|Home | About | Journals | Submit | Contact Us | Français|
Streptococcus pneumoniae may evade complement activity by binding of factor H (FH), a negative regulator of the alternative pathway, to the surface protein PspC. However, existing data on the effects of FH binding to PspC on complement activity are conflicting, and there is also considerable allelic variation in PspC structure between S. pneumoniae strains that may influence PspC-dependent effects on complement. We have investigated interactions with complement for several S. pneumoniae strains in which the gene encoding PspC has been deleted. The degree of FH binding varied between strains and was entirely dependent on PspC for seven strains. Data obtained with TIGR4 strains expressing different capsular serotypes suggest that FH binding is affected by capsular serotype. Results of immunoblot analysis for C3 degradation products and iC3b deposition assays suggested that FH bound to PspC retained functional activity, but loss of PspC had strikingly varied effects on C3b/iC3b deposition on S. pneumoniae, with large increases on serotype 4, 6A, 6B, and 9V strains but only small increases or even decreases on serotype 2, 3, 17, and 23F strains. Repeating C3b/iC3b assays with TIGR4 strains expressing different capsular serotypes suggested that differences in the effect of PspC on C3b/iC3b deposition were largely independent of capsular serotype and depend on strain background. However, data obtained from infection in complement-deficient mice demonstrated that differences between strains in the effects of PspC on complement surprisingly did not influence the development of septicemia.
Streptococcus pneumoniae is a common cause of invasive diseases such as pneumonia, meningitis, and septicemia even in immunocompetent subjects. One important component of host immunity to S. pneumoniae is the complement system, a series of approximately 30 serum and cell surface proteins organized into three enzyme cascades termed the classical, alternative, and mannan binding lectin (MBL) pathways (44). Infection experiments using complement-deficient mice have demonstrated that both the classical and alternative pathways are important for immunity to S. pneumoniae (2, 11, 15, 23, 43), and the high incidence of S. pneumoniae infections in patients with complement deficiencies confirms the relevance of complement for preventing disease in humans (3, 21). Both pathways lead to the formation of a C3 convertase that cleaves the central complement component C3, resulting in deposition of C3b on the surface of the pathogen that is further processed to iC3b. C3b and iC3b are opsonins, (44), and coating of S. pneumoniae with C3b/iC3b is vital for efficient phagocytosis of this organism by neutrophils (49). The importance of complement for immunity to S. pneumoniae is emphasized by the identification of several mechanisms by which the bacteria can inhibit complement activity. The choline binding surface protein PspA and the capsule prevent C3b/iC3b deposition on S. pneumoniae by mechanisms that remain unclear, whereas the secreted toxin pneumolysin seems to divert classical pathway activity away from the bacteria (1, 30, 32, 35, 39, 43, 48).
PspC is another choline binding surface protein that is related to PspA and is important for virulence in models of nasopharyngeal colonization, septicemia, and pneumonia (16, 22, 33, 35, 38). PspC can bind to factor H (FH), a negative regulator of the alternative pathway (8, 13, 18). PspC and the closely related proteins encoded at the same genetic locus in different strains (termed SpsA, CbpA, PbcA, and Hic) could, therefore, prevent alternative pathway complement activity against S. pneumoniae by one of three potential mechanisms (7). First, FH may speed up the degradation of C3b by promoting the factor I-dependent cleavage of C3b bound to the bacterial surface to iC3b; second, FH causes the dissociation of factor B from the alternative pathway C3 convertase C3bBb, decreasing C3b deposition on the bacteria; and third, FH may also inhibit the formation on the bacterial surface of the C3 convertase C3bBb by preferentially binding C3b, thus preventing C3b binding to factor B. However, although the affinity of PspC for FH has been clearly demonstrated and the binding sites have been identified (6, 12, 13, 17, 18, 28, 38), the effects of this interaction on complement-mediated immunity to S. pneumoniae is relatively poorly defined. Loss of PspC can cause reduced iC3b (compatible with reduced processing of C3b to iC3b) (28) and increased C3b/iC3b (compatible with reduced C3 convertase activity) deposition on a capsular serotype 2 (ST2) strain D39 or an unencapsulated mutant derived from a capsular ST3 strain (19, 35). However, in contrast to Quin et al., Lu et al. and Li et al. found little effect on total C3 bound to a ΔpspC mutant derived from the D39 strain compared to the wild type unless this mutation was combined with mutation of pspA (25, 28, 35).
