and one Rha residue with respect to the native
matogram), 2-substituted Rhap, and 3-substituted Rhap,
in a molar ratio of 0.8:1.8:1.0:0.8, suggesting a repeating
linear pentasaccharide with all residues in the pyranose
ring form.
1
The 1D H NMR spectrum of the EPS (Fig. 1)
EPS. After reduction with NaBH and dialysis of the
4
showed five signals of equal intensity in the anomeric
region (d 5.0–5.5), supporting a pentasaccharide repeat-
ing unit. The constituting monosaccharide units in the
EPS were arbitrarily named from A to E, according to
the decreasing chemical shift values of their anomeric
protons. Their d values suggest the occurrence of
sample, monosaccharide analysis confirmed the presence
of equimolar amounts of D-Glc, D-Gal, and L-Rha. The
degraded, but still polymeric material was subjected to
mild acid hydrolysis (0.5 M TFA, 80 ꢁC), showing after
30 min a major TLC band in the trisaccharide region;
fractionation on Bio-Gel P-2 yielded a main fraction,
1
Figure 1. 500-MHz 1D H NMR spectrum of the neutral EPS produced by Lactobacillus delbrueckii ssp. bulgaricus LBB.B332, recorded in D
2
O at
78 ꢁC.

I. S a´ nchez-Medina et al. / Carbohydrate Research 342 (2007) 2735–2744
2737
denoted TRI. MALDI-TOF-MS of TRI showed a
results establish the following structure for the trisaccha-
ride-alditol:
+
[
M+Na] pseudomolecular ion at m/z 510.67, together
+
with a [M+K] ion at m/z 527.16, corresponding with
Hex dHex (M = 488.17 Da). According to the periodate
2
oxidation protocol, after mild acid hydrolysis an addi-
tional glyceraldehyde group should be attached to the
reducing end. The absence of this group, as indicated
by MALDI-TOF-MS, is caused by the acid lability of
the rhamnosyl glyceraldehyde linkage under acidic
conditions. A similar situation had been previously
In order to generate supporting NMR data on the tri-
saccharide level for the NMR analysis of the EPS, TRI
was investigated in great detail. The 1D H NMR spec-
1
1
2
reported.
trum of TRI (Fig. 2) showed four major signals in the
3
Monosaccharide analysis of TRI demonstrated
anomeric proton region at d 5.188 (residue C
b
, J1,2
3
Hex dHex to be built up from D-Glc, D-Gal, and
3.9 Hz), d 5.162 (residue C , J1,2 3.9 Hz; residue Ea,
a
J1,2 <1 Hz), d 5.141 (residue D, J1,2 3.9 Hz), and d
4.877 (residue Eb, J1,2 <1 Hz). Taking into account
the d values of residues D and C, combined with the
methylation analysis data, the a-pyranose configuration
is assigned for both units. Residue E corresponds to the
2
3
3
L-Rha. Linkage analysis (Table 1) showed the presence
of terminal Glcp, 3-substituted Galp, and 3-substituted
Rhap, in a molar ratio of 1.0:1.1:0.7, in accordance with
a linear trisaccharide structure.
3
MALDI-TOF-MS analysis of NaBH -reduced TRI
4
+
+
showed [M+Na] and [M+K] pseudomolecular ions
reducing Rha unit. Splitting of C into C and C is
a b
at m/z 513.33 and 529.03, respectively, corresponding
observed as a consequence of the linkage between C
and E.
The H chemical shifts of all protons of TRI (Table 2)
were obtained by means of 2D TOCSY (mixing times,
40–100 ms), ROESY (mixing time, 300 ms), and
with Hex dHex-ol (M = 490.19 Da). Monosaccharide
2
1
analysis revealed Glc, Gal, and Rha-ol in a molar ratio
of 0.9:1.0:1.0, whereas methylation analysis, following a
methanolysis/trimethylsilylation protocol, indicated the
presence of terminal Glcp, 3-substituted Galp, and
1
13
H– C HSQC experiments. The TOCSY spectrum of
TRI with a mixing time of 100 ms is shown in Figure
3
-substituted Rha-ol in a molar ratio of 1.0:1.3:0.6. 1D
1
1
13
H NMR analysis showed two anomeric signals, that
is, terminal Glc D H-1 at d 5.120 ( J 3.9 Hz) and
Gal C H-1 at d 5.264 ( J
3, together with the ROESY spectrum. The H– C
HSQC spectrum is shown in Figure 4. Starting points
for the interpretation of the spectra were the anomeric
3
1
.2
3
3.5 Hz). The combined
.2
1
1
Figure 2. 500-MHz 1D H NMR spectrum of oligosaccharide TRI, recorded in D
2
O at 27 ꢁC.

