Skip to main content
PMC home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 1999 Jul 6;96(14):8110–8115. doi: 10.1073/pnas.96.14.8110

The mast cell tumor necrosis factor α response to FimH-expressing Escherichia coli is mediated by the glycosylphosphatidylinositol-anchored molecule CD48

Ravi Malaviya *,†, Zhimin Gao *, Krishnan Thankavel *, P Anton van der Merwe ‡, Soman N Abraham *,§
PMCID: PMC22196  PMID: 10393956

Abstract

Mast cells are well known for their harmful role in IgE-mediated hypersensitivity reactions, but their physiological role remains a mystery. Several recent studies have reported that mast cells play a critical role in innate immunity in mice by releasing tumor necrosis factor α (TNF-α) to recruit neutrophils to sites of enterobacterial infection. In some cases, the mast cell TNF-α response was triggered when these cells directly bound FimH on the surface of Escherichia coli. We have identified CD48, a glycosylphosphatidylinositol-anchored molecule, to be the complementary FimH-binding moiety in rodent mast cell membrane fractions. We showed that (i) pretreatment of mast cell membranes with antibodies to CD48 or phospholipase C inhibited binding of FimH+ E. coli, (ii) FimH+ E. coli but not a FimH− derivative bound isolated CD48 in a mannose-inhibitable manner, (iii) binding of FimH+ bacteria to Chinese hamster ovary (CHO) cells was markedly increased when these cells were transfected with CD48 cDNA, and (iv) antibodies to CD48 specifically blocked the mast cell TNF-α response to FimH+ E. coli. Thus, CD48 is a functionally relevant microbial receptor on mast cells that plays a role in triggering inflammation.

Mast cells exhibit a remarkable capacity to release a battery of inflammatory mediators when activated (1, 2). Since their discovery more than 100 years ago, the physiological role of these cells, which are preferentially located at the host–environment interface, has been the subject of much debate. Recently, several laboratories have reported that mast cells are central to protecting mice against lethal enterobacterial infections through the release of various proinflammatory mediators including tumor necrosis factor α (TNF-α), a potent neutrophil chemoattractant (3–6). Cumulatively, these studies have established the vital role of mast cells in modulating host defenses against infectious agents. Mast cells appeared to contribute to the innate immune defenses because they were activated by bacteria even in the absence of specific antibodies to the pathogens. In one study, mast cell responses were elicited through bacterial activation of the host’s complement system because the in vivo inflammatory response to enterobacteria was significantly reduced in complement-deficient mice compared with wild-type mice (4). In another study, mast cell activation in vivo appeared to be elicited through direct contact with cell surface molecules on Escherichia coli because mast cell release of inflammatory mediators was markedly higher after exposure to wild-type bacteria than to an isogenic mutant deficient in FimH, a bacterial adhesin (3). FimH is a 29-kDa mannose-binding lectin presented preferentially at the distal tips of filamentous appendages on E. coli called type 1 fimbriae (7). This molecule has been shown to mediate bacterial binding to mannosylated molecules on the surface of a variety of host cells and in mucosal secretions (8, 9). Because FimH is expressed by many members of the Enterobacteriaceae including well known pathogenic species such as E. coli, Klebsiella pneumoniae, Serratia marcescens, and Salmonella typhimurium (7), mast cells have the potential to directly bind and respond to a wide range of enteric bacteria. There is currently no information on the complementary FimH-recognizing molecule(s) on the mast cell surface. Because of its relevance in identifying the molecular events leading to the mast cell responses to bacteria, we sought to identify the putative mast cell receptor for the FimH moiety of E. coli.

MATERIALS AND METHODS

Animals and Reagents.

Male BALB/c mice, 6–8 weeks old, were purchased from Harlan–Sprague–Dawley. Animals were caged in groups of five in a pathogen-free environment, and the experimental procedures were carried out in agreement with institutional guidelines.

Fetal bovine serum was obtained from HyClone. BSA, human gamma globulin, toluidine blue, hydrogen peroxide, d-glucose, α-methyl d- mannopyranoside (MeαMan), nitro blue tetrazolium, 5-bromo-4-chloro-3-indolyl phosphate, phospholipase C (PLC), and actinomycin D were purchased from Sigma. Recombinant peptide-N-glycosidase F was obtained from Oxford Glycosystems (Rosedale, NY). Rat monoclonal antibody to mouse CD48 and mouse monoclonal antibody to rat CD48 were purchased from Serotec, and rat monoclonal antibody to mouse CD117 was purchased from PharMingen. In addition, a rat monoclonal antibody to mouse CD48 (7D1.G5) generated at the Sir William Dunn School of Pathology, University of Oxford, was also used.

Bacterial Strains.

