ABSTRACT
Glycerophosphodiester phosphodiesterase (GlpQ) metabolizes glycerophosphorylcholine from the lung epithelium to produce free choline, which is transformed into phosphorylcholine and presented on the surfaces of many respiratory pathogens. Two orthologs of glpQ genes are found in Streptococcus pneumoniae: glpQ, with a membrane motif, is widespread in pneumococci, whereas glpQ2, which shares high similarity with glpQ in Haemophilus influenzae and Mycoplasma pneumoniae, is present only in S. pneumoniae serotype 3, 6B, 19A, and 19F strains. Recently, serotype 19A has emerged as an epidemiological etiology associated with invasive pneumococcal diseases. Thus, we investigated the pathophysiological role of glpQ2 in a serotype 19A sequence type 320 (19AST320) strain, which was the prevalent sequence type in 19A associated with severe pneumonia and invasive pneumococcal disease in pediatric patients. Mutations in glpQ2 reduced phosphorylcholine expression and the anchorage of choline-binding proteins to the pneumococcal surface during the exponential phase, where the mutants exhibited reduced autolysis and lower natural transformation abilities than the parent strain. The deletion of glpQ2 also decreased the adherence and cytotoxicity to human lung epithelial cell lines, whereas these functions were indistinguishable from those of the wild type in complementation strains. In a murine respiratory tract infection model, glpQ2 was important for nasopharynx and lung colonization. Furthermore, infection with a glpQ2 mutant decreased the severity of pneumonia compared with the parent strain, and glpQ2 gene complementation restored the inflammation level. Therefore, glpQ2 enhances surface phosphorylcholine expression in S. pneumoniae 19AST320 during the exponential phase, which contributes to the severity of pneumonia by promoting adherence and host cell cytotoxicity.
INTRODUCTION
Phosphorylcholine (ChoP) is a unique feature of many respiratory bacterial species. In the cell wall of pneumococci, ChoP is a component of lipoteichoic acid and teichoic acid that anchors various choline-binding proteins (CBPs) via the choline-binding domain on the pneumococcal surface (1). These CBPs regulate autolysis and the natural transformation of Streptococcus pneumoniae (2), as well as promoting the internalization of pneumococci into pharyngeal epithelial cells to escape phagocytes (3–6).
The structure of ChoP resembles that of the platelet-activating factor (PAF), and thus, ChoP on S. pneumoniae can interact directly with the host PAF receptor (7) and allow S. pneumoniae to transit the epithelial and endothelial layers during invasion (8). The conjugation between ChoP and the PAF receptor is important for pneumococcal sequestration during immune clearance and systemic dissemination, since mice deficient in PAF receptor or treated with PAF receptor antagonist abolish pneumococcal pneumonia progression to cause sepsis and meningitis (9).
The presentation of ChoP on the bacterial surface requires an available choline source in the environment (10, 11). Free choline is metabolized and incorporated onto the S. pneumoniae surface via the lic operon (12), but the major choline source in the respiratory tract is phosphatidylcholine, which is the most abundant surfactant component that lines the lungs and that contributes to the surface activity (13, 14). The turnover of phosphatidylcholine by phospholipase A2 in the lungs generates glycerophosphorylcholine via lysophosphatidylcholine and free fatty acids (15). Glycerophosphorylcholine is utilized by bacteria via glycerophosphodiester phosphodiesterases (GlpQs), encoded by the glpQ gene, to release choline and glycerol-3-phosphate; therefore, free choline is obtained, which is utilized via the lic operon.
In Mycoplasma pneumoniae, GlpQs facilitate the cleavage of glycerophosphorylcholine into glycerol-3-phosphate, which feeds into glycolysis after oxidation, and a by-product, hydrogen peroxide, results in cytotoxicity. The inactivation of glpQ in M. pneumoniae results in the complete loss of cytotoxicity against HeLa cells (16). Moreover, the GlpQ (protein D) of nontypeable Haemophilus influenzae is an outer membrane protein that is required for the acquisition of choline from the host to present on its surface (17). Thus, GlpQ mediates the long-term colonization of the nasopharynx and infection of the middle-ear space (18). In a murine otitis media infection model, a mutation in glpQ in nontypeable H. influenzae reduced the cytotoxicity to host cells in nasopharyngeal organ cultures (19) and decreased the virulence by 100-fold (20).
