| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Channing Laboratory, and
Division of Infectious Diseases, Brigham and Women’s Hospital, Boston, MA 02115; and
Division of Infectious Diseases, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02115
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following buffers were used in this study: GVB, isotonic veronal-buffered saline (pH 7.5), containing 0.1% (w/v) gelatin; GVB+/+, GVB with 0.15 mM CaCl2 and 0.5 mM MgCl2; GVB-/-, GVB with 10 mM EDTA; DGVB, GVB diluted with an equal volume of 5% D-glucose in water; DGVB+/+, GVB+/+ diluted with an equal volume of 5% D-glucose in water containing the same concentration of Ca2+ and Mg2+; DGVB-/-, GVB-/- diluted with an equal volume of 5% D-glucose in 10 mM EDTA; DGVB*, DGVB+/+ with 3 mM CaCl2 and 10 mM MgCl2; and PBS (pH 7.4).
Chemicals
Common salts were from Fisher Scientific (Fairlawn, NJ). Dextran (m.w. 500,000), pepsin, heparin, gelatin, EGTA, EDTA, and the Sigma 104 phosphatase substrate tablets were from Sigma (St. Louis, MO). Dehydrated Todd-Hewitt broth was from Difco Laboratories (Detroit, MI); Eagle’s MEM was from BioWhittaker (Walkersville, MD).
Antibodies
Rabbit antiserum to the GBS type III capsular polysaccharide was produced in our laboratory (19). mAbs R3 and R139 to C1qRp were a gift from Dr. Andrea Tenner (University of California, Irvine, CA) (20); R3 was used at a concentration of 20 µg/ml and R139 at 50 µg/ml. The following reagents were purchased as noted: goat antiserum against human C1q (Quidel, San Diego, CA), FITC-conjugated goat IgG against rabbit IgG and alkaline-phosphatase-conjugated rabbit IgG against goat IgG (Sigma), alkaline-phosphatase-conjugated goat IgG against rabbit IgG (Tago, Burlingame, CA), and rabbit antiserum against sheep erythrocytes (Diamedix, Miami, FL).
Complement reagents
Ab-sensitized sheep erythrocytes bearing C4b (EAC4), or C4b and C1 (EAC41) were prepared as described (21). Human C1q was purchased from Quidel or, for the opsonophagocytosis inhibition experiments, purified from human plasma (see below). Human C1 and guinea pig C2 were isolated as described (22, 23). The components C4 to C9 were purchased from Quidel.
Purification, enzymatic digestion, and radiolabeling of human C1q
C1q was purified essentially according to Tenner et al. (24). Briefly, outdated human plasma, obtained from the Brigham and Women’s Hospital (Boston, MA) blood bank, was ultracentrifuged to remove lipids and clotted. The serum was applied to a Bio-Rex 70 (Bio-Rad Laboratories, Richmond, CA) cation-exchange column; the C1q-containing fractions were concentrated by ammonium sulfate precipitation and gel-filtered on a Sephacryl S-200 column (Amersham Pharmacia Biotech, Piscataway, NJ). The resulting C1q was concentrated and frozen in aliquots. Purity of the final preparation was assessed to be 90–95% by SDS-PAGE.
The N-terminal collagen-like region of C1q was prepared by pepsin digestion of C1q at pH 4.5 (25) for 20 h. After elevating pH to 7.5 to stop the digestion, the collagen-like fragment was purified from the reaction mixture by gel filtration on a Sephacryl S-200 column. SDS-PAGE was used to assess the fragment’s identity and purity.
Radiolabeling of C1q with 125I was performed with
the Iodobead system (Pierce, Rockford, IL): one bead was incubated with
0.2 mg C1q and 0.02 mCi of 125I (Amersham,
Arlington Heights, IL) in 0.25 ml of a 40 mM phosphate buffer (pH 7.4)
for 5 min at room temperature. The reaction was terminated by removal
of the bead from the solution. Iodinated C1q was separated from unbound
125I on a PD-10 gel filtration column (Pharmacia,
Piscataway, NJ). Under these conditions, 71% of
125I was incorporated in C1q, which resulted in
the sp. act. of 0.06 mCi/mg of protein, corresponding to 
135,000
cpm/µg.
