Correspondence: Franco Felici, Department of Microbiology, Genetics, and Molecular Biology, University of Messina, Salita Sperone, 31 - 98166 Messina, Italy. Tel.: +39 090 6765197; fax: +39 090 392733; e-mail: firstname.lastname@example.org
Streptococcus pneumoniae is a causative agent of otitis media, pneumonia, meningitis and sepsis in humans. For the development of effective vaccines able to prevent pneumococcal infection, characterization of bacterial antigens involved in host immune response is crucial. In order to identify pneumococcal proteins recognized by host antibody response, we created an S. pneumoniae D39 genome library, displayed on λ bacteriophage. The screening of such a library, with sera either from infected individuals or mice immunized with the S. pneumoniae D39 strain, allowed identification of phage clones carrying S. pneumoniae B-cell epitopes. Epitope-containing fragments within the families of the histidine-triad proteins (PhtE, PhtD), the choline-binding proteins (PspA, CbpD) and zinc metalloproteinase B (ZmpB) were identified. Moreover, library screening also allowed the isolation of phage clones carrying three distinct antigenic regions of a hypothetical pneumococcal protein, encoded by the ORF spr0075 in the R6 strain genome sequence. In this work, Spr0075 is first identified as an expressed S. pneumoniae gene product, having an antigenic function during infection.
Streptococcus pneumoniae (pneumococcus) is a ubiquitous human pathogen that causes significant morbidity and mortality worldwide. It is a gram-positive bacterium that colonizes the upper respiratory tract of the host, causing various invasive diseases such as pneumonia, sepsis and meningitis (Fedson et al., 1999; Tuomanen, 2000). The disease rate is particularly high in children, the elderly and patients with chronic pathologies or immunosuppressive illnesses (Ejstrud et al., 1997; Bogaert et al., 2004a). In adults, pneumococcus is the most frequent cause of community-acquired pneumonia; in children, it is the second most common cause of otitis media and meningitis after Haemophilus influenzae and Neisseria meningitidis (Dagan et al., 1994).
In the last 20 years, at least three different vaccines using S. pneumoniae capsular polysaccharides have been developed and commercialized (for a review, see Obaro, 2002). Such formulations have limited efficacy in children below 2 years and in immunocompromised individuals (Butler et al., 1999), due to the low immunogenicity of capsule polysaccharides and the lack of T-cell activation. Moreover, capsular polysaccharides are serotype-specific and thus the requirement for multivalent protection requires immunization with multiple polysaccharides. Conjugation of pneumococcal polysaccharides to carrier proteins can improve the host immune response and has been successfully introduced in clinical practice (Black et al., 2001), although conjugated vaccine production is still quite expensive. In the last decade, there has been great interest in using surface and secreted pneumococcal proteins conserved among a large number of serotypes as immunogens. Although many pneumococcal proteins, including pneumolysin, surface protein A (PspA), surface adhesin A (PsaA), surface protein C (PspC), neuraminidase and autolysin, have been proposed as potential vaccine candidates, PspA, PsaA and pneumolysin are currently the most promising (Briles et al., 2000; Bogaert et al., 2004b).
This work focuses on the identification of pneumococcal protein domains containing B-cell epitopes, which are recognized by antibodies of infected humans. To this end, we utilized a powerful approach based on λ phage display technology, which has already shown great efficacy in identifying novel antigens from human pathogens, such as Hepatitis C virus (Santini et al., 1998) and Toxoplasma gondii (Beghetto et al., 2001; Beghetto et al., 2003). We show that the challenge of a bacteriophage λ display library of pneumococcal whole genome with sera from infected patients or from S. pneumoniae immunized mice allowed identification of a large panel of pneumococcal protein fragments. The presence of B-cell epitopes within known bacterial antigens, such as the families of choline-binding proteins, histidine-triad proteins and the zinc metalloproteinases, was confirmed. Furthermore, we identified new antigenic regions matching the sequence of a previously unknown pneumococcal protein, corresponding to the ORF spr0075 in the R6 strain whole-genome sequence.