S. pneumoniae possesses two orthologs of glpQ genes: glpQ and glpQ2 (Kyoto Encyclopedia of Genes and Genomes [KEGG]). The glpQ locus, which encodes a membrane domain of glycerophosphodiester phosphodiesterase (GPDPase_memb) and a glycerophosphodiester phosphodiesterase family domain (GDPD), is widespread in S. pneumoniae strains with fully sequenced genomes. In contrast, the glpQ2 locus, with GDPD alone, is present in only a few strains, such as SPN034156, TCH8431/19A, Hungary19A-6, Taiwan19F-14, A026, ST556, and 670-6B. In S. pneumoniae, glpQ2, but not glpQ, shares high similarity with glpQ in H. influenzae and M. pneumoniae in terms of its domain structure. However, the role of glpQ2 in S. pneumoniae is still not clear.
Serotype 19A has emerged as the most common S. pneumoniae serotype that causes invasive pneumococcal diseases (21, 22) in the United States. The predominance of non-PCV7 serotypes suggests a serotype replacement effect, but increases in the incidence of serotype 19A have also occurred in countries without pneumococcal conjugate vaccine (PCV7) implementation (23, 24). Indeed, in Taiwan, one-third of children aged <5 years received PCV7 vaccination, but the overall rate of invasive pneumococcal disease still increased significantly (25). This was due to the clonal expansion of a serotype 19A sequence type 320 (19AST320) clone, which accounted for about 50% of the isolates that caused invasive pneumococcal disease and severe pneumonia among children aged <5 years (25). The increased incidence of this serotype 19AST320 clone in Taiwan highlights its specific virulence in causing severe pneumococcal disease, and it is spreading widely.
In the present study, we focused on the pathophysiological role of glpQ2 in a clinical serotype 19AST320 strain, CGMH836, to determine whether glpQ2 is an important virulence factor that mediates colonization and the severity of pneumonia.
MATERIALS AND METHODS
Ethics statement.Animal experiments were performed in strict accordance with the Taiwan regulations of the Animal Protection Act and the course on Animal Care and Use in Research and Education from the American Association for Laboratory Animal Science. All of the experiments were approved by the Institutional Animal Care and Use Committee of the National Defense Medical Center Laboratory Animal Center (permit no. IACUC-13-009).
Bacterial strains and cultures.S. pneumoniae CGMH836 was isolated from the pleural fluid of a child with necrotizing pneumonia and empyema (25). D39 (26) was purchased from the National Collection of Type Cultures, and TIGR4 (27) was purchased from the American Type Culture Collection. WU2 (28) and A66.1 were provided by David E. Briles (29). Clinical pneumococcal strains were obtained from Chang Gung Memorial Hospital. S. pneumoniae was cultured on Todd-Hewitt agar plates supplemented with 0.5% yeast extract (THY) and 5% sheep blood or in freshly prepared THY broth. Bacterial washes from plates cultured overnight were adjusted to an optical density at 580 nm (OD580) of 2, diluted 1/50 in THY broth for growth phase monitoring, and cultured until the OD580 reached 0.4 to 0.5 for the virulence assays. The antibiotic supplements were spectinomycin (500 μg/ml) and/or chloramphenicol (4 μg/ml).
Detection of glpQ and glpQ2 genes.A published pneumococcal genome database was searched to identify glpQ and glpQ2 genes using a bioinformatics tool based on lipid metabolism in the KEGG pathway. Pneumococcal DNA samples were prepared with DNAzolDirect (Molecular Research Center) according to the manufacturer's instructions. The glpQ and glpQ2 genes were determined by PCR using the primer pairs glpQF/glpQR and glpQ2F/glpQ2R, respectively (Table 1), and Phusion DNA polymerase (Thermo Scientific). The annealing temperature was 50°C, and extension was performed at 56°C.
Primers used in this studya
Construction of glpQ2 mutants and cis-complementation strains.The primers used in this study are listed in Table 1. The primer sequences were based on the TCH8431/19A genome. A primer pair, glpQ2 KOF and glpQ2 KOR, was employed to clone the glpQ2 gene from S. pneumoniae strain CGMH836. The spectinomycin resistance (Spcr) gene (30) was inserted into the EcoRI/MscI site (nucleotides [nt] 106 to 788) of the glpQ2 gene. This construct was delivered into CGMH836 via natural transformation, and the transformants were checked using the primer pair sp19A_glpQ2F and sp19A_glpQ2R. The cis-complementation strain was generated by inserting the glpQ2 gene (cloned using the primer pair glpQ2 KOF and glpQ2-comR) and the chloramphenicol acetyltransferase (CAT) gene (31) into the EcoRI site of the sp12134-sp12135 DNA fragment (cloned with the primer pair sp12134F and sp12135R). The plasmid construct was delivered into the glpQ2 mutant strain by natural transformation, and the glpQ2 complementation strain was confirmed by PCR using the primer pairs Up-sp12134F/Dn_sp12135R and glpQ2F/glpQ2R.
