Attachment of Neisseria gonorrhoeae to the cellular pilus receptor CD46: identification of domains important for bacterial adherence

Authors


*For correspondence. E-mail Helena.Kallstrom@ mtc.ki.se; Tel. (+46) 8 728 71 74; Fax (+46) 8 30 17 97.

Abstract

Pili of Neisseria gonorrhoeae mediate binding of the bacteria to human host cells. Membrane cofactor protein (MCP or CD46), a human cell-surface protein involved in regulation of complement activation, acts as a cellular pilus receptor. In this work, we examined which domains of CD46 mediate bacterial adherence. The CD46 expression was quantified and characterized in human epithelial cell lines. N. gonorrhoeae showed the highest adherence to ME180 cells, which have BC1 as the dominant phenotype. The BC isoforms of CD46 were expressed in all cell lines tested. The adherence was not enhanced by high expression of other isoforms, showing that the BC domain of CD46 is important in adherence of N. gonorrhoeae to human cells. To characterize the pilus-binding site within the CD46 molecule, a set of CD46–BC1 deletion constructs were transfected into COS-7 cells. Piliated N. gonorrhoeae attached well to CD46–BC1-expressing COS-7 cells. We show that the complement control protein repeat 3 (CCP-3) and the serine–threonine–proline (STP)-rich domain of CD46 are important for efficient adherence to host cells. Further, partial deletion of the cytoplasmic tail of CD46 results in low bacterial binding, indicating that the cytoplasmic tail takes part in the process of establishing a stable interaction between N. gonorrhoeae and host cells.

Introduction

Neisseria gonorrhoeae is a Gram-negative diplococcus causing the human-specific disease gonorrhoea. Colonization of mucosal surfaces by N. gonorrhoeae is often modelled in two steps (Nassif et al., 1999). The type IV pilus, a surface organelle that extends from the bacterial cell surface, facilitates the initial attachment. Type IV pili are expressed by numerous Gram-negative bacteria, among which are Neisseria gonorrhoeae, Neisseria meningitidis, Pseudomonas aeruginosa, Bacteriodes nodosus and Vibrio cholerae (Wall and Kaiser, 1999). After the initial pilus-mediated binding, the interaction develops into a more intimate contact between bacterial membrane components and host cell-surface receptors. Several bacteria–host cell interactions take place, leading to uptake of the bacterium and invasion of the epithelial cell layer (Nassif et al., 1999). The opacity (Opa) outer membrane proteins are associated with invasion of host cells. Opa proteins interact with the host cell receptor CD66 (Chen and Gotschlich, 1996; Virji et al., 1996a, b; Gray-Owen et al., 1997) or with heparan sulphate proteoglycans (Chen et al., 1995; van Putten and Paul, 1995).

Pili of N. gonorrhoeae are composed of highly variable major pilus subunits (PilE) that comprise the bulk of the pilus fibre and PilC that is a minor pilus-associated protein. PilC is essential for pilus biogenesis (Jonsson et al., 1991), and acts as an adhesin at the tip of the pilus fibre (Rudel et al., 1995). Upon adherence of piliated Neisseria to the host cell, the type IV pili induce calcium signalling in the host target cell. The pili trigger release of Ca2+ from intracellular stores (Källström et al., 1998).

We have shown that membrane cofactor protein (MCP), also called CD46, acts as a cellular receptor for type IV pili of pathogenic Neisseria (Källström et al., 1997). The binding of piliated bacteria to host cells was blocked with polyclonal and monoclonal antibodies against CD46 and by recombinant CD46 generated in Escherichia coli.