Binding of iodinated C1q to GBS
Concentrations from 0.1 to 20 µM of
125I- C1q were added to a suspension of GBS
(108 CFU/ml) in a total volume of 0.25 ml. The
buffer (either DGVB or PBS/gelatin) contained 0.1% gelatin to diminish
nonspecific binding of C1q. After 15 min of incubation at 37°C, the
samples were vortexed, and 50-µl aliquots were taken to determine the
total radioactivity. The bacteria were then pelleted by
microcentrifugation, and a 50-µl aliquot of the supernatant was used
to determine the free (unbound) C1q. The bound C1q was calculated as:
total C1q - free C1q. Alternatively, the bound C1q was quantified
directly by counting the radioactivity in the bacterial pellet after
two washes with DGVB. Results of both methods were similar; they were
pooled together for statistical evaluation of data. Radioactivity was
measured on a Beckman (Fullerton, CA) 
5500 counter.
GBS strains
M781, a highly encapsulated type III GBS clinical isolate (originally donated by Dr. Carol Baker, Baylor College of Medicine, Houston, TX), moderately encapsulated COH1, and its acapsular mutant COH1-13 (26) were from the culture collection at the Channing Laboratory. Before experiments, the bacteria were grown on blood agar plates overnight, then inoculated into 5 ml of Todd-Hewitt broth in a 16-mm culture tube and grown to an optical density of 0.26 at 650 nm. After three washes with PBS, the bacteria were appropriately diluted and used.
Preparation of human polymorphonuclear leukocytes (PMN)
A plasma fraction enriched in PMN was prepared from heparinized blood by dextran sedimentation, as described (27).
Serum
Serum from a normal healthy volunteer with a very low level of IgG specific for type III capsular polysaccharide was used for all the studies. The concentration of the specific anti-capsular polysaccharide IgG (0.15 ± 0.03 µg/ml) was determined by ELISA (28) in three independent measurements during the course of this work.
One part of the serum (NHS or unabsorbed serum) was immediately stored in aliquots at -70°C. Another part of the serum was "absorbed" with GBS as follows. Approximately 1010 CFU of GBS were harvested by centrifugation of a 50-ml overnight broth culture, washed with PBS, and incubated with 3–5 ml of serum at 4°C. The bacteria were then pelleted, and the serum was filtered through a 0.45-µm Acrodisc filter (Gelman Sciences, Ann Arbor, MI) and stored. In some experiments, absorption was conducted in the presence of 5 mM MgEGTA. There was no difference in opsonic potency of the serum, whether it was or was not filtered and whether it was absorbed with live or dead (heated to 60°C for 1 h) bacteria. The third part of serum was heat-inactivated (56°C for 30 min).
In some experiments, serum was selectively depleted of C1q and/or mannan-binding lectin (MBL) by affinity chromatography. A weakly acidic cation-exchange resin Bio-Rex 70 was used to deplete C1q, a highly basic protein (24). Serum was chelated with 5 mM EDTA, applied to the column, and eluted with 50 mM phosphate buffer containing 82 mM NaCl and 2 mM EDTA (pH 7.4). The degree of C1q depletion was determined by a C1q activity assay, as described below.
Depletion of MBL was achieved by passing the serum through an immobilized mannan column (Sigma) with the elution buffer containing 50 mM Tris-HCl, 1 M NaCl, 20 mM CaCl2 (pH 7.8). Before use, the MBL-depleted serum was dialyzed (tubing Spectra/Por 3, m.w. cut-off 3500; Spectrum, Houston, TX) against PBS to remove high salt. The MBL concentrations in coded serum samples were determined by ELISA in the laboratory of Dr. Alan Ezekowitz, then at Children’s Hospital (Boston, MA).