Materials and methods
Streptococcus pneumoniae strains
Streptococcus pneumoniae was maintained in Todd–Hewitt (TH) broth or tryptic soy broth (TSB) in a 5% CO2-enriched atmosphere at 37°C. Solid media were obtained by addition of 1.5% agar and 3% defibrinated horse blood (Biotech s.n.c., Grosseto, Italy) to TSB. Where necessary, streptomycin and kanamycin were used at a final concentration of 500 μg mL−1. Pneumococcal strains D39 (type 2) (Avery et al., 1944; Iannelli et al., 1999), R6 (Hoskins et al., 2001), FP22 (Pearce et al., 2002) and DP1004 (Shoemaker & Guild, 1974) were used for this study. The construction of the FP228 strain, the spr0075 isogenic mutant derived from DP1004, is described in this work.
Pneumococcal display library was affinity-selected using P1 or W3 sera. P1 serum was collected from an adult male hospitalized for acute pneumococcal pneumonia. W3 is a pool of sera from five mice (6-week-old CBA/Jico mice) immunized with the S. pneumoniae D39 strain. Briefly, 107 CFU were resuspended in Freund adjuvant and administered to animals via subcutaneous injection at days 0 and 21. At day 35, sera from immunized mice were collected and used for library screening.
Thirty serum samples were anonymously collected from blood donors and used for analysis of selected antigen fragments' immunoreactivity.
Construction and selection of S. pneumoniaeλ-display library
The S. pneumoniae D39 strain was grown in TSB, and total DNA was extracted using standard procedures (Sambrook & Russell, 2001). Using 1 ng of DNaseI (Sigma-Aldrich) at 15°C for 15 min, 5 μg of genomic DNA was randomly fragmented. The mixture of DNA fragments, ligated with specific adapters (Beghetto et al., 2003), was subjected to 1.5% agarose gel electrophoresis, and DNA inserts between 300 and 1000 bp in length were excised and cloned into vector λKM4 (Minenkova et al., 2003). The display library was packaged in vitro using the ‘Ready-To-Go Lambda Packaging kit’ (Stratagene), yielding about 2 × 107 independent clones. The average size of DNA inserts, analyzed by PCR, was between 300 and 500 bp. The affinity selection of the display library with antibodies from serum samples and isolation of recombinant phage clones were performed as described previously (Beghetto et al., 2003). Briefly, magnetic beads linked to Protein G (Dynabeads Protein-G, Dynal) were incubated for 40 min at room temperature with 10 or 50 μL of serum from humans or mice, respectively. Beads were then incubated with 1010 PFU of recombinant phages. After multiple cycles of affinity selection, pools of recombinant phages were subjected to immunoscreening, and single-phage clones were analyzed by phage-enzyme linked immunoassay (phage-ELISA) as described (Beghetto et al., 2001).
Expression of recombinant antigen fragments
DNA inserts from clones SP-cl.1, SP-cl.2, SP-cl.4, SP-cl.7, SP-cl.13, SP-cl.16, SP-cl.17, SPM-cl.10 and SPM-cl.18 were subcloned into bacterial expression vectors pGEX-SN (Beghetto et al., 2003) and pKM4 (Minenkova et al., 2003), which allowed the expression of recombinant proteins as fusion to glutathione S-transferase (GST) and λ protein D, respectively. Competent Escherichia coli cells (AD202 strain, Sambrook & Russell, 2001) were transformed with recombinant plasmids and single clones were isolated. To induce the expression of fusion proteins, transformed cells were grown in LB medium to OD600 nm=0.6–0.8. Isopropyl-β-d-thiogalactopyranoside was then added to culture (up to 0.4 mM). After induction at 37°C for 3 h, recombinant proteins were purified from the cytoplasm of bacterial cells by one-step affinity chromatography (Beghetto et al., 2003). Protein purity and quantification were assessed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) analysis and Bradford assay, respectively.