CD46 protects host tissue from complement activation by binding to C3b/C4b and serves as cofactor for factor I-mediated degradation of C3b and C4b. The protein is expressed ubiquitously by human cells, except by erythrocytes (Liszewski et al., 1991). The CD46 structure contains four complement control protein repeats (CCP-1 to CCP-4) of about 60 amino acids each, a serine–threonine–proline (STP)-rich domain, a 12-amino-acid area of undefined function, a transmembrane hydrophobic domain, a cytoplasmic anchor and a C-terminal cytoplasmic tail (Fig. 1). CCP-1, CCP-2 and CCP-4 each possess one N-glycosylation, and the STP domain is O-glycosylated. In the human body, CD46 is expressed in four major isoforms, BC1, BC2, C1 and C2 (Post et al., 1991; Purcell et al., 1991), depending on alternative splicing of the STP domain ‘B’ and the choice between one of two cytoplasmic tails (Cyt-1 and Cyt-2). The majority of human cells have similar ratios of all isoforms, but in the brain, salivary gland and kidney there is a predominance of certain isoforms (Johnstone et al., 1993). CD46 has been identified as a receptor for measles virus (Dörig et al., 1993; Naniche et al., 1993), for the M-protein of Streptococcus pyogenes (Okada et al., 1995) and for human herpesvirus 6 (Santoro et al., 1999). The crystal structure of CCP-1 and CCP-2 of CD46 has been solved (Casasnovas et al., 1999). The binding site of the measles virus has been assigned to a large, glycan-free surface that extends from CCP-1 to CCP-2 of the CD46 molecule (Manchester et al., 1997; Mumenthaler et al., 1997; Hsu et al., 1999; Patterson et al., 1999). The ligand binding site for C3b and C4b consists of CCPs 2, 3 and 4 (Adams et al., 1991; Seya et al., 1998).

Figure 1.

Schematic diagram of the four regularly expressed isoforms of CD46. The CD46 (or MCP) glycoprotein consists of four complement control protein repeats (CCP-1 to CCP-4), an alternatively spliced region designated B and C which is enriched in serines, prolines and threonines (STP domain), an area of undefined (U) function, a transmembrane domain, a cytoplasmic anchor and a cytoplasmic tail (Cyt-1 or Cyt-2). CCP-1, CCP-2 and CCP-4 each possess one N-glycosylation site. Because of alternative splicing, CD46 is expressed as four major isoforms, BC1, BC2, C1 or C2, depending on the STP content and the choice of cytoplasmic tail. The figure is reproduced with permission from Wang et al. (2000) Journal of Immunology 164, 1841. Copyright 2000gr by The American Association of Immunologists.

In this study, we analysed the adherence of piliated N. gonorrhoeae to human epithelial cell lines with different expression levels and isoform distributions of CD46. Further, we determined the domains of CD46 important for pilus recognition and bacterial adhesion. By using deletion constructs of CD46 expressed in COS-7 cells, we show that the STP domain, CCP-3 and the cytoplasmic tail Cyt-1 are crucial for adherence of piliated N. gonorrhoeae to host cells.

Results

Adherence of piliated N. gonorrhoeae to human epithelial cells correlates with the expression of BC isoforms of CD46

We measured and compared the level of bacterial adherence to the human epithelial cell lines ME180 (human cervical carcinoma), HEp-2 (human larynx), FaDu (human pharynx), HEC-1-B (human endometrium) and Chang (human conjunctiva). Piliated (P+) N. gonorrhoeae strain MS11 was allowed to bind a semiconfluent layer of each of the cell lines for 60 min. Figures 2 and 7 show the level of bacterial adherence varied among the different cell lines. N. gonorrhoeae MS11 P+ showed a stronger adherence to ME180, Chang and HEC-1-B cells than to HEp-2 and FaDu cells. The Chinese hamster ovary (CHO) cell line does not mediate adherence of the bacteria and was used as a negative control.

Figure 2.

Adherence of N. gonorrhoeae MS11 P+ to human epithelial cell lines. N. gonorrhoeae MS11 P+ (OD600 = 0.1) were allowed to adhere to a monolayer of cells for 60 min. The unbound bacteria were removed by extensive washing of the cell layer. The infected cells were saponin treated, serially diluted and spread onto GCB plates. Colony-forming units were counted. The bacterial adherence to ME180 cells was set to 100% and corresponded to 40–60 bacteria per cell. Chinese hamster ovary (CHO) cells were used as a negative control. The experiment was repeated three independent times and shown are the means ± standard deviations.

Figure 7.

Confocal image of MS11 P+ adherence to cells. The CD46-expressing COS-7 cell lines were infected with N. gonorrhoeae MS11 P+ for 60 min. The unbound bacteria were removed by extensive washing of the cell layer. The infected cells were fixed with glutaraldehyde, blocked with BSA and labelled with antiserum against non-piliated N. gonorrhoeae followed by FITC-labelled antibodies against IgG. Shown are representative images of four independent experiments. For the images, we used a 40× oil or 20× objective to scan 10 layers with a step-size of 1 µm.