Opsonophagocytosis assay
Opsonophagocytosis was assayed essentially according to Baltimore et al. (27). GBS (1 x 106 CFU) were mixed with 3 x 106 PMN in 10% human serum and Eagle’s MEM in a total volume of 0.5 ml. Aliquots for quantitative cultures were taken at the time zero and after 1 h rotation at 37°C. The results were calculated as the logarithm of decrease in CFU/ml in 1 h: log kill = log CFU at time zero - log CFU at time 1 h. Results were expressed as percentage of GBS killed or bactericidal index, BI = [1 - (CFU at 1 h)/(CFU at time 0)] x 100, or BI = (1 - 10-log kill) x 100. In each experiment, appropriate controls were included: unabsorbed heat-inactivated serum or buffer were used as controls for samples with or without serum, respectively. In these controls, bacteria usually grew through the 1-h incubation period, so that their log kill values (denoted log killo) were negative: as much as -0.3 log in the serum, but <-0.05 log in the buffer. To account for this growth and for variations in different experiments, the log kill values for experimental samples were corrected as log killcorr = log kill - log killo.
In some experiments, washed GBS were first preopsonized with the investigated substance (as, for instance, serum chelated with 5 mM MgEGTA, or purified C1q), washed two times with DGVB+/+, and only then added to the suspension of PMN in MEM for measurement of opsonophagocytosis.
Assay for the association of GBS with PMN
GBS (3 x 107 CFU) were preincubated with C1q or anti-capsular polysaccharide IgG in DGVB+/+ for 30 min, washed two times with DGVB+/+, and then mixed with 3 x 106 human PMN in MEM, total volume 0.5 ml. When desired, C1q or the anti-capsular polysaccharide IgG were added to the second suspension. After 20 min of incubation at 37°C, a 50-µl aliquot was dropped on a microscope slide (Fisher Scientific) and spread. The specimen was heat-fixed and Gram-stained. Associated (i.e., adherent or ingested) bacteria were enumerated for 100 PMN, and the median number of GBS per PMN was determined.
Complement activity assays
Modified hemolysis assays were used to measure the activities of individual CCP components C1, C1q, C3, and C4 (21, 22, 24, 29).
Coupling of the type III capsular polysaccharide to the carbolink resin
Type III capsular polysaccharide was isolated, purified, and
partially oxidized by treatment with sodium periodate (20% of
sialic acid residues had aldehyde functions introduced at the C8
position), as described previously (19). The oxidized
polysaccharide was coupled to the Carbolink Gel (Pierce) containing
hydrazide active groups, according to the manufacturer’s instructions
(see also O’Shannessy and Wilchek (30)): 25 mg of
polysaccharide were gently rotated with 8 ml of Carbolink in a 0.1 M
sodium phosphate buffer (pH 7.0) in the presence of 60 mg of sodium
cyanoborohydride for 18 h at room temperature. Efficiency of
coupling was determined qualitatively by immunofluorescence microscopy
and quantitatively by ELISA. In the first method, the Carbolink resin
beads were incubated with a rabbit antiserum against type III capsular
polysaccharide, followed by incubation with an FITC-conjugated goat IgG
against rabbit IgG. The polysaccharide resin fluoresced brightly,
whereas the control resin, with glyceraldehyde coupled instead of the
polysaccharide, did not (data not shown). In the second method, the
amount of polysaccharide remaining in the solution and in the washes
after coupling was determined by inhibition ELISA using microtiter
plates coated with poly(L-lysine) and polysaccharide
(31) or polysaccharide/human serum albumin conjugate
(28). Approximately 15 mg of polysaccharide was found to
be linked to 8 ml of the packed resin.
Determination of protein concentrations
Routinely, Pierce’s bicinchoninic acid assay was used with BSA
as a standard. When a sample volume permitted it, absorption at 280 nm
was measured and the concentration of C1q estimated using the molar
extinction coefficient 
of 2.742 x 105
cm-1M-1, which was
calculated from the published amino acid sequence (32)
according to the formula =
5700nTrp +
1300nTyr, where
nTrp and
nTyr are the numbers or tryptophans
and tyrosines, respectively, in the protein (33). This
value is within 2% of that determined by Reid et al. (34)
using a value of 410,000 for the m.w. of C1; if the currently accepted
value of 460,000 is used for the m.w. of C1q, the calculated extinction
coefficient is 14% less than that determined by Reid et al
(34).