Recombinant protein enzyme-linked immunoassays (Rec-ELISA)
An ELISA test for recombinant fusion products was performed by coating Maxisorp-plates (Nunc) with recombinant proteins at a concentration of 2.5 μg mL−1 in 50 mM NaHCO3, pH 9.6. After incubation overnight at 4°C, plates were left for 1 h at 37°C in blocking solution [5% nonfat dry milk, 0.05% Tween-20 in phosphate-buffered saline (PBS)] and subsequently incubated for 1 h at 37°C with sera diluted 1 : 200 in blocking solution. After washing plates, alkaline phosphatase-conjugated antibodies (Sigma-Aldrich) were added to each well. pNPP disodium hexahydrate (Substrate 104; Sigma-Aldrich) was used to reveal enzymatic activity. The results were recorded as the difference between the absorbance at 405 and 620 nm, detected by an automated ELISA reader (Labsystem Multiskan). Assays were performed in duplicate, and the mean values were calculated. GST carrier protein alone was used as a negative control in every assay. Cut-off values were determined for each serum as the mean of the optical density on the GST plus three times the standard deviation. The typical cut-off value was 0.2.
Immunization with recombinant proteins and purification of specific antibodies
Rabbits were immunized with GST-fusion products; 200 μg of recombinant proteins GST-SP-cl.4 and GST-SP-cl.17 were homogeneously mixed with complete Freund adjuvant and used to immunize animals via the subcutaneous route (day 0). Three and 6 weeks after vaccination, rabbits were boosted with the same amount of antigens. At day 35, animals were sacrificed, and sera were collected. In order to affinity-purify specific antibodies against fusion proteins, HisTrap NHS-activated HP columns (Amersham Biosciences/GE Healthcare) were immobilized with 0.5 mg of D-SPcl.4 and D-SPcl.17 fusion proteins according to the manufacturer's instructions. Ten milliliter of immunized rabbit serum was allowed to flow through the column overnight to permit binding of specific IgG to recombinant proteins. After washing the column with PBS and PBS containing 0.5 M NaCl, antigen-specific IgG were eluted with a solution containing 50 mM glycine pH 3, and 100 mM NaCl. The immunoreactivity of affinity-purified antibodies was confirmed by ELISA.
Construction of the spr0075 knockout mutant
An spr0075 mutant in an S. pneumoniae type 2 background was constructed by gene SOEing (Horton et al., 1990), as previously described (Iannelli et al., 1999). In the mutant, the spr0075 gene was replaced with a kanamycin-resistance cassette (aphIII) (Pearce et al., 2002). For this purpose, a 2147 bp DNA fragment containing the aphIII gene (1033 bp) flanked by the regions upstream and downstream of spr0075 (544 and 629 bp, respectively) was generated and directly used to transform the DP1004 strain. The region upstream of spr0075 was amplified from the D39 chromosome using the primer pair 75-1 (5′-CTTCAGCAGGTCAAGACCATGT-3′) and 75-2 (5′-ATCAAACGGATCCCCAGCTTGAGGATTCTCCTTAGAGTCT-3′), while the sequence downstream of spr0075 was amplified using the 75-3 (5′-CCTACGAGGAATTTGTATCTCAGCTAAGGAAATAAATGAT-3′) and 75-4 (5′-CAGTTAAATCCAGCATTTCTT-3′) oligonucleotides. The aphIII gene was amplified from plasmid pJH1 of Enterococcus faecalis (Pearce et al., 2002) using primers IF149 (5′-CAAGCTGGGGATCCGTTTGAT-3′) and IF190 (5′-GATACAAATTCCTCGTAGG-3′). The construction of the mutant was verified by PCR and sequencing. The spr0075-deficient strain was called FP228.
