The extracellular polysaccharide of Pichia (Hansenula) holstii NRRLL Y-2448: The phosphorylated side chains

Article abstract of DOI:10.1016/S0008-6215(98)00101-3

The exopolysaccharide produced by Pichia (Hansenula) holstii NRRL Y-2448 is composed of a phosphomannan core to which oligosaccharide diester phosphate side chains are appended. The oligosaccharides of the side chains were released as oligosaccharide phosphates and neutral oligosaccharides by mild hydrolysis with aqueous acetic acid and aqueous hydrogen fluoride, respectively. The liberated oligosaccharide phosphates were studied by NMR spectroscopy and by electrospray and fast atom bombardment mass spectrometry. The structures of the neutral oligosaccharides were determined by 1D and 2D NMR spectroscopic experiments. Further insight into the length of the side chains was obtained from a matrix assisted laser desorption ionisation-time of night mass spectrometric study of high and low molecular weight fragments obtained from partial acid hydrolysis of the native polysaccharide.

Full text of DOI:10.1016/S0008-6215(98)00101-3

Carbohydrate Research 309 (1998) 77±87
The extracellular polysaccharide of Pichia
(Hansenula) holstii NRRL Y-2448:
the phosphorylated side chains
Lesley A.S. Parolisa, Haralambos Parolisa,*, Lennart Kenneb,
Morten Meldalc, Klaus Bockc
a School of Pharmaceutical Sciences, Rhodes University, Grahamstown 6140, South Africa
b Department of Chemistry, Swedish University of Agricultural Sciences, S-750 07 Uppsala, Sweden
c Department of Chemistry, Carlsberg Laboratory, Gamle Carlsbergvej 10, 2500 Valby, Denmark
Received 6 January 1998; accepted 4 April 1998
Abstract
The exopolysaccharide produced by Pichia (Hansenula) holstii NRRL Y-2448 is composed of a
phosphomannan core to which oligosaccharide diester phosphate side chains are appended. The
oligosaccharides of the side chains were released as oligosaccharide phosphates and neutral oli-
gosaccharides by mild hydrolysis with aqueous acetic acid and aqueous hydrogen ¯uoride,
respectively. The liberated oligosaccharide phosphates were studied by NMR spectroscopy and by
electrospray and fast atom bombardment mass spectrometry. The structures of the neutral oligo-
saccharides were determined by 1D and 2D NMR spectroscopic experiments. Further insight into
the length of the side chains was obtained from a matrix assisted laser desorption ionisation±time
of ¯ight mass spectrometric study of high and low molecular weight fragments obtained from
partial acid hydrolysis of the native polysaccharide. # 1998 Elsevier Science Ltd. All rights
reserved
Keywords: Pichia holstii; Exopolysaccharide; Phosphomannan; Oligosaccharide side chains; NMR
spectroscopy
1. Introduction
Mild hydrolysis of the PS produces a high mol-
ecular weight phosphomannan core, PC [2±4], and
a low molecular weight oligosaccharide phosphate
fraction. Both the PC and the oligosaccharide
phosphate fraction have become valuable tools in
studies on mannose-6-phosphate receptors. The
structure of the major repeating unit of the PC has
recently been shown [5] to be the branched hexa-
saccharide phosphate.
The yeast Pichia (Hansenula) holstii NRRL Y-2448
produces a viscous, highly branched extracellular
phosphomannan, PS, when grown in a culture
medium containing an excess of orthophosphate [1].
* Corresponding author. Fax: 00 2746 1311205.
0008-6215/98/$19.00 # 1998 Elsevier Science Ltd. All rights reserved
PII S0008-6215(98)00101-3

78
L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
The low molecular weight fraction, which
accounts for approximately 90% of the PS, is
derived from the phosphorylated side chain oligo-
saccharides attached to the PC. The major oligo-
saccharide present in this fraction [3,4], and thus
the main repeating unit found in the oligosacchar-
ide side chains, is the pentasaccharide phosphate 1.
The length of the oligosaccharide diesterphosphate
side chains of the PS and the nature of the units ter-
minating these side chains have not been established,
although weak binding of the PS to mannose-6-
phosphate receptors and the presence of a signal for
diester phosphate only in the 31P spectrum of the
PS [5] indicate that units capping the side chains
are devoid of monoester phosphate. Furthermore,
the structural signi®cance of minor amounts of tet-
rasaccharide and hexasaccharide phosphate, repor-
ted to be present in low molecular weight
preparations from partial hydrolysates [6], remains
unknown.