To study the correlation between bacterial adherence and CD46 expression, we analysed CD46 expression of the different human cell lines by flow cytometry using a polyclonal CD46 antibody. As shown in Fig. 3, the CD46 expression was higher in HEp-2 than in the Chang, ME180, FaDu and HEC-1-B cells. For examination of the distribution between BC (Fig. 3, top) and C (Fig. 3, bottom) isoforms, Western immunoblotting was performed on whole cell lysates using CD46-specific antibodies. When separated by SDS–PAGE, CD46 migrates as two forms, one with a molecular mass of 59–68 kDa and one with 51–58 kDa (Liszewski et al., 1996). The higher molecular weight band consists of the BC1 and BC2 isoforms, whereas the lower molecular weight band contains the C1 and C2 isoforms of CD46. This size difference can be attributed mainly to the larger amount of O-linked sugars on the BC1 and BC2 isoforms. As shown in Fig. 4A, the distribution of isoforms varied among the cell lines examined. HEp-2 cells showed two bands of the predicted molecular weight and a pattern of approximately 40% upper band and 60% lower band. For the HEC-1-B, ME180 and FaDu cells, the upper band was prominent, indicating dominant expression of BC isoforms. Two bands were observed in Chang cell lysates, with the band corresponding to the C isoforms being predominant.

Figure 3.

The expression of CD46 in human cell lines. Flow cytometry assay showing CD46 expression in human epithelial cell lines. The cells were detached from the cell culture flask, washed and incubated with a polyclonal anti-CD46 antibody (Källström et al., 1998) followed by a FITC-labelled goat anti-rabbit IgG antibody. The fluorescence intensity of 20 000 cells was measured by flow cytometry. Shown is a representative histogram of three independent experiments.

Figure 4.

Western immunoblotting and RT-PCR analysis of CD46 isoform distribution in human epithelial cell lines.

A. Western immunoblotting. Cell lysates of epithelial cells were blotted with polyclonal antibody to CD46. The immunoblot was variably exposed and for each sample the exposure was selected which optimally demonstrates the protein phenotype. The upper band contains the BC1/BC2 isoforms and the lower band the C1/C2 isoforms of CD46. The control lane represents a mixture of the BC1, BC2, C1 and C2 isoforms, as expressed by transfected CHO cells.

B. RT-PCR analysis of human epithelial cell lines. As controls, the four CHO cell lines transfected with BC1, C1, BC2 and C2 isoforms are shown. For amplification, primers from CCP-4 and the 3′ untranslated region of CD46 were used. RT-PCR product was loaded on 3.0% agarose gel. In this set of experiments BC1 was the dominant isoform. Other experiments with ME180 have shown BC2 to be the predominant isoform (Wang et al., 2000). However, in all instances, BC isoforms were dominant and the variation in isoform expression in ME180 cells is under investigation.

The expression of the different isoforms was further analysed by reverse transcriptase PCR (RT-PCR). Figure 4B shows that in ME180 the BC1 isoform accounted for the majority of the upper band predominant phenotypic pattern. HEC-1-B cells and FaDu cells transcribed almost exclusively the BC1 and BC2 isoforms, accounting for the upper band phenotype in Fig. 4A. Chang cells transcribed all four isoforms, again corresponding with the observed phenotype. HEp-2 cells expressed high levels of all four isoforms, with C2 being predominant. Thus, the high expression of total CD46 and the moderate bacterial adherence to HEp-2 suggests that the C isoforms (C1 and C2) are less efficient in this function.

Taken together, these data argue that the level of bacterial adherence is not dependent solely upon the quantity of CD46 present on the cell surface, but is influenced by the isoform distribution. The strong predominance of the BC isoforms in ME180 cells, which display the highest level of binding, suggests that expression of the BC isoforms is sufficient for tight adherence of the bacteria.

