Glycobiology vol. 10 no. 4 pp. 339–346, 2000
The influence of glucan polymer structure and solution conformation on binding to
(1→3)-β-D-glucan receptors in a human monocyte-like cell line
Antje Mueller2,6, John Raptis2,6, Peter J.Rice3,6, John
H.Kalbfleisch4,6, Robert D.Stout5,6, Harry E.Ensley7,
William Browder2 and David L.Williams1,2,6
Departments of 2Surgery, 3Pharmacology, 4Medical Education,
5Microbiology, and 6Immunopharmacology Research Group, James H.
Quillen College of Medicine, East Tennessee State University, Johnson City,
TN 37614–0575, USA and 7Department of Chemistry, Tulane University,
New Orleans, LA 70115, USA
Received on December 21, 1998; revised on November 1, 1999; accepted on
November 7, 1999
Glucans are (1–3)-β-D-linked polymers of glucose that are
produced as fungal cell wall constituents and are also
released into the extracellular milieu. Glucans modulate
immune function via macrophage participation. The first
step in macrophage activation by (1–3)-β-D-glucans is
thought to be the binding of the polymer to specific macrophage receptors. We examined the binding/uptake of a
variety of water soluble (1–3)-β-D-glucans and control polymers with different physicochemical properties to investigate the relationship between polymer structure and
receptor binding in the CR3- human promonocytic cell
line, U937. We observed that the U937 receptors were
specific for (1→3)-β-D-glucan binding, since mannan,
dextran, or barley glucan did not bind. Scleroglucan exhibited the highest binding affinity with an IC50 of 23 nM, three
orders of magnitude greater than the other (1→3)-β-Dglucan polymers examined. The rank order competitive
binding affinities for the glucan polymers were scleroglucan>>>schizophyllan > laminarin > glucan phosphate >
glucan sulfate. Scleroglucan also exhibited a triple helical
solution structure (ν = 1.82, β = 0.8). There were two
different binding/uptake sites on U937 cells. Glucan phosphate and schizophyllan interacted nonselectively with the
two sites. Scleroglucan and glucan sulfate interacted preferentially with one site, while laminarin interacted preferentially with the other site. These data indicate that U937
cells have at least two non-CR3 receptor(s) which specifically interact with (1→3)-β-D-glucans and that the triple
helical solution conformation, molecular weight and charge
of the glucan polymer may be important determinants in
receptor ligand interaction.
Key words: binding/glucan/macrophage/receptor/polymer
1To whom correspondence should be addressed at: Department of Surgery,
James H. Quillen College of Medicine, East Tennessee State University,
Johnson City, TN 37604–0575
© 2000 Oxford University Press
Introduction
Immunologically active glucans are (1→3)-β-D-linked glucose
polymers that occur as a primary component in the cell walls of
bacteria and fungi or are secreted extracellularly by various
fungi (Williams et al., 1996). These glucose polymers can exist
as a nonbranched (1→3)-β-linked backbone or as a (1→3)-βlinked backbone with (1–6)-β-branches (Ensley et al., 1994;
Lowman et al., 1998). Glucans have been reported to stimulate
immunity and decrease infectious complications in humans
(Browder et al., 1990; Babineau et al., 1994a,b) and experimental animals (Williams et al., 1996). By way of example,
Browder et al. (1990), Felippe et al. (1993), and Babineau et
al. (1994a,b) have reported that (1–3)-β-D-glucans decrease
the incidence of infections and septic sequelae in trauma and
surgical patients. However, the underlying cellular and
molecular mechanisms by which (1–3)-β-D-glucans induce
protection have not been defined. The first step in the modulation of cellular activity by (1–3)-β-D-glucans is thought to
involve binding to a specific receptor (Mueller et al., 1996;
Battle et al., 1998). We (Mueller et al., 1996; Battle et al.,
1998) and others (Thornton et al., 1996; Vetvicka et al., 1996,
1997) have reported receptor binding of (1→3)-β-D-glucans in
both murine and human cell lines. Similar results were
obtained in both systems. Recent data also indicate that
binding of glucans to macrophage and neutrophil cell lines will
stimulate the activation and nuclear binding activity of nuclear
factor-κB (NFκB) and nuclear factor interleukin 6 (NF-IL6)
which may explain, in part, the immunomodulatory activity of
these natural product receptor ligands (Adams et al., 1997;
Battle et al., 1998).
Several reports suggest that specific physicochemical
parameters, such as primary structure, solution conformation,
molecular weight and/or polymer charge may play a role in
determining whether and with what affinity (1→3)-β-Dglucans bind to macrophage receptor(s) and modulate immune
function. However, the relationship between (1–3)-β-D-glucan
physicochemical parameters and receptor ligand interaction
has not been defined. This was due in part to the lack of wellcharacterized (1–3)-β-D-glucan polymers with varying molecular weights and conformational structures.
We have demonstrated that aqueous SEC/MALLS/DV can
be employed to establish molecular mass moments, r.m.s. radii
and polydispersity of water soluble (1–3)-β-D-glucan biological response modifiers (Mueller et al., 1995). In addition, we
have reported on the application of SEC/MALLS/DV to establish the relationship between molecular mass and polymer size
(Mark-Houwink and ν values) in order to gain insights into the
solution structure of (1–3)-β-D-glucans (Mueller et al., 1995).
The purpose of this investigation was to characterize a variety
339
A.Mueller et al.
Table I. Monosaccharide composition and predominant intrachain glycosidic
linkages in polysaccharide ligands evaluated for binding to the human
macrophage/monocyte (1→3)-β-D-glucan receptors
Polysaccharide
exception of mannan which is composed of mannose monosaccharides. The carbohydrate polymers differ in mass, size, and
viscosity (Table II). With the exception of scleroglucan, all of
the polymers exhibit relatively uniform polydispersities (I =
<1.7) (Table II). The sites of phosphorylation in glucan phosphate are limited to C-2 and C-6 and the degree of phosphorylation is <1 phosphate group/7 glucose subunits (Lowman et
al., 1998).