The extensive use of low molecular weight oli-
gosaccharide preparations, assumed to consist of
oligosaccharide 1, in research on mannose-6-phos-
phate receptors [7±13] prompted us to seek answers
to these questions. In the present paper we report
on further structural studies on the oligosaccharide
side chains of the PS.
NMR spectroscopy.ÐSamples were deuterium
exchanged by lyophilising solutions in D2O, and
then dissolved in 99.99% D2O (0.45 mL). Spectra
1
were referenced relative to acetone (ꢀ 2.230 for H
and 31.07 for 13C) and to 85% H3PO4 (ꢀ 0.0 for
31P). Spectra were recorded at either 27 or 30 ꢀC on
either Bruker AM-500, AMX-400 or AMX-600
spectrometers. The 2D NMR pulse programs used
were (a) phase sensitive COSY [14] using TPPI
with double quantum ®lter and presaturation dur-
ing relaxation delay; (b) NOESY [15] using TPPI
with presaturation during relaxation delay and a
mixing delay of 800 ms; (c) HOHAHA [16] using
TPPI with presaturation during relaxation delay and
employing the MLEV-17 sequence for mixing, with
a mixing delay of 89 ms; (d) HMQC [17] and
HMQC-TOCSY [18] using TPPI with presaturation
during relaxation delay and GARP decoupling dur-
ing acquisition; (e) HMBC [19] with no decoupling
during acquisition and a mixing delay of 60 ms.
Preparation of oligosaccharide phosphates.Ð
Small-scale hydrolysis of the PS (17 mg) in 1%
ꢀ
2. Experimental
HOAc at 100 C for 8 h, as described previously
[5], followed by GPC of the products on Bio-Gel P-
4 gave the elution pro®le shown in Fig. 1. The
yields obtained were A 11.7, B 3.0, C 0.7, and D
1.6 mg. The oligosaccharide phosphates previously
obtained from a large-scale hydrolysis of the PS [5]
were also used in this study.
General methods.ÐGel permeation chromato-
graphy (GPC) was performed on calibrated col-
umns of Bio-Gel P-2 and P-4 (1.6Â94 cm) using
0.1 M NH4HCO3 (pH 8.3) as eluent. Compounds
were detected by refractive index.

L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
79
Preparation of samples of the PS for MALDI-
TOF-MS analysis.ÐThe PS (30 mg) in aq 1%
HOAc was heated at 100 ꢀC for 2 h after which the
sample was cooled, the solvent was evaporated o€,
and the mixture was separated into a low
molecular weight fraction O (containing oligo-
saccharides) and a high molecular weight fraction
P (containing mostly polymeric material). A
sample of P was retained for MS analysis and
the remainder was hydrolysed as before with aq
1% HOAc and again separated by GPC into a
high and low molecular weight fraction. The
process was repeated a further two times yielding
in all four high and four low molecular weight
fractions which were analysed by MALDI-TOF-
MS.
Fig. 1. GPC of the hydrolysate of the PS on Bio-Gel P-4.
Mass spectrometry.ÐMALDI-TOF spectra were
obtained on a LINEAR LDI-1700XS spectrometer
in matrices of 2,4,6-trihydroxyacetophenone.
ESMS spectra were acquired in the negative mode
on a VG Quattro triple quadrupole mass spectro-
meter with aq 50% CH3CN as the mobile phase at
a ¯ow rate of 8 mL/min. Samples were prepared at
concentrations of ꢁ20 pmol/mL and 10 mL of the
sample were injected. FABMS spectra were recor-
ded on either acetylated or trideuteroacetylated
samples with a VG Analytical High Field ZAB-1F
mass spectrometer ®tted with an M-Scan FAB gun,
as previously described [20].
Aqueous HF hydrolysis of the PS.ÐThe PS
ꢀ
(200 mg) was mixed with aq 48% HF at 0 C and
was maintained at this temperature until the PS
had dissolved whereafter the mixture was placed in
a cold room at 5 ꢀC. After 96 h the hydrolysate was
poured into a slurry of CaCO3 and solid CO2 in
CH2Cl2. The mixture was extracted with H2O (4Â)
to remove the carbohydrate material (104 mg).