Natural transformation.The natural transformation procedure was modified from that of Bricker and Camilli (32). In brief, a mid-log-phase starter culture (OD580 = 0.5) was diluted 1/50 in THY broth containing 11 mM HCl and 0.5% glycine and incubated in a 5% CO2 incubator for 2 h. The culture was then supplemented, in the following order, with 10 mM NaOH, 0.2% freshly resuspended bovine serum albumin (BSA), 1 mM CaCl2, and 0.1 μg/ml competence-stimulating peptide 1 (CSP-1) (EMRLSKFFRDFILQRKK) before incubation for 15 min at 37°C and then addition of the donor DNA (5 μg/ml). After 1 h, the culture was diluted 1/4 with an enriched medium that contained 10 mM glucose and 10% fetal bovine serum before incubation for 2 h and then plating onto THY agar plates with appropriate antibiotics.
Detection of ChoP on the pneumococcal surface.A bacterial broth cultured overnight was diluted 1/50 with freshly prepared THY medium to collect bacteria growing in different phases. The bacterial broth was centrifuged and resuspended in filtered phosphate-buffered saline (PBS) containing 0.5% BSA. The primary antibody (Ab), IgA hybridoma TEPC15 (Sigma-Aldrich), was diluted 1/300 and incubated with the bacteria at 37°C for 30 min, as described previously (33). The ChoP expression level was determined using a phycoethryrin (PE)-conjugated anti-mouse IgA antibody (clone 11-44-2; eBioscience) and analyzed by flow cytometry (BD FACSCalibur).
CRP binding and complement activation assay.The binding ability of C-reactive protein (CRP) on pneumococci was examined as described previously (33). Pneumococci were incubated until the mid-log phase, harvested, and resuspended in 100 μl of 1× Veronal buffer (Lonza). Next, 10 μg of recombinant human CRP (rhCRP) (Prospec) and 2 μg of fluorescein isothiocyanate (FITC)-conjugated anti-human CRP Ab (Abcam) were used to detect pneumococcal surface-adherent CRP. Flow cytometry (BD FACSCalibur) was used for the analysis. The complement activation assay was modified from the method of Neeleman et al. (34). In brief, 2 × 108 CFU pneumococci were incubated with 100 μg/ml rhCRP at 37°C for 1 h and washed once. The pneumococci were then treated with 20% CRP-depleted human serum (BBI Solutions) or healthy human serum at 37°C for 1 h. Heat-inactivated (56°C for 30 min) sera were used as controls. The bacterial survival rate was examined based on plate counts.
CBP isolation, quantification, and analysis.CBPs were eluted from pneumococcal surfaces as described previously (35). In brief, 200 ml of pneumococci was cultured until the mid-log phase, harvested, and resuspended in PBS containing 2% choline chloride at room temperature for 30 min with mild agitation. The eluted CBPs were filtered through a 0.45-μm polyvinylidene difluoride (PVDF) membrane, followed by 20% trichloroacetic acid precipitation. The protein pellet was then dissolved in lysis buffer (2 M thiourea, 7 M urea, and 4% CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate}) and quantified using the Bradford assay (Bio-Rad). The CBP profile was analyzed by two-dimensional SDS-PAGE. The proteins were separated using an immobilized Pharmalyte gradient of pH 4 to 7 with ReadyStrip 4.7-5.9 buffer (Bio-Rad) and 10% SDS-PAGE. The proteins were then visualized by silver staining.
Adherence assay.The human lung epithelial cell lines A549 and BEAS-2B were employed to determine pneumococcal adherence ability. The cells were incubated in Dulbecco's modified Eagle's medium (Gibco; high glucose) and F-12K Nutrient Mixture Kaighn's Modification medium (Gibco), respectively, which were both supplemented with 2 mM l-glutamine and 10% fetal bovine serum (Biological Industries) but without antibiotics. Confluent epithelial cells in a 24-well plate were infected with 108 CFU pneumococci and incubated at 37°C in a 5% CO2 incubator (36). At 1 h after infection, the cells were washed three times with PBS and lysed using 1 ml of distilled water. The cell-associated bacteria were examined by plating onto blood agar plates.
Cytotoxicity assay.The cytotoxicity assay was modified from that of Schmidl et al. (16). A549 cells were cultured to confluence in a 96-well plate without antibiotics and infected with 5 × 106 CFU pneumococci for 2 h at 37°C. At least six wells were tested for each pneumococcal strain infection. Cells without bacterial infections were used as controls. The cell survival rate was determined based on 0.1% crystal violet staining followed by ethanol extraction. The cytotoxicity was evaluated as [1 − (OD595 of infected group/OD595 of control group)] × 100%.