Monosaccharide
Type of glycosidic linkage
Glucan phosphate
Glucose
β-(1→3)
Glucan sulfate
Glucose
β-(1→3)
Schizophyllan (SPG)
Glucose
β-(1→3), (1–6) a
Laminarin
Glucose
β-(1→3), (1–6) a
Solution conformation of carbohydrate polymers
Scleroglucan
Glucose
β-(1→3),
Barley glucan
Glucose
β-(1→4), β-(1→3)b
Dextran
Glucose
α-(1→6)
Mannan
Mannose
α-(1→4)
The slope of the linear relationship between log intrinsic
viscosity and log molecular mass ([η] = Kα· Mα) is known as
the Mark-Houwink or α-value for a polymer system (Mueller
et al., 1995). The slope of the linear relationship between the
log of the root mean square radius and log of the molecular
mass moment (RG = Kν· Mν) has been termed “ν”(Mueller et
al., 1995). Establishing α and ν may provide insights into the
polymer solution conformation (Mueller et al., 1995). The α
and ν values for the carbohydrate polymers are shown in Table
II. In some cases it was not possible to establish both scaling
relationships due to the narrow distribution and/or undetectable rmsz values of the polymer. The data indicate that scleroglucan has the highest ν and α values of 1.82 and 0.80,
respectively. The data suggest that scleroglucan has the most
rigid solution conformation with the other polymer systems
exhibiting increasing solution flexibility (less rigid solution
conformation) in the order of schizophyllan > glucan phosphate > laminarin = glucan sulfate = mannan > barley glucan >
dextran (Table II).
aIndicates
bIndicates
(1–6) a
the presence of (1–6)-β glycosidic side chain branches.
a mixed linkage glucan backbone.
of (1–3)-β-D-glucan and non-glucan polymers in order to accurately establish molecular mass moments and to gain insights
into the solution conformation and compare and contrast the
effect of glucan polymer structure and conformation on
receptor binding affinity.
Results
Characterization of carbohydrate polymers
The monosaccharide composition and the predominant glycosidic linkage of the eight polymers employed in this study are
listed in Table I. Specific refractive index increments, average
molecular mass moments, rmsz moments, intrinsic viscosities,
and polydispersities were established for each carbohydrate
polymer. The data are presented in Table II. In agreement with
previous reports (Pretus et al., 1991; Williams et al., 1991a;
Ensley et al., 1994; Lowman et al., 1998), the carbohydrate
polymers are composed of glucose monosaccharides, with the
13C-n.m.r.
analysis of polymers
Figure 1 shows representative 13C-n.m.r. analyses of three
(1→3)-β-D-glucans employed in the competition studies.
Carbon assignments are given above each of the major glucose
peaks. We observed six major peaks associated with the
glucose polymer backbone of the (1→3)-β-D-glucans (Figure
1). Curdulan, a nonbranched single helical (1→3)-β-D-glucan,
was selected as the glucan polymer control. The curdulan
Table II. Physicochemical parameters of water soluble polysaccharide ligands as determined by size exclusion chromatography/multi-angle laser light scattering
photometry/differential viscometry
Polysaccharide
Dn/dca (ml/g) Molecular mass
Mw (g/mol)
Glucanphosphate
0.122
1.57 × 105
Glucan sulfate
Schizophyllan (SPG)
Polydispersity I
(Mw/Mn)
Radius of the center of Intrinsic viscosity
gravity rz (nm)
[η] dl/g
νb
αc
0.302
0.65
1.67
20.3
0.182
3.70 ×
104
1.07
n.d.d
0.16
n.d.
n.d.
0.253
3.06 × 105
1.08
37.3
7.32
0.599
n.d.
Laminarin
0.164
7.70 × 103
1.17
n.d.
0.07
n.d.
0.52
Scleroglucan
0.140
1.02 × 106
3.21
35.4
1.08
1.820
0.80
Barley glucan
0.175
1.91 × 105
1.49
29.6
2.18
0.193
0.49
Dextran
0.186
7.30 × 104
1.16
18.5
0.23
0.1
n.d.
0.171
6.84 ×
1.05
n.d.
0.25
0.2
n.d.
Mannan
adn/dc,
104
0.33
Refractive index increment.
Slope of the linear relationship between the log of the root mean square radius and log of the molecular mass moment (RG = K ν·Mν) has been termed “ ν”
(Mueller et al., 1995).
cα, Slope of the linear relationship between log intrinsic viscosity and log molecular mass ([η] = Kα·Mα) is known as the Mark-Houwink or α-value for a polymer
system (Mueller et al., 1995).
dn.d., Not detectable.
bυ,
340
Influence of glucan structure on binding to (1→3)-β-D-glucan receptors
Fig. 2. Competitive displacement of tritiated glucan phosphate by unlabeled
glucan sulfate, glucan phosphate, schizophyllan, laminarin and scleroglucan.
The data are expressed as mean ± SEM and represent at least four replicates
with 4–8 data points/concentration/replicate.
Table III. Inhibitory concentration (IC50) of different (1→3)-β-D -glucans on
the binding of glucan phosphate to U937
Fig. 1. The 13C-n.m.r. spectra of four different glucans are presented. The peak
assignments are presented above the six major carbon peaks. Curdulan was
employed as a non-branched (1→3)-β-D-glucan standard. Glucan phosphate is
also a non-branched (1→3)-β-D-glucan polymer. Scleroglucan is a branched
(1:3) polymer. The scleroglucan spectrum shows two clearly defined patterns
of glucosyl signals. The major pattern appearing as C-1, C-3, C-5, C-2, C-4,
and C-6 represent the (1→3)-β-D-glucan backbone in the polymer chain. The
set of narrow line-width peaks corresponds to the side chain glucosyl units.