The role of PspC on complement-mediated immunity in different strains could also be affected by the marked allelic variation in the structure of PspC. Ianelli et al. sequenced pspC from 43 different strains, and although the derived protein sequences had the same domain organization, each protein had a unique sequence that could be divided into 11 subgroups (17). The majority of PspC alleles contain a C-terminal cell wall choline binding motif (similar to PspA), but 17 of the allelic variants contain a cell wall anchor LPTXG motif instead. FH binding requires only the N-terminal 89 amino acids of PspC from the D39 strain and is dependent on a 12-amino-acid motif which is conserved between PspC alleles from different S. pneumoniae strains (28). However, this domain is not enough for full FH binding capacity and requires additional flanking amino acids that vary with allelic variation of PspC. Indeed, it is known that the ability of S. pneumoniae to bind to FH varies between strains (36). This variation in FH binding between strains seems to be independent of serotype but is increased on strains isolated from the blood or cerebrospinal fluid (CSF), under which conditions PspC expression is increased (29, 34). Overall, the existing data suggest that the effects of PspC on complement deposition on S. pneumoniae is not clear and could vary between strains. Furthermore, several other functions have been ascribed to PspC, including direct binding to C3 (which might counteract any FH-mediated reduction in C3b deposition) (4, 40) and aiding S. pneumoniae adhesion to host cells and translocation across epithelial layers (12, 13, 37, 38, 50), which could also affect virulence independent of any effect of PspC on complement-mediated immunity. Further clarification of the effects of PspC on opsonization of S. pneumoniae with C3b/iC3b is important for a better understanding of how PspC aids S. pneumoniae virulence. To address this question, in this study we have assessed the effects of loss of PspC on C3b/iC3b deposition on a range of S. pneumoniae strains representing different capsular STs.
S. pneumoniae strains used for this study were D39 (ST2), TIGR4 (ST4), JSB6A (ST6A), JSB6B (previously M158; ST6B), JSB9V (previously Io11601; ST9V), G374 (ST17), JSB23F (previously Io11697; ST23F) (20, 22, 24, 41), 0100993, and HB565. Strains JSB6A, JSB6B, JSB9V, and JSB23F were kind gifts from B. Spratt, Imperial College London. In addition opaque-phase capsular-ST-switched strains in the TIGR4 background expressing capsular ST4 (P1702), ST6A (P1637), and ST23F (P1691) were constructed by J. Weiser using the Janus cassette as previously reported (42). Bacteria were cultured at 37°C in 5% CO2 on blood agar plates or in Todd-Hewitt broth supplemented with 0.5% yeast extract (THY medium) to an optical density (OD) of 0.4 at 580 nm (approximately 108 CFU/ml) and stored at −70°C in 10% glycerol as single-use aliquots. Bacterial phase was determined using transparent medium (tryptone soya with catalase) under magnification and oblique, transmitted illumination as previously described (45).