2738
I. S a´ nchez-Medina et al. / Carbohydrate Research 342 (2007) 2735–2744
1
13
a
Table 2. H and C NMR chemical shifts of EPS, recorded in D
2
O at 78 ꢁC, and oligosaccharide TRI, recorded in D
2
O at 27 ꢁC
Residue
Proton
EPS
TRI
Carbon
EPS
TRI
H-1
H-2
H-3
H-4
H-5
H-6a
H-6b
5.458
3.98
4.05
4.07
4.30
3.78
3.78
—
—
—
—
—
—
—
C-1
C-2
C-3
C-4
C-5
C-6
99.4 (179)
—
—
—
—
—
—
2
)-α-D-Galp-(1
75.6
70.5
71.1
72.1
62.3
A
H-1
H-2
H-3
H-4
H-5
5.239
4.09
3.92
3.51
3.84
1.307
—
—
—
—
—
—
C-1
C-2
C-3
C-4
C-5
101.5 (172)
79.5
71.4
73.6
70.5
—
—
—
—
—
—
2
)-α-L-Rha p-(1
B
CH
3
CH
3
17.7
H-1
H-2
H-3
H-4
H-5
H-6a
H-6b
5.196
4.05
4.06
4.27
4.19
3.77
3.77
5.162/5.188 (3.9)
4.06/4.01
4.01/4.06
4.26
4.21
3.75
C-1
C-2
C-3
C-4
C-5
C-6
96.8 (172)
68.1
76.2
67.1
71.9
100.2/100.0 (171)
3
)-α-D-Galp-(1
71.4
79.4
70.3
75.2
65.4
C
62.3
3.75
H-1
H-2
H-3
H-4
H-5
H-6a
H-6b
5.154
3.74
4.00
3.68
4.00
3.86
3.79
—
—
—
—
—
—
—
C-1
C-2
C-3
C-4
C-5
C-6
96.7 (174)
71.4
80.3
71.5
73.2
—
3
)-α-D-Glcp-(1
—
—
—
—
—
D
61.9
H-1
H-2
H-3
H-4
H-5
5.042
4.29
3.90
3.59
3.77
1.307
—
—
—
—
—
—
C-1
C-2
C-3
C-4
C-5
CH
103.1 (170)
68.1
76.8
71.8
70.5
—
—
—
—
—
—
3
)-α-L-Rha p-(1
E
CH
3
3
17.7
H-1
H-2
H-3
H-4
H-5
H-6a
H-6b
—
—
—
—
—
—
—
5.141 (3.9)
3.59
3.80
3.46
3.96
C-1
C-2
C-3
C-4
C-5
C-6
—
—
—
—
—
—
100.2 (171)
76.2
77.6
74.2
76.6
α-D-Glcp-(1
D
3.85
3.78
65.0
H-1
H-2
H-3
H-4
H-5
—
—
—
—
—
—
5.162 (<1)
4.15
3.90
3.56
3.93
C-1
C-2
C-3
C-4
C-5
CH
—
—
—
—
—
—
98.5 (170)
72.0
79.8
75.2
73.1
3
)-α-L-Rha p
E
CH
3
1.299
3
17.1
H-1
H-2
H-3
H-4
H-5
—
—
—
—
—
—
4.877 (<1)
4.17
3.72
3.48
3.46
C-1
C-2
C-3
C-4
C-5
CH
—
—
—
—
—
—
98.3 (n.d.)
72.2
82.0
74.8
76.7
3
)-β-L-Rha p
E
CH
3
1.322
3
17.1
3
a
1
J
1,2 and JC-1,H-1 coupling constants are included in parentheses.
1
13
13
In ppm relative to the signal of internal acetone at d 2.225 for H, and in ppm relative to the signal of external [1- C] glucose (dC-1 92.9) for C.
signals of residues C , C , D, Ea, and Eb, and the
of the chemical shifts belonging to the same spin
system.
The characteristic spin system seen on the TOCSY H-1
track of residue D (H-2,3,4,5,6a,6b) indicated a gluco-
a
b
methyl signals of rhamnose residues Ea and Eb.