Separation of this material on Bio-Gel P-2 gave the
pro®le shown in Fig. 2. The fractions were col-
lected and rechromatographed on Bio-Gel P-2 to
a€ord H (1 mg), I (2.3 mg), J (6 mg), K (59 mg), L
(10 mg), and M (9 mg). Fractions were examined
by 1H and 13C NMR spectroscopy.
3. Results and discussion
Hydrolysis of the PS with aq 1% HOAc at
ꢀ
100 C as previously described [5], followed by
GPC separation of the hydrolysate on Bio-Gel P-4
(Fig. 1), a€orded three oligosaccharide phosphate
fractions, A, B, and C, in addition to the PC frac-
tion, D. Fraction A was rechromatographed on
Bioꢀ-Gel P-2 and was then examined in D2O at
30 C by NMR spectroscopy (pD 6.05). The 31P
NMR spectrum of A showed a single resonance at
1
ꢀ 1.49 for a monophosphate ester. The H and 13C
NMR spectra (Figs. 3 and 4, respectively) clearly
identify the main component of A as a penta-
1
saccharide. The H and 13C NMR assignments for
the component residues a, b, c, d, and e (Table 1) of
the pentasaccharide were determined from COSY-
DQF [14], HOHAHA [16], and HMQC [17] 2D
NMR experiments. The sequence of the residues in
the pentasaccharide was established from a 2D
NOESY [15] experiment as a±(1-3)±c±(1-3)±e±(1-
Fig. 2. GPC of the HF hydrolysate of the PS on Bio-Gel P-2.

80
L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
Fig. 3. 1H NMR spectrum of fraction A.
Fig. 4. 13C NMR spectrum of fraction A.
3)±d±(1-2)±b. The location of the monophosphate
ester group was established at C-6 of the non-
reducing terminal residue a. This assignment fol-
lows from the down®eld location of the signal
assigned to C-6 of a, the up®eld location of C-5 of
a, the two- and three-bond 13C±31P coupling con-
stants for C-6 and C-5, respectively, and the
deshielding of the H-6a and H-6b resonances of a.
The above data establish the main component of A
as the pentasaccharide phosphate 1.

L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
81
Table 1
1H and 13C chemical shift data a for P. holstii pentasaccharide phosphate A (major anomeric component)
Residue
H-1/C-1
H-2/C-2
H-3/C-3
H-4/C-4
H-5/C-5
H-6a,b/C-6
!2)-ꢁ-ManꢁOH
b
b
a
c
e
d
5.372 (1.7) b
93.3
3.946
79.9
3.931
70.7
3.691
67.8
3.793
73.2
3.890, 3.762
61.8
!2)-ꢂ-ManꢁOH
ꢁ-Man-6-PO4
!3)-ꢁ-Man
4.905
94.3
4.070
78.6
3.751
73.2
3.598
67.4
3.389
77.5
3.901, 3.711
61.7
5.082 (1.7)
103.6
4.069 (3.4)
70.8
3.890 (10)
71.1
3.684
67.5
3.942
73.3
4.168, 3.966
65.0
5.115 (1.8)
103.3
4.329 (3.3)
70.2
3.979 (10)
80.1
3.714
66.7
3.834
74.2
3.903, 3.740
61.9
!3)-ꢁ-Man
5.114 (1.8)
102.8
4.203 (3.2)
70.6
3.984 (9)
79.1
3.781
66.7
3.769
74.1
3.861, 3.746
61.8
!3)-ꢁ-Man
5.021 (1.8)
102.8
4.221 (3.3)
70.3
3.948 (9)
78.7
3.739
67.0
3.792
74.4
3.890, 3.762
61.8
ꢀ
a Chemical shifts in ppm relative to acetone at ꢀ 2.230 for 1H and 31.07 for 13C, at 27 C.
b Coupling constants in Hz.
Linkage carbons are in bold.
Table 2
Negative ion ESMS of fractions from GPC of a partial hydrolysate of PS
Fraction
m/z
m/z
m/z
m/z
m/z
A
1069.2 (P-Man6)
979.4 (P2-Man11)
949.5 (P3-Man16)
907.2 (P-Man5)
745.2 (P-Man4)
817.6 (P2-Man9)
583.4 (P-Man3)
736.4 (P2-Man8)
421.2 (P-Man2)
655.5 (P2-Man7)
B
898.4 (P2-Man10)
C1
895.9 (P3-Man15)
893.5 a (P4-Man20)
841.8 (P3-Man14)
853.7 (P4-Man19)
787.5 (P3-Man13)
813.5 (P4-Man18)
733.2 (P3-Man12)
772.4 (P4-Man17)
C2
893.8 (P4-Man20)
892.6 a (P5-Man25)
892.0 a (P6-Man30)
891.6 a (P7-Man35)
853.7 (P4-Man19)
860.9 (P5-Man24)
865.9 (P6-Man29)
869.2 (P7-Man34)
812.9 (P4-Man18)
828.6 (P5-Man23)
838.9 (P6-Man28)
846.2 (P7-Man33)
796.2 (P5-Man22)
ꢁ811 (P6-Man27)
a Anticipated ions not unambiguously identi®ed.