E. coli ORN103(pSH2) is a recombinant strain containing a plasmid, pSH2, that encodes all the genes necessary for the expression of functional type 1 fimbriae. E. coli ORN103(pUT2002) is an isogenic FimH-minus derivative (10). The bacterial strains were cultured in Luria broth containing chloramphenicol (50 μg/ml).

Cell Culture and Growth Conditions.

Bone marrow mast cells (BMMCs) were cultured from stem cells from the bone marrow of BALB/c mice as described (11). The cells were grown in 25% WEHI-3 conditioned medium and used for experiments after 20 days in culture. Mast cells harvested from these cultures were >98% pure, as determined by toluidine blue staining, and resembled the mucosal-type mast cells (11). RBL-2H3, a rat mast cell line, was a gift from Reuben P. Siraganian (Laboratory of Microbiology and Immunology, National Institute of Dental Research, National Institutes of Health). RBL-2H3 cells were cultured in 225-cm2 flasks containing Eagle’s essential medium supplemented with 20% FBS (12). Chinese hamster ovary (CHO) cells obtained from the American Type Culture Collection were cultured in Glasgow-modified Eagle medium (GMEM) supplemented with 10% FBS. The transfected CHO-CD48 cells were maintained in GMEM supplemented with 10% FBS and G418 (Geneticin). Flow cytometry analysis of cell lines was carried out. Briefly, cells were labeled with a mouse monoclonal antibody against rat CD48 (diluted 1:10) followed by fluorescein isothiocyanate-conjugated antibodies against mouse IgG (diluted 1:100). Cells were excited on FACScan with a 15-nW laser at 488 nm. Results were presented as a histogram of the number of cells (10,000 per analysis) vs. the logarithmic scale of fluorescence intensity.

FimH/FimC Complexes and Soluble CD48.

Because of the extreme sensitivity of the recombinant bacterial FimH to degradation in the bacterial periplasm, it is necessary to coexpress FimH with FimC, the fimbrial chaperone (13). The recombinant FimH and FimC proteins form a highly stable complex in the bacterial periplasm that is readily isolated. FimC also serves to present FimH molecules in a functionally competent form (13). FimH–FimC complexes were isolated from the periplasm of the bacteria as described (13). Soluble mouse CD48 and rat CD48 (LRCD48) were obtained as described (14, 15). Deglycosylation of mouse CD48 was achieved by treatment of the denatured glycoprotein with peptide-N-glycosidase F as described by the manufacturer.

Ligand/Lectin Blotting.

After SDS/PAGE and electrophoretic transfer onto poly(vinylidene difluoride) (PVDF) membranes, the membranes were blocked overnight with 3% BSA in 10 mM Tris⋅HCl/150 mM NaCl/0.1% Tween 20 at pH 7.4 (TBST). Blots were incubated with 2 × 108 biotinylated bacteria per ml (10) or 125I-labeled FimH/FimC (50 μg/ml) in TBST containing 1 mM CaCl2, 1 mM MgCl2, and 0.4 mM MnCl2 in the presence or absence of 100 mM MeαMan at 4°C. After overnight incubation, the blots were washed three times with TBST. The blots treated with biotinylated bacteria were incubated with streptavidin linked to horseradish peroxidase (diluted 1:2,000) in TBST for 1 hr and developed with an appropriate substrate, whereas the blots incubated with 125I-labeled FimH/FimC were developed by exposing to x-ray film for 48 hr.

Bacterial Adherence Assay.

Monolayers of BMMCs (5 × 104 cells per well) in 96-well trays were blocked with 1% human gamma globulin for 1 hr, after which various concentrations of anti-CD48 or anti-CD117 were added. After 1 hr of incubation, unbound antibody was removed and then type 1 fimbriated E. coli suspended in serum-free RPMI-1640 medium was added at a bacteria/mast cell ratio of 50:1. Unbound bacteria were gently rinsed off through three washes after 1 hr of incubation at 37°C. Triton X-100 (0.1%) was added to selectively solubilize the mast cells, and the total number of mast cell-associated bacteria was determined by standard viable counts as described (11).

PLC Treatment of Mast Cells.

BMMCs suspended in RPMI medium were treated with increasing concentrations of PLC (0.1–3.0 units/ml) at 37°C for 1 hr. PLC was removed after repeated washing of the BMMCs with PBS, and the cells were resuspended in culture medium containing 1 μM actinomycin D. The cells were plated onto wells of a 96-well tissue culture plate (5 × 104 cells per well) and incubated overnight at 37°C to allow them to adhere to the bottom of the microtiter plates. The mast cell monolayers were then exposed to type 1 fimbriated bacteria and the adherence assay performed as described above.

Transfection of CHO Cells with CD48 cDNA.

CHO cells were transfected with CD48 cDNA in the vector pKG5 using the Transfectam reagent (Promega), and stably transfected cells were derived by growth in G418-containing medium.

Assays for TNF-α.