Laminarin is also branched, but the degree of branching (1:10) is much less
than that observed for scleroglucan. All samples were dissolved in DMSO-d6.
spectra show six clearly defined carbon peaks which are diagnostic for (1→3)-β-D-glucans (Williams et al., 1991a). Glucan
phosphate, a predominantly nonbranched single helical polyelectrolyte, also shows the characteristic (1→3)-β-D-glucan
pattern (Williams et al., 1991a). In agreement with previous
results (Pretus et al., 1991), the 13C-n.m.r. spectrum of scleroglucan shows two distinct patterns of glucosyl signals. The
oligosaccharide backbone appears as a set of five broad
signals, which correspond well with the curdulan spectrum
(Figure 1). The prominent C-3 triplet of scleroglucan is indicative of a highly branched (1→3)-β-D-glucan polymer. The
triplet indicates C-3 carbons in three different magnetic environments, i.e., the C-3 carbon in the glucose monomers of the
backbone, the C-3 carbon in the glucose monomers of the
backbone which have (1→6)-β-linked branches and the C-3
carbon in the (1→6)-β-linked glucose branch. The set of
narrow line-width peaks correspond to the C-2 and C-4
carbons of the side chain glucosyl units (Pretus et al., 1991).
The scleroglucan spectrum is consistent with a (1→6)-β side
chain branch on average every third glucose subunit along the
polymer backbone (Pretus et al., 1991). Laminarin is a wellcharacterized (1→3)-β-D-glucan with (1→6)-β-side chain
Glucan
IC50 ( µM)
Displacement (%) p value
Glucan phosphate
35
100
Glucan sulfate
43
37
<0.01
40
<0.005b,c
100
<0.01b,d
57
<0.01b
Scleroglucan
0.023
Schizophyllan (SPG) 11
Laminarin
21
—a
aGlucan phosphate binding was employed as the control (displacement) to
which all other glucans were compared.
bp < 0.05 vs. glucan phosphate and glucan sulfate.
cp < 0.05 vs. all other glucans.
dp < 0.05 vs. laminarin.
branching on average every tenth glucose subunit along the
polymer backbone (Williams et al., 1991a). The laminarin
13C-n.m.r. spectrum is in close agreement with previous reports
on this polymer (Williams et al., 1991a).
Competition of carbohydrate polymers for glucan binding/
uptake sites in U937 cells
We evaluated the competition of unlabeled carbohydrate polymers for binding/uptake using 3H-glucan phosphate. All of the
(1→3)-β-D-glucan polymers competed with 3H-glucan phosphate for binding and uptake into U937 cells. Barley glucan,
dextran, and mannan did not compete for binding with radiolabeled glucan phosphate. Representative competition binding
curves are shown in Figure 2. The unlabeled (1→3)-β-Dglucans competed with radiolabeled glucan phosphate with
properties that were characteristic of concentration-displacement curves. The IC50 values are shown in Table III. Scleroglucan had the highest affinity (IC50 = 23 nM) followed by
schizophyllan (IC50 = 11 µM), laminarin (IC50 = 21 µM),
glucan phosphate (IC50 = 35 µM), and glucan sulfate (IC50 = 43
µM). Competition by glucan phosphate and schizophyllan
341
A.Mueller et al.
Fig. 3. Failure of FITC-β-glucan to specifically bind U937 cells. U937 (1 ×
106) cells were incubated with 25 µg of a commercially available FITC labeled
water soluble β-glucan (Molecular Probes, Eugene, OR) for 45 min at 37°C,
5% CO2 tension prior to FACS analysis. The cells were washed five times with
isothermic RPMI-1640 medium without protein.
suggest that they interact with the same binding/uptake sites as
3H-glucan phosphate. However, there was competition for
fewer sites with scleroglucan (40%), glucan sulfate (37%), and
laminarin (57%) and these carbohydrates did not further
displace radiolabeled glucan phosphate binding at higher
concentrations (Figure 2 and Table III). These data are indicative of at least two (1→3)-β-D-glucan binding sites on U937.
The effects seen with scleroglucan at concentrations above 1
µM appear to be a solution viscosity artifact.
Absence of specific binding and competitive displacement of
FITC-labeled β-glucan in U937 cells
Previous reports indicate that commercially available FITClabeled (1→3)-β-D-glucan can be employed to assess receptor
binding of glucans using a flow cytometric approach (Ainsworth, 1994). We employed flow cytometry and a commercially available, water-soluble FITC-labeled β-glucan to
investigate binding to U937 in order to establish whether this
FITC labeled β-glucan was recognizing the same binding site
as the glucans employed in the present study. We observed that
the FITC-labeled glucan would stain U937 cells. However, the
FITC-labeled glucan was almost entirely removed from the
cells by washing three and five times with isothermic media
(Figure 3). This is indicative of non-specific staining. In addition, we observed no competition between the FITC-labeled βglucan (Molecular Probes, Eugene, OR) and any of the glucans
employed in the present study. To determine whether the lack
of FTIC-glucan binding was cell specific we conducted additional studies using other cell lines which were also known to
bind glucans, i.e., HL60 and K562. We were not able to identify specific binding of the FITC-labeled β-glucan to any cell
line tested (data not presented).
Discussion
Our data confirm and extend previous observations (Mueller et
al., 1996; Battle et al., 1998) by demonstrating not only the
specificity of the U937 receptor(s) for (1→3)-β-D-glucans, but
342
also the ability of the receptor(s) to differentiate between
(1→3)-β-D-glucans. Non-β-D-linked carbohydrate polymers
did not bind to the U937 cell line (i.e., mannan, dextran and
barley glucan). Within the group of (1→3)-β-D-glucan polymers examined, the receptor(s) showed differential affinity,
which appeared to be based primarily on solution conformation and to a lesser extent branching frequency, molecular
weight and polymer charge. We speculate that the rigid solution conformation (triple helix) is of greater importance in
receptor recognition of the polymer because the more highly
ordered (i.e., rigid) triple helical glucan structure is the
predominant form found in the cell wall of most fungi (Cabib
et al., 1993; Elorza et al., 1993; Kapteyn et al., 1995).