Murine respiratory tract colonization.Three-week-old BALB/c mice were infected intranasally (upper respiratory tract infection) or intratracheally (lower respiratory tract infection) with 5 × 107 CFU/20 μl of pneumococci (37). The level of bacterial colonization was determined based on plate counts of nasal lavage fluid, which was flushed from mice with upper respiratory tract infections after 48 h, or using homogenized lungs from mice with lower respiratory tract infection after 24 h. Histopathology examination of lung tissue was performed after staining with hematoxylin and eosin.
Statistical analysis.Differences between groups were analyzed using an unpaired Student t test (two tailed). All of the experiments were conducted three or more times.
RESULTS
Genes encoding potential GlpQs in S. pneumoniae.The KEGG bioinformatics database was used to analyze glycerophospholipid metabolism in S. pneumoniae strains. Two orthologs of glpQ encode potential GlpQs in pneumococcal strains: glpQ and glpQ2. We aligned the amino acid sequences of GlpQ (amino acids [aa] 298 to 587, excluding the GPDPase_memb domain) and GlpQ2 (aa 1 to 277) with those of GlpQs from H. influenzae and M. pneumoniae, which showed that residues Glu-46, Asp-48, His-61, and Asp-62 in GlpQ2 and Glu-369, Asp-371, His-384, and His-385 in GlpQ were strictly conserved around the active site of GDPD (Fig. 1) (16, 38). This suggests that both GlpQ and GlpQ2 are biologically active enzymes, although only the glpQ gene is found widely in published S. pneumoniae genome sequences. Indeed, the glpQ gene was detected in all of the S. pneumoniae strains listed in Table 1 by PCR screening.
Sequence alignments of GlpQ and GlpQ2 from S. pneumoniae with GlpQ orthologs in H. influenzae and M. pneumoniae. The multiple-sequence alignment was performed using Clustal Omega, and the visualization was generated with BOXSHADE v3.21. Black shading indicates >80% shared identity, and gray shading represents >80% similarity. The arrowheads indicate the strictly conserved active sites. The UniProtKB entry names of the aligned sequences are D6ZNR9_STRP0 (GlpQ2, S. pneumoniae), D6ZLF8_STRP0 (GlpQ, S. pneumoniae), GLPQ_HAEIN (GlpQ, H. influenzae), Y420_MYCPN (GlpQ, M. pneumoniae), and Y566_MYCPN (MPN566, M. pneumoniae).
Using KEGG pathway analysis, the glpQ2 gene could be found in only a few strains in the database, such as SPN034156 (serotype 3), TCH8431/19A (serotype 19A), Taiwan19F-14 (serotype 19F), and 670-6B (serotype 6B), so we analyzed the correlation between glpQ2 and serotypes. We identified the prevalence of glpQ2 among 50 S. pneumoniae strains by PCR analysis. Among the tested strains, all of the serotype 3 (3/3), 19A (6/6), and 19F (5/5) strains possessed glpQ2, whereas only 80% (12/15) of the serotype 6B isolates possessed the gene. The glpQ2 gene was also detected in strains of serotypes 6A, 9, 9V, and 23F, but not in other strains (Table 2). Thus, glpQ2 was present in most of the serotype 3, 19A, and 19F isolates that we examined, although the strains represented multiple different sequence types. However, the correlation between glpQ2 and serotypes could not be confirmed because of the low coverage of strains tested in each serotype.
Prevalences of glpQ2 among clinical S. pneumoniae strains
glpQ2 resides in a transposon-like DNA fragment.The sequence analysis showed that a ca. 10-kb DNA fragment containing the glpQ2 gene was present in TCH8431/19A (GenBank accession no. NC_014251.1) but absent from TIGR4 (GenBank accession no. NC_003028.3) (see Fig. S1 in the supplemental material). The corresponding region in TIGR4 was replaced by three open reading frames (ORFs), i.e., SP_1927 and SP_1928, which encode IS1381 transposases A and B, respectively, and SP_1929, which encodes a hypothetical protein. In TCH8431/19A, the glpQ2-containing insertion was flanked by transposases (HMPREF0837_12162/HMPREF0837_12171) and an IS4 family transposase (HMPREF0837_12163/HMPREF0837_12172), suggesting that glpQ2 resides in a transposon-like DNA fragment, which possibly facilitates genetic movement among S. pneumoniae strains.
