Glycobiology vol. 12 no. 1 pp.

114, 2002

The salivary mucin MG1 (MUC5B) carries a repertoire of unique oligosaccharides that
is large and diverse

Kristina A. Thomsson2, Akraporn Prakobphol3,

Hakon Leffler4, Molakala S. Reddy5, Michael J. Levine5,
Susan J. Fisher3,6, and Gunnar C. Hansson1,2
2Department of Medical Biochemistry, Gteborg University, Medicinaregatan
9A, 413 90 Gothenburg, Sweden; 3Department of Stomatology, University of
California San Francisco, San Francisco, CA 94143, USA; 4Department of
Molecular Medicine, University of Lund, Lund, Sweden; and 5Department of
Oral Biology, State University of New York at Buffalo, Buffalo, NY 14214,
USA; and 6Departments of Anatomy, Pharmaceutical Chemistry, and
Obstetrics, Gynecology and Reproductive Sciences, University of California
San Francisco, San Francisco, CA 94143, USA

Received on March 8, 2001; revised on August 16, 2001; accepted on August 22,
2001

The high-molecular-mass salivary mucin MG1, one of two

major mucins produced by human salivary glands, plays
an important role in oral health by coating the tooth
surface and by acting as a bacterial receptor. Here this
mucin was purified from the submandibular/sublingual
saliva of a blood group O individual. The presence of
MUC5B as the major mucin in this preparation was
confirmed by amino acid analysis and its reactivity with the
monoclonal antibody PAN H2. To structurally characterize
MG1 carbohydrates the O-glycans were released by reductive
-elimination. Nuclear magnetic resonance spectroscopy of
the nonfractionated mixture showed that (1) fucose was
present in blood group H, Lea, Lex, Leb, and Ley epitopes;
(2) NeuAc was mainly linked 2-3 to Gal or 2-6 to
GalNAcol; and (3) the major internal structures were core
1 and core 2 sequences. After this preliminary analysis the
released oligosaccharides were separated into neutral
(56%), sialylated (26%), and sulfated (19%) fractions, with an
average length of 13, 17, and 41 sugar residues, respectively.
Gas chromatographymass spectrometry and matrix-assisted
laser desorption/ionization time-of-flight mass spectrometry
of mixtures of neutral and sialylated oligosaccharides revealed
at least 62 neutral and 25 sialylated oligosaccharides
consisting of up to 20 monosaccharide residues. These
results showed that the MG1-derived oligosaccharides
were much longer than those of MG2, and only a few
species were found on both molecules. Thus, these two
mucins create an enormous repertoire of potential binding
sites for microorganisms at one of the major portals where
infectious organisms enter the body.
Key words: mass spectrometry/mucin/oligosaccharide/proton
NMR spectroscopy/salivary gland

1To

whom correspondence should be addressed

2002 Oxford University Press

Introduction
Mucins are major glycoprotein components of mucus, the
viscous gel that covers the epithelial layers of the body and
protects the underlying epithelia. These molecules are highly
glycosylated; the oligosaccharides (5080% of the weight) are
linked to serine and threonine residues clustered in mucin
domains on the protein backbone. Currently, twelve mucin
protein cores have been identified (Perez-Vilar and Hill, 1999).
There are regional differences in expression, as particular
mucin subsets variably contribute to the mucus coat that covers
the specialized epithelial layers of the body. The presence of
cell type and tissue-specific glycoforms suggests distinct
functions of mucin O-glycans, such as variations in viscosity
and adhesivity, including bacterial receptor activity.
Two major, physically distinct mucin populations have been
identified in salivathe high-molecular-mass (MG1) and the
low-molecular-mass (MG2) mucins. The MG1 mucin core, a
large oligomeric protein composed of disulfide-linked subunits, is
encoded by the MUC5B gene (Nielsen et al., 1997; Wickstrm
et al., 1998; Desseyn et al., 1997a, 1998). MUC5B is produced
by all salivary glands except the parotid (Veerman et al., 1996;
Nielsen et al., 1996). The existence of different glycoforms has
been demonstrated by the isolation of differently charged
subpopulations (Wickstrm et al., 1998; Bolscher et al., 1995;
Thornton et al., 1999). MUC5B has also been identified as one
of the two major mucin components of the mucus layer in the
respiratory tract and endocervix (Wickstrm et al., 1998;
Thornton et al., 1997). It is interesting to note that viruses
(e.g., HIV-1) and bacteria (e.g., Helicobacter pylori) can bind to
MG1, suggesting that this mucin plays a role in establishment of
the oral ecology (Bergey et al., 1994; Veerman et al., 1997).
MG2, a small, nonmultimerizing protein that is exclusively
found in salivary secretions, is encoded by the MUC7 gene
(Bobek et al., 1993). This mucin is produced by the
submandibular (SM), sublingual (SL), and palatine glands
(Bolscher et al., 1999). Two glycoforms with different levels
of fucose and sialic acid have been distinguished (Ramasubbu
et al., 1991). MG2 interacts with a large number of oral microorganisms (Nieuw Amerongen et al., 1995) and can mediate
leukocyte rolling and tethering (Prakobphol et al., 1999).
Together, MG1 and MG2 are the major glycosylated components of the acquired pellicle which coats and protects the tooth
surface (Fisher et al., 1987).
Previous descriptions of glycans expressed on MG1 and
MG2 suggested that MG1-derived oligosaccharides are more
heterogeneous than those of MG2 (Reddy et al., 1985; Levine
et al., 1987). In a recent study using nuclear magnetic
resonance (NMR) and mass spectrometry (MS) techniques, we
demonstrated that glycans isolated from highly purified MG2
1

K.A. Thomsson et al.

carry diverse structures that are largely variations of the

T antigen and Lex motifs (Prakobphol et al., 1998). In the study
presented here, a similar strategy was applied to elucidate the
structure of glycans derived from highly purified MG1. The
results showed a remarkable degree of heterogeneity, even
greater than that observed among the MG2 oligosaccharides. A
comparison of the MG1 and MG2 repertoire, with respect to
size, diversity, and expression (semiquantitative) of glycans
and terminal epitopes, showed a high level of mucin-specific
glycosylation that could have important implications for the
specialized biological functions these molecules play in the
oral cavity.

Table II. Amino acid analysis

Amino acid

MG1 (mole %)

MUC5Ba (predicted;
(mole %)

Ala

8.6

Arg

3.5

3.4

Asx

5.8

4.9

Cys

NA

8.4

4.9

Glx

6.9

6.3

Gly

7.6

7.0

His

2.5

2.2

Ile

2.4

2.2

Results

Leu

6.0

5.8

Characterization of the MG1 preparation

Lys

2.3

1.8

Purified salivary MG1 isolated from the SM/SL saliva of a

blood group O individual was analyzed for carbohydrate and
amino acid composition. The results are shown in Tables I and
II. The mucin was composed of 19% protein and 81% carbohydrate, with high amounts of Ser, Thr, and Pro (44 weight %).
With regard to carbohydrate composition, NeuAc, Fuc, Gal,
GlcNAc, and GalNAc were present in the following molar
ratios: 1:5:4:3:1 (GalNAc set to 1), respectively. As expected,
only trace amounts of mannose were detected. The amino acid
composition was almost identical to the theoretical estimate
based on translating the MUC5B nucleic acid sequence
(Desseyn et al., 1997a,b, 1998). These data suggested that
MUC5B was the major component. The binding of monoclonal antibody PAN H2 (Nielsen et al., 1997) to partially
deglycosylated MG1 confirmed the identity of the purified
mucin as MUC5B (data not shown).