Glucan sulfate and glucan phosphate are polyelectrolytes
(Williams et al., 1991a,b). Laminarin, schizophyllan and
scleroglucan are neutral polysaccharides (Pretus et al., 1991;
Mueller et al., 1995; Williams et al., 1991a). Laminarin,
schizophyllan and scleroglucan exhibit varying degrees of
(1→6)-β side chain branching (Pretus et al., 1991; Williams et
al., 1991ab). Even though we were dealing with a small
number of well-characterized polymers there were several
distinct observations that emerged. The data are characterized
by very complex receptor ligand interactions which are indicative of multiple glucan binding sites. The binding/uptake of
radiolabeled glucan phosphate was characterized by a single
dissociation constant. Unlabeled glucan phosphate and schizophyllan competed for binding with characteristics of a single
ligand-receptor interaction. Competition of radiolabeled
glucan phosphate by scleroglucan and glucan sulfate could
only inhibit ∼40% of control binding/uptake, while laminarin
inhibited ∼60%. The results of the competition experiments are
consistent with a (1→3)-β-D-glucan binding interaction occurring at two different sites. Glucan phosphate and schizophyllan
interact nonselectively with the two sites. Scleroglucan and
glucan sulfate interact preferentially with one site (40% of the
glucan phosphate binding) while laminarin appears to interact
with the other site (60% of the glucan phosphate binding).
These sites may represent separate extracellular carbohydrate
receptors, binding sites or different saturable intracellular
compartments to which carbohydrates are directed.
Glucan sulfate and glucan phosphate showed significantly
lower binding affinities than did the neutral polysaccharides,
laminarin, schizophyllan, and scleroglucan. This suggests that
the presence of the charged species on the polymer may alter
binding affinity. All of the neutral polysaccharides exhibited
branching frequencies that were greater than the polyelectrolyte glucans. In general, the receptor showed significantly greater affinity for the branched neutral
polysaccharides. Schizophyllan and scleroglucan are similar in
that both of these polymers have a branching frequency of
approximately one branch per every third glucose subunit
along the (1→3)-β-D-linked polymer backbone (Pretus et al.,
1991). A number of reports suggest the bioactivity of (1→3)β-D-glucans is related to the degree of side chain branching
(Suzuki et al., 1988; Kurachi et al., 1990; Kiho et al., 1992;
Nemoto et al., 1993; Chiba et al., 1996). However, we
observed dramatic differences between the binding affinity of
schizophyllan (IC50 = 11 µM) and scleroglucan (IC50 = 23 nM)
even though their branching frequency was very similar. When
we compared the binding affinity of laminarin (1:10
branching) versus schizophyllan (1:3 branching), we observed
Influence of glucan structure on binding to (1→3)-β-D-glucan receptors
a modest (21 µM vs. 11 µM) but significant difference. Thus,
we conclude that branching frequency may enhance the
affinity of the polymer for the U937 glucan receptor. We also
noted that the polymers with the greatest molecular weight
(i.e., schizophyllan and scleroglucan) exhibited higher binding
affinities. Kojima et al. (Kojima et al., 1986) have reported that
the anti-tumor activity of schizophyllan is molecular weight
dependent. This may reflect differences in pharmacokinetics
rather than binding affinity. In our study the contribution of
molecular weight cannot solely account for the differences
since the binding affinity of schizophyllan was 11 µM and
scleroglucan was 23 nM. By far the most significant difference
in binding affinity appeared to strongly correlate with solution
conformation. In this case, solution conformation refers to the
tertiary structure which the polymer assumes in aqueous
media. We examined the solution conformation by establishing the linear scaling relationships for each glucan as
described by Mueller et al. (1995). Scleroglucan was unique in
that it had a highly ordered solution conformation. This indicates that the predominant solution conformation of scleroglucan is a rigid triple helix as compared to the other glucans
which show scaling relationships that suggest a single helical
solution conformation (Mueller et al., 1995). Thus, the
dramatic difference between the binding affinity of the
receptor for scleroglucan and the other glucan polymers seems
to involve the more rigid solution structure. However, we
cannot discount the effects of molecular mass on this polymer
system.
Kulicke et al. (1997) investigated the correlation between
immunological activity, molar mass and molecular structure of
various (1→3)-β-D-glucans. They studied the effect of glucans
on superoxide anion production and TNFα release from human
peripheral blood mononuclear cells (Kulicke et al., 1997).
Interestingly, they reported that low molar mass glucans
increased TNFα release and superoxide anion release when
compared to high molar mass glucans (Kulicke et al., 1997).
Further they stated that “helical structures were not essential”
or advantageous for induction of immunologic activity
(Kulicke et al., 1997). However, the molar mass ranges of the
glucans studied by Kulicke et al. were narrow, i.e., 1 × 105 to
>2 × 106 g/mol, the glucan concentrations employed in the in
vitro experiments was higher than reported elsewhere (Pretus
et al., 1991; Lowman et al., 1998), and the parameters evaluated were not specific. We examined a broader molecular
weight range of water soluble glucan polymers (103 to 106).
We also examined neutral and polyelectrolyte glucans and the
effect of branching frequency. More importantly, we employed
a specific ligand receptor interaction and noted that the triple
helical solution conformation of scleroglucan may be an
important determinant with regard to glucan ligand macrophage receptor interaction.
The nature of the glucan receptor(s) is unknown. Our data
clearly demonstrate the existence of at least two specific
glucan binding site(s) on undifferentiated CR3-U937 (Mueller
et al., 1996; Battle et al., 1998). We confirmed the lack of CR3
expression on the U937 cells employed in the present study. In
addition, we have shown that glucan ligand binding to the nonCR3-U937 receptor(s) will stimulate intracellular signaling
pathways which culminate in the activation, translocation and
nuclear binding of immunoregulatory and pro-inflammatory
transcriptional activator proteins (Battle et al., 1998). This
indicates that ligation of a non-CR3 receptors by glucan ligand
has functional consequences that are consistent with modulation of immune function. Michalek et al. have confirmed and
extended this observation by reporting that a proprietary
glucan (PPG-glucan) also binds to a site distinct from CR3
(Michalek et al., 1998). Whether PPG-glucan and the glucans
described in this study bind to the same site(s) is not known.