Data analysis
Nonlinear curve fitting and statistical analyses were performed with the Prism and Instat software (GraphPad Software, San Diego, CA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Preliminary testing of serum samples from several healthy
volunteers confirmed previous observations that NHS containing very low
concentrations of specific Abs to the capsular polysaccharide mediated
60–99% killing of type III and type Ia GBS by peripheral blood
leukocytes in an in vitro assay of opsonophagocytosis
(11, 12, 13). Serum depleted of 98% of its IgG by passage
over a protein A affinity column mediated 89% killing of type III GBS,
compared with 97% before IgG depletion (M. B. Marques and M.
R. Wessels, unpublished observation). In the present investigation, one
serum selected for study was found to contain 0.15 ± 0.03 µg/ml
of IgG specific to type III GBS capsular polysaccharide by quantitative
ELISA (28). Opsonic activity of the serum could be removed
by absorption with type III GBS, which resulted in classical pathway
activation and depletion from the serum of C3 and C4 (Table I
). However, sufficient complement
activity remained in the absorbed serum to support >90%
opsonophagocytic killing of type III GBS in the presence of capsular
polysaccharide-specific Abs (1% rabbit antiserum to type III
polysaccharide-tetanus toxoid conjugate vaccine). To rule out the
possibility that the opsonic activity of the NHS was due to the small
amount of specific Abs, we tested whether opsonic activity of the
GBS-absorbed serum could be restored by adding a sample of the same
serum not absorbed with GBS but previously heated to 56°C to
inactivate complement but preserve Abs. In this experiment, unabsorbed
serum mediated 61 ± 4% kill (mean ± range,
n = 2), 0 ± 0% kill after absorption, and 5
± 2% kill after adding heat-inactivated, unabsorbed serum to the
absorbed serum. These results indicate that the amount of specific Abs
in the serum used for the present studies was not sufficient to support
opsonophagocytic killing under the conditions used in the assay.
|
The cascade of CCP activation usually starts with binding of its first component C1q to the activating entity, which can be an Ag-Ab complex or a polymeric molecule. Therefore, we examined the ability of C1q to bind to the GBS surface.
The binding of native, unlabeled C1 to GBS was determined in modified
C1 transfer experiments (21). GBS were first exposed to
serum, washed, and then mixed with sensitized sheep erythrocytes
carrying the C4b component of complement (EAC4). Under favorable
conditions, namely higher ionic strength, any C1 previously adsorbed on
GBS would transfer onto the EAC4 cells and, upon addition of the other
CCP components, would cause their hemolysis. The data in Table II show that incubation of GBS with NHS
resulted in deposition of C1 on the bacterial surface, reflected in
subsequent transfer to EAC4.
|
documents binding of
radiolabeled C1q to the GBS strain M781 in buffers of normal (PBS) and
decreased (DGVB) ionic strength. Binding parameters were obtained by
nonlinear curve fitting to the data points. The dissociation constants
Kd were 2.0 ± 0.8 nM and 5.6 ±
2.0 nM, while Bmax values were 2400 ±
200 C1q/CFU and 1600 ± 200 C1q/CFU in DGVB and PBS, respectively.
Thus, increasing ionic strength decreased binding of C1q to
GBS.
|
). The data were fitted with the
one-site competition equation y = 100/[1 + 10(log
x - log C)], where y is the degree of binding in %,
x is the concentration of the competing molecule, and C is a constant
(called IC50) denoting the concentration at which
binding is 50%. The value of log C was found to be 7.17 ± 0.09.
Taking into account the concentration of radiolabeled C1q (1.9 nM), the
50% inhibition of binding was reached at 35-fold excess of
unlabeled C1q. From Fig. 2, it is clear that
C1qcoll is much less efficient in inhibiting the
125I-C1q binding than the whole C1q molecule.
|
). One-way ANOVA showed
significant differences in the Bmax values:
the acapsular mutant COH1–13 bound significantly more C1q molecules
than its parent strain COH1 or the highly encapsulated strain M781. In
addition, there was a statistically nonsignificant trend of stronger
binding (decreasing Kd) with decreasing
capsule expression.
|
To test whether the pure capsular polysaccharide can deplete NHS
of CCP activity and of opsonic power to the same extent as the whole
GBS organism, we linked the polysaccharide to a Carbolink resin as
described in Materials and Methods. If it is the capsule
that is responsible for the interaction with CCP, passing the serum
over this column would mimic the serum absorption with the whole GBS.