Met

0.4

1.2

Experimental strategy
The MG1 glycans were released by treatment with alkaline
borohydride. During this procedure the GalNAc attached to
Ser/Thr of the peptide core was converted to GalNAcol. The
percent of saccharides that were recovered after release from

Table I. Chemical composition of MG1

MG1 a
Weight %
Protein

19

Carbohydrate

81

mol/mg mucin
Fuc

1.5

Gal

1.2

Man

0.02

GlcNAc

0.88

GalNAc

0.33

NeuAc

0.46

a The monosaccharide composition analysis is the mean of two separate

analyses.

Phe

2.3

2.3

Pro

11.6

9.9

Ser

11.8

11.1

Thr

20.6

20.7

Trp

NA

1.2

Tyr

1.7

1.7

Val

5.8

5.3

NA, not analyzed.

aBased on DNA sequence data from references (Desseyn et al., 1997a,b,
1998).

the protein backbone was estimated based on monosaccharide

compositional analyses of the whole mucin and of the neutral,
sialylated, and sulfated fractions. The yield was approximately
50%, typical for this procedure. After the desalting step, the
MG1 oligosaccharide mixture was first analyzed by NMR
spectroscopy. Then the oligosaccharides were fractionated into
pools of neutral, sialylated, and sulfated species, from which
aliquots were withdrawn for monosaccharide compositional
analyses. Neutral and sialylated oligosaccharides were
permethylated and analyzed by gas chromatography (GC),
matrix-assisted laser desorption/ionization time-of-flight
(MALDI-TOF) MS, and GC-MS. Finally, the sulfated
oligosaccharides were subjected to gel chromatography as an
additional method for estimating their molecular weight.
NMR spectroscopy
Proton NMR spectra of a large number of O-linked glycans
have been compiled (Kamerling and Vliegenthart, 1992) and
gathered in a database (Sugabase) of the proton chemical
shifts. Typically, these chemical shifts are very sensitive to the
presence of the immediate saccharide neighbors and relatively
less sensitive to more distant residues. In the current study, we
have used 1D and 2D double quantum-filtered correlated
spectroscopy (DQF-COSY) and total correlated spectroscopy
(TOCSY) proton NMR spectroscopy to analyze the unfractionated
mixture of O-linked glycans released from MG1. The 1D NMR
spectra provided the highest sensitivity, but in some areas
crowding precluded positive identification of important

Glycosylation of the MG1 salivary mucin

Fig. 1. 1D NMR spectrum of the total mixture of MG1 oligosaccharides recorded at 25C. O-linked oligosaccharides were released from purified MG1 by
reductive -elimination in 0.05 M NaOH/1 M NaBH4, desalted on Dowex 50WX8, and transferred to D20 (see Materials and methods). (A) The entire 1D
spectrum; assignments of the major resolved signals are indicated. (B) Expansion of the Fuc H1 region with annotations for type 1 chain (Gal13GlcNAc)based
structures (top) and type 2 chain (Gal14GlcNAc)based structures (bottom).

signals. DQF-COSY, which is less sensitive, resolved these

signals; groups of cross-peaks with characteristic fine structures
permitted identification of the same type of structural reporter
group within multiple environments.
The results showed that the MG1 NMR spectra (Figures 1
and 2) were far more complex than those obtained from analysis
of the MG2 oligosaccharides (Prakobphol et al., 1998).
Despite this complexity, different classes of NMR signals
(Table III) could be readily distinguished and interpreted in
light of previously published spectra of purified reference
compounds (see Materials and methods). In agreement with
the composition analysis (Table I), -linked fucose residues
were prominent components of the MG1 glycans. They were
identified by downfield shifted H1 signals (> 4.9 ppm) and H2
signals at 3.8 ppm. The Fuc residues were also identified by
their H5 and H6 (CH3) signals (Figure 1A) and their H5/H6
cross-peaks clearly visible in the 2D spectra (Figure 2A and
Table III). H1/H2 cross-peaks from -linked Gal confirmed the
presence of several fucosylated determinants (Figures 1, 2A,
2B and Table III). Other possible -linked monosaccharides
(Gal, GalNAc, and GlcNAc) were not detected. Thus, Ley
[(Fuc12)Gal14(Fuc13)GlcNAc)] was the major
fucosylated determinant on MG1. The Leb determinant

[(Fuc12)Gal13 (Fuc14)GlcNAc)], and all possible

combinations of a single Fuc linked to either C-2 of Gal
(H type 1 and 2; Fuc12Gal1 3/4GlcNAc), C-3 of GlcNAc
[Lex; Gal14(Fuc13)GlcNAc], or C-4 of GlcNAc [Lea;
Gal13(Fuc14) GlcNAc], were also found.
NeuAc residues were less abundant than Fuc, but easily
identified in the NMR spectra by their H3ax and H3eq signals.
About two-thirds were found in NeuAc23Gal (H3ax/H3eq
at about 1.8/2.76 ppm) and one-third in NeuAc26GalNAcol
(H3ax/H3eq at about 1.7/2.73 ppm), with small amounts in
NeuAc26Gal (H3ax/H3eq at about 1.72/2.67 ppm)
(Figures 1A, 2A, 2C and Table III). The cross-peaks from Gal
H1/H3 and Gal H1/H4 (Figure 2D) were observed in the
TOCSY spectrum in an area that did not contain other major
peaks. When substituted with NeuAc at C-3, the Gal H3
signals shifted downfield to > 4.11 ppm. Three partially
resolved groups of signals could be assigned to Gal H1/H3 in
the TOCSY spectrum and hence gave positive confirmation of
the presence of NeuAc23Gal. All had typical H1 signals at
about 4.524.54 ppm. For the top cluster, H3 was at 4.08 ppm,
indicative
of
Gal
in
the
sLex
determinant
[NeuAc23Gal14(Fuc13)GlcNAc] and for Gal in
NeuAc23Gal 13GlcNAc. An H1/H2 cross-peak from
3

K.A. Thomsson et al.

Table III. Assignments of signals (ppm) in 3H-NMR analysis of alditol mixture released from MG1a

aThe

reporter groups from each type of saccharide residue are marked in bold. The chemical shift values were determined from the cross-peaks identified
by 2D COSY and TOCSY at 25C. Each value given represents a range of about 0.01 ppm because most cross-peaks derive from more than one
compound and, hence, are slightly broadened due to this heterogeneity.
bSemquantitative estimates + +++.
cThe observed chemical shift for H3ax (about 1.87 ppm) was slightly lower than reported for Sda/CAD determinants (about 1.93 ppm).