Comparison of TOCSY spectra with increasing mixing
times allowed the assignment of the sequential order

I. S a´ nchez-Medina et al. / Carbohydrate Research 342 (2007) 2735–2744
2739
Figure 3. 2D TOCSY (mixing time, 100 ms) and ROESY (mixing time, 300 ms) spectra of the oligosaccharide TRI, recorded in D
2
O at 27 ꢁC. The
3
CH signals were observed at 1.30 and 1.32 ppm, but were not included in the picture. Cross-peaks belonging to the same scalar-coupling network are
indicated near a dotted line starting from the corresponding diagonal peaks; TOCSY: D1 corresponds to the diagonal peak of residue D H-1; D2
refers to a cross-peak between D H-1 and D H-2, etc.; ROESY: D1 corresponds to the diagonal peak of residue D H-1; D1,2 refers to an intra-residue
cross-peak between D H-1 and D H-2, and D1,C4 indicates an inter-residue connectivity between D H-1 and C H-4, etc.
configuration, whereas those found on the TOCSY H-1
79.4; a-D-Galp1Me, dC-3 70.5), and a 3-substituted
a-Rhap E unit (Ea C-3, d 79.8; a-L-Rhap, dC-3 71.1;
tracks of residues C and C (H-2,3,4) a galacto-configu-
a
b
1
4
ration. Note the typical upfield positions of Glc H-2 and
H-4, and the typical downfield position of Gal H-4. The
finding of only one cross-peak (H-2) on the TOCSY H-1
tracks of both residues Ea and Eb is in agreement with a
rhamno-configuration. In both cases a correlation with
Eb C-3, d 82.0; b-L-Rhap, dC-3 73.8) could be verified.
Finally, sequence analysis data followed from inter-
1
13
residue ROESY cross-peaks (Fig. 3) and H– C HMBC
long-range couplings. The D(1!3)C linkage is reflected
by the ROESY cross-peaks D H-1/C H-3,4 and the
1
13
the related upfield CH signal is detected.
H– C HMBC connectivity observed between D C-1
3
1
3
The C resonances of TRI (Table 2) were determined
by interpretation of the H– C HSQC spectrum
Fig. 4), whereas a 2D H– C HMBC spectrum
and Cb H-3. The C(1!3)E linkage fits the ROESY
1
13
cross-peaks C H-1/Ea H-2,3 and C H-1/Eb H-2,3. A
a
b
1
13
1
13
(
H– C HMBC connectivity between Ca C-1 and Ea
1
1
revealed the J
coupling constants. The J
H-3 confirms this linkage.
C-1,H-1
C-1,H-1
values of 171 Hz for Galp C and Glcp D confirmed their
Taking together all information obtained, the struc-
ture of TRI can be formulated as
1
3
a configuration.
1
3
By comparison of the C chemical shifts of residues
C , C , D, Ea, and Eb with those of (methyl) aldo-
a
b
1
4
sides, going for typical downfield shifts, the occurrence
of a terminal a-Glcp D unit (compare with a-D-
Glcp1Me), a 3-substituted a-Galp C unit (C C-3, d

2740
I. S a´ nchez-Medina et al. / Carbohydrate Research 342 (2007) 2735–2744
2
.3. 2D NMR spectroscopy of the native polysaccharide
Rha residues. The TOCSY results for residue D indicate
a gluco-configuration, whereas the spin systems of resi-
dues A and C, with downfield positions of H-4, are in
agreement with a galacto-configuration.
1
13
The complete assignment of the H and C chemical
shifts of the native EPS (Table 2) was made by means
of 2D TOCSY (mixing times, 40–100 ms), NOESY
Evaluation of the C-1 chemical shifts (Table 2; Fig. 6)
1
13
1
(
mixing time, 150 ms), and H– C HSQC experiments.
The TOCSY spectrum (100 ms) of the EPS is shown in
Figure 5, together with the NOESY spectrum. The
and the J
coupling constants, deduced from 2D
C-1,H-1
1
13
H– C HMBC experiments, confirmed that all residues
occur in a-pyranosyl form: residue A/Gal, d 99.4,
JC-1,H-1 179 Hz; residue B/Rha, d 101.5, JC-1,H-1
172 Hz; residue C/Gal, d 96.8, J
D/Glc, d 96.8, J
103.1, JC-1,H-1 170 Hz. The relatively high JC-1,H-1
value of residue A had been reported previously for
1
3
1
13
1
1
H– C HSQC spectrum is shown in Figure 6. Starting
1
points for the interpretation of the spectra were the ano-
meric signals of the residues A–E, and the methyl signals
of the Rha residues B and E. Comparison of TOCSY
spectra with increasing mixing times allowed the assign-
ment of the sequential order of the chemical shifts
belonging to the same spin system.