For the in-frame deletion of pspC (Sp2190), a construct was created in which 572 bp of flanking DNA 5′ to the pspC ATG, starting with the stop codon of the Sp2191 open reading frame (ORF; primers Ery-Sp2191F and Sp2191R) and 1,042 bp of flanking DNA 3′ to the pspC ORF (ending with the stop codon of the preceding Sp2189 ORF; primers Sp2189F and Ery-Sp2189R) were amplified by PCR and fused with the erm cassette from pACH74 (primers EryF and EryR; a suicide vector carrying erm for selection in S. pneumoniae was a kind gift from J. Paton) by overlap extension PCR (27). These constructs were transformed into S. pneumoniae by homologous recombination and allelic replacement using competence-stimulating peptide (CSP-2 for ST4 strains or CSP-1 for all other strains) and standard protocols (14, 24). The identity and accuracy of the mutations were confirmed by PCR analysis (data not shown). All the ΔpspC mutant strains were stable after two 8-h growth cycles in THY medium without antibiotic. The JSB6AΔpspC, JSB6BΔpspC, and JSB9VΔpspC mutant strains were constructed by transformation with genomic DNA from an existing D39ΔpspC strain (kind gift from J. Paton) and confirmed by PCR (10).
Genomic DNA was isolated from S. pneumoniae strains using a Wizard genomic DNA isolation kit following the manufacturer's instructions except that cells were incubated with 0.1% deoxycholic acid (Sigma) at 37°C for 10 min before extraction. The ΔpspC mutant strains and primers used for this study and their derivation, capsular ST, PspC type (if known), and source are listed in Tables Tables11 and and2.2. pspC alleles were amplified and sequenced using the PCR primers IF30, IF43, cbpAseqF1, and cbpAseqR1 (16).
Serum samples from healthy volunteer controls were obtained according to institutional guidelines and stored as single-use aliquots at −70°C. Human serum deficient in factor B (Bf− serum) was supplied by Calbiochem (46, 49). FH binding in 10% or 20% serum and C3b/iC3b or iC3b deposition on bacteria in 20% serum were assessed using previously described flow cytometry assays (2, 46, 47, 48). Briefly, bacteria (107 CFU/ml) were incubated in 10 μl of human serum (20% or 10% diluted in phosphate-buffered saline [PBS]) or 10 μg/ml purified human FH (Calbiochem). After samples were washed, FH binding was detected using a polyclonal goat anti-human FH (Calbiochem) antibody (1/200) and a fluorescein isothiocyanate (FITC)-labeled donkey anti-goat IgG (Serotec) secondary antibody (1/300 diluted); C3b deposition was detected using an FITC-conjugated polyclonal goat anti-human C3 antibody (ICN), and iC3b deposition was detected using a monoclonal mouse anti-human iC3b (Technoclone) antibody (1/200) and an FITC-labeled rabbit anti-mouse IgG (Dako) secondary antibody (1/200). Samples were analyzed using a FACSCalibur flow cytometer (BD Biosciences), using forward and side scatter parameters to gate on at least 25,000 bacteria, as previously described (2, 46, 47, 48). Results were compared using the fluorescence index (FI; defined as the proportion of bacteria positive for C3b/iC3b, iC3b, or FH multiplied by the geometric mean fluorescence intensity [MFI]) (9, 49).
Capsule thickness was evaluated using mid-log-phase S. pneumoniae cells prepared for electron microscopy (EM) using a ruthenium red and London resin protocol, and viewed using a Jeol 1010 transmission electron microscope (100 kV). Image J software was used to determine capsule thickness by obtaining the cross-sectional area of the whole bacterium including and excluding the capsule, and these measurements were then used to calculate the radii and hence the average width of the capsule layer. Data were obtained for 10 or more randomly chosen bacteria for each strain investigated.
C3 activation in human serum was investigated by incubating different concentrations of bacteria in 1 ml of 10% pooled human serum for 20 min at 37°C, followed by centrifugation at 13,000 rpm for 15 min. The supernatants were removed, diluted to a final serum running concentration of 1% in sodium dodecyl sulfate (SDS) sample buffer with 2-mercaptoethanol, boiled at 95°C for 10 min, separated on a 10% SDS-PAGE gel, transferred to nitrocellulose membranes using standard methods (Maniatis), and probed using an anti-C3 (ICN) conjugated with horseradish peroxidase (HRP).