Western blot analysis
Streptococcus pneumoniae strains R6, D39, FP22 and FP228 were grown in 15 mL of TH under standard conditions. Cells were collected by centrifugation at 3000 g for 15 min at 4°C and washed twice with PBS. Bacteria were resuspended in a final volume of 1 mL of SDS-sample buffer (10 mM Tris-HCl, 1 mM EDTA, 1% SDS, 10 mM dithiothreitol, 10% glycerol, 0.01% bromophenol blue) and boiled for 5 min at 95°C. Samples were subjected to SDS-PAGE electrophoresis on a 12% acrylamide gel. Proteins were transferred onto a nitrocellulose membrane (BioTrace NT, Pure Nitrocellulose Blotting Membrane, Pall Life Science) and filters were incubated overnight at 4°C with rabbit affinity-purified antibodies (1 : 100 dilution). Alkaline phosphatase-conjugated goat antirabbit IgG (Sigma-Aldrich) was used as a secondary antibody. Filters were developed with suitable choromogenic substrates, nitroblue tetrazolium (NBT; Sigma-Aldrich) and 5-bromo-4-chloro-3-indosyl phosphate (BCIP; Sigma-Aldrich).
Construction of the S. pneumoniae display library and selection of specific B-cell epitopes
Streptococcus pneumoniae genomic DNA was randomly digested using DNaseI enzyme, and the resulting fragments were cloned as fusion products with the λ bacteriophage protein D (Beghetto et al., 2003; Minenkova et al., 2003). The resulting display library contains 2 × 107 independent clones, whose inserts have an average size of 300–500 nucleotides. In order to identify pneumococcal protein fragments containing B-cell epitopes, the display library was affinity-selected using serum samples either from a patient hospitalized for pneumococcal pneumonia, or from mice immunized with S. pneumoniae D39 strain. Three rounds of affinity-selection were performed with each serum and the resulting phage population was analyzed, after every round of selection, for its immunoreactivity. Recombinant phages were then filter-screened with the same serum used for affinity selection, and their reactivity was subsequently confirmed by phage-ELISA (data not shown).
At the end of the selection procedures, several phage clones bearing distinct protein regions were identified (Table 1). Upon database search, all sequences from selected phage clones match the sequence of S. pneumoniae genes. Clones SP-cl.1 and SP-cl.2, whose inserts are, respectively, 123 and 152 amino acids (aa) in length, matched the sequence of PhtE protein (Adamou et al., 2001). SP-cl.1 corresponds to the N-terminal region of PhtE protein between aa 55 and 178, while SP-cl2 is downstream of SP-cl.1, between aa 235 and 387. Clone SP-cl.7 matches the sequence of PhtD protein (Adamou et al., 2001), with an insert of 278 aa. Clone SP-cl.16 codes for 90 aa, corresponding to the central region of CbpD protein (Gosink et al., 2000). Clones SP-cl.20 and SPM-cl.18 match the sequence of pspA gene (McDaniel et al., 1992). The first insert is 356 aa long and comprises seven of 10 repeated motifs of the choline-binding region. Clone SPM-cl.18 matches the N-terminal region of SP-cl.20 before the start of a choline-binding motif. Clone SPM-cl.10 encodes 308 aa of ZmpB protein (Novak et al., 2000). Three clones, namely SP-cl.4, SP-cl.13 and SP-cl.17, match the sequence of the hypothetical Spr0075 gene product (GenBank accession No. AE008391). Clone SP-cl.4 matches the protein sequence between aa 133 and 334; clone SP-cl.13 is located between aa 443 and 606; and clone SP-cl.17 corresponds to the sequence between aa 851 and 1133. Figure 1 shows the alignments of the recombinant pneumococcal antigen fragments with the sequence of the entire corresponding product.
Table 1. Affinity selection of the λ display library of Streptococcus pneumoniae D39 genome with sera of an infected human, or mice immunized with D39 strain
Serum used for affinity selection
Number of different phage clones isolated
S. pneumoniae genes identified
Human (pneumococcal respiratory disease)
PhtE, PhtD, CbpD, Spr0075, PspA
Mice (Immunized with D39 strain)
Expression of recombinant fusion protein and ELISA reactivity
DNA inserts of clones SP-cl.1, SP-cl.2, SP-cl.4, SP-cl.7, SP-cl.13, SP-cl.16, SP-cl.17, SPM-cl.10 and SPM-cl.18 were subcloned into expression vector pGEX-SN, directing the expression of GST fusion proteins in E. coli cells. With the exception of clone SP-cl.13, all antigen fragments were purified in large amounts from the cytoplasm of bacterial cells. Figure 2 shows the SDS-PAGE analysis of recombinant proteins purified by one-step affinity chromatography. The large-scale purification of GST fusion products yields from 5 to 10 mg L−1 of culture broth, with a purity of at least 90%.