Piliated N. gonorrhoeae adhere to COS-7 cells expressing human CD46

The above studies suggest that the initial pilus-mediated adherence of N. gonorrhoeae to epithelial cells is dependent on the isoform expression pattern of CD46 rather than the amount of CD46 on the cell surface. We have previously shown that piliated Neisseria adhere better to CHO cells transfected with the CD46–BC1 isoform than with the other isoforms (Källström et al., 1997). COS-7 cells were transfected with the CD46–BC1 isoform and deletion and substitution constructs of BC1 (Table 1). These were then analysed for their ability to interact with N. gonorrhoeae. COS-7 was used rather than CHO because levels of bacterial adherence were higher in transfected COS-7 cells than in transfected CHO cells. Expression of CD46 on the surface of the transfected COS-7–BC1 cells was confirmed by flow cytometry using a polyclonal CD46 antibody (Fig. 5). As shown in Fig. 5, all the cell lines prepared in COS-7 expressed similar amounts of CD46, except the BC1 cells and the d7–12 cells in which expression was higher.

Table 1.  CD46 mutants.
Cell lineMutationaSource or reference
  • a

    . N, asparagine; Q, glutamine.

ΔSTPDeletion of the STP domainLiszewski et al. (1998)
Tmδ1Deletion of the CCP-1Manchester et al. (1995)
Tmδ2Deletion of the CCP-2Manchester et al. (1995)
Tmδ3Deletion of the CCP-3Liszewski et al. (2000)
Tmδ4Deletion of the CCP-4Liszewski et al. (2000)
NQ1Q was substituted for N in CCP-1Liszewski et al. (1998)
NQ2Q was substituted for N in CCP-2Liszewski et al. (1998)
NQ4Q was substituted for N in CCP-4Liszewski et al. (1998)
d1–6Deletion of amino acids 1–6 of Cyt-1Liszewski et al. (1994)
d7–12Deletion of amino acids 7–12 of Cyt-1Liszewski et al. (1994)
d13–16Deletion of amino acids 13–16 of Cyt-1Liszewski et al. (1994)
Figure 5.

Expression of CD46 in the different COS-7 constructs. Flow cytometry assay showing CD46 expression in COS-7 cells and CD46-transfected COS-7 cells. The cells were detached from the cell culture flask, washed and incubated with a polyclonal anti-CD46 antibody (Källström et al., 1998) followed by a FITC-labelled goat anti-rabbit IgG antibody. Here, the fluorescence intensity of 20 000 cells was measured by flow cytometry. Shown is a representative histogram of three independent experiments.

Adherence of N. gonorrhoeae is dependent on the CCP-3 domain, the STP domain and the cytoplasmic tail of CD46

We further analysed the adherence of N. gonorrhoeae MS11 P+ to the COS-7 cells expressing the different constructs of CD46. N. gonorrhoeae MS11 P+ were allowed to bind a semiconfluent layer of the CD46-expressing COS-7 cells for 60 min. As shown in Figs 6 and 7, N. gonorrhoeae MS11 P+ adhered well to the CD46–BC1-expressing COS-7 cells. In contrast, non-piliated MS11 did not bind (data not shown).

Figure 6.

Adherence of N. gonorrhoeae MS11 P+ to CD46-expressing COS-7 cells. The CD46-expressing COS-7 cell lines were infected with N. gonorrhoeae MS11 P+ for 60 min. The unbound bacteria were removed by extensive washing of the cell layer. The infected cells were saponin treated, serially diluted and spread onto GCB plates. Colony-forming units were counted. COS-7 cells were used as a negative control. The experiment was repeated three independent times and shown are the means ± standard deviations.

Figures 6 and 7 show that MS11 P+ bound at highest numbers to intact BC1 cells and to NQ1 cells. The deletion of the STP domain abolished bacterial adherence, and the Tmδ3 cells showed reduced interaction with MS11 P+, indicating that the CCP-3, and particularly the STP domain, is important for adherence. Further, MS11 P+ scarcely bound to NQ4, which lacks the N-glycosylation site of CCP-4. Finally, deletion of amino acids 7–12 and 13–16 of the 16-amino-acid cytoplasmic tail Cyt-1 greatly reduced the adherence of MS11 P+. Deletions of amino acids 1–6 of the Cyt-1 also resulted in significant, but less pronounced, reductions in bacterial binding.