Gal in sLex at 4.52/3.52 was also positively identified. The

other two signals with H3 around 4.11 and 4.12 ppm are
typical of the NeuAc23Gal13GalNAcol and
NeuAc23Gal14GlcNAc structures, respectively. In
addition, evidence for NeuAc23, as in blood group Sda/Cad
determinants, was detected by a weak NeuAc H3ax/H3eq
cross-peak in the COSY spectrum at 2.68/1.87 ppm and an H1/H3
Gal cross-peak in the TOCSY spectrum at about 4.53/4.15 ppm
(Figures 2C and D, respectively). Because H3 of Gal in sLea,
usually found at about 4.04 ppm, was missing, this epitope was
either absent or scarce in the MG1 sample.
The cross-peak in the TOCSY spectrum at 4.45/4.10 is
typical of the H1/H4 signals of branched Gal residues, substituted
with GlcNAc at both C-3 and C-6 (Figure 2D). The same
signal from Gal substituted with GlcNAc only at C-3, which is
found above 4.12 ppm (often at about 4.15), was missing.
Hence, the Gal-GlcNAc backbones of the MG1 glycans were
predominantly extended by branching, although smaller
amounts of linear extension may contribute to the bottom part
of the cross-peak.
The region of the 1D NMR spectra between 4 and 4.8 ppm
was very complex and could not be fully interpreted (Figure 1A).
Nevertheless, a great deal of information about core structure
was deduced. Among others were signals from H1 of -Gal
and -GlcNAc, as well as various signals from GalNAcol. A
strong cluster of signals from H2 of GalNAcol was identified
at about 4.39 ppm based on the position and fine structure of
4

the cross-peak with the two H1 signals. This chemical shift is

typical of core structures containing Gal13GalNAcol (core
1 and 2). Because the H2 GalNAcol signal from core 1 and 2
was clearly visible in the 1D spectrum and no other signal was
found in the 4.40 ppm region, we could use these as a reference
to compare the amount of GalNAcol with other saccharides.
The relative size of the NMR signals agreed with the molar
ratios of constituent saccharides estimated in the composition
analysis (Table I). The H1/H2 and H3/H4 cross-peaks of
GlcNAc13GalNAcol (as in core 3 and 4containing
saccharides) were weak (Table III) and overlapped with other
signals. Cross-peaks from H5 and H6, although typically
weaker, permitted identification of both -GlcNAc and NeuAc at
C-6 of GalNAcol and unsubstituted C-6 of GalNAcol (Table III).
GalNAcol saccharides, free or substituted at C-6 with NeuAc,
were identified by their H3/H4 cross-peak (Table III).
Monosaccharide composition of neutral, sialylated, and
sulfated MG1-derived glycans
The released oligosaccharides where then separated into
neutral, sialylated, and sulfated fractions and analyzed for
monosaccharide composition (Table IV). For comparison
purposes, the table also includes the composition of the corresponding glycan fractions purified from MG2, information that
we previously published (Prakobphol et al., 1998). The molar
ratios of the monosaccharides were calculated relative to
GalNAcol, the reduced product of the GalNAc that had been

Glycosylation of the MG1 salivary mucin

Fig. 2. 2D DQF-COSY and TOCSY NMR spectra of the total mixture of MG1 oligosaccharides. (A) The entire 2D DQF-COSY spectrum; assignments of the major resolved
signals are indicated. (B) A region of DQF-COSY spectrum showing cross-peaks between Fuc H1 and H2; annotations for type 1 chain (Gal13GlcNAc)based structures
and type 2 chain (Gal14GlcNAc)based structures are labeled. (C) A region of the DQF-COSY spectrum showing cross-peaks between NeuAc H3ax and
H3eq. (D) A region of the TOSCY spectrum showing cross-peaks between Gal H1 and H3 or H4.

Table IV. Monosaccharide composition analysis of neutral, sialylated, and sulfated oligosaccharides from MG1 and MG2
MG1a

Chains (mole

%)d

MG2b

Neutralc

Sialylated

Sulfated

Neutral

Sialylated

Sulfated

56

26

19

49

40

11

Fuce

4.6

4.6

13.5

0.8

1.6

1.1

Gal

3.8

4.8

11.4

1.2

2.4

2.5

Man

ND

0.4

ND

GlcNAc

3.0

3.9

10.3

GalNAc

0.2

0.5

1.1

ND

NeuAc

NA

GalNAcol
Average chain

lengthf

ND
1.0

1.4

NA

12.6

16.6

41.3

4.0

ND

ND

2.6

2.4

ND

ND

2.5

NA

1
10.1

1
7.0

ND, not detected; NA, not analyzed.

results are the mean of two separate analyses.
bMG2 purification, MG2 glycan release, and fractionation are described elsewhere (Prakobphol et al., 1998).
cOligosaccharides were subfractionated into neutral, sialylated, and sulfated species as described in Materials and methods.
dPercentage molar amounts of GalNAcol distributed in the neutral, sialylated, and sulfated fractions.
eMolar ratios of individual monosaccharide residues are given relative to the linkage sugar GalNAcol.
fAverage monosaccharide chain length, based on measured molar amounts of Fuc, Gal, GlcNAc, GalNAc, and NeuAc relative to GalNAcol.
aThe

K.A. Thomsson et al.

linked to the mucin protein core during the release reaction.

The molar distribution of neutral, sialylated, and sulfated MG1
oligosaccharides was 56, 26, and 19 mole%, respectively.
MG2 contained a similar percentage of neutral species
(49 mole%), but a much higher percentage of sialylated
oligosaccharides (40 mole%) and a lower percentage of
sulfated species (11 mole%). In this way calculated average
MG1 oligosaccharide chain length was substantially longer
than that of MG2: 13 versus 4 in the neutral fraction, 17 versus
10 in the sialylated fraction, and, most strikingly, 41 versus 7
in the sulfated group. These oligosaccharide lengths are longer
than when estimated by MS (see below). This discrepancy is
always found and largely due to MS favoring low masses, but
it can also be attributed to a possible underestimate of the
GalNAcol during monosaccharide analyses. However, the
relative differences between MG1 and MG2 are valid, as are
the general trends in the length of the oligosaccharides.
MS analyses of the neutral and sialylated fractions
The neutral and sialylated oligosaccharide fractions were
permethylated and analyzed by MALDI-TOF MS (Figure 3).
Fifty-five molecular ions of neutral oligosaccharides of 20
monosaccharide residues were detected. The deduced
monosaccharide composition of the individual components
showed the composition seen for polylactosamine structures,

approximately equal amounts of Gal and GlcNAc (Table V).

The composition of 15 of the annotated ions had masses
suggesting two more Hex than HexNAc residues. This is
unusual, and no structures or epitopes supporting this unusual
composition were detected by GC-MS or NMR. However,
further exploration of the nature of the ions at m/z 1171.6
(labeled N5.3) and m/z 1345.6 (N6.3) from both MG1 and
MG2 by nano-electrospray ionization tandem mass spectrometry (ESI-MS/MS) generated product ions corresponding to
the saccharide structures given (data not shown). No definitive
structures could be predicted, but were suggestive of -HexHexNAcol-Hex core structures.
In the sialylated oligosaccharide fraction, 24 molecular ions
consisting of up to 19 residues were detected (Table VI). Only
five of the sialylated oligosaccharides contained a second
NeuAc, and most compounds were also fucosylated.
Permethylated neutral and sialylated oligosaccharides were
fractionated by gel filtration, the smaller oligosaccharides were
collected and analyzed by GC and GC-MS (Figure 4). Sixteen
neutral and eight sialylated compounds were possible to
characterize by GC-MS (Tables VII and VIII). The interpretation
was based on the mass spectra obtained from each peak, where
most fragment ions detected were oxonium and inductive ions
(Prakobphol et al., 1998; Carlstedt et al., 1993; Karlsson et al.,
1989). The majority of the oligosaccharides had core 1
(Hex13HexNAc) or core 2 [Hex13(HexNAc16)HexNAc]
type sequences. The neutral oligosaccharides were nearly all
fucosylated, with Fuc linked to Gal, or to GlcNAc in H, Leb
and/or Ley-like epitopes. NeuAc was found linked either to Gal
residues or to C-6 of GalNAcol.
Considered together, the results from MS and NMR were
consistent and indicated that the structures of the MG1
oligosaccharides were much more diverse than those of the
MG2 (Prakobphol et al., 1998). Core 1 and 2 structures were
shown to be prominent by both MS and NMR. Three core 3
sequences, N3.3, N4.3, and N5.4 (Table VII, Figure 3) were
detected as minor components by GC-MS as well as in the
NMR spectra by weak signals. Small amounts of core 4
sequences were only detected by NMR. The Gal-GlcNAc
backbone was predominantly extended by branching. The
neutral oligosaccharides were complex and highly fucosylated.
The sialylated compounds were less complex and contained
both the NeuAc23Gal and the NeuAc26GalNAcol
sequences. In contrast to MG2, which carried fewer fucosylated
determinants, MG1 carried several fucosylated determinants,
including H type 1 and 2 (Fuc12Gal13/4GlcNAc),
Lea [Gal13(Fuc14)GlcNAc], Leb [(Fuc12)Gal1
3(Fuc14)GlcNAc)], Lex [Gal14(Fuc13) GlcNAc],
sLex [NeuAc23Gal14 (Fuc13)GlcNAc], and Ley
[(Fuc12)Gal14 (Fuc13)GlcNAc)]. Finally, free
GalNAcol or GalNAcol carrying a single NeuAc substitution
at C-6 were detected by both GC/MS, NMR spectroscopy.
Size fractionation of sulfated MG1-derived glycans