Before applying serum, both the polysaccharide and mock columns were
washed with 0.1% gelatin in PBS to diminish nonspecific protein
adsorption. After passage of the serum over the columns, C1q activity
was found to be essentially the same in both: the titers were 143 and
125 for the polysaccharide and mock column, respectively. The opsonic
power of the serum, however, was reduced to a modest extent: while
untreated serum mediated 88% killing of type III GBS, serum passed
over the polysaccharide column mediated 60% killing, and serum passed
over the mock column (matrix without polysaccharide) mediated 81%
killing (Fig. 3). By comparison,
absorption of the serum with whole type III GBS organisms reduced
killing to 6% in this experiment.
|
In an attempt to further define the role of C1q in
opsonophagocytosis of GBS by NHS, we measured the extent of bacterial
killing with serum devoid of C1q. C1q was selectively and
quantitatively removed from NHS by cation exchange chromatography on a
column of Bio-Rex 70. After passage over the column, C1q activity
remaining in the serum was <1% of that in the control, untreated
serum. However, the C1q-depleted serum still supported opsonophagocytic
killing at a level of 81% of that mediated by the control serum (Fig. 4). This result suggested that, apart
from C1q, there is another molecule in NHS that is deposited on the
surface of GBS and is able to either activate CCP or interact with a
receptor on PMN, or both. Another serum protein implicated in classical
pathway activation is MBL (previously called mannose- or mannan-binding
protein, MBP). MBL has been shown to interact with bacterial
polysaccharides (35) and activate CCP (36).
To test the involvement of MBL in CCP activation and opsonophagocytic
killing of GBS, NHS was depleted of MBL by affinity chromatography. A
mannan affinity column removed >96% of MBL, but the serum retained
63% of its opsonic activity, defined as the ratio of bactericidal
indices after and before MBL depletion (Fig. 5). Furthermore, serum subjected to
absorption with type III GBS organisms lost 94% of its opsonic power,
while it showed only 9% reduction in the MBL concentration. In a
separate experiment, serum depleted of both C1q and MBL still exhibited
80% of opsonic activity compared with the untreated serum (data not
shown). Together, these results indicated that MBL does not bind to GBS
to a significant extent and is not an opsonic ligand for type III
GBS.
|
|
The previous experiments demonstrated that C1q binds to the GBS
surface. Therefore, we tested whether the killing activity of the
GBS-absorbed serum could be restored by addition of C1q. The results
(data not shown) were negative: C1q up to a concentration of 50 µg/ml
failed to restore the killing activity of the GBS-absorbed serum.
Moreover, we noticed in those experiments that excess C1q added to the
control, untreated serum inhibited killing activity of that serum. When
GBS were incubated with 10% NHS and PMN in the presence of increasing
amounts of C1q, a concentration-dependent inhibition of
opsonophagocytic killing was observed (data not shown). While no
discernible effect on bactericidal index was seen up to 40 µg/ml,
opsonophagocytosis was inhibited by 50% at 100 µg/ml C1q, which
corresponds to 1 mg/ml of serum or 10-fold excess over the usual C1q
concentration in serum (Fig. 6).
|
At least two PMN membrane proteins may be involved in C1q-mediated
responses. Complement receptor 1 (CR1, CD35) can directly bind C1q and
its collagen-like domain (37). In addition, another
protein termed the C1q receptor of phagocytosis, C1qRp, participates in
C1q-enhanced phagocytosis of IgG-coated sheep erythrocytes by monocytes
(38). C1qRp is also expressed on PMN, but its role in PMN
phagocytosis is less well defined. Therefore, we tested the effect of
the purified collagen-like fragment of C1q on opsonophagocytosis. The
data in Fig. 6
show that 35 µg/ml of the fragment inhibited
opsonophagocytosis to the same extent as an equimolar concentration
(100 µg/ml) of the whole molecule. The fact that an excess of soluble
C1q or its collagen-like fragment inhibited opsonophagocytosis in NHS
suggested some involvement of a PMN receptor for C1q in the
opsonophagocytic killing. However, addition to NHS of either of two
mAbs (R3 or R139) to C1qRp failed to inhibit opsonophagocytic killing
of GBS (data not shown).
Inability of C1q to act as a primary opsonin
To investigate whether C1q itself can act as an opsonin in the absence of other complement components, we preincubated GBS either with purified C1q (40 µg/ml) or 10% NHS (as a control) in DGVB+/+ and, after two washes with DGVB+/+, added PMN and continued incubation in the absence of serum for 1 h. In this assay (n = 14), the bactericidal index for GBS preopsonized with NHS was 60 ± 5, whereas those for GBS preopsonized with the preabsorbed serum or with C1q alone were low and indistinguishable (12 ± 5 or 8 ± 5, respectively). Thus, there was very little, if any, phagocytic killing mediated by C1q in the absence of other serum components. Thus, C1q does not appear to act as a primary opsonin for GBS.
Association of GBS with PMN
C1q by itself may not be able to support opsonophagocytic killing
in a serum-free system, but may participate in adherence of C1q-coated
bacteria to PMN, thereby enhancing the effects of other opsonins. The
role of C1q, alone or in conjunction with specific Ab, in the
association of GBS with PMN was examined in the next experiment. The
bacteria were first preincubated with either C1q or anti-capsular
polysaccharide IgG and, after two washes, mixed with PMN in the
presence of either the anti-capsular polysaccharide IgG or C1q,
respectively. The GBS/PMN association was examined microscopically. The
association data are compared with the bactericidal indices (BI) of the
respective samples in Fig. 7. The
negative control, labeled "no IgG/no C1q", only contained GBS and
PMN without added opsonins. This sample exhibited a BI of 12 ± 5,
and the median number of GBS/PMN was 1.5. The sample labeled "C1q/no
IgG" confirmed that, in the absence of other complement proteins and
Ab, C1q is unable to support opsonophagocytosis of GBS: the BI was
15 ± 6 and median GBS/PMN 2, values not different from those of
the negative control. However, when GBS were exposed to both C1q and
the Ab, irrespective of the order of incubations, BI increased to
35, and the median number of GBS/PMN increased above 3. It is
interesting to compare the samples of the IgG-preopsonized GBS mixed
with PMN in the absence ("IgG/no C1q") and presence ("IgG/C1q")
of C1q. The sample with C1q shows a slightly lower BI (30 ± 9 vs
47 ± 9, statistically NS), but higher association of GBS with PMN
(3.5 vs 1). These results indicate that, although C1q does not promote
phagocytic killing directly, it can act synergistically with IgG to
enhance association of GBS with PMN.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In the present study, we investigated the role of C1q in IgG-independent opsonization of type III GBS. First, C1 activity was demonstrated on the bacteria exposed to NHS. Second, CCP components were depleted from NHS absorbed with the bacteria, presumably due to activation of CCP by deposited C1q. These data suggested a role for C1q in the IgG-independent opsonization of type III GBS. It was then quite surprising that serum completely devoid of C1q still exhibited significant opsonic power, a result that indicated that there must be another serum protein able to replace C1q in GBS opsonization and/or activation of CCP. One possibility is that, in addition to direct binding of C1q, C2 or C4 binds directly to GBS and activates the classical pathway; such a mechanism has been reported for pathogenic mycobacteria (40). Another candidate protein is MBL, which, like C1q, has the capacity to activate CCP (35, 36). In our studies, however, levels of MBL did not correlate with the serum opsonic ability. Therefore, MBL appears not to be a major opsonic ligand for type III GBS. This conclusion is in accord with the previously reported low affinity of radioactively labeled MBL for GBS (41). The modest effect of depleting C1q and/or MBL leaves open the possibility that other serum factors may contribute to opsonophagocytosis of GBS as direct opsonins or in the activation of CCP.