Infection experiments conformed to institutional and government guidelines for working with animals. C3−/− or C1qa−/− C2−/− Bf−/− complement-deficient mice were a kind gift from M. Botto, Imperial College London, and wild-type C57B/6 mice were obtained from commercial sources. Mice aged over 6 weeks were inoculated by intraperitoneal (i.p.) injection with 5,000 CFU diluted in PBS of a 50/50 mix of the wild-type and corresponding ΔpspC strains and culled at 24 h, and serial dilutions of blood and spleen homogenates were plated onto plain and antibiotic-containing media to calculate the bacterial CFU present for each strain. Competitive indices (CIs) were calculated using the following formula: ratio of mutant to wild-type strain recovered from mice divided by the ratio of mutant to wild-type strain in the inoculum (48). A CI of less than 1 indicates that the mutant strain is attenuated in virulence compared to the wild-type strain, and the lower the CI, the more attenuated is the mutant strain.
Data presented are representative of results obtained from repeated independent experiments, and each data point represents mean and standard deviations (SD) for to three to five replicas. Results for FH binding and C3b deposition in Bf− serum were compared using one-way analysis of variance (ANOVA) tests with Dunnett's posttest analysis. Comparisons of FH binding and C3b deposition between wild-type and unencapsulated or ΔpspC strains and between transparent and opaque-phase variants were analyzed using two-tailed Student's t tests.
FH binding to S. pneumoniae has previously been shown to be completely dependent on PspC for an ST2 and an ST3 strain (5, 7, 18, 33). To extend these observations and to confirm that FH binding is dependent on PspC for the strains used for our experiments, FH binding in human serum was assessed using a flow cytometry assay with eight ΔpspC strains (five created by deletion of pspC and three created by insertional duplication mutagenesis) representing different S. pneumoniae capsular serotypes and at least five different allelic variants of PspC, including a strain containing two copies of PspC (G374ΔpspC) (Table (Table1).1). In the wild-type strains FH bound to between 55 and 92% of wild-type bacteria incubated in 10% serum (Fig. (Fig.1A).1A). However, for seven strains a similar percentage of ΔpspC bacteria were positive for FH as the control samples (wild-type bacteria incubated in PBS alone) (Fig. (Fig.1),1), demonstrating that there is no significant FH binding to ΔpspC bacteria. For the remaining strain, an ST6A strain, there was a low level of FH binding in serum to the ΔpspC strain compared to bacteria incubated on PBS although the biological significance of this is uncertain. As has been recently demonstrated for a large number of clinical isolates (36), the degree of FH binding (represented by a fluorescence index) (9, 49) varied significantly between the strains (Fig. (Fig.1D).1D). The ST9V strain had a particularly high and the ST6B strains had relatively low levels of FH binding, with the remaining strains having intermediate levels of FH binding. As access to cell wall proteins may be affected by capsule expression, we measured the thickness of the capsule layer for selected strains. However, the capsule layer was actually thicker for two strains (mean and SD for 0100993, 265 ± 16 nm; TIGR4, 185 ± 19 nm) with increased FH binding compared to the D39 strain (104 ± 11 nm), suggesting that a thinner capsule layer is unlikely to explain why certain strains have greater levels of FH binding.
To characterize the effect of loss of PspC on complement activity against S. pneumoniae, a flow cytometry assay was used to investigate C3b/iC3b deposition on ΔpspC strains compared to their corresponding parental wild-type strains. The results were markedly varied between strains. Loss of PspC resulted in a massive increase in C3b/iC3b deposition on the ST6A strain, moderate increases on the ST4, ST6B, and ST9V strains, and a small increase on the ST2 strain; but there was no significant increase on the ST17 and ST23F strains, and seemingly there was a decrease on the 0100993 ST3 strain (Fig. (Fig.2).2). Because the decrease in C3b/iC3b deposition on the ST3 ΔpspC strain was unexpected, C3b/iC3b deposition was also investigated for an additional ST3 strain (H565), and this strain showed similar results to the 0100993 strain (Fig. (Fig.2).2). The variation in degree of FH binding between strains could potentially explain the differences in the effects of loss of PspC on C3b/iC3b between strains. However, there was no significant correlation between degree of FH binding to each S. pneumoniae strain to the effects of loss of PspC on C3b/iC3b deposition (Fig. (Fig.33).