Monolayers of mast cells in 96-well trays in serum-free RPMI-1640 medium containing 15 mM Hepes were exposed to bacteria for 1 hr at 37°C or were sensitized with anti-dinitrophenyl (DNP) IgE and subsequently challenged with DNP-BSA at 50 ng/ml for 1 hr at 37°C as described (16). After the incubation period, TNF-α in the cell-free supernatant was measured by a standard cytotoxicity assay (3).

RESULTS

FimH-Expressing E. coli Recognize a 45-kDa Component of Mast Cell Membranes.

In our previous studies, we had used either IL-3-dependent mouse BMMCs or freshly isolated peritoneal mast cells to study mast cell–enterobacteria interactions (3, 11, 17). However, we chose to isolate the FimH receptor from the rat mast cell line RBL-2H3 because, unlike BMMCs or mouse peritoneal mast cells, these cells could be readily cultured to generate large numbers of cells (12). Our studies have shown that FimH-expressing E. coli bound in comparable numbers to RBL-2H3 cells or to BMMCs (401 ± 19 and 436 ± 8 bacteria per 50 mast cells, respectively). Moreover, the binding interactions were inhibitable by MeαMan, indicating that the nature of the reaction with FimH-expressing bacteria was likely to be identical in both types of mast cells. Our strategy to isolate the FimH receptor in mast cells is summarized in Fig. 1. We reasoned that because the FimH receptor contained mannose, we could use the mannose-binding Con A lectin to enrich for mannose-containing molecules from the mast cell membrane preparations. Pooled batch cultures of approximately 1011 cells were prepared and a Triton X-100-soluble membrane fraction (100,000 × g fraction) was obtained as described (18). This membrane fraction was passed through a Sepharose-Con A affinity column to isolate candidate mannosylated compounds. Because Con A also binds to glucose-containing molecules, we eliminated the membrane material that was binding to the column by glucose residues by first washing with 100 mM glucose. None of the glucose-eluted material aggregated FimH-expressing E. coli on glass slides, confirming that no putative receptors were lost with this step. Next, 100 mM MeαMan was used to elute all materials bound via their mannose residues. The resulting eluate was dialysed to remove MeαMan, then concentrated, and subjected to SDS/PAGE. Many mannose-containing membrane components bound to the Con A column as evidenced by the large number of bands seen after staining of gels with Coomassie blue (Fig. 2, lane 1). To identify the putative FimH-recognizing moiety among the mannoside-eluted material, we electrophoretically transferred the material after SDS/PAGE onto PVDF membranes. The immobilized material was then exposed to 125I-labeled recombinant E. coli FimH, which was presented in a complex with its chaperone, FimC (13). The chaperone is required to stabilize the recombinant FimH protein and to present it in a functionally competent manner (13). As shown in Fig. 2, the FimH probe specifically bound to a 45-kDa band in the absence (lane 2) but not in the presence (lane 3) of MeαMan. Furthermore, when the blot was exposed to FimH-expressing E. coli ORN103(pSH2) and mutant FimH-deficient E. coli ORN103(pUT2002), only the former bound to the 45-kDa band (Fig. 2., lanes 4 and 5). The binding reaction of E. coli ORN103(pSH2) could be inhibited by 100 mM MeαMan (Fig. 2, lane 6). These data indicate that E. coli FimH binds specifically to a 45-kDa mast cell membrane component and that the binding is via mannose residues on the membrane component because FimH is not glycosylated.

Figure 1.

Figure 1

Open in a new tab

Strategy used to isolate and identify the putative FimH receptor on mast cells.

Figure 2.

Figure 2

Open in a new tab

Specific binding of recombinant FimH, FimH-expressing bacteria, or CD48-specific antibody to CD48 in mast cell membrane preparations. Lanes: 1, MeαMan-eluted material from the Con A column after SDS/PAGE and Coomassie blue staining; 2 and 3, PVDF membrane blots of Con A-eluted material probed with 125I-labeled FimH–FimC complex in the absence or presence, respectively, of 100 mM MeαMan; 4 and 5, PVDF membrane blots of Con A-eluted material probed with biotinylated FimH-expressing E. coli or biotinylated FimH-minus E. coli, respectively; 6, PVDF membrane blot of Con A-eluted material probed with biotinylated FimH-expressing E. coli in the presence of 100 mM MeαMan; 7, PVDF membrane blot of Con A-eluted material probed with rat CD48-specific mouse monoclonal antibody; 8, PVDF membrane blot of a whole cell lysate of BMMCs probed with rat monoclonal antibody to mouse CD48.

The 45-kDa Moiety Recognized by FimH-Expressing E. coli on Mast Cells Is the Immune Recognition Molecule CD48.