Construction of glpQ2 mutants and complementation strains in S. pneumoniae CGMH836.To identify the possible virulence effects of glpQ2, a serotype 19AST320 strain, CGMH836, which was isolated from the pleural fluid of a child with necrotizing pneumonia and empyema (25), was employed to identify the pathophysiological role of glpQ2 both in vitro and in vivo. The construction of the glpQ2 mutants and cis-complementation strains is represented in Fig. 2A. The glpQ2 mutant strains, M1 and M2, were confirmed by using the primer pair sp19A_glpQ2F/sp19A_glpQ2R to detect the inserted spectinomycin resistance (Spcr) gene and by using the primer pair glpQ2F/glpQ2R to confirm glpQ2 truncation (Fig. 2B). The glpQ2 cis-complementation strains, C1 and C2, were checked using the primer pair Up_sp12134F and Dn_sp12135R to confirm the insertion of glpQ2 and the CAT gene between ORF12134 and ORF12135 (Fig. 2C).
Construction of the glpQ2 mutants and glpQ2 cis-complementation strains in CGMH836. (A) The construction process is described in detail in the text. (B) The glpQ2 mutant strains were confirmed by PCR. The primer pairs sp19A_glpQ2F/sp19A_glpQ2R and glpQ2F/glpQ2R were used to distinguish the wild type (lanes 1) and glpQ2 mutants (M1, lanes 2; M2, lanes 3). (C) The glpQ2 cis-complementation strains (C1, lanes 2; C2, lanes 3) were first confirmed with glpQ2F/glpQ2R and distinguished from the wild type (lanes 1) using the primer pairs sp19A_glpQ2F/sp19A_glpQ2R and Up_sp12134F/Dn_sp12135R.
Role of glpQ2 in ChoP expression on pneumococcal surfaces.To evaluate the relationship between glpQ2 and ChoP expression on pneumococcal surfaces, CGMH836 (wild type [WT]) and its glpQ2 isogenic mutants (M1 and M2) or complementation (C1 and C2) strains were harvested during the exponential phase, which was divided into phase I (early to mid-log phase), phase II (mid- to late log phase), and phase III (late log phase to stationary phase), to monitor ChoP expression (Fig. 3A).
Effects of glpQ2 on ChoP expression by S. pneumoniae during the exponential phase. (A) Bacterial growth was monitored based on the OD580. The culture broth was collected during the exponential phase and divided into phase I (early to mid-log phase), phase II (mid- to late log phase), and phase III (late log to stationary phase). (B) ChoP expression was analyzed using anti-ChoP antibody (TEPC15) and flow cytometry. The histograms compare the ChoP expression levels on the surfaces of CGMH836, glpQ2 mutants (M1 and M2), and glpQ2 complementation strains (C1 and C2). The ratios indicate the highly ChoP-expressed bacteria among the total bacteria collected. The data are representative of three independent experiments. The TIGR4 strain, which lacked the glpQ2 gene, is also compared. (C) The ChoP level on CGMH836 and its isogenic strains during phases I, II, and III were compared based on three independent experiments (**, P < 0.01; *, P < 0.05; Student's t test). (D) The CRP binding level was examined after incubation with 100 μg/ml rhCRP and detected using a secondary Ab, FITC-conjugated anti-human CRP Ab, and flow cytometry. CGMH836 incubated with the secondary Ab only was used as the negative control. The geomean value indicates the average amount of CRP bound on the pneumococci. The percentage in the gated region represents the fraction of pneumococci that were strongly bound to CRP among the total bacteria analyzed.
The results showed consistently that the glpQ2 mutant strains (M1 and M2) could not achieve a population density as high as that of the glpQ2-expressing strains (WT versus C1 and C2) during the stationary phase, although the growth rates of the strains were similar (Fig. 3A). The ChoP expression analysis detected a consistently high level on CGMH836 throughout the exponential phase, but it decreased markedly on the glpQ2 mutants, particularly in phase I and phase II (Fig. 3B and C). Restoration of glpQ2 restored the expression of ChoP on the surfaces of the cis-complementation strains. Interestingly, the TIGR4 strain, which lacked the glpQ2 gene, also exhibited decreased ChoP expression during the exponential phase compared with CGMH836 (Fig. 3B). These results suggest that the efficient utilization of a choline source due to the additional glycerophosphodiester phosphodiesterase encoded by the glpQ2 gene contributed to the increased population density and to enhanced ChoP expression.