It is also interesting to observe that N. gonorrhoeae form microcolonies on the human cells during the 60 min of adherence, but not on the CD46-expressing COS-7 cells (Fig. 7). Taken together, these data show that the STP domain, CCP-3, glycosylation of CCP-4 and the cytoplasmic tail of CD46 are required for optimal adherence of piliated N. gonorrhoeae to host cells.

CCP-3 is crucial for the interaction between bacteria and host cells

To prove further that CD46 is directly involved in attachment of Neisseria gonorrhoeae to host cells, inhibition assays with soluble CD46 and mutant soluble CD46 were performed. Soluble CD46 (sCD46 or sMCP) and its CCP deletion mutant ΔCCP-3 were produced in COS-7 cells, analysed by immunoblotting and used to inhibit binding of piliated MS11 to COS-7 cells expressing the BC1 isoform. Preincubation of the bacteria with sCD46 resulted in much reduced adherence compared with the control (Fig. 8). The deletion mutant lacking CCP-3 did not block bacterial adherence, supporting that CCP-3 is crucial for the bacteria–host cell contact. Thus, inhibition of bacterial adherence using soluble CD46 constructs demonstrate that CCP-3 is clearly involved in the process of Neisseria adherence to target receptors.

Figure 8.

Inhibition of bacterial adherence to COS-7 cells expressing CD46–BC1. Piliated N. gonorrhoeae MS11 were preincubated with medium, sCD46 or sδ3 for 30 min and were used to infect CD46-expressing COS-7 cells for 60 min. Unbound bacteria were removed by extensive washing of the cell layer, the infected cells were saponin treated, serially diluted and spread onto GCB plates. Colony-forming units were counted. COS-7 cells were used as a negative control. The experiment was repeated three independent times and shown are the means ± standard deviations.

Discussion

The initial adherence of piliated pathogenic Neisseria to human cells is mediated by the human-specific cell-surface receptor CD46 (Källström et al., 1997). In this work, we identified regions of CD46 required for adherence of N. gonorrhoeae to host target cells. Although HEp-2 cells express more CD46 than other cell lines examined, piliated N. gonorrhoeae MS11 adhere better to ME180, Chang and HEC-1-B cells than to HEp-2 cells. However, the isoform distribution among these cell lines is different. ME180 expresses predominantly the upper band BC isoforms whereas HEp-2 cells have nearly equal amounts of BC and C isoforms, implying that the bacteria interact primarily with the BC isoforms. This correlates with previously presented data of the adherence of N. gonorrhoeae to CD46-expressing CHO cells, in which the binding was best to CHO cells expressing the BC1 and BC2 isoforms and was very low to C1 and C2 isoforms (Källström et al., 1997). However, the binding to CD46-expressing CHO cells was unstable, and therefore we used a system in which CD46 was expressed in COS-7 cells. Because the CD46 expression level was equal in both CHO and COS-7 cells, it is likely that other factors, such as glycosylation or other proteins, influenced bacterial adherence. It cannot be excluded that cell type-specific differential glycosylation in the cell lines influences the accessibility of the protein receptor.

Infection of ME180, HEp-2, FaDu and Chang cells by N. gonorrhoeae MS11 P+ resulted in bacterial aggregation at 60 min after infection (Fig. 7). Such microcolony formation has been shown to appear as an early stage of localized adherence of Neisseria (Nassif et al., 1999; Pujol et al., 1999). Interestingly, the microcolony formation was absent in the CD46-expressing COS-7 cells, indicating that microcolony formation of the bacteria is also dependent on other factors of the host cell. It is also possible that microcolony formation depends on the phenotype of the bacterial strain.

We show that the interaction between MS11 P+ and CD46 is dependent on CCP-3, the STP domain, N-glycosylation of CCP-4 and the cytoplasmic tail of CD46. CCP-3, along with CCP-2 and CCP-4, are required for C3b and C4b binding and for cofactor activity. In addition, if expressed on a variety of cells, CCPs 2–4 are required for MCP to mediate protection against the human classic and alternative pathway-mediated cytolysis (Adams et al., 1991; Iwata et al., 1995). Therefore, it is possible that the pilus affinity for CCP-3 not only mediates adherence of the bacterium but also affects the C3b/C4b binding and cofactor activity of CD46. Further, studies using soluble CD46 demonstrated that CD46 blocks adherence of N. gonorrhoeae to COS-7 cells expressing human CD46. Further, a CD46 deletion mutant lacking CCP-3 did not inhibit bacterial binding to cells, supporting the importance of this region in the bacteria–host cell interaction.