Fig. 3. MALDI-TOF MS spectra of permethylated neutral and sialylated

oligosaccharides from MG1. Molecular ions occur as lithium adducts [M+Li] +.
Contaminants are marked with an asterisk. The deduced oligosaccharide
compositions are listed in Table V. Only masses above m/z 1000 are shown.

Detailed structural analysis of the sulfated oligosaccharides, the

least abundant fraction with the most complex oligosaccharides,
awaits purification of larger amounts of MG1. Here we can
only report the results of preliminary analyses (data not
shown). We estimated the sizes of the oligosaccharide by gel
filtration on two serially coupled Superdex Peptide columns
that have a fractionation range of 1009000 Da for peptides.

Glycosylation of the MG1 salivary mucin

Table V. Permethylated neutral oligosaccharides from MG1 and MG2 detected by MALDI-TOF MS
Composition
Compound

Molecular ion
[M+Li]+

N1.1

NA

N2.1

518.4

N3.1

692.4

N3.2,N3.3

763.4

N4.3,N4.4

937.5

Fuc

Relative peak height (%)

Hex

HexNAc

GalNAcol

MG1

MG2

NA

NA

2.1

14.3

10.6

8.8
4.9

2.4

6.3

N4.1,N4.2

967.5

4.1

N4.5

1008.5

1.6

10.4

N5.4

1111.4

2.1

N5.1, N5.2, N5.5, N5.6, N5.7

1141.6

7.6

10.7

N5.3a

1171.6

1.6

4.3

N5.8

1212.6

1.1

N6.1, N6.4, N6.5

1315.7

10.3

2.8

N6.3a

1345.7

3.2

2.1

N6.6

1386.7

1.3

1.2

N6.2

1416.7

3.2

N7.1, N7.5

1489.8

10.6

3.0

N7.3a

1519.8

3.2

1.2

N7.6

1560.8

1.9

N7.2

1590.8

1.0

N7.4a

1620.8

1.3

N7.7

1661.9

1.9

N8.3a

1693.9

0.5

N8.4

1734.9

1.2

N8.1

1764.9

2.2

N8.2

1835.9

N8.5

1866.0

N9.3

1939.0

3.8

N9.4a

1969.0

0.7

N9.1

2040.1

N9.2a

2070.0

N10.3

2113.1

4.4

N10.4a

2143.1

1.1

N10.5

2184.1

0.7
0.5

N10.1

2214.1

N10.2a

2244.0

4.7

2.5
2.1
5.2

7.0
2.5

2.9
1.2

N11.2

2287.2

1.4

N11.3a

2317.2

0.5

N11.4

2358.2

0.7

N11.1

2388.2

0.9

N11.5a

2418.2

0.2

N12.1

2532.3

0.3

N12.2

2562.3

1.5

N12.3a

2592.3

0.4

N13.1

2736.4

1.6

0.8

K.A. Thomsson et al.

Table V. Continued
Composition

Relative peak height (%)

Compound

Molecular ion
[M+Li]+

Fuc

Hex

HexNAc

GalNAcol

MG1

N13.2a

2766.4

0.6

N13.3

2807.4

0.3

N13.4

2837.5

0.2

N14.1

2910.5

1.1

N14.2a

2940.5

0.4

N14.3

2981.5

0.3

N14.4

3011.6

0.3

N15.1

3084.6

0.3

N15.2a

3114.6

0.2

N15.3

3155.6

0.2

N15.4

3185.6

0.5

N16.1

3359.7

0.7

N16.2a

3389.7

0.2

N16.3

3533.8

0.5

N17.1

3563.8

0.2

N17.2a

3604.9

0.1

N18.2

3707.9

0.1

N18.3

3808.9

0.2

N19.1

3983.0

0.3

N19.2a

4013.7

0.1

N20.1

4157.1

0.3

Sum

100%

MG2

100%

NA, not analyzed.

aThe calculated composition of the annotated ions includes 2 + n Hex and n HexNAc residues. No structures or epitopes supporting this composition were
detected by GC-MS or NMR, but ESI-MS/MS of the m/z 605 and 692 [M+2Na]2+ for N5.3 and N6.3, respectively, suggested the presence of several components
with -Hex-HexNAcol-Hex containing core structures (not shown).

Aliquots of individual fractions were analyzed for monosaccharide composition. The oligosaccharides eluted as a
single broad peak between 15 and 22 ml, that is, close to the
void volume (15.5 ml). For comparison and calibration
purposes, oligosaccharides released from porcine small
intestinal mucins, with an average chain length of five
residues, eluted between 25 and 35 ml; salt eluted at 38 ml.
These results suggested that the average oligosaccharide is
very large and might have a mass of up to 8 kDa. This could
suggest a chain length of about 40 sugar residues, lengths also
suggested from the results of the monosaccharide compositional
analysis (Table IV).

Discussion
Molecular cloning techniques have recently revealed information
about the MG1 protein core (apomucin) (Desseyn et al., 1997a,
1998). The gene contains 48 exons, encodes 6000 amino
acids, and produces a polypeptide of Mr 600,000. The
MUC5B central exon, the largest ever reported for a vertebrate
8

gene, encodes four super-repeats of 528 amino acids. In turn,

each super-repeat contains a region of imperfectly conserved
29-residue tandem repeats, a conserved region without repeats,
and a Cys-rich subdomain.
Very little is known about MG1 glycosylation, which
accounts for approximately 80% of the mucins molecular
mass (Levine et al., 1987; Loomis et al., 1987). To our knowledge this is the first report on the structure of the O-glycans of
purified MG1. The results from MALDI-TOF MS, GC-MS,
and NMR spectroscopy showed the presence of highly fucosylated oligosaccharides. Fuc was present in a large variety of
terminal linkages that included blood group H epitopes (types
1 and 2), as well as Lea, Leb, Lex, and Ley determinants. A
portion of these results is in agreement with the data of Klein
et al. (1992). These investigators characterized, by using
NMR, the high-molecular-weight oligosaccharides that were
present in a pronase digest of a pooled sample of whole saliva
obtained from 20 blood group O donors. The Fuc-containing
epitopes we detected in association with purified MG1 were
among those previously reported, an expected result because

Glycosylation of the MG1 salivary mucin

Table VI. Permethylated sialic acidcontaining oligosaccharides from MG1 analyzed by MALDI-TOF MS and GC-MS
Composition
Compound
S2.1

Molecular ion
688.4

[M+Li]+

NeuAc

Relative peak height (%)