Thornton et al. (1996), Vetvicka et al. (1996), and colleagues
have reported a CR3 (CD11b/CD18) glucan binding site on
macrophages, neutrophils and NK cells. The glucan binding is
reported to be through one or more lectin sites located outside
the CD11b I domain (Thornton et al., 1996; Vetvicka et al.,
1996). Duan et al. (1994) have also reported a β-glucan
binding lectin on NK cells which contributes to NK cell
mediated cytotoxicity. Zimmerman et al. reported that lactosylceramide binds PPG-glucan and that this glycosphingolipid
may be a leukocyte glucan binding moiety (Zimmerman et al.,
1998). Dushkin et al. (1996), Vereschagin et al. (1998), and
colleagues have reported that a carboxymethylated glucan
binds to the macrophage scavenger receptor. Thus, there may
be multiple glucan binding sites on macrophages, neutrophils,
and NK cells. Additional studies are required to determine the
nature of the glucan receptor(s) and which receptor(s) are
essential to the expression of the various immunobiological
effects ascribed to (1→3)-β-D-glucans.
Previous reports in the literature indicate that FITC-labeled
glucans can be employed to assess receptor binding by flow
cytometry (Ainsworth, 1994). We examined a commercially
available FITC-labeled glucan ligand and were unable to document specific binding not only to U937, but also to HL-60,
K562, and J774a.1 cells (data not presented). Therefore, we
conclude that this FITC-glucan does not specifically stain/bind
to the glucan receptor(s). This is also supported by the fact that
we did not observe any competition by unlabeled glucans
following flow cytometric analysis (data not shown). In an
attempt to further address this issue, we prepared several FITC
derivatives of glucan. None of the fluoresceinated glucans
specifically bound to U937. The precise reasons for this failure
to document binding of the FITC-labeled glucans are not clear.
We speculate that FITC derivatization of glucan polymers
results in extensive substitutions along the polymer backbone,
probably at the C-6 hydroxyl, which may block specific
binding sites along the polymer.
Stahl (1992) and Fearon and Locksley (1996) have reviewed
the role of carbohydrate recognition by immunocyte receptors
as an important component of innate immune recognition.
These authors speculate that carbohydrate recognition may
have evolved as a component of innate immunity because
complex carbohydrates are common constituents of microbial
cell walls and membranes and are distinct from carbohydrates
produced in mammalian cells (Fearon and Locksley, 1996).
Indeed, Stahl has stated that complex carbohydrates may be
“ideal for specific recognition” in the induction of innate
immune responses (Stahl, 1992). In agreement with the
concepts of Stahl (1992) and Fearon and Locksley (1996), we
propose that (1→3)-β-D-glucan receptors may be important
components of the human innate immune response recognition
system. In support of this concept, Obayashi (Obayashi et al.,
1992; Obayashi, 1997) and others (Tamura et al., 1994, 1997;
Miyazaki et al., 1995) have reported high levels of glucans in
the serum of patients with systemic fungemia and deep tissue
343
A.Mueller et al.
mycoses. Further, they observed significant differences in
serum glucan levels between patients with fungal infections
and patients infected with bacteria which do not produce
glucans (Obayashi et al., 1992; Tamura et al., 1994, 1997;
Miyazaki et al., 1995; Obayashi, 1997). Since (1→3)-β-Dglucans are not produced in mammalian systems it is assumed
that the glucans detected in these patients was released from
the fungal cell wall. Okamoto et al. (1998) have recently
reported the presence of circulating glucans in patients with
Pneumocystis carinii infections. From the perspective of
innate immune response to infection these observations
suggest that the (1→3)-β-D-glucan receptor may be involved
in the recognition of circulating (1→3)-β-D-glucan and may
play a role in the immune response to systemic fungal and
perhaps parasitic infections.
In conclusion, these data indicate the existence of at least
two (1→3)-β-D-glucan receptors on undifferentiated CR3U937. We also observed that scleroglucan, glucan sulfate and
laminarin appear to interact selectively with a single site, while
other glucan phosphate and schizophyllan interact non-selectively with both sites. The data also indicate that the human
U937 (1→3)-β-D-glucan receptor(s) are specific for (1→3)-βlinked glucopyranoses and the receptor(s) exhibit differential
affinities for glucans with varying physical properties.
Whether the differences in binding affinity of the receptor(s)
relate to differences in expression of biological activity is not
presently known. However, we now have the well-characterized ligands to investigate this intriguing question.
obtained from Pharmacia (Piscataway, NJ). Barley glucan was
kindly provided by Dr. Peter Wood and Dr. Barry McCleary
(Megazyme Ltd., Sydney, Australia). The curdulan which was
used as a (1→3)-β-D-glucan polymer control was a gift from
Dr. Ragnar Rylander (Dept. of Environmental Medicine,
University of Gothenburg). SEC/MALLS/DV analysis of the
curdulan showed linear scaling relationships that were
consistent with a single helical conformation. Table I lists the
predominant glycosidic linkage and monosaccharide composition
of the polysaccharides. The primary structure of each carbohydrate polymer was confirmed by variable temperature FT13C-n.m.r. in DMSO
d6 at a concentration of 50 mg/ml as
described previously (Ensley et al., 1994; Lowman et al.,
1998). For the competitive displacement studies, stock
solutions of the polysaccharides were prepared in RPMI 1640
cell culture media, filtered, and subsequently diluted over a
concentration range. An FITC labeled β-glucan (Molecular
Probes, Eugene, OR) was employed for the flow cytometry
experiments.
Determination of the specific refractive index increment (dn/dc)
The dn/dc values were determined with an Optilab 903 interferometric refractometer (Wyatt Technology, Santa Barbara,
CA) at 25°C in 50 mM sodium nitrite mobile phase.
Characterization of water soluble polysaccharides by size
exclusion chromatography/multi-angle laser light scattering/
differential viscometry (SEC/MALLS/DV)
We used the human promonocytic cell line U937. This cell line
has been extensively utilized by our laboratory group in (1–3)β-D-glucan receptor binding studies (Mueller et al., 1996;
Battle et al., 1998). U937 was maintained in RPMI-1640
medium with 10% serum protein supplement at 37°C and 5%
CO2 tension. The lack of CR3 (CD11b/CD18) expression on
the cells employed for the binding studies was confirmed by
flow cytometry. Specifically, U937 cells were cultured and
harvested according to Mueller et al. (1996). To block nonspecific FcR binding of the primary antibody U937 cells were
treated with anti-FcRγII Mab or 0.1% serum albumin. Cells
were incubated with fluorescein conjugated antibodies to
CD11b and CD18. All antibodies were in FACS buffer and the
incubations were for 30 min on ice. Following staining the
cells were washed 2× in buffer containing 1% bovine serum
albumin and 1 × 104 cells were analyzed on a FACStar (Becton
Dickinson).