To determine the identity of the 45-kDa band, we purified the protein from the Con A-eluted fraction to homogeneity by FPLC using a Bio-Rad HQ column. The eluted proteins from this column were subjected to SDS/PAGE and stained with Coomassie blue to visualize the 45-kDa band. The band of interest was transferred to PVDF membranes and subjected to microsequencing. A sequence of 12 amino acid residues in the N terminus were identified as 100% homologous to rat CD48 protein. CD48 is a glycosylphosphatidylinositol (GPI)-linked molecule that has been reported to be present primarily on cells of hematopoietic lineage (19, 20). Further confirmation that the band recognized by FimH-expressing bacteria was indeed CD48 comes from the finding that a mouse monoclonal antibody directed at rat CD48 reacted with the 45-kDa band on a Western blot (Fig. 2, lane 7). It is also noteworthy that when a whole cell lysate of BMMCs was probed with a rat monoclonal antibody (7D1.G5) to mouse CD48, an immunoreactive band corresponding to CD48 was seen in the preparation (Fig. 2, lane 8), confirming the presence of this molecule on BMMCs.

Pretreatment of Mast Cells with PLC or CD48-Specific Antibody Results in Reduced Binding of FimH-Expressing E. coli.

Because GPI-linked moieties are cleaved off the surface of host cells with PLC (21), we incubated BMMCs with increasing amounts of PLC before exposure to FimH expressing E. coli. PLC pretreatment of mast cells was found to inhibit bacterial binding in a dose-dependent fashion (Fig. 3A). More direct evidence implicating CD48 as the putative E. coli FimH receptor comes from the observation that pretreatment of BMMCs with rat monoclonal antibodies to mouse CD48 inhibited the adherence of FimH expressing E. coli in a dose-dependent fashion, whereas rat monoclonal antibodies to mouse CD117 (c-kit), a well known mast cell membrane marker, did not (Fig. 3B). Thus, the putative FimH receptor on mast cells is the GPI-anchored moiety CD48.

Figure 3.

Figure 3

Open in a new tab

Inhibition of mast cell association with FimH-expressing bacteria after pretreatment with increasing concentrations of PLC and antibody to CD48. (A) BMMC suspensions were pretreated with indicated concentrations of PLC for 1 hr and then PLC was removed. The PLC treated cells were seeded in wells of a 96-well tissue culture plate in the presence of 1 μM actinomycin D. After overnight incubation, bacterial adherence assays were carried out on these cells. (B) Monolayers of BMMCs in a 96-well tissue culture plate were coated with 1% human gamma globulin for 1 hr at 37°C. The monolayers were then incubated with indicated concentrations of antibody against either CD48 or CD117 (control) for 1 hr at 37°C. Bacterial adherence assays were then done, and the number of mast cell associated bacteria was determined by standard viable counts on agar plates. The data points represent the mean ± SEM values obtained from three experiments.

The Mannosylated Residues on CD48 Play a Critical Role in Mediating Binding to FimH-Expressing E. coli.

We next examined the binding of FimH-expressing E. coli to soluble mouse CD48 in the presence and absence of MeαMan. As shown in Fig. 4 (lanes 1 and 2), bacterial binding to CD48 was almost completely abolished in the presence of MeαMan. That the glycosylated portion of CD48 was involved in binding was further confirmed by the observation that FimH-expressing bacteria failed to bind soluble CD48 after deglycosylation with peptide-N-glycosidase F (Fig. 4, lane 3). We also examined bacterial binding to an exclusively unprocessed form of soluble rat CD48 bearing high-mannose N-linked carbohydrates (15) and observed that bacterial binding was three times higher (as determined by densitometry) than binding levels seen with normally processed CD48 (Fig. 4, lanes 1 and 4). These experiments show that FimH-expressing E. coli can bind to recombinant CD48 in a cell-free system by specifically recognizing the mannosylated region of the molecule.

Figure 4.

Figure 4

Open in a new tab

Binding of FimH-expressing E. coli to processed and unprocessed recombinant CD48. Equal concentrations (5 μg) of processed (lanes 1 and 2), processed and deglycosylated (lane 3), and unprocessed (lane 4) soluble CD48 were subjected to SDS/PAGE and then transferred onto PVDF membranes. The corresponding strips were exposed for 1 hr to biotinylated FimH-expressing E. coli in the absence (lanes 1, 3, and 4) or presence (lane 2) of 100 mM MeαMan. Bound bacteria were detected by the addition of streptavidin-peroxidase and an appropriate substrate. Similar results were obtained in two experiments.

Cell Surface Expression of CD48 on CHO Cells Increases Binding of FimH-Expressing E. coli.