The TOCSY A H-1 track (d 5.458) showed cross-
peaks with A H-2,3,4. The resonances for A H-5,6a,6b
were found in the HSQC spectrum. Using the TOCSY
B H-2 track, found via the B H-1 track (d 5.239), the res-
172 Hz; residue
C-1,H-1
1
174 Hz; and residue E/Rha, d
C-1,H-1
1
1
1
2,15,16
2-substituted monosaccharides,
and is probably
due to the 2-substitution. Note that the a-anomeric
configuration of the two Rha residues follows also from
a comparison of the chemical shifts of their H-5 atoms (B
H-5, d 3.84; E H-5, d 3.77) with those of a-L-Rhap1Me
1
7
(H-5, d 3.66) and b-L-Rhap1Me (H-5, d 3.39).
1
3
onances for B H-3,4,5,CH were detected. The TOCSY
Taking into account the published C chemical shift
3
1
4
C H-1 track (d 5.196) revealed the cross-peaks with C
H-2,3,4,5, whereas on the C H-5 track the cross-peak
with C H-6a,b was found. The TOCSY D H-1 track (d
.154) showed cross-peaks with D H-2,3,4. The assign-
ment of the D H-5,6a,6b resonances was made via the
H– C HSQC spectrum. Finally, via the TOCSY E
H-1 track (d 5.042) the E H-2 track was found, which
showed the cross-peaks with E H-3,4,5,CH3.
As is evident from the TOCSY results, residues B and
E, with the short H-1 tracks (only H-2 is seen) and the
typical H-6 signals for 6-deoxyhexoses, represent the
data of methyl aldosides, and the methylation analysis
data of the EPS (Table 1), residue A was assigned as
2-substituted a-Galp (downfield shift of A C-2, d 75.6;
a-D-Galp1Me, dC-2 69.2), residue B as 2-substituted a-
Rhap (downfield shift of B C-2, d 79.5; a-L-Rhap1Me,
dC-2 71.0), residue C as 3-substituted a-Galp (downfield
shift of C C-3, d 76.2; a-D-Galp1Me, dC-3 70.5), residue
D as 3-substituted a-Glcp (downfield shift of D C-3, d
80.3; a-D-Glcp1Me, dC-3 74.1), and residue E as 3-substi-
tuted a-Rhap (downfield shift of E C-3, d 76.8; a-L-
Rhap1Me, dC-3 71.3).
5
1
13
1
13
Figure 4. 2D H– C HSQC spectrum of the oligosaccharide TRI, recorded in D
1
2
3
O at 27 ꢁC. The CH signals were observed at 1.30/17.1 and 1.32/
7.1 ppm, but were not included in the picture. D1 corresponds to the cross-peak between D H-1 and D C-1, etc.

I. S a´ nchez-Medina et al. / Carbohydrate Research 342 (2007) 2735–2744
2741
Figure 5. 2D TOCSY (mixing time, 100 ms), and NOESY (mixing time, 150 ms) spectra of EPS, recorded in D
2
O at 78 ꢁC. The CH
3
signals were
observed at 1.31 ppm, but were not included in the picture. Cross-peaks belonging to the same scalar-coupling network are indicated near a dotted
line starting from the corresponding diagonal peaks; TOCSY: D1 corresponds to the diagonal peak of residue D H-1; D2 refers to a cross-peak
between D H-1 and D H-2, etc.; NOESY: D1 corresponds to the diagonal peak of residue D H-1; D1,2 refers to an intra-residue cross-peak between
D H-1 and D H-2, and D1,C4 indicates an inter-residue connectivity between D H-1 and C H-4, etc.
The establishment of the sequence of the monosaccha-
ride residues within the repeating unit of the EPS was
made by the assignment of the inter-residue cross-peaks
in the 2D NOESY spectrum (Fig. 5) and the relevant
long-range couplings in the HMBC spectrum (Table 3).
Inspection of the NOESY spectrum showed on the E
H-1 track an inter-residue cross-peak with B H-2, indi-
cating a E(1!2)B linkage. On the B H-1 NOESY track,
a connectivity was found between B H-1 and A H-2,
leading to the assignment of a B(1!2)A linkage. The in-
ter-residue A H-1,D H-3 connectivity supports the
occurrence of an A(1!3)D linkage. The observed
NOESY cross-peaks between D H-1 and C H-3,4, com-
residue C (vide supra) demonstrated a D(1!3)C link-
age. Finally, the inter-residue connectivity C H-1/E
H-3 allowed the assignment of the C(1!3)E linkage.