The interaction of piliated N. gonorrhoeae and CD46 is highly dependent on the STP domain of the CD46 molecule. The importance of the STP domain is also demonstrated by the bacterial preference for binding to the BC isoforms compared with the C isoforms. The STP domain of CD46 has been suggested to serve as a functional modulator to CD46 because the presence of this domain alters complement regulatory activity (Iwata et al., 1995). The O-glycans of the STP domain influence, but are not necessary for, the CD46-mediated protection against classic pathway-mediated cytolysis (Liszewski et al., 1998). Deletions of the STP region of decay-accelerating factor (DAF or CD55), a highly homologous protein to CD46, abolished its cytoprotective capacity (Coyne et al., 1992). Another example of STP contribution has been shown in the case of Dr-fimbriated E. coli internalization into DAF-transfected CHO cells (Lublin and Coyne, 1991; Selvarangan et al., 2000). DAF acts as a cellular receptor for Dr-fimbriated E. coli, and the STP region has been shown to serve as a non-specific spacer to project the CCP-3 region for binding. It is possible that the STP region acts in a similar fashion in the case of gonococcal adherence to CD46.

The N-glycans of CCP-2 and CCP-4 are important for CD46-mediated protection against cytolysis (Liszewski et al., 1998). Interestingly, interruption of the N-glycosylation site on CCP-4 almost completely abrogates bacterial binding, yet Tmδ4 cells, which entirely lack CCP-4, are still capable of mediating strong bacterial adherence. The apparent importance of the glycosylation site in CCP-4 is interesting because previous experiments have demonstrated that recombinant CD46 made in E. coli was able to inhibit bacterial adherence (Källström et al., 1997). This would suggest that the interaction between pili and CD46 are direct protein–protein interactions. A possible explanation for the above findings could be that carbohydrates on CD46 and the pilus have a stabilizing role in the interaction. For example, a deletion of CCP-4 places the pilus binding sites of CCP-3 and the STP domain closer together, possibly circumventing the need for the stabilizing interaction conferred by the N-glycosylation on CCP-4.

CD46 is a transmembrane protein with a cytoplasmic tail and has been suggested to transduce signals into the cell (Karp et al., 1996; Liszewski et al., 1996; Wong et al., 1997; Källström et al., 1998; Seya et al., 1998; Wang et al., 2000). The cytoplasmic tail is most probably required for the signal transduction events during adherence of pathogenic Neisseria to target cells. Upon interaction between pili and the surface-exposed domains of CD46, it is possible that a signal is transduced into the cytoplasm, which leads to induction of intimate bacteria–cell adhesion. Most probably, the deletions of the tail result in a weak interaction and the bacteria are washed away. It has been shown that a signal is transduced through DAF, which also protects cells from complement-mediated lysis by preventing or dissociating the C3 convertase. GPI-(glycosyl phosphatidylinositol) anchored DAF was able to transduce signals leading to interleukin 2 (IL-2) production and tyrosine phosphorylation (Shenoy-Scaria et al., 1992).

Although the adherence of piliated N. gonorrhoeae to the human pilus receptor CD46 is thoroughly investigated in this study, the actual role of PilC in this interaction is still under investigation. Experiments about the role of PilE variation and PilC expression in this interaction are currently being performed. Further, the data shown in this study hold for strain MS11mk (P+).

To summarize, pilus-mediated adhesion of N. gonorrhoeae to host cells is enhanced by expression of the BC isoforms of CD46. Further, the CCP-3, the STP domain and the N-linked glycosylation of CCP-4 are needed for proper adherence. Also, particularly the amino acids 7–12 and 13–16 of the cytoplasmic tail of CD46–BC1 play a role in the bacterial adherence. The function of the tail is most probably in signalling into the eukaryotic cell upon gonococcal adhesion. Investigations concerning the cross-talk between the bacterium and the host are currently being performed.

Experimental procedures

Bacterial strains and growth conditions

N. gonorrhoeae MS11mk (P+, PilC2+, PilC1) has been described previously (Swanson et al., 1986) and is referred to as MS11 P+ in the text. Piliated phenotypes were distinguished by colony morphology under a binocular microscope. The bacteria used did not express detectable levels of Opa, as detected by SDS–PAGE of outer membrane preparations. Bacteria were grown on GCB [GC medium base (Difco)]agar plates containing Kellogg's supplement (Kellogg et al., 1968) at 37°C under 5% CO2 and were passaged every 18–20 h.

Cell lines and growth conditions

All cell culture media were supplemented with 10% inactivated fetal bovine serum (FBS) and 2 mM l-glutamine. ME180 (ATCC HTB33), an epithelial-like human cell line from cervical carcinoma, was maintained in McCoy's 5A medium and was grown in Dulbecco's modified Eagle medium (DMEM). Chang conjunctiva (ATCC CCL 20.2), an epithelial-like human conjunctival cell line, was grown in Medium 199. The following cell lines were all maintained in DMEM: COS-7 (fibroblast-like African green monkey kidney; ATCC CRL 1651), FaDu (epithelial-like human pharynx carcinoma; ATCC HTB 43), HEp-2 (epithelial-like human laryngeal carcinoma; ATCC CCL23) and HEC-1-B (epithelial-like endometrial adenocarcinoma; ATCC HTB113). Transfected COS-7 cells were cultured in DMEM containing 1 mg ml−1 of G418 (Sigma). Chinese hamster ovary (CHO) cells were cultured in Ham's F12 medium and were grown at 37°C under 5% CO2. CHO cells transfected with the four isoforms of CD46 (CHO–BC1, CHO–BC2, CHO–C1 and CHO–C2) have been described previously (Liszewski and Atkinson, 1996). The experiments were carried out in medium free from FBS, l-glutamine and antibiotics. Media and growth supplements were purchased from Life Technologies. Cell culture materials were purchased from Costar.

Adherence assays and inhibition assays

Cells were grown to a semiconfluent layer in tissue culture plates. After washing of the cell layer, N. gonorrhoeae MS11 P+ (OD600 = 0.1) were added and incubated with the cells for 60 min at 37°C under 5% CO2 to allow the bacteria to adhere to the cells. For inhibition assays, bacteria were preincubated with soluble CD46 or its mutants for 30 min, and were then added to a washed cell layer. The wells were extensively washed in medium until no unbound bacteria were seen. The infected cells were treated with 1% saponin for 5 min, serially diluted, spread onto GCB plates and incubated overnight at 37°C, 5% CO2. Colony-forming units were counted the next day.

Flow cytometry measurements

Cells were grown in 75 cm3 flasks until a semiconfluent layer was obtained. The monolayer was washed in PBS (pH 7.4) and treated with 0.2% trypsin for 1 min. The detached cells were washed in 10 ml PBS followed by a gentle centrifugation at 200 g for 15 min. After washing, the cells were incubated with 100 µl of rabbit anti-CD46 immunoglobulin G (IgG) antibody diluted 1:50 (Källström et al., 1998) on ice for 30 min. The cells were washed twice as described above and were incubated with 100 µl of a fluorescein isothiocyanate (FITC)-labelled goat anti-IgG antibody (diluted 1:100) for an additional 30 min on ice. The cells were washed twice in PBS, once in 10 ml FBS and twice with PBS, resuspended in 1 ml of PBS and analysed for fluorescence intensity by FACScan.

Western immunoblotting

Cell lysates were prepared as described previously (Liszewski et al., 1996). Briefly, cell lysates were prepared by solubilizing cells (2 × 107 cells ml−1) in cell lysis buffer [1% Nonidet P-40, 0.05% SDS in Tris-buffered saline (TBS)] for 15 min at 4°C, followed by collection of the supernatant after centrifugation at 12 000 g for 10 min. The lysates were electrophoresed by 10% SDS–PAGE and transferred to a nitrocellulose membrane using a Bio-Rad semidry transfer system. Membranes were incubated overnight at 4°C with 5% non-fat dry milk in TBS with 0.05% Tween-20. Membranes were immunoblotted using rabbit polyclonal CD46 antiserum (1:3000). Blots were then washed and incubated with horse radish peroxidase (HRP)-conjugated donkey anti-rabbit IgG (1:10000) (Amersham–Pharmacia Biotech). After washing, the blots were developed using the Supersignal West Pico Luminol reagents (Pierce).

RT-PCR

RNA was harvested from cell monolayers using the RNeasy mini protocol (Qiagen), as per the recommendations of the manufacturer. cDNAs were prepared using the SuperScript One-Step RT-PCR System (Life Technologies). RNA (1 µg) in a total volume of 50 µl was combined with 10 pmol each of the primers 5′-GTGGTCAAATGTCGATTTCCAGTAGTCG-3′ and 5′-CAAGCCACATTGCAATATTAGCTAAGCCACA-3′. Cycling conditions consisted of a step for cDNA synthesis at 51°C for 30 min followed by predenaturation of 2 min at 94°C. The cDNA was amplified by 40 cycles at 94°C for 15 s, 54°C for 30 s and 71°C for 1 min, followed by one final extension cycle of 72°C for 10 min. PCR products were analysed on a 3.5% agarose gel.

Transfection

Plasmids containing cDNA from the BC1, ΔSTP, NQ1, NQ2 and NQ4 mutants have previously been described (Liszewski et al., 1998). Deletion constructs d1–6, d7–12 and d13–16 (Liszewski et al., 1994), and Tmδ1 and Tmδ2 (Manchester et al., 1995) of CD46 have also been described previously. CD46 mutants Tmδ3 and Tmδ4 were made from soluble CD46 mutants (Adams et al., 1991) by adding the transmembrane and cytoplasmic tail (Cyt-1) of CD46. These were cloned into the EcoRI site of the vector pSG5 (Stratagene) (Liszewski et al., 2000). Purified plasmids were transfected into COS-7 cells using lipofectin (Life Technologies) according to the protocol of the manufacturer. Briefly, COS-7 cells were cultured in a six-well culture dish until 50% confluence was obtained. Plasmid DNA (5 µg) was added to 300 µl serum-free DMEM, and, in another tube, 35 µl of Lipofectin was mixed with 300 µl serum-free DMEM. The two solutions were incubated separately at 20°C for 45 min, gently mixed and then left for another 10 min. The mixture was added together with 2.4 ml of DMEM to washed COS-7 cells and was incubated for 5 h at 37°C under 5% CO2. The medium was replaced with serum containing DMEM, and after 2 days G418 (Sigma) was added to the medium to the final concentration of 1 mg ml−1.

Transfection of soluble CD46 (sCD46) constructs was performed as described by Adams et al. (1991), and the mutant proteins were checked by immunoblotting. A cDNA insert encoding sCD46 (sMCP) and its mutant deleted of CCP-3 (ΔCCP-3) were used. All constructs producing soluble protein are deleted of the transmembrane domain and cytoplasmic tail.

Confocal imaging

Cells were subcultured in chamber slides (Lab-Tek). N. gonorrhoeae MS11 P+ (OD600 = 0.1) were allowed to bind to the cells for 60 min. The cell layer was then extensively washed until no unbound bacteria were seen. The infected cells were fixed with 1% glutaraldehyde for 30 min and blocked with 2% BSA for 2 h. Adherent bacteria were detected with rabbit antiserum against non-piliated MS11 (1:200) (Jonsson et al., 1991) and with goat anti-rabbit IgG–FITC (1:500). Images of adherent N. gonorrhoeae were made using a MultiProbe 2001TM CLSM confocal laser scanning system (Molecular Dynamics) equipped with a diaphot 200 inverted microscope (Nikon). An excitation filter of 488 nm and the emission filter 510EFLP were used. The images were visualized by a 40 × 1.4 oil objective. Data were collected in a stack of 10 layers with a Z-step size of 1 µm.

Acknowledgements

This work was supported by grants from the Swedish Medical Research Council (Dnr 10846), the Swedish Cancer Society and from Lars Hiertas Stiftelse and Åke Wibergs Stiftelse. Financial support was also provided by funding from the National Institutes of Health (RO1 AI 37618). H. K. was supported by grants from Vårdalstiftelsen, the Swedish Society for Medical Research and the Karolinska Institute.

Ancillary