Our data show that C1q binds to the surface of GBS. Binding was not inhibited by the collagen-like fragment of C1q, which indicates that C1q binds to GBS through the globular head region, a mode of binding that was observed for C1q deposition on other microorganisms (42). C1q binding alone did not mediate opsonophagocytic killing of GBS by PMN, as we incorrectly concluded from preliminary data (43). However, binding of C1q together with IgG to GBS promoted their association with PMN, which may enhance phagocytosis mediated by interactions between PMN and other opsonins, such as C3b and/or C4b.
The polysaccharide capsule is a conspicuous surface structure of GBS that has been identified as an important virulence factor (8, 10, 11, 44). High concentrations of soluble type III capsular polysaccharide in serum inhibit opsonophagocytosis of GBS (44). However, in our hands, even very high concentrations did not accomplish complete inhibition: 1.9 mg of polysaccharide per 1 ml of serum still allowed for 68% killing of the bacteria (data not shown). One could argue that the soluble polysaccharide conformation and/or affinity constant for complement proteins may differ from those of the polysaccharide on the GBS surface. Indeed, polysaccharide immobilized on Carbolink beads, perhaps a better model of GBS capsule than soluble macromolecules, decreased the serum’s bactericidal index from 88 to 60. However, absorption of serum with whole type III GBS organisms reduced killing to 6% in the same experiment. Thus, interaction between a serum molecule and the GBS capsule accounts for at most 32% (= 1 - 60/88) of the opsonophagocytic killing of GBS. It is reasonable to expect nonspecific electrostatic binding between the negatively charged polysaccharide and the positively charged subcomponent C1q of the first component of CCP. However, a polysaccharide affinity column did not deplete C1q when compared with serum passed over a mock column. This finding is consistent with that of Baker et al. (12), who reported evidence of C1q binding by immunofluorescence to a moderately encapsulated type Ia strain, but no detectable C1q deposition on a highly encapsulated strain. Our own preliminary experiments attempting to measure the interaction between C1q and the polysaccharide using surface plasmon resonance, electron microscopy, and ELISA all failed to demonstrate appreciable binding (our unpublished observations). Our data show that the acapsular GBS mutant COH1–13 binds C1q even better than the encapsulated strains COH1 and M781. The Kd values of the C1q/GBS interactions in our experiments (3–6 nM) are very similar to those describing interaction between C1q and a Klebsiella porin (1.5 nM) (42). The data suggest that the target of C1q on the surface of GBS is probably not the capsular polysaccharide, but rather a protein, as is the case for other microorganisms (15, 16, 17, 18, 45). The capsule, on the contrary, seems to protect the bacteria against direct deposition of C1q on their surface.
While C1q did not support phagocytosis in the absence of other opsonins, inhibition of opsonophagocytic killing of GBS by an excess of free C1q suggests a direct interaction of C1q with a receptor on PMN. C1q binding function has been described for PMN (46, 47), although the molecular identity of the receptor(s) is an area of ongoing investigation (48). In our studies, Abs to C1qRp had no effect on serum-mediated opsonophagocytosis of GBS; however, these Abs also failed to inhibit C1q-stimulated oxidative burst in neutrophils, so it is not clear that they can block C1q binding to C1qRp on this cell type (20). Another candidate receptor for C1q is complement receptor type 1 (CR1, CD35), a membrane protein recently shown to bind C1q in addition to its previously described ligands C3b and C4b (37). Abs to CR1 and to CR3 inhibit opsonophagocytosis of GBS in NHS (49, 50). An appealing hypothesis to explain the inhibition of phagocytic killing by excess C1q observed in the present study is that free C1q (in contrast with C1q deposited on the bacterial surface) may compete with other opsonins for binding to CR1 on PMN. Inhibition of the interaction between GBS-bound C3b and CR1 would be expected to interfere with phagocytic killing, since CR1 (together with complement receptor type 3, CR3) has been shown to mediate opsonophagocytosis of GBS both in the presence and absence of anticapsular Abs (49, 50). Another potential mechanism through which excess C1q could inhibit opsonophagocytosis is by competing with C1 for binding to GBS, thereby preventing classical pathway activation.