To confirm whether FH binding had functional consequences on C3b processing despite variable effects on total C3b/iC3b deposition on the bacterial surface, immunoblotting for C3 breakdown products in serum incubated with wild-type and ΔpspC strains was done. Three different strains were investigated: the ST3 (decrease in C3b/iC3b on the ΔpspC strain), ST6A (massive increase in C3b/iC3b on the ΔpspC strain), and ST23F (no increase in C3b/iC3b on the ΔpspC strain) strains. For all three strains incubation of the ΔpspC strains in serum resulted in greater levels of the C3 breakdown products C3d and iC3b detected by immunoblotting (Fig. (Fig.4),4), suggesting that loss of PspC results in greater complement activation, independent of the strain background. In addition, iC3b deposition (investigated using flow cytometry and an iC3b-specific antibody) was reduced on the surface of the ΔpspC mutants compared to the wild-type for the ST4, ST6A, and ST23F strains (ST3 strains were not investigated because of the decrease in the total C3b/iC3b on the corresponding ΔpspC strains) (Fig. (Fig.5).5). These data suggest that FH bound to PspC was able to cleave C3b to iC3b and was therefore still functionally active on these strains despite the differences in the effect of PspC on total C3b/iC3b deposition.
As FH bound to PspC would prevent alternative pathway-dependent C3b/iC3b deposition on S. pneumoniae, differences between strains in the relative importance of the alternative pathway for C3b/iC3b deposition could explain why loss of PspC had different effects on total C3b/iC3b deposition depending on the strain. Our previously published data have compared C3b/iC3b deposition in human serum depleted of factor B (Bf− serum, in which no alternative pathway activity is possible) to the results for serum depleted of C9, a terminal complement pathway component which has no effect on C3b/iC3b deposition (46, 49) for the ST2 (small increase in C3b/iC3b on the ΔpspC strain), ST4 (moderate increase in C3b/iC3b on the ΔpspC strain), and ST23F (no increase in C3b/iC3b on the ΔpspC strain) strains. Loss of alternative pathway activity decreased C3b/iC3b deposition by 14.4% (SD, 8.9%) for the ST2 strain and 33.8% (SD, 2.7%) for the ST4 strain, matching the relative effects of loss of PspC on C3b/iC3b deposition on these strains. However, alternative pathway activity was responsible for 41.3% (SD, 12%) of C3b/iC3b deposition on the ST23F strain despite the lack of effect of loss of PspC on C3b/iC3b deposition on this strain. Hence, differences in the importance of alternative pathway activity are unlikely to explain all the variation in the effects of loss of PspC on C3b/iC3b deposition between S. pneumoniae strains.
To assess the role of PspC-complement interactions during disease, mixed infection experiments were performed in which an inoculum of 50% wild-type and 50% ΔpspC bacteria were inoculated i.p. into mice, and a CI was calculated for bacteria recovered from spleens 24 h later. Experiments were performed for three strains known to be virulent in mice (ST2, ST3, and ST4 strains) and using wild-type and genetically modified complement-deficient mice (either due to loss of C3 or due to loss of C1q, C2, and Bf, with both strains functionally unable to opsonize bacteria with C3) as any impaired virulence of the ΔpspC strains in wild-type mice caused by loss of the effects of PspC on complement should be partially restored in complement-deficient mice (as previously shown for PspA and pneumolysin) (48). In this mouse model of S. pneumoniae sepsis, the ΔpspC strains had a modest reduction in virulence in wild-type mice (CI, 0.15 to 0.23) but were markedly more attenuated in complement-deficient mice (Cis, 0.057 to 0.025) (Table (Table3).3). These data suggest that the interactions of PspC with complement do not strongly affect virulence in this model of sepsis.