To establish molecular mass and size, polydispersity, (weightaverage molecular mass Mw/number-average molecular mass
Mn) and intrinsic viscosity, the polysaccharides were analyzed
by SEC/MALLS/DV as previously reported (Mueller et al.,
1995). The samples (∼3 mg/ml) were dissolved in 50 mM
sodium nitrite mobile phase. Three Ultrahydrogel SEC
columns (2000, 500, and 120 Waters Corp.) were connected in
series and the columns were maintained at 30°C with continuous flow of mobile phase. The system was calibrated using
narrow-band pullulan and dextran standards. The weightaverage molecular mass and the z-average radius of the center
of gravity as an index of molecular size of the samples were
determined by on-line MALLS photometry employing a
DAWN-DSP-argon-ion (488 nm) MALLS photometer
(Mueller et al., 1995). Intrinsic viscosity was determined by inline differential viscometry with a Viscotek model 200 differential viscometer (Viscotek, Houston, TX) (Mueller et al.,
1995). The Mark-Houwink or α-value for each polymer
system was established with Unical software (v. 4.03,
Viscotek, Houston, TX). The ν-value for each polymer system
was established with EASI software (v. 7.02, Wyatt Technology, Santa Barbara, CA) (Mueller et al., 1995).
Carbohydrate polymers
Radiolabeling of a water soluble (1–3)-β-D-glucan phosphate
We evaluated eight water soluble carbohydrate polymers.
Glucan phosphate and glucan sulfate were prepared from water
insoluble (1→3)-β-D-glucan, isolated from S.cerevisiae as
previously described (Williams et al., 1991a). Schizophyllan
(SPG, derived from S.commune) was obtained in sterile water
(10 mg/ml) from Kaken Chemical Co. (Tokyo, Japan). Laminarin and mannan were purchased from Sigma Chemical Co.
(St. Louis, MO). Water soluble scleroglucan was prepared
according to the protocol of Pretus et al. (1991). Dextran was
Water soluble (1–3)-β-D-glucan phosphate was radiolabeled as
previously described by our group (Mueller et al., 1996).
Briefly, ∼100 mg (1–3)-β-D-glucan phosphate was dissolved in
1.5 ml DMSO overnight at 45°C. This solution is added to a
vial containing tritiated NaB3H 4 (ICN Biomedicals Inc., Irvine,
CA; 25 mCi, 718 mCi/mmol). Since the tritium was introduced
by reduction of the reducing terminus of the glucan phosphate
polymer, a maximum of one tritium per glucan phosphate
polymer was introduced.
Materials and methods
Human cell line
344
Influence of glucan structure on binding to (1→3)-β-D-glucan receptors
Receptor-binding assays
Receptor binding was evaluated using the Millipore Multiscreen Assay System with 96-well-GF/C glass fiber filter
plates (Millipore Corp., Bedford, MA). Displacement binding
was determined in the present of a constant amount of radiolabeled ligand (15 µg/well) and increasing concentrations of
unlabeled polysaccharide. The total volume was 200 µl/well.
This volume was employed for all binding studies. After incubation at 37°C for 90 min, the plates were vacuum filtered and
washed five times with warm serum-free RPMI 1640. The
filters were then harvested and dried, and the radioactivity was
determined by liquid scintillation counting (LSC 1409 Wallac
Inc., Gaithersburg, MD) with a typical counting efficiency for
tritiated glucan phosphate of 45–50%.
Flow cytometry
U937 (1 × 106) cells were incubated with 25 µg of a commercially available FITC labeled water soluble β-glucan (Molecular Probes, Eugene, OR) for 45 min at 37°C, 5% CO2 tension
in a humidified environment. The cells were washed either
three or five times with isothermic RPMI-1640 without serum
protein. The control groups consisted of cells that were not
treated with labeled β-glucan or cells which were incubated
with the labeled glucan and not washed with media. Competitive displacement binding was determined in the present of a
constant amount of FITC-labeled ligand (25 µg) and increasing
concentrations of unlabeled polysaccharides. The cells were
harvested by centrifugation and suspended in isothermic
RPMI-1640 without serum proteins. Fluorescence was excited
using an argon-ion laser (488 nm); FITC emissions were
distinguished by passage of emitted light through a 560 nm
dichroic mirror and a 530/25 bandpass filter. Detection was
triggered by forward-angle light scatter signals.
Data analysis
Binding displacement data for (1→3)-β-D-glucans were
analyzed by unweighted non-linear regression using models of
one and two site competitive displacement (GraphPad Prism v.
2.1, San Diego, CA). Maximum (100%) binding was fixed to
that seen in the absence of competing carbohydrates and
nonspecific binding ≥0. Models were used to estimate the
apparent binding parameters (% binding, IC50) for each site.
Sequential F-testing was used to decide if a more complex
model (e.g., two vs. one site) was justified. Following a significant F-test in a 1-way ANOVA, IC50 values for different
glucan groups (Table III) were compared with the least significant difference procedure. A p value of ≤0.05 was considered
significant.
Acknowledgments
This work was supported, in part, by NIHGM535322 to DLW
and a VA Merit Review Grant to WB.
Abbreviations
α, Slope of the linear relationship between log intrinsic
viscosity and log molecular mass ([η] = Kα·Mα) is known as
the Mark-Houwink or α-value for a polymer system (Mueller
et al., 1995); ν, slope of the linear relationship between the log
of the root mean square radius and log of the molecular mass
moment (RG = Kα·Mα) has been termed “ν” (Mueller et al.,
1995); CR3, Type 3 complement receptor (CD11b/CD18);
DMSO, dimethyl sulfoxide; dn/dc, refractive index increment;
DV, differential viscometry detector; FITC, fluorescein isothiocyanate; IC50, inhibitory concentration for 50% competition;
MALLS, multi-angle laser light scattering detector; n.m.r.,
nuclear magnetic resonance; SEC, size exclusion chromatography.