Our results indicate that C1q plays a role in opsonization and phagocytosis of GBS by at least two mechanisms operating in parallel. C1q binds directly to the GBS surface, a consequence of which is CCP activation with deposition of C3b. In addition, bound C1q may act synergistically with IgG to increase the association between GBS and PMN, thereby facilitating opsonophagocytosis mediated by IgG and C3b. In the absence of specific Abs, C1q deposited on the GBS surface (in contrast to soluble C1q) could enhance the association of GBS with PMN, perhaps through interaction with the CR1 complement receptor. Excess free C1q may inhibit opsonophagocytic killing of GBS opsonized with C3b or other opsonins by blocking CR1 on PMN or by interfering with classical pathway activation. Neither C1q nor MBL is required for opsonophagocytosis of GBS. This observation suggests that additional serum factors, which previously were not thought to have this function, may opsonize GBS directly or through a novel mechanism of CCP activation. It is possible that these factors could be found among members of the collectin or ficolin families of serum proteins (51, 52, 53).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: Department of Chemistry and Biochemistry, University of Southern Mississippi, Hattiesburg, MS 39406.
3 Address correspondence and reprint requests to Dr. Michael R. Wessels, Channing Laboratory, 181 Longwood Avenue, Boston, MA 02115. E-mail address:
4 Abbreviations used in this paper: GBS, group B streptococcus; CCP, classical complement pathway; MBL, mannan-binding lectin; NHS, normal human serum; PMN, polymorphonuclear leukocytes.
Received for publication March 26, 1999. Accepted for publication June 21, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Hamilton, D. L. Popham, D. J. Carl, X. Lauth, V. Nizet, and A. L. Jones Penicillin-binding protein 1a promotes resistance of group B streptococcus to antimicrobial peptides. Infect. Immun., November 1, 2006; 74(11): 6179 - 6187. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Yuste, S. Ali, S. Sriskandan, C. Hyams, M. Botto, and J. S. Brown Roles of the Alternative Complement Pathway and C1q during Innate Immunity to Streptococcus pyogenes J. Immunol., May 15, 2006; 176(10): 6112 - 6120. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Aoyagi, E. E. Adderson, J. G. Min, M. Matsushita, T. Fujita, S. Takahashi, Y. Okuwaki, and J. F. Bohnsack Role of L-Ficolin/Mannose-Binding Lectin-Associated Serine Protease Complexes in the Opsonophagocytosis of Type III Group B Streptococci J. Immunol., January 1, 2005; 174(1): 418 - 425. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Levy, R. M. Jean-Jacques, C. Cywes, R. B. Sisson, K. A. Zarember, P. J. Godowski, J. L. Christianson, H.-K. Guttormsen, M. C. Carroll, A. Nicholson-Weller, et al. Critical Role of the Complement System in Group B Streptococcus-Induced Tumor Necrosis Factor Alpha Release Infect. Immun., November 1, 2003; 71(11): 6344 - 6353. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Brown, T. Hussell, S. M. Gilliland, D. W. Holden, J. C. Paton, M. R. Ehrenstein, M. J. Walport, and M. Botto The classical pathway is the dominant complement pathway required for innate immunity to Streptococcus pneumoniae infection in mice PNAS, December 24, 2002; 99(26): 16969 - 16974. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. T. Watford, M. B. Smithers, M. M. Frank, and J. R. Wright Surfactant protein A enhances the phagocytosis of C1q-coated particles by alveolar macrophages Am J Physiol Lung Cell Mol Physiol, November 1, 2002; 283(5): L1011 - L1022. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. S. Grace, R. A. Bronson, and B. Ghebrehiwet Surface Expression of Complement Receptor gC1q-R/p33 Is Increased on the Plasma Membrane of Human Spermatozoa after Capacitation Biol Reprod, March 1, 2002; 66(3): 823 - 829. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||
) or DGVB
(ionic strength one half of normal; •). Curves are the best
hyperbolic fits to the data. Data points represent means ± SEM of
two independent experiments in duplicate.