To determine whether expression of CD48 on cells that do not normally express this molecule can promote binding of FimH-expressing E. coli, we stably transfected CHO cells with the full-length cDNA encoding rat CD48 (22). As shown in Fig. 5A, the flow cytometry analysis using mouse anti-rat CD48 as probe confirmed that virtually 100% of the transfected CHO cells expressed CD48, whereas none of the cells transfected with control cDNA (encoding the C terminus of conglutinin) expressed CD48. Further, we examined the transfectants for their capacity to bind FimH-expressing E. coli. The association of bacteria with these CD48-expressing cells was at least 4-fold higher than the number associated with CHO cells transfected with control cDNA (Fig. 5B). Thus, CD48 molecules on transfected CHO cells are functional as FimH receptors.

Figure 5.

Figure 5

Open in a new tab

Induced expression of CD48 on CHO cells increases binding of FimH-expressing E. coli. (A) Expression of CD48 on transfected CHO cells assessed by flow cytometry analysis. The open peak represents control transfected CHO cells (with cDNA encoding the C terminus of conglutinin) and the solid peak represents CHO cells transfected with cDNA encoding rat CD48. The cells were labeled with CD48-specific mouse monoclonal antibody and fluorescein isothiocyanate-labeled second antibody. Similar results were obtained in three experiments. (B) Binding of FimH-expressing bacteria to CD48-expressing CHO cells and control CHO cells. CHO-cell-associated bacteria were quantitated by standard viability assays. The results are expressed as percent of control. The data points represent the mean ± SEM values obtained from three experiments.

CD48 Is the Determinant on Mast Cells That Is Responsible for Triggering the TNF-α Response to FimH-Expressing E. coli.

We have already shown that CD48-specific antibody inhibits FimH-mediated bacterial binding to mast cells (Fig. 3B). It is known that the early mast cell TNF-α response to bacteria plays a critical role in triggering the innate immune response to bacteria. We sought to demonstrate the role of CD48 in eliciting the mast cell TNF-α response to FimH-expressing bacteria after (i) blocking mast cell surface CD48 with CD48-specific antibody and (ii) removing CD48 from mast cell surfaces with PLC. Antibody to CD48, but not antibody to CD117, blocked mast cell TNF-α release in a dose-dependent manner (Fig. 6A). Further confirmation of the critical role of CD48 in the mast cell TNF-α response comes from the finding that pretreatment of BMMCs with increasing concentrations of PLC significantly reduced the mast cell’s capacity to release TNF-α after exposure to FimH-expressing E. coli (Fig. 6B). To ensure that either pretreatment did not, through some other mechanism, impair the mast cell TNF-α response, we examined the effect of both treatments on the mast cell’s ability to release TNF-α after stimulation by IgE and antigen. Either pretreatment (tested at the highest concentrations used in Fig. 6 A and B) did not reduce the mast cell TNF-α response to IgE and antigen (Fig. 6C). Thus, these observations provide definitive evidence that the mast cell TNF-α response to FimH-expressing bacteria is mediated by CD48 molecules present on the mast cell surface. Although there are likely to be other mannosylated moieties on the mast cell membrane capable of binding FimH-expressing E. coli, these studies show that CD48 is the biologically relevant receptor.

Figure 6.

Figure 6

Open in a new tab

Blocking of the mast cell TNF-α response to FimH-expressing E. coli but not to IgE and antigen by pretreatment of BMMCs with antibodies to CD48 and with PLC. (A) Human gamma globulin-treated monolayers of BMMCs were incubated with indicated concentrations of antibody against either CD48 or CD117 (control) for 1 hr at 37°C, and then bacteria were added to the wells. TNF-α release from cell-free supernatants was measured after 1 hr. (B) BMMC suspensions were pretreated with indicated concentrations of PLC for 1 hr. The PLC was removed by repeated washing, and the BMMCs were seeded in wells of a 96 well tissue culture plate in the presence of 1 μM actinomycin D to form monolayers. After overnight incubation, the BMMC monolayers were exposed to FimH-expressing bacteria and the TNF-α released from cell-free supernatants was assayed after 1 hr. (C) To examine the TNF-α response of mast cells to IgE and antigen, untreated control BMMCs and BMMCs pretreated with PLC at 3 units/ml or CD48-specific antibody at 30 μg/ml were sensitized with anti-DNP IgE and subsequently challenged with DNP-BSA (antigen) at 50 ng/ml for 1 hr at 37°C. TNF-α released from cell-free supernatants was assayed thereafter. Data are the mean ± SEM values obtained from three experiments.