References
Adams,D.S., Pero,S.C., Petro,J.B., Nathans,R., Mackin,W.M. and
Wakshull,E. (1997) PGG-Glucan activates NF-κB-like and NF-IL-6-like
transcription factor complexes in a murine monocytic cell line. J. Leukocyte Biol., 62, 865–873.
Ainsworth,A.J. (1994) A β-glucan inhibitable zymosan receptor on channel
catfish neutrophils. Vet. Immunol. Immunopathol., 41, 141–152.
Babineau,T.J., Hackford,A., Kenler,A., Bistrian,B., Forse,R.A., Fairchild,P.G., Heard,S., Keroack,M., Causha,P. and Benotti,P. (1994a) A Phase
II multicenter, double-blind, randomized, placebo-controlled study of
three dosages of an immunomodulator (pgg-glucan) in high-risk surgical
patients. Arch. Surgery, 129, 1204–1210.
Babineau,T.J., Marcello,P., Swails,W., Kenler,A., Bistrian,B. and Forse,R.A.
(1994b) Randomized phase I/II trial of a macrophage-specific immunomodulator (PGG-glucan) in high-risk surgical patients. Ann. Surg., 220,
601–609.
Battle,J., Ha,T., Li,C., Della Beffa,V., Rice,P., Kalbfleisch,J., Browder,W. and
Williams,D. (1998) Ligand binding to the (1→3)-β-D-glucan receptor
stimulates NFκB activation, but not apoptosis in U937 cells. Biochem. Biophys. Res. Commun., 249, 499–504.
Browder,W., Williams,D., Pretus,H., Olivero,G., Enrichens,F., Mao,P. and
Franchello,A. (1990) Beneficial effect of enhanced macrophage function
in the trauma patient. Ann. Surg., 211, 605–613.
Cabib,E., Mol,P.C., Shaw,J.A. and Choi,W.J. (1993) Biosynthesis of cell wall
and septum during yeast growth. Arch. Med. Res., 24, 301–303.
Chiba,N., Ohno,N., Terui,T., Adachi,Y. and Yadomae,T. (1996) Effect of
highly branched (1->3)-β-D-glucan, OL-2, on zymosan-mediated hydrogen peroxide production by murine peritoneal macrophages. Pharmaceutical Pharmacol. Lett., 6, 12–15.
Duan,X., Ackerly,M., Vivier,E. and Anderson,P. (1994) Evidence for involvement of beta-glucan-binding cell surface lectins in human natural killer cell
function. Cell Immunol., 157, 393–402.
Dushkin,M.I., Safina,A.F., Vereschagin,E.I. and Schwartz,Y.S. (1996) Carboxymethylated β-1,3-glucan inhibits the binding and degradation of
acetylated low density lipoproteins in macrophages in vitro and modulates
their plasma clearance in vivo. Cell Biochem. Funct., 16, 185–215.
Elorza,M.V., Garcia de la Cruz,F., San Juan,R., Marcilla,A., Rico,H., Mormeneo,S. and Sentandreu,R. (1993) Linkages between macromolecules in
Candida albicans cell wall. Arch. Med. Res., 24, 305–310.
Ensley,H.E., Tobias,B., Pretus,H.A., McNamee,R.B., Jones,E.L., Browder,I.W. and Williams,D.L. (1994) NMR spectral analysis of a water-insoluble (1–3)-β-D-glucan isolated from Saccharomyces cerevisiae.
Carbohydr. Res., 258, 307–311.
Fearon,D.T. and Locksley,R.M. (1996) The instructive role of innate immunity in the acquired immune response. Science, 272, 50–53.
Felippe,J., Silva,M., Maciel,F.M., Soares,A. and Mendes,N.F. (1993) Infection prevention in patients with severe multiple trauma with the immunomodulator beta 1–3 polyglucose (glucan). Surg. Gynecol. Obstet., 177,
383–388.
Kapteyn,J.C., Montijn,R.C., Dijkgraaf,G.J.P., Van den Ende,H. and Klis,F.M.
(1995) Covalent association of β-1, 3-glucan with β-1, 1–6 glucosylated
mannoproteins in cell walls of Candida albicans. J. Bacteriol., 177, 3788–
3792.
Kiho,T., Katsuragawa,M., Nagai,K. and Ukai,S. (1992) Structure and antitumor activity of a branched (1–3)-β-D-glucan from the alkaline extract of
Amanita muscaria. Carbohydr. Res., 224, 237–243.
Kojima,T., Tabata,K., Itoh,W. and Yanaki,T. (1986) Molecular weight
dependence of the antitumor activity of Schizophyllan. Argic. Biol. Chem.,
50, 231–232.
345
A.Mueller et al.
Kulicke,W.-M., Lettau,A.I. and Thielking,H. (1997) Correlation between
immunological activity, molar mass and molecular structure of different
(1–3)-β-D -glucans. Carbohydr. Res., 297, 135–143.
Kurachi,K., Ohno,N. and Yadomae,T. (1990) Preparation and antitumor activity of hydroxyethylated derivatives of 6-branched (1→3)-β- D-glucan,
SSG, obtained from the culture filtrate of Sclerotinia sclerotiorum IFO
9395. Chem. Pharm. Bull. Tokyo, 38, 2527–2531.
Lowman,D., Ensley,H. and Williams,D. (1998) Identification of phosphate
substitution sites by NMR spectroscopy in a water-soluble phosphorylated
(1–3)-β-D -glucan. Carbohydr. Res., 306, 559–562.
Michalek,M., Melican,D., Brunke-Reese,D., Langevin,M., Lemerise,K., Galbraith,W., Patchen,M. and Mackin,W. (1998) Activation of rat macrophages by Betafectin PGG-glucan requires cross-linking of membrane
receptors distinct from complement receptor three (CR3). J. Leukocyte
Biol., 64, 337–344.