Fuc

Hex

HexNAc

S3.1, S3.2

892.5

S3.3

933.5

S4.4

1066.6

S4.1

1137.7

S4.2

1266.7

S4.3

1311.7

14.2

MG2

15.9

3.5

11.2

6.6

6.8

3.1

0.9

2.4

1
1

MG1

1
1
1

GalNAcol
1

23.8

S5.1, S5.2

1341.7

4.0

S6.5

1485.9

2.1

S6.1, S6.2, S6.4

1515.8

11.6

20.9

S7.2

1689.9

14.6

2.3

S6.3

1715.9

S7.1

1890.0

0.8

7.7

S8.2

1935.0

0.4

S8.1

1965.0

0.4

2.7

S9.1

2139.1

1.6

1.1

S10.2

2313.2

2.8

S10.1

2414.2

S11.3

2487.3

S11.1

2588.3

S12.2

2762.4

S11.2

2788.4

S13.1

2936.5

S12.1

2962.5

S14.1

3110.6

0.8

S15.1

3385.7

0.2

S16.1

3559.8

0.4

S17.1

3733.9

0.3

S19.1

4183.2

0.2

Sum

MG1 was undoubtedly a component of the pooled sample.

However, we also noted significant differences, a result we
also expected because, in addition to MG1, saliva contains
many other highly glycosylated components. For example, we
primarily found core 1 and 2 structures, wherease Klein et al.
(1992) report a substantial number of cores types 3 and 4.
One of our more interesting observations were that the
sulfated fraction contained oligosaccharides of an extraordinary lengththe average size was approximately 40
monosaccharide residues, accounting for 19 mole% of the
glycans expressed on the protein. Because there is very little
information about such structures, it is difficult to know
whether they are unique to MG1 or found on other mucin
cores. Our groups have not detected these unusual structures

22.8

3.4

3.2
1.3
1.4
0.5
0.9
1.0
0.9

100%

100%

on any other highly glycosylated salivary components

(Prakobphol et al., 1998; Gillece-Castro et al., 1991).
However, long sulfated oligosaccharides, with up to 200
residues, have been suggested on mucins isolated from the
sputum of a cystic fibrosis patient (Sangadala et al., 1993).
Previously we characterized the glycans of MG2, the other
major mucin found in human saliva (Prakobphol et al., 1998).
The experimental strategy included all the basic elements that
were applied in the present study. Briefly, we purified (to
homogeneity) MG2 from the SM/SL saliva of a single individual.
The O-glycans were chemically released and their structures
determined by using a combination of NMR and MS techniques.
Accordingly, we were able to compare the O-glycosylation
patterns of the two mucins. Here we report major differences in
the glycosylation of these two mucins. These differences
9

K.A. Thomsson et al.

Fig. 4. High-temperature GC of permethylated neutral and sialylated

oligosaccharides. The peaks were labeled with numbers and letters. These
labels were then used to construct the deduced oligosaccharide sequences that
are listed in Tables VII and VIII.

included the much longer chain length of the MG1 sulfated

species. The chain length of the neutral and sialylated glycans

was also longer, evident from both the monosaccharide

composition analyses (Table IV) and the spectra of the
oligosaccharides obtained by MALDI-TOF MS (Figure 3).
These subsets of MG1 oligosaccharides carried up to 20 residues,
whereas components of the neutral and sialylated MG2 fractions
had a maximum chain length of 12 monosaccharides.
The differences in glycosylation between the two salivary
mucins affected not only length but also terminal epitopes. For
example, MG1 carried a wide variety of the fucosylated
terminal epitopes, which included blood group H, Lea, Lea,
Leb, Lex, and Ley, whereas MG2 carried fewer fucosylated
epitopes. Finally, our previous study indicated that MG1, but
not MG2, carried the ABO blood group structures (Prakobphol
et al., 1993). Since, the MG1 sample that we analyzed in this
study was obtained from an individual with blood type O, we
were unable to determine which species carried these
additional modifications.
With regard to oligosaccharides sequences GC-MS showed
that MG1 and MG2 shared only 6 out of 22 neutral species
(Table VII) and 4 out of 13 sialylated oligosaccharides
(Table VIII). The data from our study also indicated that MG1
carries a much more diverse repertoire of oligosaccharide
structures than MG2. MALDI-TOF MS revealed 79 molecular
ions, as compared to 38 for MG2. If the GC-MS results are

Table VII. Comparison of neutral oligosaccharides from MG1 and MG2 analyzed with GC-MS
No.

Neutral oligosaccharidesa

N1.1

GalNAcol

Molecular massb

MG1

MG2

307.2

N2.1

Gal3GalNAcol

511.3

N3.1

FucGal3GalNAcol

685.4

N3.2

Gal3(GlcNAc6)GalNAcol

756.4

N3.3

GalGlcNAc3GalNAcolc

756.4

N4.1

Gal3(GalGlcNAc6)GalNAcolc

960.5

N4.2

Gal4GlcNAcGal3GalNAcol

960.5

x
x

N4.3

FucGalGlcNAc3GalNAcol

930.5

N4.4

FucGal3(GlcNAc6)GalNAcol

930.5

N5.1

Gal3(Gal4(Fuc)GlcNAc6)GalNAcol

1134.6

N5.2

Gal4(Fuc)GlcNAcGal3GalNAcol

1134.6

N5.4

FucGal(Fuc)GlcNAc3GalNAcol

1104.6

N5.5

Gal3(FucGalGlcNAc6)GalNAcol

1134.6

N5.6

FucGal3(GalGlcNAc6)GalNAcol

1134.6

N5.7

FucGalGlcNAcGal3GalNAcol

1134.6

N6.1

FucGal(Fuc)GlcNAcGal3GalNAcol

1308.7

N6.2

GalGlcNAc(GalGlcNAc)Gal3GalNAcol

1409.7

x
x

N6.4

Gal3(FucGal(Fuc)GlcNAc6)GalNAcol

1308.7

N6.5

FucGal3(FucGalGlcNAc 6)GalNAcol

1308.7

N7.1

FucGal(Fuc)GlcNAc(Fuc)Gal3GalNAcol

1482.8

N7.2

Gal(Fuc)GlcNAc(GalGlcNAc)Gal3GalNAcol

1583.8

N7.5

FucGal3(FucGal(Fuc )GlcNAc6)GalNAcol

1482.8

aThe saccharides marked in bold are located on C-6 of GalNAcol. The following assumptions were made based on monosaccharide composition analysis: hexose
residues are Gal, N-acetylhexosamine residues are GlcNAc, deoxyhexose residues are Fuc and N-acetyl hexosaminitol residues are GalNAcol.
bPermethylated oligosaccharides, calculated monoisotopic mass.
cSuggested to be type 2 by the abundant ion at m/z 182 in the MG1 sample.

10

Glycosylation of the MG1 salivary mucin

Table VIII. Comparison of sialylated oligosaccharides from MG1 and MG2 analyzed with GC-MS
No.