To investigate whether differences between strains in FH binding were related to capsular serotype or to other genetic differences between strains (such as allelic variation in PspC structure), ΔpspC strains were constructed in opaque-phase variants of strains derived from the TIGR4 capsular ST4 background expressing different capsular serotypes (ST6A and ST23F) or reconstituted ST4 (45). FH binding to the TIGR4(−)+4, TIGR4(−)+6A, and TIGR4(−)+23F strains (in which the genes for the ST4 capsule were removed from the TIGR4 strain [TIGR4(−)] and replaced [+] by the indicated STs) was investigated using flow cytometry and bacteria incubated in serum. In serum, FH binding levels to the TIGR4(−)+4 and TIGR4(−)+23F strains were very similar, with FIs of 1,600 (SD, 65) and 1,369 (SD, 33), respectively, whereas FH binding to the TIGR4(−)+6A strain was markedly higher, with an FI of 3515 (SD, 288) (P < 0.01 versus TIGR4(−)+4 strain, one-way ANOVA with Dunnett's posttest comparison). Very similar results were obtained when the TIGR4(−)+ strains were incubated in 10 μg/ml of purified human FH (Fig. (Fig.6).6). EM demonstrated no difference in the thickness of the capsular layers between the TIGR4(−)+ strains [mean ± SD for TIGR4(−)+6A, 96 ± 15; TIGR4(−)+23F, 94 ± 9; TIGR4(−)+4, 86 ± 13). These data suggest that FH binding to S. pneumoniae can be affected by capsular serotype.
To investigate whether differences between strains in the effects of loss of PspC on C3b/iC3b deposition were related to capsular serotype, C3b/iC3b deposition on the otherwise isogenic ΔpspC TIGR4(−)+4, TIGR4(−)+6A, and TIGR4(−)+ 23F strains were compared to the corresponding pspC+ strain using flow cytometry. Loss of PspC significantly increased C3b/iC3b deposition on all three capsular-ST-switched strains (Fig. (Fig.7).7). However, the results contrasted to those obtained with clinical isolates expressing the same capsular serotypes. Deletion of pspC increased C3b/iC3b deposition on the TIGR4(−)+6A and TIGR4(−)+23F capsular-ST-switched strains by 178% and 563%, respectively, but for the ΔpspC strains derived from ST6A and ST23F clinical isolates, C3b/iC3b deposition increased by 3412% and 43%, respectively (Fig. (Fig.2D2D and and6D).6D). The results for the TIGR4(−)+4 strain were similar to those for the TIGR4 strain (C3b/iC3b deposition increased by 293% and 271%, respectively). Hence, the differences in the effects of PspC on C3b/iC3b deposition between strains of different serotypes are reduced when the bacterial strains are otherwise isogenic apart from the capsular locus, suggesting that the differences in C3b/iC3b deposition are influenced by noncapsular genetic variation between strains.
Several pathogenic bacteria are thought to evade alternative pathway-mediated immunity by binding to the host complement regulator protein FH (26). A large body of data has demonstrated that the S. pneumoniae choline binding protein PspC binds to FH, indicating that S. pneumoniae could also prevent complement attack by this strategy (5, 12, 13, 17, 18, 28, 38), but there are only limited data on the effects of PspC on complement-mediated immunity. Data for a capsular serotype 2 strain showing that FH bound to PspC promotes breakdown of C3b to iC3b suggests that FH on the surface of S. pneumoniae retains its complement-modulating activity. However, the effects of loss of PspC have been variously reported as increasing or having no effect on C3b deposition for the ST2 strain D39 and as increasing C3b deposition on an unencapsulated ST3 strain (19, 25, 28, 35). Furthermore, PspC has marked allelic variation in structure, and the amount of FH binding to S. pneumoniae varies between different strains (17). These observations prompted our investigation of C3b/iC3b deposition on a range of ΔpspC strains, especially as PspC has important roles in promoting adhesion to and translocation across epithelial layers (12, 13, 37, 38, 50), and its importance for virulence may therefore be independent of its effects on complement-mediated immunity.