Miyazaki,T., Kohno,S., Mitsutake,K., Maesaki,S., Tanaka,K., Ishikawa,N.
and Hara,K. (1995) Plasma (1→3)-β- D-glucan and fungal antigenemia in
patients with candidemia, aspergillosis and cryptococcosis. J. Clin. Microbiol., 33, 3115–3118.
Mueller,A., Pretus,H., McNamee,R., Jones,E., Browder,I. and Williams,D.
(1995) Comparison of the carbohydrate biological response modifiers
Krestin, schizophyllan and glucan phosphate by aqueous size exclusion
chromatographty with in-line argon-ion multi-angle laser light scattering
photometry and differential viscometry detectors. J. Chromatogr., 666,
283–290.
Mueller,A., Rice,P.J., Ensley,H.E., Coogan,P.S., Kalbfleisch,J.H., Kelley,J.L.,
Love,E.J., Portera,C.A., Ha,T., Browder,I.W. and Williams,D.L. (1996)
Receptor binding and internalization of water-soluble (1→3)- β-D-glucan
biologic response modifier in two monocyte/macrophage cell lines. J.
Immunol., 156, 3418–3425.
Nemoto,J., Ohno,N., Saito,K., Adachi,Y. and Yadomae,T. (1993) Expression
of interleukin 1 family mRNAs by a highly branched (1→3)-β- D-glucan,
OL-2. Biol. Pharm. Bull., 16, 1046–1050.
Obayashi,T. (1997) (1→)-β-D-Glucanemia in deep mycosis. Mediat. Inflamm.,
6, 271–273.
Obayashi,T., Yoshida,M., Tamura,H., Aketagawa,J., Tanaka,S. and Kawai,T.
(1992) Determination of plasma (1–3)-β-D -glucan: a new diagnostic aid to
deep mycosis. J. Med. Vet. Med., 30, 275–280.
Okamoto,K., Yamamoto,T., Nonaka,D., Shin,H., Okada,K., Hayashi,J. and
Kashiwagi,S. (1998) Plasma (1→3)-β-D-glucan measurement and
polymerase chain reaction on sputum as practical parameters in Pneumocystis carinii pneumonia. Intern. Med., 37, 618–621.
Pretus,H.A., Ensley,H.E., McNamee,R.B., Jones,E.L., Browder,I.W. and Williams,D.L. (1991) Isolation, physicochemical characterization and preclinical efficacy evaluation of soluble scleroglucan. J. Pharmacol. Exp. Ther.,
257, 500–510.
346
Stahl,P.D. (1992) The mannose receptor and other macrophage lectins. Curr.
Opinion Immunol., 4, 49–52.
Suzuki,I., Hashimoto,K. and Yadomae,T. (1988) The effects of a highly
branched β-1,3-glucan, SSG, obtained from Sclerotinia sclerotiorum IFO
9395 on the growth of syngeneic tumors in mice. J. Pharmacobio-Dyn.,
11, 527–532.
Tamura,H., Arimoto,Y., Tanaka,S., Yoshida,M., Obayashi,T. and Kawai,T.
(1994) Automated kinetic assay for endotoxin and (1→3)-β-D -glucan in
human blood [letter]. Clin. Chim. Acta, 226, 109–112.
Tamura,H., Tanaka,S., Ikeda,T., Obayashi,T. and Hashimoto,Y. (1997)
Plasma (1→3)-β-d-glucan assay and immunohistochemical staining of
(1→3)-β-D-glucan in the fungal cell walls using a novel horseshoe crab
protein (T-GBP) that specifically binds to (1→3)-β-D-glucan. J. Clin. Lab.
Anal., 11, 104–109.
Thornton,B.P., Vetvicka,V., Pitman,M., Goldman,R.C. and Ross,G.D. (1996)
Analysis of the sugar specificity and molecular location of the beta-glucanbinding lectin site of complement receptor type 3 (CD11b/CD18). J.
Immunol., 156, 1235–1246.
Vereschagin,E.I., Van Lambalgen,A.A., Dushkin,M.I., Schwartz,Y.S., Polyakov,L., Heemskerk,A., Huisman,E., Thijs,L.G. and Van den Bos,G.C.
(1998) Soluble glucan protects against endotoxin shock in the rat: the role
of the scavenger receptor. Shock, 9, 193–198.
Vetvicka,V., Thornton,B.P. and Ross,G.D. (1996) Soluble beta-glucan
polysaccharide binding to the lectin site of neutrophil or natural killer cell
complement receptor type 3 (CD11b/CD18) generates a primed state of the
receptor capable of mediating cytotoxicity of iC3b-opsonized target cells.
J. Clin. Invest., 98, 50–61.
Vetvicka,V., Thornton,B.P., Wieman,J. and Ross,G.D. (1997) Targeting of
natural killer cells to mammary carcinoma via naturally occurring tumor
cell-bound iC3b and beta-glucan-primed CR3. J. Immunol., 159, 599–605.
Williams,D.L., McNamee,R.B., Jones,E.L., Pretus,H.A., Ensley,H.E., Browder,I.W. and Di Luzio,N.R. (1991a) A method for the solubilization of a
(1–3)-β- D-glucan isolated from Saccharomyces cerevisiae. Carbohydr.
Res., 219, 203–213.
Williams,D.L., Mueller,A. and Browder,W. (1996) Glucan-based macrophage
stimulators: A review of their anti-infective potential. Clin. Immunother.,
5, 392–399.
Williams,D.L., Pretus,H.A., McNamee,R.B., Jones,E.L., Ensley,H.E., Browder,I.W. and Di Luzio,N.R. (1991b) Development, physicochemical characterization and preclinical efficacy evaluation of a water soluble glucan
sulfate derived from Saccharomyces cerevisiae. Immunopharmacology,
22, 139–156.
Zimmerman,J.W., Lindermuth,J., Fish,P.A., Palace,G.P., Stevenson,T.T. and
DeMong,D.E. (1998) A novel carbohydrate-glycosphingolipid interaction
between a β- (1–3)-glucan immunomodulator, PGG-glucan and lactosylceramide of human leukocytes. J. Biol. Chem., 273, 22014–22020.