Sialylated oligosaccharidesa

Molecular massb

MG1

MG2

S2.1

NeuAc6GalNAcol

681.4

S3.1

NeuAcGal3GalNAcol

885.5

885.5

1130.6

x
x

S3.2

Gal3(NeuAc6)GalNAcol

S4.1

NeuAcGal3(GlcNAc6)GalNAcol

S4.2

NeuAcGal3(NeuAc6)GalNAcol

1259.7

S4.4

FucGal3(NeuAc6)GalNAcol

1059.6

S5.1

NeuAcGal3(GalGlcNAc6)GalNAcol

1334.7

S5.2

Gal3(NeuAcGalGlcNAc6)GalNAcol

1334.7

S6.1

NeuAcGal3(Gal4(Fuc)GlcNAc 6)GalNAcol

1508.8

S6.2

Gal3(NeuAcGal(Fuc)GlcNAc6)GalNAcol

1508.8

S6.3

NeuAcGal3(NeuAcGalGlcNAc6)GalNAcol

1708.9

S6.4

NeuAcGal3(FucGalGlcNAc6)GalNAcol

1508.8

S7.1

NeuAcGal3(NeuAcGal(Fuc)GlcNAc6)GalNAcol

1883.0

x
x
x

aThe saccharides marked in bold are located on C-6 of GalNAcol. The following assumptions were made based on monosaccharide composition analysis: Hex
residues are Gal, N-acetylhexosamine residues are GlcNAc, deoxyhexose residues are Fuc, and N-acetyl hexosaminitol residues are GalNAcol.
bPermethylated oligosaccharides, calculated monoisotopic mass.

added, then MG1 carries 87 different glycan structures versus

the 43 isolated from MG2. These numbers are underestimates
because MALDI-TOF MS only shows molecular ions and does
not distinguish between isomeric oligosaccharides. This is
important because the complexity and number of potential
isomers grows quickly as a function of the number of sugar
residues. For example, Table VII shows six different structures
with identical masses that are built from the same five sugar
residues. Taking into account the number of isomers and the
large sulfated oligosaccharides that have not been characterized in detail, it is likely that the MG1 mucin carries several
hundred different oligosaccharide structures.
The functions of salivary mucins have been reviewed
(Tabak, 1995). In addition to well-recognized roles in alimentation, they form the pellicle that coats the resident hard (Fisher
et al., 1987; Al-Hashimi and Levine, 1989) and soft (Slomiany
et al., 1989) oral tissues. Here, mucins interact with cells that
traffic into and out of the oral cavity. It has been appreciated
for some time that microorganisms attach to host glycans
(Karlsson, 1989). Together, the results of our recent work
show a high degree of diversity in expression of glycans by
MG1, few of which are shared with MG2. The end result is an
enormous repertoire of potential binding sites for microorganisms in the oral cavity. A few of these interactions have
been described. For example, the observation that MG1
interacts with the gastrointestinal pathogen H. pylori (Veerman
et al., 1997) is consistent, first, with reports that this organism
adheres via saccharides with Leb epitopes (Boren et al., 1993),
and second, with the evidence presented here that MG1 carries
these substituents. With regard to oral bacterial species,
Gibbons and Qureshi (1978) found that several strains of
Streptococcus mutans bound to a blood group-reactive mucin
in whole saliva, presumably MG1. Finally, there are also
reports that Haemophilus influenza strains bind to MG1
(Veerman et al., 1995).

Our previous work also showed that the salivary mucins

(e.g., MG1 and MG2) are L-selectin ligands (Prakobphol et al.,
1998), a possibility that was suggested by our detection of Lex
and sLex epitopes among the low-molecular-weight salivary
mucin carbohydrate structures. Subsequently, we proved the
functional significance of this interaction, as additional experiments showed that MG2 can mediate leukocyte rolling and
tethering in a parallel plate flow chamber that assays adhesion
as a function of shear stress (Prakobphol et al., 1999). Because
MG1 oligosaccharides also carry Lex epitopes, as well as
complicated sulfate-containing glycans, it is possible that the
high-molecular-mass salivary mucin, like MG2, can mediate
leukocyte adhesion under flow. Although the significance of
the leukocyte interactions with mucins is not yet known, it is
interesting to note that bacteria can adhere via sugar sequences
other than Lex and sLex, suggesting that mucins may coordinate
adhesion of these different cell types, and thereby play a role in
the immunological processes modulating the oral ecology
(Prakobphol et al., 1999).
Finally, information about salivary mucin structure will
allow us to determine whether alterations in the oligosaccharide repertoire are associated with particular disease
states. Although the absence of saliva has devastating
consequences for oral health, little is known about the role of
individual salivary components, including the mucins. Content
and structure would appear to be particularly relevant. Mucin
content of human saliva obtained from older patients (6585
years) is significantly lower than that of younger individuals
(1835 years; Denny et al., 1991), raising the interesting
possibility that when levels of these molecules fall below a
critical threshold, certain age-related pathologies may ensue.
With regard to structure, several variations are possible. A
preliminary report suggests that MUC5B, the gene that encodes
MG1, may show little polymorphism (Debailleul et al., 1998).
However, glycoforms of this mucin have been identified
11

K.A. Thomsson et al.

(Prakobphol et al., 1993; Bolscher et al., 1995; Thornton et al.,

1997), suggesting that glycosylation may be an important
variable. The results reported here give us the necessary background information to begin studies if individuals who express
certain oligosaccharide epitopes on specific mucins are
predisposed to either oral health or disease.
Material and methods
Isolation of MG1
Human MG1 was purified from SM/SL saliva, collected as the
ductal secretion, from a single donor with blood type O, using
the method described by Ramasubbu et al. (1991). The purified
mucin had the expected electrophoretic characteristics, that is,
appeared as a heterodisperse band with an Mr > 1,000,000.
Immunostaining of MG1 with monoclonal antibody PAN H2
Serial dilutions of purified MG1 (0.16 mg/300 l) and, as a
negative control, guanidinium chloride insoluble porcine small
intestine mucin (solubilized by reduction), were immobilized
on polyvinylidene difluoride membrane strips. The samples
were partially deglycosylated by treatment with a mixture of
trifluoromethanesulfonic acid (5 ml) and toluene (300 l) for
5 h at 20C, followed by rinsing with 25 mM TrisHCl, pH 8,
and distilled water. Then the membranes were incubated with
2% bovine serum albumin (BSA) overnight at room temperature
before exposure to the monoclonal antibody PAN H2 (Nielsen
et al., 1997) diluted (1:100 v/v) in 2% BSA/phosphate buffered
saline for 2 h at room temperature. After washing, the bound
IgG was detected by incubation with goat anti-mouse IgG
conjugated to alkaline phosphatase (Dako, Denmark) and
visualized by developing with nitro-blue tetrazolium and
bromo-chloro-indolyl-phosphate (p-toluidine salt).
Amino acid analysis
Amino acid analysis was performed as described elsewhere
(Spackman et al., 1958) on an Alpha Plus amino acid analyzer
(Pharmacia Biotech, Uppsala, Sweden).
Release of oligosaccharides
The O-linked oligosaccharides were released from MG1
(4.5 mg) by reductive -elimination in 0.05 M potassium
hydroxide and 1 M sodium borohydride (Carlstedt et al.,
1993). The total mixture of saccharides was desalted on
Dowex 50WX8, followed by co-distillation of borate methyl
ester with MeOH/1% acetic acid to remove borate.
Proton NMR spectroscopy
The released and desalted oligosaccharide mixture was
analyzed by 1D and 2D DQF-COSY and TOCSY proton
NMR. Deuterium exchange was performed twice in 0.5 ml
99.95% D2O (30 min) followed by 0.5 ml 99.98% D2O
(30 min). The sample was lyophilized between each exchange.
The oligosaccharides were transferred to a 5-mm NMR tube
and analyzed with a Varian Innova 600 MHz instrument [5 mm
1H (13C,X) triple resonance pulsed field gradient probe] at
25C, using acetone as an internal standard. The NMR spectra
were interpreted by comparison with spectral data published
for a large number of O-linked glycans and other relevant
saccharides. Overviews are given by Kamerling and Vliegenthart
12