DISCUSSION

Mast cells are found in relatively large numbers at the host–environment interface, and recent studies have shown that these cells play a crucial role in microbial recognition and in modulating the innate immune response to these infectious agents. To effect this role, mast cells must possess the capacity to recognize microorganisms in the absence of microbe-specific antibodies. Two separate recognition mechanisms appear to exist, one of which is mediated through other host opsonins, such as the iC3b fragment of complement. For example, the CR3 moiety on the surface of mast cells recognized S. typhimurium and the parasite Schistosoma mansoni after they were coated with iC3b component of complement (23, 24). The second recognition mechanism involves the direct interaction of bacterial surface molecules with complementary molecules on the mast cell surface without a need for opsonins. So far, the best-described paradigm of these opsonin-independent interactions involves enterobacteria expressing type 1 fimbriae (17). FimH, a mannose-binding subunit, located preferentially at the type 1 fimbrial tips is the specific determinant recognized by the mast cell. Evidence of the role of the FimH moiety in mediating bacterial adhesion includes the finding that a FimH-negative E. coli mutant derivative exhibited limited mast cell binding but that the parental type 1 fimbriated (FimH+) E. coli bound avidly to mast cells (3, 17). Herein, we report that the complementary receptor on mast cells for bacterial FimH is CD48. This is based on the following evidence: (i) antibodies to CD48 specifically blocked binding of FimH-expressing bacteria to the surface of mast cells; (ii) FimH-expressing, but not FimH- deficient, bacteria mediated mannose- sensitive binding to recombinant CD48; and (iii) FimH-expressing E. coli bound in appreciably higher numbers to transfected CHO cells expressing CD48 compared with control-plasmid-transfected non-CD48-expressing CHO cells. CD48 is also a biologically active receptor because antibodies directed at CD48 specifically blocked the TNF-α response to bacteria. Thus CD48 is a physiologically relevant receptor for bacteria on mast cells.

CD48 has been referred to as BCM1 in mice, OX 45 in rats, and Blast-1 in humans (14, 20, 22). The expression of CD48 is restricted to cells of hematopoietic lineage, particularly lymphocytes, monocytes, granulocytes, and mast cells. CD48 was discovered as a cell surface molecule expressed by human B lymphocytes in response to Epstein–Barr virus infection (25), but its physiologic role in the body is still unclear. Its interaction with CD2, particularly in rodents, suggests that it plays a central role in T cell activation (26). Recently, its interaction with human epithelial cells was demonstrated, suggesting a role for CD48 in mediating interactions between lymphoid cells and the epithelium (27). The CD48 ligand identified on epithelial cells has now been shown to be heparan sulfate (28). Thus, the involvement of CD48 in bacterial recognition and in triggering TNF-α release in inflammatory cells represents a distinct function for this molecule. Rodent CD48 is mannosylated and the primary structure of CD48 reveals several potential sites of glycosylation (22). The importance of the sugar moiety on CD48 and specifically its mannose residues is indicated by the observation that binding of FimH-expressing E. coli to a highly glycosylated and unprocessed form of CD48 was markedly higher than to processed CD48. Moreover, we showed that MeαMan inhibits the binding of both FimH-expressing E. coli and recombinant FimH to CD48. We have recently observed that human mast cells readily bind FimH expressing type 1 fimbriated E. coli but not its isogenic FimH-minus mutant derivative (29). There is appreciable homology in covalent structure between human and rodent CD48 (30). However, because it is the glycosylation pattern on CD48 rather than its protein configuration that is critical for bacterial recognition, it is difficult at this time to predict whether or not CD48 will serve as the FimH receptor on human mast cells.

To date, several distinct receptors have been identified for E. coli type 1 fimbriae on different host cells, indicating considerable heterogeneity in the molecules mediating FimH recognition in these host cells. This is not surprising because it is the glycosylation pattern rather than the protein component of the receptor that is relevant to bacterial recognition. These receptors include CD66 and CD67 on granulocytes (31), CD11b/CD18 on neutrophils (32), uroplakin on uroepithelial cells (33), and CD48 on macrophages (34). Moreover, a number of constituents in mucosal secretions have been identified as receptors for type 1 fimbriated E. coli, including Tamm–Horsfall protein in urine (35). CD48 joins a growing class of GPI-anchored cell surface molecules that serve as receptors for microbes and their toxins. Some notable examples include CD14 for lipopolysaccharide (36), Thy-1 for aerolysin (37), and CD55 for echovirus, group B coxsackie viruses, and Dr-fimbriated E. coli (38). Engagement of these receptors by microbes or their products has been shown to trigger cellular responses (36–38). Engagement of CD48 by FimH-expressing E. coli triggers the mast cell TNF-α response because inhibition of FimH–CD48 interactions by a CD48-specific antibody abrogates the mast cell TNF-α response. How CD48 and other GPI-anchored receptors, which are only linked to the exoplasmic leaflet of the lipid bilayer of the plasma membrane actually transduce intracellular signals is not clear. However, many GPI-anchored moieties including CD48 are typically found in special glycolipid-enriched microdomains in the plasma membranes of cells (39). These microdomains are rich in signaling molecules such as the heterotrimeric GTP-binding proteins that can potentially mediate signal transduction from GPI-anchored proteins. G proteins are involved in many signal-transduction pathways, including stimulation of adenylate cyclase, regulation of Ca2+ channels, stimulation of phospholipase A2, stimulation of phosphatidylinositol 3-kinase, and stimulation of PLC (40). A recent immunochemical study has shown that CD48 in lymphocytes is physically associated with GTP-binding proteins (40). Thus, engagement of CD48 could potentially trigger a mast cell response via a signaling pathway involving GTP-binding proteins. Engagement of CD48 has also been shown to activate Src family member tyrosine kinases, which are also important effectors of signal transduction found associated with glycolipid-enriched microdomains of the plasma membrane (41).