Using ΔpspC mutant strains representing common capsular STs, we have confirmed and extended to new strains the results of previous studies showing that binding of FH to S. pneumoniae is almost totally dependent on PspC, with no detectable FH binding to six of the seven ΔpspC strains we have investigated and only a low level of binding to the seventh, a capsular ST6A strain. Recently, S. pneumoniae histidine triad proteins have been shown to bind to FH with a lower affinity than PspC, but our data suggest that these proteins do not contribute significantly to FH binding to whole bacteria, at least under the conditions for our assay (in human serum) (31). We have confirmed earlier reports that the degree of FH binding varies between strains, and measuring the capsule layer thickness for some strains suggests that this is unlikely to be simply related to thinner capsule layers in some strains, allowing better access of FH to PspC on the bacterial cell wall. Our data using the capsular-ST-switched TIGR4(−)+ strains demonstrated a clear increase in FH binding to the ST6A strain, suggesting that capsular serotype can influence FH binding. In addition, differences in levels of PspC expression could explain variations in FH binding between strains and need to be investigated. However, correlating the results of pspC expression experiments to FH binding levels will be confounded by differences in capsular serotype, making interpretation difficult. Furthermore, allelic variation of PspC will affect the avidity of antibody to each allelic variant and means that different sets of primers for real-time PCR will be necessary, complicating accurate investigation of PspC expression. Allelic variations in PspC structure (17, 36) may also affect FH binding and could be investigated by exchanging pspC alleles between strains.
The effects of loss of PspC on C3b/iC3b deposition were strikingly varied, with a massive increase in C3b/iC3b deposition on the ST6A ΔpspC strain, large increases on the ST4 and ST6B ΔpspC strains, only a small increase on the D39 ST2 ΔpspC strain, no consistent effect on the ST17 and ST23F ΔpspC strains, and even a decrease for two ST3 ΔpspC strains. Direct binding of PspC to C3 on cell surfaces (4, 40) has been reported and could potentially explain why loss of PspC had only a weak or no detectable effect on C3b/iC3b deposition on the bacterial surface for some strains. However, incubating the ST2 strain with purified human C3 in the absence of other complement factors demonstrated no detectable binding to fluid-phase C3 (data not shown). iC3b deposition assay results for three strains confirm earlier findings for the D39 capsular ST2 strain (28) and suggest that FH bound to PspC retains its physiological function of increasing the breakdown of C3b to iC3b even in strains where loss of PspC has little effect on opsonization of S. pneumoniae with C3b/iC3b. Furthermore, immunoblot analysis of C3 breakdown shows that loss of PspC results in greater complement activation in serum. These data suggest that FH bound to PspC is physiologically active even if this has variable consequences, depending on strain, for C3b/iC3b deposition on the bacterial surface. Such large variation in the effects of loss of PspC on C3b/iC3b deposition could be due to a large number of factors, including the relative role of the alternative pathway for C3b/iC3b deposition on different strains, capsular serotype, the relative level of FH binding, allelic variation in PspC, and other genetic differences between strains that could influence complement activity. However, perhaps surprisingly, the relative level of FH binding did not correlate to the effects of loss of PspC on C3b/iC3b deposition among different strains, showing that there is no obvious association between the quantity of FH on the bacterial surface and effects on C3b/iC3b deposition. Furthermore, our data suggest that differences between strains in alternative pathway-dependent complement activation do not explain variations in PspC effects on C3b/iC3b deposition.