(1992) and in Sugabase, a database available on the Web

(www.boc.chem.ruu.nl/sugabase/sugabase.html).
Fractionation of oligosaccharides
After NMR analyses, the oligosaccharides were separated into
neutral, sialic acidcontaining, and sulfate-containing fractions
(Karlsson et al., 1995). Briefly, the released and desalted
oligosaccharides were applied to an anion exchange column,
and the neutral oligosaccharides were eluted with MeOH.
After an on-column derivatization of the carboxyl groups to
methyl esters, the sialylated species were eluted with MeOH
and the sulfated species were eluted with pyridinium acetate.
The methyl esters of the sialylated oligosaccharides were
converted to methyl amides (Karlsson et al., 1995).
Monosaccharide composition analyses
Monosaccharides were analyzed as described (Karlsson and
Hansson, 1995), by using a Dionex GP40 pump with an ED40
pulsed amperometric detector (Dionex, Sunnyvale, CA).
NeuAc was quantified after reversion of the methyl esterified
residues to carboxyl groups by saponification in 0.1 M sodium
hydroxide (50C, 2 h). After neutralization, the samples were
hydrolyzed in 100 l of 0.1 M HCl at 80C or 1 h and analyzed
by high pH anion exchange chromatography, using pulsed
amperometric detection with muraminic acid as an internal
standard. The column (CarboPac PA1, 4 250 mm; Dionex)
with a guard column (CarboPac PA1, 4 50 mm) was eluted
isocratically at 1 ml/min with 0.1 M sodium acetate and 0.05 M
sodium hydroxide (from a 50% solution, J.T. Baker, The
Netherlands).
MS
The neutral and sialylated oligosaccharides from MG1 and
MG2 were permethylated and analyzed with MALDI-TOF MS
as previously described (Prakobphol et al., 1998). Prior to
analysis by high-temperature GC and GCelectron impact MS
(Karlsson et al., 1995), the permethylated oligosaccharides of
MG1 were fractionated into low- (17 monosaccharide
residues) and high-molecular-mass species ( 7 residues) by
gel chromatography on a Sephadex LH20 column eluted with
MeOH (Thomsson et al., 1998).
Gel chromatography of sulfated oligosaccharides
The high-performance liquid chromatographysystem consisted
of a Pharmacia-LKB 2150 pump and a Pharmacia 2212
Helirac fraction collector. The sulfated oligosaccharides were
dissolved in 200 l H2O before injection. The sample was
applied to two serially connected Superdex peptide columns
HR 30/10 (1 cm 30 cm; Amersham Pharmacia Biotech) and
eluted with 25 mM ammonium bicarbonate at 0.2 ml/min.
Fractions (0.5 ml) were collected and pooled, and their
monosaccharide composition was determined.

Acknowledgments
This work was supported by the Swedish Medical Research
Council (7461), the Swedish NMR center, the Glycoconjugates in Biological Systems program sponsored by the
Swedish Foundation for Strategic Research, the IngaBritt and
Arne Lundbergs Foundation, and the National Institutes of

Glycosylation of the MG1 salivary mucin

Health (DE07244). The mass spectrometers were obtained by

grants from F.R.N. and the Knut and Alice Wallenberg
Foundation. Charlotta Damberg, Tomas Larsson, and Hasse
Karlsson are acknowledged for technical assistance. We also
thank Nancy J. Phillips and Karoline Scheffler for helpful
discussion.
Abbreviations
BSA, bovine serum albumin; DQF-COSY, double quantumfiltered correlated spectroscopy; ESI-MS/MS, nano-electrospray ionization tandem mass spectrometry; GC, gas chromatography; MALDI-TOF MS, matrix-assisted laser desorption/
ionization time-of-flight mass spectrometry; NMR, nuclear
magnetic resonance; MS, mass spectrometry; SM/SL, submandibular/sublingual; TOCSY, total correlated spectroscopy.
References
Al-Hashimi, I. and Levine, M.J. (1989) Characterization of in vivo salivary-derived
enamel pellicle. Arch. Oral Biol., 34, 289295.
Bergey, E.J., Cho, M.I., Blumberg, B.M., Hammarskjold, M.L., Rekosh, D.,
Epstein, L.G., and Levine, M.J. (1994) Interaction of HIV-1 and human
salivary mucins. J. Acquir. Immune Defic. Syndr., 7, 9951002.
Bobek, L.A., Tsai, H., Biesbrock, A.R., Levine, M.J. (1993) Molecular
cloning, sequence, and specificity of expression of the gene encoding the
low molecular weight human salivary mucin (MUC7). J. Biol. Chem., 268,
2056320569.
Bolscher, J.G.M., Groenink, J., van der Kwaak, J.S., van den Keijbus, P.A.M.,
vant Hof, W., Veerman, E.C.I., and Nieuw Amerongen, A.V. (1999)
Detection and quantification of MUC7 in submandibular, sublingual,
palatine and labial salive by anti-peptide antiserum. J. Dent. Res., 78,
13621369.
Bolscher, J., Veerman, E., Nieuw Amerongen, A.V., Tulp, A., and Verwoerd, D.
(1995) Distinct populations of high-M(r) mucins secreted by different
human salivary glands discriminated by density-gradient electrophoresis.
Biochem. J., 309, 801806.
Boren, T., Falk, P., Roth, K.A., Larson, G., and Normark, S. (1993)
Attachment of Helicobacter pylori to human gastric epithelium mediated
by blood group antigens. Science, 262, 18921895.
Carlstedt, I., Herrmann, A., Karlsson, H., Sheehan, J., Fransson L.-., and
Hansson, G.C. (1993) Characterization of two different glycosylated
domains from theinsoluble mucin complex of rat small intestine. J. Biol.
Chem., 268, 1877118781.
Debailleul, V., Laine, A., Huet, G., Mathon, P., dHooghe, M.C., Aubert, J.P.,
and Porchet, N. (1998) Human mucin genes MUC2, MUC3, MUC4,
MUC5AC, MUC5B, and MUC6 express stable and extremely large
mRNAs and exhibit a variable length polymorphism. J. Biol. Chem., 273,
881890.
Denny, P.C., Denny, P.A., Klauser, D.K., Hong, S.H., Navazesh, M., and
Tabak, L.A. (1991) Age-related changes in mucins from human whole
saliva. J. Dent. Res., 70, 13201327.
Desseyn, J.-L., Guyonnet-Deprat, V., Porchet, N., Aubert, J.-P., and Laine, A.
(1997a) Human mucin gene MUC5B, the 10.7-kb large central exon
encodes various alternate subdomains resulting in a super-repeat. J. Biol.
Chem., 272, 31683178.
Desseyn, J.-L., Aubert, J.P., van Seuningen, I., Porchet, N., and Laine, A.
(1997b) Genomic organization of the 3 region of the human mucin gene
MUC5B. J. Biol. Chem., 272, 1687316883.
Desseyn, J.L., Buisine, M.P., Porchet, N., Aubert, J.P., and Laine, A. (1998)
Genomic organization of the human mucin gene MUC5B-cDNA and
genomic sequences upstream of the large central exon. J. Biol. Chem., 273,
3015730164.
Fisher, S.J., Prakobphol, A., Kajisa, L., and Murray, P.A. (1987) External
radiolabelling of components of pellicle on human enamel and cementum.
Arch. Oral Biol., 32, 509517.
Gibbons, R.J. and Qureshi, J.V. (1978) Selective binding of blood groupreactive salivary mucins by Streptococcus mutans and other oral
organisms. Infect. Immunol., 22, 665671.