In summary, we have identified CD48 as a mast cell membrane molecule that is capable of recognizing E. coli via their FimH fimbrial adhesin. As a result of this binding interaction, CD48 triggers TNF-α release in the mast cell. This CD48-initiated and mast cell-mediated activity could be especially important in the modulation of immune responses to pathogenic E. coli in naive or immunocompromised individuals who are likely to be deficient in E. coli-specific antibodies.

Acknowledgments

We are grateful to Dr. J. Gierse (Monsanto, St. Louis, MO) for help with FPLC systems. This work was supported in part from research grants from the National Institutes of Health (AI 35678 and DK 50814). P.A.V. is supported by the Medical Research Council.

ABBREVIATIONS

TNF-α

tumor necrosis factor α

GPI

glycosylphosphatidylinositol

BMMC

bone marrow mast cell

CHO cells

Chinese hamster ovary cells

DNP

2,4-dinitrophenyl

PLC

phospholipase C

MeαMan

α-methyl d-mannopyranoside

PVDF

poly(vinylidene difluoride)

Footnotes

This paper was submitted directly (Track II) to the Proceedings Office.

References

  • 1.Galli S J. N Engl J Med. 1993;328:257–265. doi: 10.1056/NEJM199301283280408. [DOI] [PubMed] [Google Scholar]
  • 2.Wasserman S I. Ann Allergy. 1989;63:546–550. [PubMed] [Google Scholar]
  • 3.Malaviya R, Ikeda T, Ross E, Abraham S N. Nature (London) 1996;381:77–80. doi: 10.1038/381077a0. [DOI] [PubMed] [Google Scholar]
  • 4.Prodeus A P, Zhou X, Maurer M, Galli S J, Carroll M C. Nature (London) 1997;390:172–175. doi: 10.1038/36586. [DOI] [PubMed] [Google Scholar]
  • 5.Echtenacher B, Mannel D N, Hultner L. Nature (London) 1996;381:75–77. doi: 10.1038/381075a0. [DOI] [PubMed] [Google Scholar]
  • 6.Abraham S N, Malaviya R. Infect Immun. 1997;65:3501–3508. doi: 10.1128/iai.65.9.3501-3508.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Abraham S N, Sun D, Dale J B, Beachey E H. Nature (London) 1988;336:682–684. doi: 10.1038/336682a0. [DOI] [PubMed] [Google Scholar]
  • 8.Abraham S N, Jaiswal S. In: Escherichia coli: Mechanisms of Virulence. Sussman M, editor. New York: Cambridge Univ. Press; 1997. pp. 169–192. [Google Scholar]
  • 9.Ofek I, Sharon N. Curr Top Microbiol Immunol. 1990;151:91–113. doi: 10.1007/978-3-642-74703-8_5. [DOI] [PubMed] [Google Scholar]
  • 10.Thankavel K, Madison B, Ikeda T, Malaviya R, Shah A H, Arumugam P M, Abraham S N. J Clin Invest. 1997;100:1123–1136. doi: 10.1172/JCI119623. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Malaviya R, Abraham S N. Methods Enzymol. 1995;253:27–43. doi: 10.1016/s0076-6879(95)53005-3. [DOI] [PubMed] [Google Scholar]
  • 12.Okazaki H, Zhang J, Hamawy M M, Siraganian R P. J Biol Chem. 1997;272:32443–32447. doi: 10.1074/jbc.272.51.32443. [DOI] [PubMed] [Google Scholar]
  • 13.Jones C H, Pinkner J S, Nicholes A V, Slonim L N, Abraham S N, Hultgren S J. Proc Natl Acad Sci USA. 1993;90:8397–8401. doi: 10.1073/pnas.90.18.8397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Wong Y W, Williams A F, Kingsmore S F, Seldin M F. J Exp Med. 1990;171:2115–2130. doi: 10.1084/jem.171.6.2115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Davis S J, Puklavec M J, Ashford D A, Harlos K, Jones E Y, Stuart D I, Williams A F. Prot Engin. 1998;6:229–232. doi: 10.1093/protein/6.2.229. [DOI] [PubMed] [Google Scholar]
  • 16.Malaviya R, Malaviya R, Jakschik B A. J Biol Chem. 1993;268:4939–4944. [PubMed] [Google Scholar]
  • 17.Malaviya R, Ross E A, MacGregor J I, Ikeda T, Little J R, Jakschik B A, Abraham S N. J Immunol. 1994;152:1907–1914. [PubMed] [Google Scholar]