All of the NMR spectra of A showed several
minor resonances in addition to those attributed to
1 indicating additional minor components in A. In
order to establish the nature of the minor oligo-
saccharide components, A was investigated by
electrospray mass spectrometry (ESMS). Table 2
lists the [M H] ions observed in the negative ion
spectrum of A. In addition to the anticipated major
[M H] ion at m/z 907.2 for the pentasaccharide
phosphate 1, peaks for the [M H] ions for a tet-
rasaccharide phosphate (m/z 745.2), a trisaccharide
phosphate (m/z 583.4), a disaccharide phosphate
(m/z 421.2), and a hexasaccharide phosphate (m/z
1069.2) were observed. The ratio of the intensities
of the ions at m/z 745.2 and 907.2 suggested that
the tetrasaccharide phosphate constituted 20±25%
of A. The other three oligosaccharide phosphates
appeared to be present in very small amounts. The
proportion of the tetrasaccharide phosphate was
greater than was anticipated from the earlier report
mentioned above [6]. We were therefore concerned
that, although ES is a soft ionisation technique
that would not be expected to cause glycosidic
cleavage, the tetrasaccharide phosphate and the
previously unreported tri- and disaccharide phos-
phates may have resulted from such cleavage of the
pentasaccharide phosphate. The ESMS experi-
ments were therefore repeated at lower ionisation
energies and the ratio of the ions at m/z 745.2 and
907.2 was monitored. As there was no signi®cant
variation in the ratio of these ions in all of the
experiments performed it was concluded that the
tetrasaccharide phosphate was a genuine compo-
nent of A and had not resulted from the loss of
hexose from the pentasaccharide phosphate. The
most signi®cant change in the ESMS spectra was
the dramatic increase in the [M 2H]2 ion at m/z
453.4 as the ionisation energy was decreased.

82
L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
Indeed it was the base peak in the experiment with
the lowest ionisation energy. Further proof that the
tetra- and lower oligosaccharide phosphates were
genuine components of A was obtained from
FABMS experiments on trideuteroacetylated A
[20]. The major molecular ions in the positive ion
spectrum occurred at m/z 1691 and 1394. These
represent the [M+NH4]+ ions for the penta-
saccharide and tetrasaccharide phosphate, respec-
tively, in which the phosphate groups are
monodeuteroacetylated. The respective [M+Na]+
ions occurred at m/z 1696 and 1399, respectively.
Smaller [M+NH4]+ and [M+Na]+ ions occurred
at m/z 1097 and 1102, respectively, con®rming the
presence of trisaccharide phosphate in A. No
molecular ions were observed for a disaccharide
phosphate in the positive ion FAB spectra. Very
small [M+Na]+ ions were observed at m/z 1993,
2290, 2587, 2884, and 3181 indicating the presence
of hexa-, hepta- octa-, nona-, and decasaccharide
phosphates in fraction A. The negative ion FAB
spectra showed major pairs of [M H] ions in
which the phosphate groups were monodeutero-
acetylated and nondeuteroacetylated at m/z 1673
and 1628, and 1376 and 1331 for a pentasaccharide
phosphate and tetrasaccharide phosphate,
a
respectively. In addition similar pairs of [M H]
ions, but of very low intensity, were observed for
di-, tri-, hexa-, hepta-, octa-, nona- and dec-
asaccharide phosphates. Both the positive and
negative ion FAB spectra suggested a ratio of 1:4
for the tetrasaccharide phosphate and penta-
saccharide phosphate in A.
The oligosaccharide fraction B was rechromato-
graphed on Bio-Gel P-4 and was then examined by
NMR spectroscopy in D2O at 30 ꢀC (pD 5.68). The
proton decoupled 31P spectrum of B showed signals
at ꢀ 1.52 and 1.29 for a mono- and diester phos-
phate, respectively. The 1H NMR spectrum (Fig. 5)
showed H-1 resonances at ꢀ 5.682 (1 H, J1,2 1.3, JH-P
7.9 Hz), 5.385 (0.9 H, J1,2 1.5 Hz, ꢁ-anomer of
reducing Man residue), 5.132 (2 H), 5.126 (1 H),
5.121 (1 H), 5.109 (1 H, J1,2 1.8 Hz), 5.103 (1 H, J1,2
1.8 Hz), 5.064 (1 H, J1,2 1.8 Hz), 5.037 (1 H, J1,2
1.8 Hz), and 4.930 (0.1 H, ꢂ-anomer of reducing
Man residue). These results together with the NMR
data for the pentasaccharide phosphate (P-Man5)
identify B as the decasaccharide diphosphate,
Fig. 5. 1H NMR spectrum of fraction B.

L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
83
Fig. 6. 13C DEPT spectrum of fraction B.
P-Man5±P-Man5, in which the reducing terminal
Man of one P-Man5 unit is linked through an ꢁ-
glycosidic bond to the 6-phosphate group of
another P-Man5 unit. The 13C DEPT spectrum of
B (Fig. 6), which shows, inter alia, signals at
95.5 ppm (2JC-P 5.3 Hz) for C-1 of the Man residue
involved in the glycosyl phosphate linkage, 66.0
(2JC-P 4.3 Hz) and 65.0 ppm (2JC-P 4.3 Hz) for two
phosphorylated C-6, 73.1 (3JC-P 7.2 Hz) and 73.3 ppm
(3JC-P 7.6 Hz) for two C-5, 79.0 ppm (3JC-P 8.1 Hz)
for C-2 of the 2-substituted Man residue involved
in the glycosylphosphate linkage, is consistent with
the suggested decasaccharide diphosphate.
Examination of fraction B by ESMS showed the
peaks listed in Table 2. The major peaks at m/z
898.4 and 817.6 represent the [M 2H]2 ions of
the P-Man5±P-Man5 and P-Man4±P-Man5 (or P-
Man5±P-Man4) oligosaccharides, respectively.
These ions occurred in the approximate ratio of
1.5:1 which is consistent with the approximate 4:1
ratio of pentasaccharide phosphate to tetra-
saccharide phosphate found in fraction A. The
much smaller peaks at m/z 979.4, 736.4, and 655.5
represent oligosaccharides with the composition
shown in parentheses in Table 2. The negative ion
FAB spectrum of acetylated B con®rmed the
ESMS results showing prominent [M H] ions at
m/z 3099.8 and 2811.5 for P2-Man10 and P2-Man9,
respectively, and very small peaks for P2-Man7
(m/z 2233.9), P2-Man8 (m/z 2523.5), and P2-Man11
(m/z 3387.5). In addition minute [M H] ions
were observed for P2-Mann where n=12±15. All
the [M H] ions above were accompanied by ions
which were 42 amu higher and which corresponded
to monoacetylation of the monoester phosphate
groups.
The oligosaccharide material in fraction C
(Fig. 1) was collected in two parts designated C2,
which contained the material collected from the
leading edge of the peak to the middle of the peak,
and C1, which contained the material collected
from the middle to the trailing edge of the peak.
1H NMR examination of C1 and C2 indicated
that they consisted of higher multiples of the
pentasaccharide phosphate. The spectrum of C1
showed a signal at ꢀ 5.683 (3JH-P 7.3 Hz) for H-1 of
the Man involved in the glycosyl phosphate link-
age, and at ꢀ 5.387 (ꢁ) and 4.934 (ꢂ) for H-1 of the
reducing Man. The integration of the signals was
consistent with a pentadecasaccharide (P3-Man15).
Examination of fraction C1 by ESMS identi®ed the
ions listed in Table 2. The major peaks at m/z 895.9
and 841.8 corresponded to the [M 3H]3 ions for
the oligosaccharides P3-Man15 and P3-Man14,

84
L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
respectively, and were present in an approximate
ratio of 1.9:1. These corresponded to the oligo-
saccharides P-Man5±P-Man5±P-Man5 and P-
Man4±P-Man5±P-Man5 (and/or the other two
variations). The other [M 3H]3 ions (Table 2)
found in C1 were present in much smaller amounts,
as were the [M 4H]4 ions. It was not possible to
identify an [M 4H]4 ion for the oligosaccharide
P4-Man20 (m/z 893.5) as such a peak would have
been swamped by the major ion at m/z 895.9. The
negative ion FAB spectrum of acetylated C1 con-
®rmed the ESMS results showing major [M H]
ions at m/z 4621.5 and 4333.1 for P3-Man15 and
P3-Man14, respectively. The spectrum also showed
a small peak at m/z 6143 con®rming the suspected
presence of P4-Man20 in C1.
for P5-Man23, P5-Man24, P6-Man28, P6-Man29, P7-
Man33, and P7-Man34. The negative ion FAB
spectrum of acetylated C2 identi®ed P4-Man20 (m/z
6143) and P4-Man19 (m/z 5855) as the most pro-
minent oligosaccharides in the mixture and also
proved conclusively the presence of P5-Man25.
Oligosaccharides of higher molecular masses could
not be detected by FABMS.
The MS data obtained on fractions A±C indicate
quite clearly that the separation of the components
of the partial hydrolysate by GPC was strongly
in¯uenced by the number of phosphate ester
groups present in the oligosaccharides. It was thus
not possible to separate the mixture of oligo-
saccharide phosphates present in fraction A by
GPC for further structural studies due to the
retarding in¯uence of the phosphate group on the
separation. The FABMS results revealed that all
the oligosaccharides with a monoester phosphate
group were eluted together irrespective of their
molecular weights. In order, therefore, to separate
and determine the structures of the tetra and other
oligosaccharides detected by the MS studies, the
PS was treated with aqueous 48% HF at 0 ꢀC and the
neutral oligosaccharides produced were separated
1
The H NMR spectrum of C2 showed signals as
for C1 at ꢀ 5.683 (3JH-P 7.3 Hz) for H-1 of the Man
involved in the glycosyl phosphate linkage, and at ꢀ
5.387 (ꢁ) and 4.934 (ꢂ) for H-1 of the reducing
Man. The ions observed in the ESMS spectrum of
C2 are listed in Table 2 together with the composi-
tion of the oligosaccharides from which they arise.
The major peak at m/z 893.9 is consistent with the
[M 4H]4 ion for (P-Man5)4. The peak was
approximately 6 amu in width at the base and cov-
ered the range 892.4±895.4 amu at half height.
Thus the anticipated ions for [M 5H]5 at m/z
892.6 for P5-Man25, for [M 6H]6 at m/z 892.0 for
P6-Man30, and for [M 7H]7 at m/z 891.6 for P7-
Man35, could not be distinguished. Their presence
was suggested by the appearance of prominent ions
1
by GPC on Bio-Gel P-2 (Fig. 2). The H NMR
spectra of the isolated oligosaccharides, designated
H, I, J, and K, are shown in Fig. 7 while the NMR
data are collected in Table 3. The structures of the
oligosaccharides were established by 2D NMR
spectroscopy (phase sensitive COSY, HOHAHA,
HMQC, and NOESY and/or HMBC) as

L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
85
The residues a±e are labelled in the same manner
as for the pentasaccharide phosphate 1.
In order to probe further the composition and
length of the phosphorylated side chains attached
to the PC, the PS was hydrolysed with aqueous
1% CH3CO2H at 100 C for 2 h and the mixture
obtained was separated on Bio-Gel P-4 into an
1H and 13C NMR spectra of fraction L (Fig. 2)
indicated it to be primarily a mixture of the
decasaccharide diester phosphate ꢁ-d-Man-(1!3)-
ꢁ-d-Man-(1!3)-ꢁ-d-Man-(1!3)-ꢁ-d-Man-(1!2)-
ꢁ-d-Man-1-O-PO2H-O-6-ꢁ-d-Man-(1!3)-ꢁ-d-
Man-(1!3)-ꢁ-d-Man-(1!3)-ꢁ-d-Man-(1!2)-d-
Man and the corresponding nonasaccharide diester
ꢀ
oligosaccharide fraction (O), and
a
poly-
saccharide fraction (P), most of which eluted in
the void volume. The two fractions were then
examined by MALDI-TOF-MS. The higher
molecular weight fraction P showed clusters of
ions for P-Man4±6, P2-Man8±11, P3-Man13±16, P4-
1
phosphate. The H NMR spectrum of L showed a
signal at ꢀ 5.677 (JH-P 7.7 Hz) for H-1 of an ꢁ-
mannosyl diester phosphate linkage, a H-1 signal
(0.9 H) for a reducing ꢁ-mannopyranosyl residue
at ꢀ 5.382 and a H-1 signal (0.1 H) for a ꢂ-manno-
pyranosyl residue at ꢀ 4.930. The 13C NMR spec-
trum of L showed a C-1 signal at 95.5 ppm (2JC-P
5.7 Hz), a signal for C-6 at 66.1 ppm (2JC-P 3.8 Hz),
and a signal for C-5 at 73.1 ppm (3JC-P 8.1 Hz). The
1H NMR spectrum of fraction M (Fig. 7) sug-
gested that this contained the PC [5] with residual
diester phosphate side chains.
Man18±21, P5-Man23±26, P6-Man28±31, P7-Man33±36
,
and probably for P8-Man38±41, P9-Man43±46, and
P10-Man48±51. While the molecular masses of the
latter two clusters of ions could not be deter-
mined accurately, ions with m/z ꢁ8000 and 8800
were de®nitely present. The lower molecular
weight fraction O showed prominent ions for P-
Man4 and P-Man5 and ions of much lower
intensity for P-Man2±3 and P-Man6±10. Prominent
ions were also observed for P2-Man9 and P2-
Fig. 7. 1H NMR spectra of the oligosaccharides H±K obtained from HF hydrolysis of the PS.

86
L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
Table 3
1H and 13C chemical shift data a for P. holstii derived oligosaccharides H±K (major anomeric component)
Residue
H-1/C-1
H-2/C-2
H-3/C-3
H-4/C-4
H-5/C-5
H-6a,b/C-6
Disaccharide H
ꢁ-Man
a
b
5.047
103.0
5.380
93.3
4.076
70.8
3.958
79.9
3.856
71.1
3.951
70.8
3.645
67.8
3.670
67.6
3.754
74.0
3.780
73.2
61.8 b
61.7 b
!2)-ꢁ-ManꢁOH
Trisaccharide I
ꢁ-Man
a
b
c
5.148
103.0
5.379
93.3
5.037
102.9
4.072
70.8
3.963
79.9 f
4.229
70.4
3.890
71.1
3.951
70.8
3.968
78.7
3.648
67.8
3.702
67.7
3.758
67.0
3.79
74.1
3.79
73.3
3.79
74.2
61.7 c
61.7 c
61.9
!2)-ꢁ-ManꢁOH
!3)-ꢁ-Man
Tetrasaccharide J
ꢁ-Man
a
b
c
5.141
103.1
5.378
93.3
5.127
102.9
5.039
102.8
4.076
70.8
3.960
79.9
4.230
70.5
4.231
70.4
3.890
71.1
3.949
70.8
4.021
78.9
3.958
78.8
3.645
67.7
3.703
67.8
3.763
67.0 e
3.755
67.0 e
3.795
74.2
3.81
73.3
3.82
74.3
3.78
74.1
61.9
!2)-ꢁ-ManꢁOH
!3)-ꢁ-Man
61.8 d
61.7 d
61.7 d
!3)-ꢁ-Man
d
Pentasaccharide K
ꢁ-Man
a
b
c
d
e
5.143
103.2
5.380
93.4
5.129
102.9
5.039
102.9
5.121
103.1
4.077
70.8
3.954
79.9
4.234
70.5
4.233
70.4
3.889
71.1
3.953
70.8
4.020
78.9
3.964
78.8
3.644
67.7
3.703
67.8
3.762
67.0
3.753
67.0
3.799
74.3
3.809
73.3
3.913, 3.751
61.9
3.852
!2)-ꢁ-ManꢁOH
!3)-ꢁ-Man
61.7
74.4
61.7
3.716
61.7
!3)-ꢁ-Man
74.2
3.780
74.4
!3)-ꢁ-Man
4.237
70.5
4.026
79.0
3.760
67.0
61.7
ꢀ
a Chemical shifts in ppm relative to acetone at ꢀ 2.230 for 1H and 31.07 for 13C, at 30 C.
b,c,d,e Values may have to be interchanged.
f Linkage carbons are in bold.
Man10 and ions of much lower intensity for P2-
Man8 and P2-Man11.
terminal 6-O-phosphate groups since the PS shows
very weak binding to mannose-6-phosphate recep-
tors. An indication of the nature of the units ter-
minating the side chains was forthcoming from a
comparison of the elution volumes of the neutral
oligosaccharides H±K with the minor component
eluted after fraction A (Fig. 1). This comparison
indicated the presence of at least H and K in the
aqueous acetic acid hydrolysate of the PS. As it is
unlikely that the aqueous acetic acid caused
hydrolysis of linkages other than glycosyl phos-
phate, the presence of H and K in the hydro-
lysate of the PS strongly suggests that these are
chain capping units. The presence of oligo-
saccharide residues other than repeating penta-
saccharide monophosphate in the side chains
indicates that, under the stress conditions of
growth in excess orthophosphate, the biosynthetic
control of the assembly of the side chain is not
stringent.
4. Conclusion
The accumulated MS data suggest that the
majority of the chains attached to the core phos-
phomannan PC are composed of up to at least ten
repeating pentasaccharide phosphate residues. In
other chains pentasaccharide phosphate residues
are replaced by a tetrasaccharide phosphate and to
a much lesser extent by hexa-, tri-, and disaccharide
phosphates. The penta- to tetrasaccharide phos-
phate ratio observed in fractions A and B (Fig. 1)
suggests that some chains may contain more than
one tetrasaccharide phosphate residue. The MS
data unfortunately did not furnish de®nitive infor-
mation on the nature of the residues capping the
chains. Such residues are not expected to contain

L.A.S. Parolis et al./Carbohydrate Research 309 (1998) 77±87
87
Acknowledgements
[7] R.J. Youle, G.J. Murray, and D.M. Neville, Jr.,
Proc. Natl. Acad. Sci. USA, 76 (1979) 5559±5562.
[8] G.J. Murray and D.M. Neville, Jr., J. Biol. Chem.,
255 (1980) 11942±11948.
[9] T. Yoshida and Y.C. Lee, Carbohydr. Res., 251
(1994) 175±186.
[10] A.M. Leichtner and M. Krieger, J. Cell Sci., 68
(1984) 183±194.
[11] S.A. Weston and C.R. Parish, J. Immunol., 146
(1991) 4180±4186.
[12] R.W. Loveless, G. Floyd-O'Sullivan, J.G. Raynes,
C.-T. Yuen, and T. Feizi, EMBO J., 11 (1992) 813±
819.
[13] S.A. Weston and C.R. Parish, Eur. J. Immunol., 22
(1992) 1975±1981.
[14] U. Piantini, O.W. Sùrensen, and R.R. Ernst, J.
Am. Chem. Soc., 104 (1982) 6800±6801.
[15] R. Baumann, G. Wider, R.R. Ernst, and K.
Wuthrich, J. Magn. Reson., 44 (1981) 402±406.
[16] A. Bax and D.G. Davis, J. Magn. Reson., 65 (1985)
355±360.
The authors thank the Foundation for Research
Development (Pretoria) for a sabbatical grant and for
continued ®nancial support to H.P. We are grateful
to Professors A. Dell and H. Morris and Mr S.
Haslam (Imperial College, London) for the prepara-
tion of derivatives for FABMS and for the acquis-
ition of the FAB mass spectra. The FABMS work
was supported by the Wellcome Trust and the
BBSRC.
References
[1] A. Jeannes, J.E. Pittsley, P.R. Watson, and R.J.
Dimler, Arch. Biochem. Biophys., 92 (1961) 433±
450.
[2] M.E. Slodki, Biochim. Biophys. Acta, 57 (1962)
525±533.
[3] R.K. Bretthauer, G.J. Kaczorowski, and M.J.
Wiese, Biochemistry, 12 (1973) 1251±1256.
[4] M.E. Slodki, in M. Yalpani (Ed.), Industrial Poly-
saccharides: Genetic Engineering, Structure/Prop-
erty Relations and Applications, Elsevier,
Amsterdam, 1987, pp 109±119.
[17] A. Bax, R.H. Gri€ey, and B.L. Hawkins, J. Magn.
Reson., 55 (1983) 301±315.
[18] L. Lerner and A. Bax, J. Magn. Reson., 69 (1986)
365±380.
[5] L.A.S. Parolis, J.é. Duus, H. Parolis, M. Meldal,
and K. Bock, Carbohydr. Res., 293 (1996) 101±117.
[6] M.E. Slodki, R.M. Ward, and J.A. Boundy, Bio-
chim. Biophys. Acta, 304 (1973) 449±456.
[19] A. Bax and M.F. Summers, J. Am. Chem. Soc., 108
(1986) 2093±2094.
[20] A. Dell and P.R. Tiller, Biochem. Biophys. Res.
Commun., 135 (1986) 1126±1134.