The observed intra-residue NOE connectivities were in
accordance with the assigned anomeric configurations.
Interestingly, two more inter-residue cross-peaks were
detected in the NOESY spectrum, namely, A H-1/B
H-5 and B H-1/E H-5, of importance for future confor-
mational studies of the EPS.
Combining the various data of the EPS analysis, sup-
ported by the structural determination of the generated
trisaccharide fragment, demonstrates the polysaccharide
to be built up from the following pentasaccharide
repeating unit:
1
3
bined with the methylation analysis/ C NMR data for

2742
I. S a´ nchez-Medina et al. / Carbohydrate Research 342 (2007) 2735–2744
1
13
Figure 6. 2D H– C HSQC spectrum of EPS, recorded in D
2
O at 78 ꢁC. The CH
3
signals were observed at 1.31/17.7 ppm, but were not included in
12,15,18–20
the picture. D1 corresponds to the cross-peak between D H-1 and D C-1, etc.
1
13
Table 3. Long-range H– C couplings found in the HMBC spectrum
for the anomeric signals of the residues of the EPS
branched EPSs produced by S. thermophilus
and
16
Lactococcus lactis ssp. cremoris strains. All these EPSs
have the same pentameric backbone consisting of a !3)-
a-D-Galp-(1!3)-a-L-Rhap-(1!2)-a-L-Rhap-(1!2)-a-
D-Galp-(1!3)-Hexp-(1! sequence, differing in the fifth
residue: an a-D-Glcp unit in the case of Lactococcus lactis
ssp. cremoris B39, S. thermophilus Sfi12, and Lb. del-
brueckii ssp. bulgaricus LBB.B332; an a-D-Galp unit in
the case of S. thermophilus Rs, Sts, OR 901 (three
Residue
d
H-1/C-1
Connectivities
80.3
Residue
A
5.458
D C-3
A C-5
A C-3
D H-3
7
2.1
7
0.5
9
9.4
4.00
16
19
B
C
5.239
79.5
B C-2
A C-2
B C-3
B C-5
A H-2
75.6
71.4
70.5
3.98
15
15
18
2
0
identical repeating units), and MR-1C; or a b-D-Galp
1
2
unit in the case of S. thermophilus S3. Furthermore, they
also keep the fourth position of the 3-substituted a-L-
Rhap unit as the branching point, with the exception of
the EPS of S. thermophilus S3, which has it at the sixth
position of the 3-substituted a-D-Galp. However, more
variations are found in the structure of the side chains.
1
01.5
5.196
76.8
E C-3
C C-3
C C-5
E H-3
7
7
6.2
1.9
96.8
3.90
D
E
5.154
80.3
D C-3
C C-3
D C-5
1
9
12
7
7
6.2
3.2
They can vary from single b-D-Galp or b-D-Galf units
1
6
to disaccharides such as b-D-Galp-(1!4)-b-D-Glcp or
1
5,18
b-D-Galp-(1!6)-b-D-Galp.
The latter is also found
5.042
79.5
B C-2
E C-3
E C-5
E C-2
E H-2
B H-2
in the EPS of S. thermophilus MR-1C, together with a sec-
76.8
70.5
68.1
4.29
2
0
ond side chain, consisting of a L-Fuc unit. This could
mean that these bacteria have a very similar pathway
for the biosynthesis of their corresponding EPSs.
1
03.1
4
.09
A: !2)-a-D-Galp-(1!; B: !2)-a-L-Rhap-(1!; C: !3)-a-D-Galp-(1!;
D: !3)-a-D-Glcp-(1!; E: !3)-a-L-Rhap-(1!.
3. Experimental
3.1. Production, isolation, and purification of the
exopolysaccharide
2
.4. Final remarks
The structure establishedfor the EPS of Lb. delbrueckii ssp.
bulgaricus LBB.B332 shows some similarities with other
The Lb. delbrueckii ssp. bulgaricus LBB.B332 strain, iso-
lated from home-made yoghurt, was obtained from the

I. S a´ nchez-Medina et al. / Carbohydrate Research 342 (2007) 2735–2744
2743
LBB collection of LB Bulgaricum Plc. (Sofia, Bulgaria).