ChoP can also bind CRP on the pneumococcal surface (39) to activate the classical pathway in human serum (40, 41). A high level of CRP binding to CGMH836 was detected, which agreed with the high-level expression of ChoP on the surface (Fig. 3D), and the geomean value was reduced 5- to 7-fold when CRP bound to the glpQ2 mutants. The fraction of CRP strongly binding glpQ2 mutants (25 to 33%) was reduced by 2- to 4-fold compared with their glpQ2-expressing counterparts (69 to 82%), suggesting that the glpQ2 gene facilitated CRP binding to CGMH836. However, the enhanced CRP binding to CGMH836 did not induce effective bacterial killing by complement in human serum compared with the glpQ2 mutant or compared with the heat-inactivated-serum control (data not shown).
glpQ2 facilitated CBP anchorage on pneumococcal surfaces.ChoP facilitates the anchorage of CBPs on pneumococcal surfaces. ChoP expression was decreased in the glpQ2 mutant strains, so we eluted CBPs from WT and M1 strains with comparable bacterial numbers at the mid-log phase using 2% choline chloride, and we found that there was a 2-fold reduction in the total CBPs obtained from the glpQ2 mutant (194 to 227 μg) compared with the WT (379 to 472 μg), according to the Bradford assay. However, the CBP profile determined by two-dimensional SDS-PAGE demonstrated similarity between the wild-type and mutant strains (Fig. 4A), even when they were loaded with same amount of CBPs (data not shown).
Effect of the glpQ2 gene on CBP anchorage. (A) CBPs from 1010 CFU of CGMH836 (WT) and the glpQ2 mutant (M1) were eluted with 2% choline chloride and analyzed by two-dimensional SDS-PAGE (pH 4 to 7; 10%) using silver staining. Approximately 160 and 77 μg of CBPs from the WT and M1, respectively, were analyzed. (B) Colony morphologies of CGMH836 (WT), M1/M2, and C1/C2 after 15-h incubation on THY plates, which were compared using a dissecting microscope under a 10× lens objective (Leica). (C) Autolysis was examined during the stationary phase. The growth curve was monitored based on the OD580 (solid lines), and live bacteria were confirmed after plate counts (dashed lines). Refreshed cultures of the WT (no symbol), M1 (×), M2 (asterisk), C1 (■), and C2 (⧫) were monitored for 24 h. (D) The natural transformation efficiencies of the WT and M1 were compared. The CAT gene flanked by ORF12134 and ORF12135 was delivered into pneumococcal strains under CSP-1 induction, and the efficiency is represented as follows: [no. of CFU on the THY plate containing 4 μg/ml chloramphenicol]/[no. of CFU on a plain THY plate] × 100% (n = 4; **, P < 0.01; Student's t test).
Some CBPs are responsible for cell wall autolysis, and they may potentially facilitate natural transformation (42, 43). Therefore, we investigated the effects of the glpQ2 gene on the capacity for autolysis and the natural transformation efficiency of CGMH836.
S. pneumoniae often forms colonies with a crater-like appearance because of autolysis on agar plates after overnight culture (44). We found that CGMH836 formed colonies with a collapsed center and a raised periphery, based on observations using a dissecting microscope (Fig. 4B). In contrast, the M1/M2 strains lacked this typical “draftsman colony” appearance, and the capacity for autolysis was restored in the C1/C2 strains in a manner similar to that of CGMH836. Monitoring the live bacteria during the stationary phase produced similar results. The bacterial counts for the WT and C1/C2 decreased dramatically after 2 to 4 h during the stationary phase; however, the glpQ2 mutation (M1/M2) alleviated autolysis, even when incubation was extended for 17 h in the stationary phase (Fig. 4C).
The glpQ2 gene also affected natural transformation efficiency in S. pneumoniae. A suicide vector that carried the CAT gene flanked by the homologous regions of ORF12134 and ORF12135 was delivered into the WT and M1 strains by CSP-1 induction to examine the natural transformation efficiency based on the ratios of chloramphenicol-resistant transformants. glpQ2 deficiency reduced the DNA transformation capacity by around 2.22-fold compared with the parent strain (Fig. 4D), suggesting that glpQ2 facilitated the natural transformation of S. pneumoniae.
Effects of glpQ2 on the adherence and cytotoxicity of S. pneumoniae to human lung epithelial cells.GlpQ is an important virulence factor in some respiratory pathogens, where it mediates colonization or dissemination. To identify the effects of glpQ2 on virulence, we determined the interactions between CGMH836 isogenic strains with or without the glpQ2 gene and lung epithelial cell lines.
The human lung adenocarcinoma epithelial cell line A549 and the human bronchial epithelial cell line BEAS-2B were employed to determine pneumococcal adherence ability. The cultured cells were infected with the pneumococcal strains for 1 h, and the M1/M2 strains exhibited significantly reduced associations with lung epithelial cell lines compared with CGMH836, whereas the capacity for adherence was restored in the C1/C2 strains (Fig. 5A).