Gillece-Castro, B.L., Prakobphol, A., Burlingame, A.L., Leffler, H., and

Fisher, S.J. (1991) Structure and bacterial receptor activity of a human
salivary proline-rich glycoprotein. J. Biol. Chem., 266, 1735817368.
Kamerling, J.P. and Vliegenthart, J.F.G. (1992) High-resolution 1-H-nuclear
magnetic resonance spectroscopy of oligosaccahride-alditols released from
mucin-type O-glycoproteins. In Berliner, L., and Reben, J. (eds.), Biological
magnetic resonance, vol. 10. New York: Plenum Press, pp. 1287.
Karlsson, H., Carlstedt, I., and Hansson, G.C. (1989) The use of gas
chromatography and gas chromatography-mass spectrometry for the
characterization of permethylated oligosaccharides with molecular mass
up to 2300. Anal. Biochem., 182, 438446.
Karlsson, K.-A. (1989) Animal glycosphingolipids as membrane attachment
sites for bacteria. Annu. Rev. Biochem., 58, 309350.
Karlsson, N.G. and Hansson, G.C. (1995) Analysis of monosaccharide
composition of mucin oligosaccharide alditols by high-performance anionexchange chromatography. Anal. Biochem., 224, 538541.
Karlsson, N.G., Karlsson, H., and Hansson, G.C. (1995) Strategy for the
investigation of O-linked oligosaccharides from mucins based on the
separation into neutral, sialic acid- and sulfate-containing species. Glycoconj. J., 12, 6976.
Klein, A., Carnoy, C., Wieruszeski, J.-M., Strecker, G., Strang, A.-M.,
van Halbeek, H., Roussel, P., and Lamblin, G. (1992) The broad diversity
of neutral and sialylated oligosaccharides derived from human salivary
mucins. Biochemistry, 31, 61526165.
Levine, M.J., Reddy, M.S., Tabak, L.A., Loomis, R.E., Bergey, E.J.,
Jones, P.C., Cohen, R.E., Stinson, M.W., and Al-Hashimi, I. (1987)
Structural aspects of salivary glycoproteins. J. Dent. Res., 66, 436441.
Loomis, R.E., Prakobphol, A., Levine, M.J., Reddy, M.S., and Jones, P.C.
(1987) Biochemical and biophysical comparison of two mucins from
human submandibular-sublingual saliva. Arch. Biochem. Biophys., 258,
452464.
Nielsen, P.A., Bennett, E.P., Wandall, H.H., Therkildsen, M.H., Hannibal, J.,
and Clausen, H.H (1997) Identification of a major human high molecular
weight salivary mucin (MG1) as tracheobronchial mucin MUC5B. (1997)
Glycobiology, 7, 413419.
Nielsen, P.A., Mandel, U., Therkildsen, M.H., and Clausen, H. (1996)
Differential expression of human high-molecular-weight salivary mucin
(MG1) and low-molecular-weight salivary mucin (MG2). J. Dent. Res., 75,
18201826.
Nieuw Amerongen, A.V., Bolscher, J.G.M., and Veerman, E.C.I. (1995)
Salivary mucins: protective functions in relation to their diversity. Glycobiology, 5, 733740.
Perez-Vilar, J. and Hill, R.L. (1999) The structure and assembly of secreted
mucins. J. Biol. Chem., 274, 3175131754.
Prakobphol, A., Leffler, H., and Fisher, S.J. (1993) The high-molecular weight
human Mucin is the primary salivary carrier of ABH, Le(a), and Le(b)
blood group antigens. Crit. Rev. Oral Biol. Med., 4, 325333.
Prakobphol, A., Thomsson, K.A., Hansson, G.C., Rosen, S.D., Singer, M.S.,
Phillips, N.J., Medzihradszky, K.F., Burlingame, A.L., Leffler, H., and
Fisher, S.J. (1998) Human low-molecular-weight salivary mucin expresses
the sialyl Lewis x determinant and has L-selectin ligand activity.
Biochemistry, 37, 49164927.
Prakobphol, A., Tangemann, K., Rosen, S.D., Hoover, C.I., Leffler, H., and
Fisher, S.J. (1999) Separate oligosaccharide determinants mediate
interactions of the low-molecular-weight salivary mucin with neutrophils
and bacteria. Biochemistry, 38, 68176825.
Ramasubbu, N., Reddy, M.S., Bergey, E.J., Haraszthy, G.G., Soni, S.-D., and
Levine, M.J. (1991) Large-scale purification and characterization of the
major phosphoproteins and mucins of human submandibular-sublingual
saliva. Biochem. J., 280, 341352.
Reddy, M.S., Levine, M.J., and Prakobphol, A. (1985) Oligosaccharide
structure of the low-molecular-weight salivary Mucin from a normal
individual and one with cystic fibrosis. J. Dent. Res., 64, 3336.
Sangadala, S., Bhat, U.R., Mendicino, J. (1993) Structures of sulphated
oligosaccharides in human trachea Mucin glycoprotein. Mol. Cell.
Biochem., 126, 3747.
Slomiany, B.L., Murty, V.L., Mandel, I.D., Zalesna, G., and Slomiany, A.
(1989) Physico-chemical characteristics of mucus glycoproteins and lipids
of the human oral mucosal mucus coat in relation to caries susceptibility.
Arch. Oral Biol., 34, 229237.
Spackman, D.H., Stein, W.H., and Moore, S. (1958), Automatic recording
apparatus for use in the chromatography of amino acids. Anal. Chem., 30,
11901206.

13

K.A. Thomsson et al.

Tabak, L.A. (1995) In defense of the oral cavity: structure, biosynthesis and
function of salivary mucins. Annu. Rev. Physiol., 57, 547564.
Thomsson, K.A., Karlsson, N.G., Carlstedt, I., Karlsson, H., and Hansson, G.C.
(1998) Different O-glycosylation of respiratory mucin glycopeptides from a
patient with cystic fibrosis. Glycoconj. J., 15, 823833.
Thornton, D.J., Howard, M., Khan, N, and Sheehan, J.K. (1997) Identification
of two glycoforms of the MUC5B mucin in human respiratory mucus
evidence for a cysteine-rich sequence repeated within the molecule. J. Biol.
Chem., 272, 95619566.
Thornton, D.J., Khan, N., Mehrotra, R., Howard, M., Veerman, E.,
Packer, N.H., and Sheehan, J.K. (1999) Salivary mucin MG1 is comprised
almost entirely of different glycosylated forms of the MUC5B gene
product. Glycobiology, 9, 293302.

14

Veerman, E.C.I., Bank, C.M.C., Namavar, F., Appelmelk, B.J., Bolscher, J.G.M.,
and Nieuw Amerongen, A.W. (1997) Sulfated glycans on oral mucin as
receptors for Helicobacter pylori. Glycobiology, 7, 737743.
Veerman, E.C.I., Ligtenberg, A.J., Schenkels, L.C., Walgren-Weterings, E., and
Nieuw Amerongen, A.W. (1995) Binding of human high-molecular-weight
salivary mucins (MG1) to Hemophilus parainfluenzae. J. Dent. Res., 74,
351357.
Veerman, E.C.I., Van den Keybus, P.A.M., Vissink, A., and
Nieuw Amerongen, A.V. (1996) Human glandular salivas: their separate
collection and analysis. Eur. J. Oral Sci., 104, 346352.
Wickstrm, C., Davies, J., Eriksen, G.V., Veerman, E.C.I., and Carlstedt, I.
(1998) MUC5B is a major gel-forming, oligomeric mucin from human
salivary gland, respiratory tract and endocervix: identification of glycoforms and C-terminal cleavage. Biochem. J., 334, 685693.