3.4. Methylation analysis
An aliquot of an activated bacterial culture was used to
inoculate 1 L of sterile (121 ꢁC, 7 min) reconstituted
skimmed milk powder in water (10% w/v; E. Merck,
Darmstadt, Germany), and the strain was grown for
Samples (native EPS or oligosaccharide) were permethyl-
ated using methyl iodide and solid sodium hydroxide in
2
5
dimethyl sulfoxide as described previously. For the
work-up two protocols were followed. On one hand,
the permethylated material was methanolyzed and
analyzed as trimethylsilylated methyl glycosides by
GLC–MS (monosaccharide analysis protocol, see
Section 3.3). On the other hand, the permethylated
material was hydrolyzed with 2 M TFA (2 h, 120 ꢁC),
and the partially methylated monosaccharides obtained
2
4 h at 42 ꢁC. After incubation, proteins were removed
from the culture medium by adding 80% (w/v) trichloro-
acetic acid (150 mL/L) and subsequent centrifugation at
1
0,000g for 10 min. After discarding the pellet, the EPS
in the supernatant was precipitated with 3 vol in abs
EtOH overnight at ꢀ18 ꢁC, and collected by centrifuga-
tion at 10,000g for 10 min. A soln of the pellet in 40 mL
hot distilled water (90 ꢁC) was extensively dialyzed for
were reduced with NaBD . Conventional work-up,
4
7
2 h against distilled water at 4 ꢁC, then lyophilized.
comprising neutralization and removal of boric acid
by co-evaporation with MeOH, followed by acetylation
The freeze-dried sample was redissolved in 50 mM so-
dium phosphate buffer, pH 6.0, and an aliquot (1 mL;
with 1:1 pyridine–Ac O (30 min, 120 ꢁC) yielded
2
3
–10 mg of carbohydrate) was applied to a C16/20 col-
mixtures of partially methylated alditol acetates, which
2
2
umn (Pharmacia Biotech, Uppsala, Sweden), packed
with the weakly basic anion exchanger DEAE-Trisacryl
Plus M (Sigma Aldrich Chemie GmbH, Taufkirchen,
Germany). The neutral EPS was eluted with 50 mM so-
dium phosphate buffer, pH 6.0 (40 mL), at a flow rate of
were analyzed by GLC–MS.
3.5. Gas–liquid chromatography and mass spectrometry
Quantitative GLC analyses were performed on a
Chrompack CP9002 gas chromatograph, equipped with
an EC-1 column (30 m · 0.32 mm, Alltech, Deerfield,
IL) using a temperature program of 140–240 ꢁC at
0
.5 mL/min, monitored at 280 nm with a UV-1 detector
(
Pharmacia Fine Chemicals, Uppsala, Sweden). The
fractions containing the neutral EPS were pooled, de-
salted by dialysis against distilled water for 48 h at
4
ꢁC/min and flame-ionization detection. GLC–MS
analyses were carried out on a GC8060/MD800 system
Fisons instruments, Interscience; 70 eV), using an
AT-1 column (30 m · 0.25 mm, Alltech) at the same
4
ꢁC, and lyophilized.
(
3
.2. Molecular mass determination
2
2
temperature program.
The average molecular mass of the EPS was determined
by gel filtration chromatography on a Sephacryl S-400
HR C16/100 column (Amersham Pharmacia Biotech,
Uppsala, Sweden), calibrated with dextran standards
Matrix-assisted laser desorption ionization time-of-
flight mass spectrometry (MALDI-TOF-MS) experi-
ments were performed using a Voyager-DE PRO mass
spectrometer (Applied Biosystems, Foster City, CA)
equipped with a nitrogen laser (337 nm, 3 ns pulse
width). Positive-ion mode spectra were recorded using
the reflectron mode and delayed extraction (100 ns).
The accelerating voltage was 20 kV with a grid voltage
of 75.2%; the mirror voltage ratio was 1.12, and the
acquisition mass range 500–3000 Da. Samples were
prepared by mixing on the target 1 lL oligosaccha-
ride soln with 1 lL of 2,5-dihydroxybenzoic acid
(
M 1800, 750, 410, 150, 50, and 25 kDa; Fluka Chemie
w
GmbH, Buchs, Switzerland), using 50 mM phosphate
buffer, pH 6.0, containing 150 mM NaCl as eluent.
The flow rate was 0.2 mL/min and the fraction size
2
mL. The carbohydrate content of each fraction was
2
1
determined by the phenol–sulfuric acid assay.
3
.3. Monosaccharide analysis
(10 mg/mL) in 50% aqueous acetonitrile as matrix
soln.