The glpQ2 gene confers adherence and cytotoxicity to lung epithelial cells on S. pneumoniae. (A) Adherence assay. The confluent lung epithelial cell lines A549 and BEAS-2B were cultured in 24-well plates and infected with 108 CFU of the indicated pneumococcal strains for 1 h before washing to remove nonadherent bacteria. The number of cell-associated bacteria was then determined by plate counts (n > 3; *, P < 0.05; **, P < 0.01; Student's t test). (B) Cytotoxicity assay. The confluent A549 cell line was cultured in a 96-well plate and infected with 5 × 106 CFU of the indicated pneumococcal strains for 2 h. The ratio of live cells was calculated by crystal violet staining and determined based on the OD595. The cytotoxicity was determined by the percentage of killed cells (n = 3; *, P < 0.05; **, P < 0.01; Student's t test).
In addition, the confluent A549 cell line was cultured in a 96-well plate with pneumococcal strains for 2 h, and the cytotoxicity was determined by crystal violet staining. CGMH836 exhibited a cytotoxic capacity of around 51% in the A549 cell line, but the cytotoxicity was reduced to 18 to 22% in the M1/M2 strains. Complementation with glpQ2 restored the cytotoxicity to 48 to 50% (Fig. 5B, C1/C2), indicating that glpQ2 is a virulence gene that contributes to adherence and cytotoxicity in human lung epithelial cells.
glpQ2 mediates pneumococcal colonization and inflammation in the murine respiratory tract.To elucidate further the role of glpQ2 in infections, 3-week-old BALB/c mice were treated intranasally or intratracheally with 5 × 107 CFU bacteria to define the pathophysiological role of glpQ2. At 24 h after intratracheal infection, the mice were sacrificed, and their lungs were homogenized to determine the number of colonizing bacteria by plating (Fig. 6A). We found that glpQ2 deficiency led to reduced colonization in the lung, and the glpQ2 mutants appeared to be cleared from a few mice within 24 h. Complementation with the glpQ2 gene restored the colonization capacity, and it was comparable to that of the wild type. In the upper respiratory tract infection model, lack of the glpQ2 gene reduced the bacterial load obtained by nasopharyngeal flushing compared with the wild type, but restoration of glpQ2 increased bacterial association with the nasopharynx (Fig. 6B). These results indicate that the glpQ2 gene facilitates colonization of the respiratory tract by S. pneumoniae.
The glpQ2 gene facilitates S. pneumoniae colonization of the respiratory tract in a murine model. Three-week-old BALB/c mice were treated intratracheally (A) or intranasally (B) with 5 × 107 CFU of the CGMH836 (WT), glpQ2 mutant (M1), or glpQ2 complementation (C1) strain. The number of colonizing bacteria was determined based on plate counts of homogenized lung at 24 h after intratracheal infection (A) or using flushed nasal washes at 48 h after intranasal infection (B) (**, P < 0.01; Student's t test).
At 48 h after infection, histological examination of pulmonary sections from young BALB/c mice wth intratracheal infection using CGMH836 detected marked inflammation (Fig. 7A). In contrast, there was no obvious infiltration of leukocytes in the pulmonary tissue infected by the glpQ2 mutant strain (Fig. 7B). Complementation with glpQ2 restored pathogenicity and led to inflammatory infiltration (Fig. 7C), suggesting that glpQ2 contributes to the severity of the pneumonia caused by CGMH836.
The glpQ2 gene contributes to pneumonia caused by CGMH836 in a young-mouse model. Three-week-old BALB/c mice were treated intratracheally with 5 × 107 CFU of CGMH836 (A), the glpQ2 mutant strain (B), and the glpQ2 cis-complementation strain (C). The lungs were removed at 48 h after infection, and the histology was examined with hematoxylin and eosin staining. The arrowheads indicate areas with marked inflammatory infiltration. Original magnifications, ×40 (left) and ×400 (right).
DISCUSSION
In this study, we identified a virulence gene, glpQ2, that is prevalent in serotype 3, 6B, 19A, and 19F strains of S. pneumoniae. We found that the glpQ2 gene facilitated CBP anchorage to bacterial surfaces, as well as enhancing autolysis and the natural transformation efficiency in a serotype 19AST320 strain, CGMH836, which was isolated from a patient with severe pneumococcal pneumonia. Mutation of the glpQ2 gene decreased the expression of ChoP and the capacity for CRP binding to CGMH836 during the exponential phase, as well as reducing adherence and cytotoxicity. Complementation with glpQ2 restored the expression of ChoP on pneumococcal surfaces, as well as the infectivity of epithelial cells and colonization ability in murine respiratory tract infection models. In conclusion, our results demonstrate that glpQ2 is important for respiratory tract colonization and that it contributes to pathogenesis in severe lung infections.