The immmunoreactivity of GST fusion proteins was confirmed in ELISA using sera used during affinity selection (Table 2a). In addition, we analyzed the immunoreactivity of the recombinant proteins against IgG present in the sera from 30 healthy adults (Table 2b). All antigen fragments reacted with human antibodies in more than 40% of samples. GST-SP-cl.16 protein, corresponding to a fragment of pneumococcal CbpD protein, showed the lowest reactivity (43%), while GST-SP-cl.7 protein, corresponding to a PhtD protein fragment, showed the highest (83%).
Table 2a. Reactivity of GST recombinant proteins against sera used in affinity selection
Values indicate the number (and percentage) of reactive sera (mean OD>cut-off) using recombinant proteins.
GST-SP-cl. 17 (Spr0075)
GST-SPM-cl. 18 (PspA)
Molecular characterization of the Spr0075 gene product
Three of the selected antigen fragments matched the sequence of a hypothetical pneumococcal protein, named Spr0075. The spr0075 ORF in the R6 genome of S. pneumoniae encodes a putative protein of 1161 aa (GenBank accession no. NP357669), having an expected molecular mass of 123 kDa. Analysis of the Spr0075 protein sequence reveals the presence of: (a) a putative signal peptide, located between amino acids 1 and 40 (putative cleavage site aa 41), (b) six adjacent repeated regions (152 aa long) and (c) an LPxTG anchoring motif (Schneewind et al., 1993), in the C-terminal region (residues 1148–1152 aa). Comparative analysis of the spr0075 gene from the R6 strains with the gene sequence from different pneumococcal serotypes reveals that the protein is well preserved among all investigated strains (type 19F, 6B, 2, 4, 23F), although the number of repeated regions may vary (data not shown).
In order to analyze the expression of the hypothetical protein Spr0075 in S. pneumoniae, specific antibodies against GST-SP-cl.4 and GST-SP-cl.17 were raised in rabbits. The presence of specific IgG antibodies in sera from immunized rabbits was assayed in ELISA, using Spr0075 antigen fragments expressed as a fusion to λ capsid protein D.
The presence of Spr0075 in cell lysates from S. pneumoniae strains D39, R6 and FP228 was assessed by Western blot, using antigen-specific immunopurified IgG. Figure 3 shows a unique protein band with a molecular mass of c. 120 kDa reacted with anti-Spr0075 antibodies in both encapsulated (D39) and unencapsulated (R6) type 2 pneumococci. In the FP228 mutant strain, in which the spr0075 gene was deleted, such a protein band is not visible, demonstrating that the hypothetical protein is indeed encoded by an spr0075 ORF in the R6 genome sequence (Hoskins et al., 2001).
Streptococcus pneumoniae is a transient member of normal bacterial flora that generally colonizes the upper respiratory tract. Although pneumococcal colonization is generally asymptomatic, it may progress to respiratory or invasive diseases (Bogaert et al., 2004a).
The efficacy of antipneumococcus capsular polysaccharide-based vaccines has been extensively debated, as the protection elicited by capsule polysaccharides is stringently serotype-specific (Hausdorff et al., 2005) and often unable to induce long-term memory response. In the last generation of vaccines (Prevnar/Prevenar, 7-valent pneumococcal conjugate vaccine), purified capsular polysaccharides of seven S. pneumoniae strains were coupled with a protein carrier, in order to exceed the above limitations. The vaccine is effective in 97% of invasive diseases caused by vaccine serotypes and offers some protection against otitis media and pneumococcal carriage (Bogaert et al., 2004b). Owing to the limited serotype coverage, risks of serotype replacement and the high cost of pneumococcal glycoconjugated vaccines, great interest in the development of formulations based on pneumococcal protein antigens has emerged in the last decade (Bogaert et al., 2004b).
The present study focused on identifying antigenic regions of S. pneumoniae proteins through a novel approach, namely, by challenging a pneumococcal genome display library on bacteriophage λ with antibodies from humans or animals infected by the bacterium. We previously demonstrated the potential of λ display technology in identifying antigen fragments involved in the human immune response against Toxoplasma gondii infection, showing that these antigenic regions are involved in both humoral and cell-mediated immunity in the human host (Beghetto et al., 2003; Di Cristina et al., 2004; Beghetto et al., 2005). Here, we show that, by screening an S. pneumoniae genome display library, a large number of recombinant phage clones carrying pneumococcal B-cell epitopes are selected.
The display library affinity-selection permits characterization of ten distinct phage clones, with DNA inserts including sequences corresponding to both known and previously unknown pneumococcal antigens. Among known proteins, three antigenic regions within PhtD and PhtE proteins were identified. Interestingly, previous studies in mice demonstrated that immunization with full-length PhtD protein confers protection from sepsis due to virulent pneumococcal strains (Adamou et al., 2001). In contrast, in rabbits, the N-terminal region of PhtE (aa 21–509) induces high antibody levels that do not protect against infection in passive immunization experiments (Adamou et al., 2001; Hamel et al., 2004). We also identified one immunodominant region within CbpD pneumococcal protein (Gosink et al., 2000). CbpD-deficient mutant strains show a significant reduction in colonization of the nasopharynx in mice, although they are similar to the parental strain in efficiency of genetic transformation, lysis in stationary phase and sensibility to β-lactam antibiotics. Recent studies identify CbpD as a key component in competence-induced cell lysis (Kausmally et al., 2005). Moreover, we identified two immunoreactive epitopes inside PspA and ZmpB pneumococcal proteins. It was demonstrated that a zmpB mutant is largely attenuated in a mouse pneumonia model, suggesting that ZmpB may play a role during respiratory tract infection by the pneumococcus (Chiavolini et al., 2003). In addition to known protein antigens, some antigen fragments matching the sequence of a hypothetical pneumococcal protein, named Spr0075, were identified. The results obtained using purified anti-Spr0075 antibodies in Western blot analysis on whole-cell extracts demonstrate that this protein is really encoded by an spr0075 ORF in the R6 genome, located between nucleotides 80186 and 83671.
Most of the antigen fragments selected from S. pneumoniae display library were efficiently expressed as fusion proteins in the cytoplasm of bacterial cells, and their immunoreactivity was assayed in ELISA using a panel of sera from healthy adults. The immunoreactivity of GST fusion proteins, as assessed with human IgG antibodies, highlights the broad recognition of the selected antigen fragments by the human B-cell response. Notably, these results agree with previous reports that illustrate a broad and specific recognition of pneumococcal protein antigens PspA, PsaA and pneumolysin by IgG antibodies in the healthy population (Rapola et al., 2000; Virolainen et al., 2000). Interestingly, antigenic regions of Spr0075 protein reacted with more than 60% of sera, indicating a broad recognition of this protein antigen.
In perspective, there is great interest in further investigating the ability of these antigen-specific IgG antibodies in conferring protective immunity against S. pneumoniae infection. In particular, the presence of human epitopes in the novel Spr0075 antigen strongly encourages testing protection in animal models of infection, with the aim of assessing its potential as a vaccine candidate for the development of protein-based vaccines against pneumococcal diseases.
We are extremely grateful to Francesco Iannelli for helpful discussion and advice. We acknowledge Emiliano Pavoni, Andrea Spadoni, Katia Genovese and Giovanna Tuscano for their participation in some of the experimental work. We also thank Luca Bruno and Velia Braione for excellent technical support and Marlene Deutsch for linguistic revision of the manuscript.