Oligo/polysaccharide was subjected to methanolysis
methanolic 1 M HCl; 18 h, 85 ꢁC). The resulting
mixtures of methyl glycosides were trimethylsilylated
1:1:5 hexamethyldisilazane–trimethylchlorosilane–pyri-
dine; 30 min, room temperature), then quantitatively
(
3.6. Periodate oxidation
(
To a soln of polysaccharide (30 mg) in 0.1 M NaOAc
buffer (35 mL; pH 3.9), NaIO was added to a final
2
2
analyzed by GLC as described. In addition, the
absolute configurations of the monosaccharides were
determined by GLC analysis of the trimethylsilylated
4
concentration of 50 mM, and the mixture was kept in
the dark for 5 days at 4 ꢁC. Excess of periodate was
destroyed by the addition of ethylene glycol (2 mL) after
which the soln was dialyzed against tap water. Then,
the oxidized polysaccharide was reduced with NaBH4
(2 h; 20 ꢁC), neutralized with 4 M HOAc, dialyzed
2
3,24
(
ꢀ)-2-butyl glycosides.
For both analyses, the
identities of the monosaccharides were confirmed by
gas–liquid chromatography/mass spectrometry (GLC–
2
2
MS).

2744
I. S a´ nchez-Medina et al. / Carbohydrate Research 342 (2007) 2735–2744
against tap water, and lyophilized. The obtained mate-
rial was subsequently hydrolyzed (0.5 M TFA; 30 min,
References
1
2
3
4
5
6
. De Vuyst, L.; Degeest, B. FEMS Microbiol. Rev. 1999, 23,
8
0 ꢁC), and the progress of the hydrolysis was moni-
1
53–177.
. Laws, A. P.; Gu, Y.; Marshall, V. M. Biotechnol. Adv.
001, 19, 597–625.
. Uemura, J.; Itoh, T.; Kaneko, T.; Noda, K. Milchwis-
senschaft 1998, 53, 443–446.
. Grobben, G. J.; Smith, M. R.; Sikkema, J.; de Bont, J. A.
M. Appl. Microbiol. Biotechnol. 1996, 46, 279–284.
. Faber, E. J.; Kamerling, J. P.; Vliegenthart, J. F. G.
Carbohydr. Res. 2001, 331, 183–194.
. Harding, L. P.; Marshall, V. M.; Hernandez, Y.; Gu, Y.;
Maqsood, M.; McLay, N.; Laws, A. P. Carbohydr. Res.
tored by TLC (Merck Kieselgel 60 F254 sheets; 2:1:1
n-butanol–acetic acid–water; orcinol/sulfuric acid stain-
ing). After concentration, the residue was fractionated
on a Bio-Gel P-2 column (60 · 1.5 cm), eluted with
2
1
0 mM NH HCO at a flow rate of 3 mL/min at room
4 3
temperature, while monitored by differential refraction
index detection. The sugar-containing fraction was
lyophilized. A part of the material was directly used
for structural analysis. The other part was reduced with
2
005, 340, 1107–1111.
NaBH (2 h; 20 ꢁC), neutralized with 4 M HOAc, and
4
7
8
9
. Cerning, J.; Bouillanne, C.; Desmazeaud, M. J.; Landon,
M. Biotechnol. Lett. 1986, 8, 625–628.
. Grobben, G. J.; Sikkema, J.; Smith, M. R.; de Bont, J. A.
M. J. Appl. Bacteriol. 1995, 79, 103–107.
. Gruter, M.; Leeflang, B. R.; Kuiper, J.; Kamerling, J. P.;
Vliegenthart, J. F. G. Carbohydr. Res. 1993, 239, 209–226.
purified by Bio-Gel P-2.
3
.7. NMR spectroscopy
Resolution-enhanced 1D/2D 500-MHz NMR spectra
1
0. Harding, L. P.; Marshall, V. M.; Elvin, M.; Gu, Y.; Laws,
were recorded in D O on a Bruker DRX-500 spectro-
A. P. Carbohydr. Res. 2003, 338, 61–67.
2
meter (Bijvoet Center, Department of NMR Spectro-
11. Chaplin, M. F.; Kennedy, J. F. Carbohydrate Analysis, a
Practical Approach; IRL Press: Oxford, 1986, pp 71–75
and references cited therein.
scopy) at
oligosaccharides and 78 ꢁC for the polysaccharide. Prior
to analysis, samples were exchanged twice in D O
a
probe temperature of 27 ꢁC for
1
2. Faber, E. J.; van den Haak, M. J.; Kamerling, J. P.;
2
Vliegenthart, J. F. G. Carbohydr. Res. 2001, 331, 173–182.