S. pneumoniae possesses two orthologs of glpQ, both of which have conserved residues in active sites. In CGMH836, ChoP expression was found to be consistently high during the exponential to stationary phases. The isogenic glpQ2 mutant expresses ChoP on its surface at a moderately low level during the early log phase, but there was a dramatic reduction in the mid- to late log phases. TIGR4 displays reduced, but not absent, ChoP expression during mid- to late log phase, which suggests the other ortholog, the glpQ gene, might be complementary to glpQ2 during this stage in glpQ2 mutants of CGMH836. However, the biological significance of ChoP fluctuation mediated by glpQ or the regulation of the glpQ gene remained unknown. The expression of ChoP on bacterial surfaces mediates invasive infection by binding to epithelial cells. One receptor implicated in adherence is the PAF receptor, and its natural ligand, PAF, is structurally related to ChoP. The adherence of S. pneumoniae to cytokine-activated epithelial cells is reduced after PAF receptor antagonist treatment (7), and the role of the PAF receptor in invasive pneumococcal diseases has been reviewed recently (45). The PAF receptor is induced after S. pneumoniae infection (8, 46), which suggests that ChoP upregulation in S. pneumoniae is advantageous for anchoring to the respiratory epithelia before invasive infection. In addition, ChoP facilitates the anchorage of CBPs on pneumococci. In the present study, the mutation of glpQ2 resulted in a 2-fold reduction in the CBPs on the surface, implying that these CBPs, such as PspC (47) and PavA (48), might facilitate colonization and participate in the pathogenesis of CGMH836.
The presence of the glpQ2 gene enhanced CRP binding on CGMH836, but it did not induce efficient complement activation in human serum to kill bacteria compared with heat-inactivated serum (data not shown). Interference with complement killing is possibly mediated by phosphoglycerate kinase (49), enolase (50), and PspC (51), etc., although we found that glpQ2-mediated ChoP upregulation did not activate complement killing as a consequence of CRP recruitment in vitro. Interestingly, CRP can bind directly to the phospholipid bilayer (52), which suggests that upregulated ChoP can recruit more CRP to bind the cell membrane. This highlights the importance of glpQ2 in secondary pneumococcal pneumonia following influenza virus infection, and it will be investigated further.
The glpQ2 gene in CGMH836 contributed to cytotoxicity, but glpQ2-mediated cytotoxicity might be independent of oxidative stress, because the levels of H2O2 produced in infected cell culture supernatants did not differ in the wild-type, glpQ2 mutant, and glpQ2 complementation strains (data not shown). A previous study showed that pneumolysin, a key virulence factor of S. pneumoniae, led to the cytolysis and apoptosis of cells (53). Interestingly, pneumococcal autolysis facilitates pneumolysin release (54), which suggests that the glpQ2-expressing strains promoted cytotoxicity.
S. pneumoniae is the leading cause of community-acquired pneumonia worldwide (55, 56). Serotype 19AST320 has emerged in many countries, where it causes invasive disease, severe pneumonia, acute otitis media, and hemolytic-uremic syndrome (21, 23, 57–60). In the respiratory tract infection murine model, we found that the glpQ2 gene was responsible for colonization and that it contributed to the severity of pneumonia, which suggests that glpQ2 facilitates the invasive ability of CGMH836, thereby causing severe pneumococcal pneumonia in clinical settings.
However, the presence of clinical strains with virulence but without glpQ2 reflects the involvement of other virulence factors. From the nose to the lung, genes related to adherence, metabolism, transcriptional regulation, etc., are also essential for pneumococcal pneumonia (61–63). Thus, the overall capacity for colonization, invasion, and dissemination contributes to pathogenesis, and glpQ2 is one of the genes found in S. pneumoniae that confers virulence.
In this study, we identified a virulence gene in an S. pneumoniae strain, glpQ2, that facilitates the persistent expression of ChoP on pneumococci during the exponential phase and confers a greater capacity for colonization during infection. In S. pneumoniae, the GlpQ2 protein is homologous to protein D in H. influenzae. Protein D has been employed as the protein conjugate in a 10-valent pneumococcal polysaccharide vaccine to extend its protective efficacy in pediatrics. Because colonization is the initial step in the pathogenesis of all pneumococcal diseases (64), it may explain the emergence of serotype 19AST320, which is associated with severe pneumococcal diseases, and thus, it may be a valuable target for prevention or treatment in the future.
ACKNOWLEDGMENTS
This research was supported by the Ministry of Science and Technology, Taiwan, Republic of China (NSC 102-2320-B-016-013-MY2).
We are grateful to J. R. Chen for his assistance in animal experiments and to Y. B. Chuang for scheme preparation.
FOOTNOTES
- Received 16 July 2014.
- Returned for modification 12 August 2014.
- Accepted 21 November 2014.
- Accepted manuscript posted online 24 November 2014.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.02357-14.
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