Repeating the C3b/iC3b deposition assays with ΔpspC strains derived from the capsular-ST-switched TIGR4 strains demonstrates that strain genetic background does have a strong influence on the effects of loss of PspC independent of capsular serotype. For the clinical isolates, loss of PspC resulted in a massive increase in C3b deposition on the ST6A strain but had no consistent effect on the ST23F strain. In contrast, C3b/iC3b deposition was only moderately increased on the ΔpspC TIGR4(−)+6A strain and was also significantly increased on the TIGR(−)+23F strains. Although the factors that can affect PspC interactions with complement are likely to be very complex, overall our results indicate that strain background has a major influence on the effects of loss of PspC on C3b/iC3b deposition, independent of capsular serotype. The effects of strain background could be mediated by the extensive allelic variation in PspC structure (17) and perhaps other causes of noncapsular genetic variation between S. pneumoniae strains such as allelic variation in PspA or other surface proteins, the presence of absence of regions of diversity, or multiple single gene deletions and insertions (41). Further characterization of whether allelic variation of PspC explains these results would require detailed structure-function analysis using strains expressing different PspC alleles in an otherwise isogenic background. Of note, there are were large differences in the C3b/iC3b deposition results between the wild-type TIGR4(−)+ strains, suggesting that capsular serotype can affect complement activity independent of other genetic variation between strains.
The functional consequences of variations between strains on the effects of PspC on complement interactions with S. pneumoniae are at present not clear. Data from infection experiments have shown that in a model of S. pneumoniae pneumonia (for which complement is a vital component of host immunity), loss of PspC has a variable effect on virulence, depending on the S. pneumoniae strain. Interestingly, the ST4 ΔpspC strain, which had a 270% increase in C3b/iC3b deposition, was attenuated in virulence, whereas the ST2 ΔpspC strain, which had only a 70% increase in C3b/iC3b deposition, did not significantly differ from the wild-type strain in virulence (22). This may suggest that differences in effects on C3b/iC3b deposition between strains may have functional effects on virulence, especially during systemic infection when complement is vital for immunity (48). However, using mixed infection experiments, we have shown that deletion of pspC resulted in similar, mild reductions in virulence at 24 h in a sepsis model for the ST2, ST3, and ST4 strains. Furthermore, the virulence of the ΔpspC strains was surprisingly even further reduced in mice depleted of complement, the reverse of what would be expected if interactions of PspC with complement are important for virulence and the opposite result of previously published data for PspA and pneumolysin mutants (48). These results suggest that differences in the effects on C3b/iC3b deposition of loss of PspC on these strains are not particularly relevant during infection, at least for this model. The fall in CI for mixed infections with wild-type and ΔpspC strains in complement-deficient mice is likely to be due to the more rapid progress of infection in these mice (2), increasing differences between wild-type and ΔpspC strains, and suggests that complement-independent physiological roles of PspC are important for virulence. The mechanisms by which PspC can aid virulence at different anatomical sites require further investigation and clarification.
In summary, we have presented data confirming that FH binding to S. pneumoniae is dependent on PspC for a range of capsular STs but varies in intensity between strains and can be affected by capsular serotype. Although FH binding by PspC inhibits complement activity, the effects of loss of PspC on C3b/iC3b deposition on the S. pneumoniae surface varied markedly between strains and were partially dependent on noncapsular genetic variation between strains. Experiments using complement-deficient mice suggested that the interactions of PspC with complement have surprisingly little effect in a systemic model of infection. These data provide further evidence that the role of important virulence factors may need to be assessed in different strains of the same pathogen to provide a more complete understanding of their contribution toward disease pathogenesis.
This work was funded by grants from the British Lung Foundation (grant reference P02/52) and the Wellcome Trust (grant reference 076442) and was undertaken at UCLH/UCL, which received a proportion of funding from the Department of Health's NIHR Biomedical Research Centre's funding scheme. C.J.H. is supported by the Astor Foundation and Glaxo Smith Kline through the University College London MB-Ph.D. program.
We are grateful for Marco Oggioni's assistance in identifying PspC alleles for some of the strains used in this report.
Editor: A. Camilli
Published ahead of print on 2 November 2009.