  • 18.Lennartz M R, Wileman T E, Stahl P D. Biochem J. 1987;245:705–711. doi: 10.1042/bj2450705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Staunton D E, Thorley-Lawson D A. EMBO J. 1987;6:3695–3701. doi: 10.1002/j.1460-2075.1987.tb02703.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Yokoyama S, Staunton D, Fisher R, Amiot M, Fortin J J, Thorley-Lawson D A. J Immunol. 1991;146:2192–2200. [PubMed] [Google Scholar]
  • 21.Low M, Saltiel R. Science. 1988;239:268–275. doi: 10.1126/science.3276003. [DOI] [PubMed] [Google Scholar]
  • 22.Killeen N, Moessner R, Arvieux J, Willis A, Williams A F. EMBO J. 1988;7:3087–3091. doi: 10.1002/j.1460-2075.1988.tb03174.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Sher A. Nature (London) 1976;263:334–336. doi: 10.1038/263334a0. [DOI] [PubMed] [Google Scholar]
  • 24.Sher A, Hein A, Moser G, Caulfield J P. Lab Invest. 1979;41:490–499. [PubMed] [Google Scholar]
  • 25.Thorley-Lawson D A, Schooley R T, Bhan A K, Nadler L M. Cell. 1982;30:415–425. doi: 10.1016/0092-8674(82)90239-2. [DOI] [PubMed] [Google Scholar]
  • 26.Davis S J, van der Merwe P A. Immunol Today. 1996;17:177–187. doi: 10.1016/0167-5699(96)80617-7. [DOI] [PubMed] [Google Scholar]
  • 27.Ianelli C J, Edson C M, Thorley-Lawson D A. J Immunol. 1997;159:3910–3920. [PubMed] [Google Scholar]
  • 28.Ianelli C J, DeLellis R, Thorley-Lawson D A. J Biol Chem. 1998;273:23367–23375. doi: 10.1074/jbc.273.36.23367. [DOI] [PubMed] [Google Scholar]
  • 29.Arock M, Ross E, Lai-Kuen R, Averlant G, Gao Z, Abraham S N. Infect Immun. 1998;66:6030–6034. doi: 10.1128/iai.66.12.6030-6034.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Staunton D E, Fisher R C, LeBeau M M, Lawrence J B, Barton D E, Francke U, Dustin M, Thorley-Lawson D A. J Exp Med. 1989;169:1087–1099. doi: 10.1084/jem.169.3.1087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Sauter S L, Rutherfurd S M, Wagener C, Shively J E, Hefta S A. Infect Immun. 1991;59:2485–2493. doi: 10.1128/iai.59.7.2485-2493.1991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Gbarah A, Gahmberg C G, Ofek I, Jacobi U, Sharon N. Infect Immun. 1991;59:4524–4530. doi: 10.1128/iai.59.12.4524-4530.1991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Wu X R, Sun T T, Medina J J. Proc Natl Acad Sci USA. 1996;93:9630–9635. doi: 10.1073/pnas.93.18.9630. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Baorto D M, Gao Z, Malaviya R, Dustin M L, van der Merwe A, Lublin D M, Abraham S N. Nature (London) 1997;389:636–639. doi: 10.1038/39376. [DOI] [PubMed] [Google Scholar]
  • 35.Orskov I, Orskov F, Birch-Andersen A. Infect Immun. 1980;27:657–666. doi: 10.1128/iai.27.2.657-666.1980. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Fenton M J, Golenbock D T. J Leukocyte Biol. 1998;64:25–32. doi: 10.1002/jlb.64.1.25. [DOI] [PubMed] [Google Scholar]
  • 37.Diep D B, Nelson K L, Raja S M, Pleshak E N, Buckley J T. J Biol Chem. 1998;273:2355–2360. doi: 10.1074/jbc.273.4.2355. [DOI] [PubMed] [Google Scholar]
  • 38.Clarkson N A, Kaufman R, Lublin D M, Ward T, Pipkin P A, Minor P D, Evans D J, Almond J W. J Virol. 1995;69:5497–5501. doi: 10.1128/jvi.69.9.5497-5501.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Simons K, Ikonen E. Nature (London) 1997;387:569–572. doi: 10.1038/42408. [DOI] [PubMed] [Google Scholar]
  • 40.Solomon K R, Rudd C E, Finberg R W. Proc Natl Acad Sci USA. 1996;93:6053–6058. doi: 10.1073/pnas.93.12.6053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Cebecauer M, Cerny J, Horejsi V. Biochem Biophys Res Commun. 1998;243:706–710. doi: 10.1006/bbrc.1998.8149. [DOI] [PubMed] [Google Scholar]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES