Resistance of Neisseria meningitidis to Human Serum Depends on T and B Cell Stimulating Protein B

Abstract

The ability of the human bacterial pathogen Neisseria meningitidis to cause invasive disease depends on survival in the bloodstream via mechanisms to suppress complement activation. In this study, we show that prophage genes coding for T and B cell stimulating protein B (TspB), which is an immunoglobulin-binding protein, are essential for survival of N. meningitidis group B strain H44/76 in normal human serum (NHS). H44/76 carries three genes coding for TspB. Mutants having all tspB genes inactivated did not survive in >5% NHS or IgG-depleted NHS. TspB appeared to inhibit IgM-mediated activation of the classical complement pathway, since survival of the tspB triple knockout was the same as that of the parent strain or a complemented mutant when the classical pathway was inactivated by depleting NHS of C1q and was increased in IgM-depleted NHS. A mutant solely carrying tspB gene nmbh4476_0681 was as resistant as the parent strain, while mutants carrying only nmbh4476_0598 or nmbh4476_1698 were killed in ≥5% NHS. The phenotype associated with TspB is formation of a matrix containing TspB, IgG, and DNA that envelopes aggregates of bacteria. Recombinant proteins corresponding to particular subdomains of TspB were found to have human IgG Fcγ- and/or DNA-binding activity, but only TspB derivatives containing both domains formed large, biofilm-like aggregates when combined with purified IgG and DNA. Recognizing the role of TspB in serum resistance may lead to a better understanding of why strains that carry tspB genes are associated with invasive meningococcal disease.

INTRODUCTION

Invasive Neisseria meningitidis is a major cause of bacterial meningitis and sepsis worldwide. The reasons for why some N. meningitidis strains cause disease and others do not are not well understood. With the exception of periodic epidemics occurring mainly in sub-Saharan Africa, disease caused by pathogenic N. meningitidis is relatively rare. However, asymptomatic carriage is comparatively common, ranging from ∼5% to >80% depending on the population studied (1, 2). Host factors associated with increased risk of disease include complement deficiencies, carriage state, genetics, social behavior, and geographic location (reviewed in reference 3). Strains causing invasive disease, on the other hand, appear to be limited to those from a few, so-called hypervirulent lineages (2). However, not many specific characteristics of N. meningitidis strains have been identified that can be linked directly to disease. In an epidemiological study comparing disease-causing isolates with carriage isolates by microarray analysis during an outbreak of meningococcal disease in the Czech Republic, Bille et al. found a statistically significant association of the presence of prophage DNA with isolates that caused disease (4). However, the reasons for the link between the prophage DNA and invasive disease were not determined. Recently, we showed that the prophage gene ORF6 codes for an IgG-binding protein specific for the Fc portion of a human IgG2 paraprotein (5). The protein, which is known as T and B cell stimulating protein B (TspB), mediates formation of a biofilm-like matrix that contains TspB, IgG, and DNA (5).

IgG-binding proteins (Igbps) are produced by several human-pathogenic bacterial species (6–8). The presence of Igbps on the bacterial surface has been shown to promote survival in the human host by inhibiting opsonophagocytosis and providing resistance to complement-dependent bacteriolysis (6–8). For example, Escherichia coli immunoglobulin-binding proteins (Eibs), which are similar to N. meningitidis TspB in being encoded by prophage DNA occurring as multiple copies in the bacterial genome, were shown to promote survival in normal human serum (NHS) (8). However, the amino acid sequences of meningococcal TspBs have no significant homology to those of the Eib proteins or other members of the Igbp family of proteins. Also, Eib proteins are autotransporters, while TspB does not appear to have a similar functional activity.

The aim of the present study was to address the question of whether TspB is functionally important in promoting resistance to bacteriolysis in human serum and, if so, what mechanism is involved.

MATERIALS AND METHODS

Ethics statements.

Donated human blood used in this study was obtained from adult donors under a protocol approved by the UCSF Benioff Children's Hospital Oakland Institutional Review Board, with written informed consent obtained from all donor participants.

All procedures involving animals were performed in the UCSF Benioff Children's Hospital Oakland Research Institute (CHORI) Animal Research Facility, which is an AAALAC (Association for Assessment and Accreditation of Laboratory Animal Care International)-accredited facility. The investigators adhered to the Guide for the Care and Use of Laboratory Animals (9). Protocols involving the use of animals were approved by the CHORI Institutional Animal Care and Use Committee.

Bacterial strains.

N. meningitidis strain H44/76 (B:P1.7,16:fHbp ID 1:ST 32) is a wild-type N. meningitidis strain obtained in 1976 from a patient who died from meningococcal meningitis and septicemia during an epidemic in Norway (10). Strains with knocked-out tspB genes were constructed in H44/76 as described previously (5). The complemented mutant with reintroduced tspB genes was made using the same methods as those for the tspB triple knockout, except that nmbh4476_0681 was knocked out first, followed by nmbh4476_0598 and nmbh4476_1698. Wild-type tspB genes were reintroduced by phage transduction with phage produced in wild-type H44/76, the complemented mutant was selected by survival in 20% NHS, and the identities of reintroduced tspB genes were determined by PCR as described previously (5).

The nonencapsulated SiaD (11) and factor H-binding protein (fHbp) (12) knockout mutants of H44/76 were provided by Serena Giuntini (CHORI).

Serum survival.

Bacterial strains were grown overnight on chocolate agar plates and then cultured in chemically defined medium (CDM) supplemented with 1 mg/ml of Cohn fraction IV (CFIV) from human serum (Sigma-Aldrich, St. Louis, MO). Previously, we had shown that culturing bacteria in CDM plus CFIV resulted in surface exposure of TspB and promoted bacterial aggregation and matrix formation in the presence of human IgG (5). The CDM was Catlin 6 medium (13) modified to contain 5.5 mM glucose, 4 mM d,l-lactate, 50 μM cysteine, and 150 μM cystine. The stock cysteine and cystine solutions were prepared just prior to use. The modifications made to Catlin 6 medium were found to be essential to support growth rates of N. meningitidis strains equivalent to those in rich medium. Cultures were started at an optical density at 620 nm (OD₆₂₀) of ∼0.15 and grown to an OD₆₂₀ of ∼0.6. The bacteria were then diluted 1:20,000 in Dulbecco's phosphate-buffered saline (DPBS) with Ca²⁺ and Mg²⁺ (Mediatech, Manassas, VA) containing 1% (wt/vol) human serum albumin (HuSA; Sigma-Aldrich). Bacteria (10 μl) were mixed with NHS, IgG-depleted NHS, IgM-depleted NHS, or C1q-depleted NHS, with the remaining volume made up with DPBS-1% HuSA, in the wells of a 96-well flat-bottomed microtiter plate, such that the final concentration of complement was 0%, 5%, 20%, or 60% in a total volume of 40 μl. IgG-depleted NHS was prepared as described previously (14). IgM-depleted NHS was prepared by incubating 1 ml of NHS containing 10 mM EDTA and 1 M NaCl with 2 ml of anti-human IgM agarose (Sigma-Aldrich) equilibrated with phosphate-buffered saline (PBS) with the same concentrations of EDTA and NaCl for 45 min at 4°C. The IgM-depleted NHS was buffer exchanged with DPBS by dialysis and concentrated (10-kDa Spin-X UF unit; Corning) at 4°C to achieve the same hemolytic complement activity as that of NHS prior to IgM depletion. As a control, IgM eluted from the column was added back to the IgM-depleted NHS. Purified IgM was obtained by eluting the anti-IgM column with 0.2 M histidine, pH 2.8, containing 24 mM sucrose and 0.01% Tween 20, adjusting the pH to 7 with 1 M Tris, and concentrating the serum (100-kDa Spin-X UF unit) to one-fourth the original volume of NHS applied to the column. The IgM depletion procedure removed 99% of IgM compared to that in the original NHS preparation as determined by a capture-based enzyme-linked immunosorbent assay (ELISA). Hemolytic complement activities of NHS and depleted NHS were determined using an EZ Complement CH50 test kit (Diamedix Corp., Miami, FL). C1q-depleted NHS and purified C1q were obtained from Complement Technologies, Inc. (Tyler, TX). Control wells were set up by the same procedure, but with 60% heat-inactivated NHS or depleted NHS. The plate was then placed in an incubator at 37°C on a rotating platform for 1 h. After 1 h, 10 μl from each well was dripped onto a chocolate agar plate and incubated overnight at 37°C. Each assay was performed with three biological replicates.

Fluorescence microscopy.

Complemented and tspB knockout mutants were grown in CDM supplemented with 1 mg/ml Cohn fraction IV of human serum and 50 μg/ml purified human IgG. After washing the cells in PBS containing 1% bovine serum albumin (BSA), a 1:200 dilution of sera from mice immunized with adjuvant only (negative control) or anti-TspB 1628 CR (5) was added to the same buffer and incubated for 1 h at ambient temperature. The cells were washed, and bound mouse IgG was detected with Alexa Fluor 488-conjugated goat anti-mouse IgG(H+L) F(ab′)₂ secondary antibody (Life Technologies, Grand Island, NY). After a wash in the same buffer to remove unbound secondary antibody, the cells were fixed with 0.5% formaldehyde in PBS and placed as spots onto coverslips, and the coverslips were mounted on slides for fluorescence microscopy (Zeiss Axioplan) as described previously (5).

Expression and purification of TspB derivatives.

The TspB NT and CT subdomains were constructed from a pQE30 expression plasmid containing an nmb_1628 gene segment coding for the CR domain from N. meningitidis B strain MC58. The plasmid was digested with BmgB1, followed by digestion with either BamHI or HindIII, and religated after purification of the digested plasmid in an agarose gel to generate a plasmid for expression of the CT or NT derivative, respectively. The CT construct was found to contain an added base pair, which put the CT gene out of frame. This was corrected by digesting the plasmid with BamHI and HindIII, purifying the CT-encoding fragment, and ligating it into pQE32 digested with the same enzymes. The resulting construct coded for a protein with the amino-terminal sequence MRGSHHHHHHGIRPF, where the underlined segment corresponds to the beginning of the amino acid sequence of CT as encoded by nmb_1628. The NT derivative had the same sequence as the CR derivative reported previously (5) but ended at the carboxyl terminus with the sequence MYRLALA. The four TspB derivatives (i.e., FL, CR, NT, and CT) (see Fig. 3) were expressed and purified as described previously, except that only fractions eluted from a Ni²⁺-nitrilotriacetic acid (Ni²⁺-NTA) column in 0.2 M acetic acid containing 6 M guanidine-HCl were retained for further purification on a CM Sephadex column. The combined 0.2 M acetic acid–6 M guanidine-HCl-eluted fractions were dialyzed in 50 mM sodium phosphate buffer, pH 6.5, containing 2.5 mM cetyltrimethylammonium bromide (CTAB) and then loaded onto a CM Sephadex ion-exchange column (HiTrap CM FF; GE Healthcare). The column was eluted with a linear gradient of 0 M to 0.7 M NaCl in 50 mM sodium phosphate-2.5 mM CTAB buffer and then with a step gradient to 1 M NaCl. Protein eluted with 1 M NaCl was dialyzed (Spectra/Por S/P 7 RC 1-kDa pore membrane; Fisher Scientific) in 10 mM Tris, pH 7, 2.5 mM CTAB buffer and then concentrated to ∼2 mg/ml by packing of the dialysis bag in polyethylene glycol 8000 (PEG 8000) powder (Sigma-Aldrich). Protein concentrations were determined using a Direct Detect spectrometer (EMD Millipore).

FIG 3.

IgG-, Fab-, and Fcγ-binding activities of recombinant TspB derivatives and live bacteria. (A) Schematic representation of the putative domains of TspB and corresponding recombinant proteins. (B) TspB NT (lane 1) and CT (lane 2) derivatives resolved in a 4% to 12% SDS-PAGE gel and stained with Simply Blue. (C) Human IgG1 (circles), IgG2 (squares), and IgG3 (triangles) paraprotein binding to recombinant NT (filled symbols, solid lines) and CT (open symbols, dashed lines) TspB derivatives by ELISA. (D) Whole-molecule human IgG2 paraprotein (filled circles, solid line) and Fab (filled squares, dashed line) and Fc (filled triangles, dashed line) fragment binding to the recombinant NT TspB derivative by ELISA. The controls (open symbols) for the ELISA binding experiments show the respective levels of IgG paraprotein binding to irrelevant proteins purified from E. coli carrying a plasmid with an out-of-frame tspB gene by the same purification method as that for the TspB derivatives. (E) Human polyclonal IgG binding (solid line) compared to Fc (dashed line) and Fab (dotted line) binding to live N. meningitidis B strain H44/76 by flow cytometry. The shaded histogram is for a negative control with HuSA substituted for IgG or IgG fragments to show binding by the secondary antibody alone.

IgG binding by ELISA.

The ability of the TspB NT and CT derivatives to bind to human IgG paraproteins (EMD Millipore) and IgG2 Fab and Fc fragments was determined by ELISA as described previously (5). IgG2 and polyclonal IgG Fab and Fc fragments were prepared as described by Müller et al. (5). Purified polyclonal human IgG was prepared during IgG depletion of NHS as described previously (14). Human IgG1 and IgG2 subclass paraproteins and IgG3 purified from human plasma, used for the ELISA shown in Fig. S2 in the supplemental material, were obtained from Meridian Life Sciences, Memphis, TN. Biotinylated mouse anti-human IgG subclass-specific reagents and alkaline phosphatase-conjugated streptavidin were obtained from Life Technologies.

Flow cytometry.

Measurements of IgG, Fc, and Fab binding to live bacteria were performed as described previously (5), using Dylight 488-conjugated donkey anti-human IgG(H+L) F(ab′)₂ (Jackson ImmunoResearch, West Grove, PA).

DNA binding and analysis of DNA binding data.

TspB binding to DNA was determined using gel mobility shift assays. The following two DNA derivatives were used: linear plasmid DNA [CsCl gradient-purified pSK(+) digested with EcoRI; New England BioLabs, Ipswich, MA] and a biotin-UMP-labeled 61-bp oligonucleotide prepared by PCR. The oligonucleotide produced by PCR (using a New England BioLabs kit) required a mixture of deoxynucleoside triphosphates (dNTPs) and biotin-UTP (Fisher Scientific) and was amplified from H44/76 genomic DNA by using the primers 5′-GGAATTTTCTGATAGCAACACCGA-3′ and 5′-GCTGGTTCTGCAAAACTAAGAGA-3′. The plasmid DNA was purified by agarose gel electrophoresis (QIAquick gel extraction kit; Qiagen, Valencia, CA), and oligonucleotides were purified by use of Centri-Spin columns (Princeton Separations, Adelphia, NJ). For plasmid DNA binding experiments, each TspB derivative was serially diluted in DPBS containing Ca²⁺ and Mg²⁺ (MediaTech), 32 ng/ml plasmid DNA, and 10% glycerol. The samples were loaded directly without tracking dye into a 1% agarose submarine gel in Tris-borate-EDTA (TBE) buffer. DNA was stained with ethidium bromide (1 μg/ml), and the gel image was recorded using an AlphaImager (Protein Simple, Santa Clara, CA) photo documentation system. Binding to oligonucleotides was performed similarly, except that the biotin-labeled DNA concentration was 0.2 μg/ml and the complexes were resolved in either 5% or 15% polyacrylamide-TBE gels (Criterion; Bio-Rad, Hercules, CA). The DNA was transferred to a Hybond-N+ nylon membrane (GE Healthcare, Pittsburgh, PA) by use of a Trans-Blot SD (Bio-Rad) semidry blot transfer cell. After UV cross-linking (Stratalinker 2400; Fisher Scientific), the membrane was blocked using Odyssey blocking buffer (Li-Cor, Lincoln, NE) containing 1% sodium dodecyl sulfate, and the DNA was detected after staining with IRDye 800CW-labeled streptavidin (Li-Cor) on a Li-Cor Odyssey infrared imaging system.

DNA oligonucleotide gel mobility shift data were analyzed by least-squares fitting to the following equation (15) for cooperative, nonspecific binding of TspB NT and CT derivatives to DNA:

$$\frac{R}{L}=\left[K\left(1-cR\right)\right]\times {\left\{\frac{\left(2\omega -1\right)\left(1-cR\right)+R-H}{\left\{2\left(\omega -1\right)\left(1-cR\right)\right\}}\right\}}^{\left(c-1\right)}{\left[\frac{1-\left(c+1\right)R+H}{2\left(1-cR\right)}\right]}^2$$

Where

$$H=\sqrt{{\left[1-\left(c+1\right)R\right]}^2+\left[4\omega R\left(1-cR\right)\right]}$$

Complex/aggregate formation.

Purified recombinant TspB derivatives (10 μM), protein G-purified polyclonal human IgG (10 μg/ml), and pSK(+) plasmid DNA (32 ng/ml) were incubated at ambient temperature for 1 h in a 10-μl total volume of DPBS with Ca²⁺ and Mg²⁺ containing 1% human serum albumin (DPBS-HuSA). The solution was placed onto a glass coverslip, dried, and fixed with 4% (wt/vol) paraformaldehyde. DNA was stained with Hoechst 33258, and human IgG with Dylight 488-conjugated goat anti-human IgG(H+L) (Jackson ImmunoResearch, West Grove, PA), and fluorescence micrographs were recorded using a Zeiss Axioplan microscope as described previously (5).

RESULTS

tspB genes are required for N. meningitidis resistance to human serum.

Group B strain H44/76 carries three complete tspB genes, annotated nmbh4476_0598, nmbh4476_0681, and nmbh4476_1698 in GenBank, and two incomplete genes homologous to the 5′ end of the full-length (FL) genes. Mutants of H44/76 in which combinations of the complete tspB genes were knocked out were tested for the ability to survive in increasing concentrations of NHS. We showed previously that a mutant of H44/76 with the three complete tspB genes knocked out did not express TspB, have human IgG-binding activity, or form large bacterial aggregates in a biofilm-like matrix (5). The effect of knocking out tspB genes on serum survival was compared to the effects in capsular polysaccharide and fHbp knockout mutants, which are known to have impaired survival in NHS (reviewed in reference 16). To control for the potential effect of antibiotic resistance genes in the tspB knockout strains, a tspB triple knockout strain was complemented by reintroducing tspB genes via transduction with phage produced by wild-type H44/76. The resulting complemented strain retained triple antibiotic resistance but contained reintroduced tspB genes nmbh4476_0681 and nmbh4476_1698, with an unknown copy number per cell.

At 60 min, there was ∼100% survival of the wild-type parent strain and the complemented tspB triple knockout with reintroduced tspB genes when the strains were incubated with up to 60% NHS compared to survival in heat-inactivated negative-control NHS (which was set at 100%) (Fig. 1A). In contrast, the triple tspB, capsular polysaccharide, and fHbp knockout mutant strains were each killed in ≥5% NHS (Fig. 1A).

FIG 1.

Survival of N. meningitidis B strain H44/76 wild-type (WT), nonencapsulated (Cap KO), fHbp knockout (fHbp KO), and TspB knockout (TspB triple KO) strains and of the complemented TspB triple KO strain (complemented mutant) in normal human serum (NHS). (A) Survival of WT (filled circles, solid line), complemented mutant (open circles, dashed line), TspB triple KO (filled squares, solid line), Cap KO (open triangles, dashed line), and fHbp KO (inverted filled triangles, solid line) strains in NHS. (B) Survival of TspB triple KO (open squares, dashed line) and complemented mutant (open circles, dashed line) strains in C1q-depleted NHS to inactivate the classical complement pathway or in C1q-replete NHS (filled squares and circles, respectively, with solid lines). (C) Survival of WT (filled circles, solid line), complemented mutant (open circles, dashed line), TspB triple KO (filled squares, dashed line), Cap KO (open triangles, solid line), and fHbp KO (inverted filled triangles, solid line) strains in IgG-depleted NHS. (D) Survival of TspB triple KO in IgM-depleted (open squares, dashed line) or IgM-replete (filled squares, solid line) NHS. (E) Survival of the WT (filled circles, solid line) compared to double mutants carrying a single complete nmbh4476_0598 (open squares, solid line), nmbh4476_0681 (open triangles, dashed line), or nmbh4476_1698 (filled diamonds, dashed line) tspB gene in NHS. All bacteria were cultured in CDM supplemented with 1 mg/ml of Cohn fraction IV prepared from human serum. Percent survival was calculated by using the number of CFU after 60 min of incubation in 60% heat-inactivated serum as 100% survival. Error bars represent the ranges for three biological replicates.

TspB is an IgG-binding protein and therefore might afford resistance to complement-dependent bacteriolysis by inhibiting antibody activation of the classical complement pathway. When the immunoglobulin-dependent classical pathway of complement was inactivated by depleting NHS of C1q, the tspB triple knockout mutant survived as well as the complemented strain in 60% complement (Fig. 1B). When bacteriolytic activity was restored by adding back C1q, the tspB triple knockout mutant did not survive in ≥5% C1q-replete NHS. There was 100% survival of the complemented mutant control strain in 60% C1q-depleted or -replete NHS (Fig. 1B).

We also tested survival of the tspB triple knockout mutant in human serum that had been depleted of IgG by use of protein G or of IgM by use of anti-human IgM agarose. The depletion procedure reduced the concentration of IgG from ∼10 mg/ml to ∼10 μg/ml and, separately, the measured concentration of IgM from ∼0.5 mg/ml to ∼6 μg/ml. The wild-type strain and the complemented mutant survived in 60% IgG-depleted NHS (Fig. 1C), while the TspB, capsular polysaccharide, and fHbp knockout mutants survived only in 5%, not 20% or higher, IgG-depleted NHS. However, there was 50% survival of the tspB triple knockout in 20% IgM-depleted NHS but <14% survival in 5% IgM-replete NHS (Fig. 1D), suggesting that the increase in serum resistance of the tspB knockout mutant resulted from the removal of IgM.

To determine whether more than one tspB gene was required to promote survival in NHS and if there were differences in the abilities of specific tspB genes to promote survival, we prepared mutants in which two of the three genes were knocked out, leaving only a single complete gene. We found by using quantitative PCR that all three tspB genes were expressed at similar levels in the wild-type parent strain grown in CDM-CFIV, using expression of mRNA coding for glyceraldehyde phosphate dehydrogenase as the internal control (N. E. Moe and G. R. Moe, unpublished observations). The tspB double mutant having only the tspB gene nmbh4476_0681 was as resistant to serum as the wild-type strain, but the mutants with only nmbh476_0598 or nmbh4476_1698 did not survive in >5% NHS (Fig. 1E).

To determine whether the differences in serum resistance of the mutants carrying single tspB genes resulted in changes of the characteristic bacterial aggregates in a matrix containing TspB, aggregates of bacteria containing a single tspB gene were compared to those of the complemented mutant and tspB triple knockout strains for the presence of TspB. As shown in Fig. 2, TspB was abundant in the complemented mutant, absent from the tspB triple knockout, present but decreased relative to that in the complemented mutant for the nmbh4476_0681 strain, and decreased further in the nmbh4476_0598 mutant. The TspB level of the nmbh4476_1698 mutant was comparable to that observed for the nmbh4476_0598 mutant (data not shown). By flow cytometry, TspB was not present on the surfaces of bacteria of wild-type H44/76 or any of the mutants and was present in bacterial aggregates only when bacteria were grown in CDM-CFIV with human IgG (data not shown). Since the anti-TspB polyclonal serum is reactive with diverse TspB CR sequences, the results suggest that there are differences in the abilities of the TspB proteins to form the matrix or to accumulate outside bacteria. The latter might result from differences in protein production or secretion from the bacteria.

FIG 2.

Reactivities of anti-TspB in aggregates of the complemented mutant, the TspB triple KO, and mutants carrying single tspB gene nmbh4476_0681 or nmbh4476_0598 by fluorescence microscopy. The negative control was the complemented mutant with sera from mice immunized with adjuvant only. Bacteria were grown in CDM plus CFIV plus purified human IgG. Bound anti-TspB mouse IgG was detected with Alexa Fluor 488-conjugated anti-mouse IgG secondary antibody.

A subdomain of TspB has IgG Fcγ-binding activity.

Previously, we showed that a recombinant protein corresponding to a highly conserved region of TspB, referred to as TspB CR, had IgG-binding activity specific for the Fc fragment of a human IgG2 paraprotein (5). Based on sequence analysis, it appeared that TspB CR could be divided further, into two subdomains, with one containing the amino-terminal 111 amino acids of the CR domain (designated NT) and the other containing the remaining 142 amino acids of the carboxyl-terminal end of CR (designated CT). The subdomains are shown schematically in Fig. 3A, and the sequences are compared to the full-length protein in Fig. S1A in the supplemental material. Recombinant proteins corresponding to the NT and CT portions were prepared from sequences cloned from N. meningitidis B strain MC58 tspB gene nmb_1628, which is representative of the most common TspB amino acid sequence and is 98% identical to any of the corresponding sequences of the three TspB proteins produced by strain H44/76 (the CR domains are ≥96% identical [see Fig. S1B]). The purified recombinant proteins were observed as single bands by SDS-PAGE (Fig. 3B, arrows), although the CT derivative had an apparent mass that was slightly lower than predicted and was very poorly stained by Simply Blue after dialysis in buffer containing the detergent CTAB.

The NT and CT subdomains were tested for IgG-binding activity by ELISA with the TspB derivatives as the solid-phase antigen. As shown in Fig. 3C, the NT protein bound to the IgG2 but not IgG1 or IgG3 human paraprotein standards from EMD Millipore. IgG2 binding to the NT protein was mediated via Fcγ (Fig. 3D), as observed previously for the TspB CR derivative (5). In contrast, the CT protein did not bind to whole IgGs of any of the three IgG subclass paraproteins tested (Fig. 3C). The controls in Fig. 3D show binding of the respective IgG paraproteins to proteins from E. coli transformed with a plasmid containing a gene coding for TspB CR with a frameshift mutation such that no TspB protein was produced, prepared using the same methods as those for the recombinant TspB derivatives.

Recombinant TspB derivatives bind to normal polyclonal IgG purified from human serum but not to IgA or IgM (5). In addition, as shown in Fig. S2 in the supplemental material, recombinant TspB CR (i.e., NT plus CT) binds to human polyclonal IgG2 and IgG3 purified from donor serum. However, it did not bind to the IgG3 paraprotein standards obtained from EMD Millipore, as seen, for example, in Fig. 3C. The reason for variable reactivity of TspB CR with paraproteins may be related to the aberrant glycosylation known to occur with human paraproteins (17). For example, TspB CR binding to the IgG2 paraprotein used in Fig. 3C increased when the paraprotein was treated with a sialidase (5).

To show that the specificity for IgG Fcγ observed for the recombinant TspB derivatives, which were based on the sequence of TspB encoded by nmb_1628, was the same for TspB expressed by H44/76, we compared binding of purified human polyclonal IgG with Fab and Fcγ fragments prepared from the same polyclonal IgG to live bacteria grown in CDM, using flow cytometry. As shown in Fig. 3E, Fcγ but not Fab fragments were bound by the bacteria. In data not shown, the bacteria cultured in CDM containing IgG or Fcγ fragments formed aggregates enveloped in a matrix of TspB, IgG or Fcγ, and DNA similar to those described previously (5) and shown in Fig. 2.

TspB derivatives bind nonspecifically to DNA.

To determine whether the full-length (FL), CR, NT, and CT TspB derivatives bind nonspecifically to DNA, each of the proteins was combined with linear plasmid DNA, and the resulting complexes were identified by an agarose gel mobility shift assay (Fig. 4A). TspB derivative-DNA binding, as indicated by decreased mobility compared to that of plasmid DNA alone, was observed for all of the derivatives (Fig. 4A). To estimate relative DNA binding constants, TspB derivatives were combined with a biotin-labeled 61-bp DNA fragment, and the protein-DNA complexes were resolved in 5% polyacrylamide gels. An example blot of a DNA mobility shift in a 5% polyacrylamide gel with the TspB FL derivative is shown in Fig. 4B. The figure shows how the gel mobility shift (R) is determined as the ratio of the distance traveled by the complex (d) to the distance traveled by DNA in the absence of protein (d₀). Curves of R values versus TspB derivative concentrations observed in 5% polyacrylamide gels are shown in Fig. 4C. Based on the protein concentration for half-maximal DNA mobility shift, all TspB derivatives tested had association constants that were between 10⁶ M⁻¹ and 10⁸ M⁻¹. The CT derivative had an ∼100-fold greater DNA-binding activity than the NT derivative, with the FL derivative having the intermediate activity expected for a protein containing both NT and CT domains that compete for the same DNA-binding sites. Scatchard plots of the gel mobility shift data for the FL derivative were complex and could not be fitted using a simple model for noncooperative or cooperative binding to a homogenous lattice (15; data not shown). Scatchard plots for the NT and CT derivative-DNA complexes resolved in 15% polyacrylamide gels (Fig. 4D, upper and lower panels, respectively) were curved, indicating that nonspecific binding was cooperative. Fits of the data (Fig. 4D) using the equation given in Materials and Methods provided estimates of the binding constants and cooperativity factors (Table 1). All binding experiments shown were performed in DPBS that contained Ca²⁺ and Mg²⁺. The results were the same with DPBS without Ca²⁺ and Mg²⁺ (data not shown), indicating that TspB derivative binding was not dependent on calcium and magnesium cations.

FIG 4.

DNA-binding activities of recombinant TspB derivatives measured by gel mobility shift assay. (A) Agarose gel mobility shifts observed for recombinant TspB FL, CR, NT, and CT derivatives with linear plasmid DNA [pSK(+)]. The protein concentrations ranged from 10 μM to 4.6 nM, with a 3-fold dilution in each successive lane. The last lane on the right contained DNA with buffer only. (B) Example of TspB FL gel mobility shift assay using a biotin-UMP-labeled 61-bp oligonucleotide resolved in a 5% polyacrylamide gel. Each lane shows the gel mobility shift for a 3-fold serial dilution of TspB FL, beginning at 10 μM. The last lane on the right contained DNA with the mock-purified E. coli proteins described as a negative control in the legend to Fig. 3. The solid line indicators show how the gel mobility shift (R) was determined as the ratio of the distance traveled by the complex (d) to the distance traveled by DNA in the absence of protein (d₀). (C) Plot of R values for a 61-bp biotin-labeled oligonucleotide resolved in a 5% polyacrylamide gel as a function of the TspB FL (filled circles, dashed line), NT (filled triangles, solid line), or CT (X's, solid line) derivative concentration. (D) Scatchard plots of observed (filled circles, solid lines) and calculated (open circles, dashed lines) gel mobility shift data for NT and CT TspB derivative binding to a 61-bp biotin-labeled oligonucleotide resolved in a 15% polyacrylamide gel. The calculated curves were obtained by least-squares fitting (Excel; Microsoft) of the observed data to a previously described equation (15) (see Materials and Methods).

TABLE 1.

DNA-binding parameters determined by fitting gel mobility shift data for TspB NT and CT derivatives^a

TspB derivative	ω	K (M−1)	ωK (M−1)	R∞
NT	19	0.26 × 106	5 × 106	0.5
CT	8.6	2.55 × 106	22 × 106	0.5

^aK, association constant; ω, cooperativity factor; R∞, DNA mobility shift R at saturating protein concentration.

TspB derivatives form a biofilm-like matrix when combined with purified DNA and IgG.

It is unclear whether the characteristic aggregates of N. meningitidis enveloped in a matrix of TspB, IgG, and DNA observed in the presence of NHS (5) or CDM supplemented with IgG (Fig. 2) depend on the presence of additional molecules produced by this bacterium. To address this question, the TspB derivatives were combined with purified IgG and plasmid DNA alone or in combinations. Complexes formed were visualized by fluorescence microscopy after DNA and IgG staining with Hoechst 33258 and Dylight 488-conjugated anti-human IgG, respectively. As shown in the fluorescence micrographs in Fig. 5, purified recombinant TspB FL and NT derivatives formed aggregates with IgG, but only the FL derivative formed an interconnected, biofilm-like matrix when combined with DNA and IgG (Fig. 5). Although the CT derivative did not bind to human IgG (Fig. 3C and 5), it did form relatively small IgG-containing aggregates when combined with DNA (Fig. 5), but not the much larger, biofilm-like aggregates observed for the FL derivative.

FIG 5.

Fluorescence micrographs of complexes formed by purified recombinant TspB derivatives (indicated at left), purified polyclonal IgG, and plasmid DNA. Fluorescence (light areas) indicates DNA and human IgG staining.

Like DNA, group B capsular polysaccharide is a polyanionic polymer. However, none of the TspB derivatives formed biofilm-like complexes with IgG combined with N. meningitidis B capsular polysaccharide (data not shown), suggesting that the formation of large complexes requires the specific combination of TspB, IgG, and DNA.

DISCUSSION

N. meningitidis must survive and continue to divide in the bloodstream in order to cause meningitis, sepsis, and other clinical manifestations of invasive disease. The mechanism for N. meningitidis survival in human blood involves resistance to complement activation that would otherwise lead to bacteriolysis and/or opsonophagocytosis (reviewed in reference 16). Factors known to be important for N. meningitidis resistance to complement activation include capsular polysaccharide, certain lipooligosaccharide (LOS) structures (18–20), factor H-binding protein (fHbp) (21), and other proteins that bind factor H (22–24). Although epidemiological studies established a link between the presence of prophage DNA and strains that cause invasive meningococcal disease (4, 25), the possible role of specific prophage genes in serum resistance has not been described previously.

The results presented here show that prophage tspB genes promote survival of N. meningitidis B strain H44/76 in NHS. Of the three tspB genes present in H44/76, nmbh4476_0681 alone was sufficient to promote survival in up to 60% NHS. Reasons for the predominant importance of one tspB gene over the others are not known but do not appear to be the result of differences in the amino acid sequences of the IgG/DNA-binding CR domain. These domains are nearly identical to each other (96% identical to that of the 0681 protein), the differences are conservative, and recombinant TspB CR derivatives having as little as 75% identity to the most common TspB CR (encoded by nmb_1628) were found to be functionally equivalent with respect to immunoglobulin (5) and DNA (P. Q. Richards and G. R. Moe, unpublished observations) binding (see Fig. S1 for sequence comparisons). In contrast to the CR domains, the remaining segments of TspB are highly variable, particularly at the amino-terminal end. These variable sequences may affect protein production, secretion, or, possibly, incorporation into phage particles. It is unclear whether the matrix containing TspB, IgG, and DNA is formed from TspB alone or as a component of a phage particle. In any case, the relative amount of TspB in the matrix (Fig. 2) was correlated with survival in NHS (Fig. 1E).

Increased survival of the tspB triple knockout strain in C1q- and IgM-depleted NHS indicates that TspB promotes serum resistance by inhibiting IgM-mediated activation of the classical pathway of complement. Since recombinant TspB derivatives bind poorly to IgM antibodies (5) (see Fig. S2 in the supplemental material), the effect of TspB on complement-mediated killing appears to be indirect.

There are many examples of Igbps expressed by pathogenic bacteria, including staphylococcal protein A and M proteins, streptococcal protein G, and E. coli Eib proteins (reviewed in reference 26). Binding to different immunoglobulins and Ig-like segments enables bacteria to evade innate and adaptive immune responses by a variety of mechanisms. As a result, Igbps are important virulence factors. Igbp activity is generally directed to constant domains of any of several Ig classes and subclasses but can also be limited to a single class or subclass. In addition to Ig-binding activity, Igbps can be multifunctional, with distinct domains for binding to Igs and other serum proteins, such as factor H, serum albumin, plasmin, and fibrinogen. The binding of Igbps to Igs and other serum proteins, such as the staphylococcal M proteins, can result in the formation of matrices that surround bacteria. For Gram-positive bacteria, the effect of immunoglobulin binding is to inhibit opsonophagocytosis and the alternative pathway of complement.

Thus, some characteristics of TspB-mediated serum resistance are similar to those for other Igbps, but TspB is unique in other respects. Like that of other Igbps, TspB binding was specific for the Fcγ portion of human IgG. However, TspB appears to inhibit the classical pathway of complement rather than the alternative pathway and does not appear to be bound directly to the surfaces of bacteria in amounts detectable by flow cytometry or fluorescence microscopy. Additionally, TspB has at least one other functional activity, namely, binding to DNA. The data in Fig. 4 suggest that this occurs in a cooperative, nonspecific manner. Both of these functional activities are located within discrete subdomains, with the NT domain having both IgG Fcγ- and DNA-binding activities, while the CT domain binds only DNA. Curiously, the protein sequence of TspB has no significant homology to any other known Igbp, and to date, there are no other Igbps that have been shown to bind DNA. The four TspB derivatives tested all showed binding to DNA, even though the NT and CT derivatives (the smallest effective units tested) have no sequence homology to each other, no obvious segments of high positive charge, and an isoelectric point predicted to be <7 based on amino acid sequence (i.e., an overall negative charge at pH 7.4). When combined with human IgG and DNA, recombinant TspB formed large aggregates similar to the matrix observed for N. meningitidis bacteria cultured in NHS (5) or CDM+CFIV+IgG (Fig. 2). This implies that TspB is the mediator of IgG and DNA interactions that leads to formation of a matrix that increases resistance to antibody-mediated, complement-dependent bacteriolysis.

It has been established in many studies that complement-dependent bacteriolysis is the major correlate of protection against invasive N. meningitidis disease (reviewed in reference 16). To survive in the bloodstream of the human host, N. meningitidis strains employ several mechanisms to prevent complement activation and/or amplification of complement activation via the alternative pathway. These include capsular polysaccharide, sialylated LOS, and at least three proteins that bind fH directly, including fHbp (27), NspA (22), and porin B2 (24). Our working hypothesis is that the formation of the TspB/IgG/DNA matrix that surrounds aggregates of bacteria is key to TspB-mediated serum resistance. For example, the IgG-containing matrix might facilitate nonproductive activation of complement far from the bacterial surface, or TspB may have other functional activities, not yet identified, that are enhanced by matrix formation. Importantly, capsular polysaccharide-, fH binding-, or TspB-dependent mechanisms individually were not sufficient to promote survival in NHS, since mutants lacking any one of these three factors were not able to survive in >5% NHS or IgG-depleted NHS. While it is not known whether the results obtained for H44/76 can be applied to all N. meningitidis strains, prophage DNA, which carries the tspB genes, has been shown to be prevalent among strains causing invasive disease (4). However, the plain presence of a tspB gene in invasive or even noninvasive strains does not necessarily imply that it contributes to serum resistance, as shown for nmbh4476_0598 and nmbh4476_1698, whose presence or absence individually was not sufficient to promote survival in NHS.

In summary, the results presented in this study suggest that prophage-encoded TspB may have an important role in shifting normally commensal N. meningitidis strains to disease-causing invasive strains, in part through its effect on IgG binding, through mediating aggregate/matrix formation, and through inhibiting the classical complement pathway. A better understanding of the role of TspB in meningococcal disease and immune evasion may improve the identification of potentially invasive N. meningitidis strains, as well as point toward new approaches to the treatment and prevention of meningococcal disease.