(
99.9 at % D, Cambridge Isotope Laboratories, Inc.,
13. Bock, K.; Pedersen, C. J. Chem. Soc., Perkin Trans. 2
1974, 293–297.
Andover, MA) with intermediate lyophilization, and
then dissolved in 0.6 mL D O. Chemical shifts are
expressed in parts per million by reference to internal
acetone (d 2.225) for H and/or to the a-anomeric signal
1
4. Bock, K.; Pedersen, C. Adv. Carbohydr. Chem. Biochem.
983, 41, 27–66.
2
1
1
5. Faber, E. J.; Zoon, P.; Kamerling, J. P.; Vliegenthart, J. F.
1
G. Carbohydr. Res. 1998, 310, 269–276.
16. Van Casteren, W. H. M.; Dijkema, C.; Schols, H. A.;
1
3
13
of external [1- C]glucose (d
92.9) for C. Suppres-
C-1
sion of the HOD signal was achieved by applying a
Beldman, G.; Voragen, A. G. J. Carbohydr. Res. 2000,
2
6
3
24, 170–181.
7. Bock, K.; Thøgersen, H. Annu. Rep. NMR Spectrosc.
982, 13, 1–57.
WEFT pulse sequence for 1D experiments and by a
pre-saturation of 1 s during the relaxation delay for
1
1
1
2
7
2
D experiments. 2D TOCSY spectra were recorded
8. Bubb, W. A.; Urashima, T.; Fujiwara, R.; Shinnai, T.;
Ariga, H. Carbohydr. Res. 1997, 301, 41–50.
19. Lemoine, J.; Chirat, F.; Wieruszeski, J.-M.; Strecker, G.;
2
8
using an MLEV-17 mixing sequence with spin-lock
times of 40–100 ms. 2D ROESY experiments were
performed at a mixing time of 300 ms for the oligosac-
charide; 2D NOESY experiments were performed with
a mixing time of 150 ms for the polysaccharide. Natural
Favre, N.; Neeser, J.-R. Appl. Environ. Microbiol. 1997,
6
3, 3512–3518.
2
0. Low, D.; Ahlgren, J. A.; Horne, D.; McManon, D. J.;
Oberg, C. J.; Broadbent, J. R. Appl. Environm. Microbiol.
1998, 64, 2147–2151.
1
13
abundance 2D H– C HSQC and HMBC experiments
were recorded with and without decoupling, respec-
21. Dubois, M.; Gilles, K. A.; Hamilton, J. K.; Rebers, P. A.;
1
Smith, F. Anal. Chem. 1956, 28, 350–356.
tively, during acquisition of the H FID. Resolution
2
2. Kamerling, J. P.; Vliegenthart, J. F. G. Carbohydrates. In
Mass Spectrometry; Lawson, A. M., Ed.; Clinical Bio-
chemistry—Principles, Methods, Applications; Walter de
Gruyter: Berlin, 1989; Vol. 1, pp 175–263.
enhancement of the spectra was performed by a
Lorentzian-to-Gaussian transformation or by multipli-
cation with a squared-bell function phase shifted by
p/(2.3) for 2D spectra, and when necessary, a fifth order
polynomial baseline correction was performed. All
NMR data were processed using in-house developed
software (J. A. van Kuik, Bijvoet Center, Utrecht
University).
23. Gerwig, G. J.; Kamerling, J. P.; Vliegenthart, J. F. G.
Carbohydr Res. 1978, 62, 349–357.
4. Gerwig, G. J.; Kamerling, J. P.; Vliegenthart, J. F. G.
Carbohydr. Res. 1979, 77, 1–7.
5. Ciucanu, I.; Kerek, F. Carbohydr. Res. 1984, 131, 209–217.
26. H a˚ rd, K.; van Zadelhoff, G.; Moonen, P.; Kamerling, J. P.;
2
2
Vliegenthart, J. F. G. Eur. J. Biochem. 1992, 209, 895–915.
2
7. H a˚ rd, K.; Vliegenthart, J. F. G. Nuclear Magnetic
Resonance Spectroscopy of Glycoprotein-Derived Carbo-
hydrate Chains. In Glycobiology, a Practical Approach;
Fukuda, M., Kobata, A., Eds.; Oxford University Press:
Oxford, 1993; pp 223–242.
Acknowledgment
The authors thank Dr. Bas R. Leeflang for assistance
with the use of the NMR spectrometer.
28. Bax, A.; Davies, D. G. J. Magn. Reson. 1985, 65, 355–360.