The strains of S. pneumoniae and S. mitis used in this study are listed in Table 1, some of which have been previously described (Camberlein et al., 2015; Kilian and Tettelin, 2019). Bacteria were cultured at 37°C in 5% CO₂ in air on Colombia blood agar (Oxoid, United Kingdom) plates supplemented with 5% defibrinated horse blood (TCS Biosciences, United Kingdom) or in Todd-Hewitt broth supplemented with 0.5% yeast extract (THY) (Oxoid, United Kingdom). Growth in liquid medium was assessed by optical density. Bacterial stocks were grown to approximately mid-log phase and stored in 10% glycerol as single-use aliquots.

Under sterile conditions, 1 × 10⁶ CFU/ml were inoculated into 1 ml of THY broth in a sterile cuvette and incubated at 37°C for 8 h. The OD was measured every hour at 600 nm, following agitation using a pipette. Each strain was tested in triplicate. For assessing growth in serum, 2 × 10⁶ CFU of each S. pneumoniae or S. mitis strain were inoculated into 1 ml of serum that had previously been heat inactivated by incubation at 56°C for 30 min or in fresh human blood. The blood was prevented from clotting using a final concentration of 0.1 M sodium citrate. After a 4-h incubation at 37°C, 5% CO₂, serial dilutions were plated onto Colombia blood agar, incubated overnight and colonies counted to calculate bacterial numbers measured as CFU/ml. Data are presented as the percentage of CFU per ml compared to the initial inoculum count.

Complement C3 and FH interactions with S. pneumoniae and S. mitis were assessed using established flow cytometry assays as previously described (Brown et al., 2002; Hyams et al., 2010a, 2011, 2013). Briefly, 5 × 10⁶ CFU of S. pneumoniae and S. mitis were pelleted and resuspended in either baby rabbit complement (BRC) or pooled human sera and incubated at 37°C for 20 min. Samples were then washed twice with 0.1% PBS Tween-20. For FH binding, a sheep anti-factor H primary antibody was added at a 1:300 dilution and incubated for 30 min before washing in PBS 0.1% Tween-20 and addition of FITC-labelled donkey anti-sheep secondary antibody (FH binding) or a fluorescein-conjugated goat IgG fraction to rabbit complement C3 (C3b/iC3b deposition). After incubation on ice for 30 min and further washes with PBS 0.1% Tween-20 samples were fixed with 10% neutral buffered formalin and analysed by flow cytometry using a BD FACSVerse flow cytometer.

Phagocytosis was investigated using an established flow cytometry assay (Lu et al., 2008; Hyams et al., 2010a). Briefly, S. pneumoniae and S. mitis fluorescently labelled with 6-carboxyfluorescein succinimidyl ester (FAMSE; Molecular Probes) were incubated in pooled human serum heat-inactivated (56°C for 30 min), pooled human serum or baby rabbit complement for 30 min at 37°C then added to 10⁵ neutrophils extracted from fresh human blood at a multiplicity of infection (MOI) of 10–1, incubated for 30 min, then analysed using a BD FACSVerse flow cytometer. A minimum of 10,000 cells were analysed by flow cytometry, and the results are presented as a fluorescence index (FI: the percentage of positive bacteria multiplied by the geometric mean fluorescence index).

For heterologous expression of S. pneumoniae pspC by S. mitis, pspC was inserted downstream of the lactate dehydrogenase (ldh) gene under the control of the native promoter. RNAseq data has shown ldh is highly expressed constitutively (unpublished data). An antibiotic resistance cassette for kanamycin (km^r) was inserted into an intergenic spacer (IGS) site upstream of pspC within the S. pneumoniae TIGR4 strain genome. Small flanking fragments either side of the IGS in S. pneumoniae were amplified (primers: 885/886, 887/888) (Table 2), and ligated to km^r using separate bipartite ligations and then overlap extension PCR (primers 885 and 888) to generate a whole construct consisting of both 5′ and 3′ flanking regions and km^r. This construct was then used to transform S. pneumoniae TIGR4. Small flanking fragments either side of the IGS downstream of ldh in S. mitis were amplified (primers: 875/876, 879/889) and then ligated to pspC km^r amplified from the S. pneumoniae mutant. Two separate bipartite ligations were then again independently performed, and the two resulting ligation products linked by overlap extension PCR amplification (primers 875 and 889) to generate a construct consisting of both 5′ and 3′ flanking fragments, pspC and km^r. This ligation product was then used to transform S. mitis SK142 (Supplementary Figure 1).

Transformation was carried out as previously described (Rukke et al., 2014) with minor modifications. Briefly, S. mitis pre cultures were diluted 1:12 in TSB and allowed to grow to an OD₆₀₀ 0.03–0.04 prior to the addition of 5 mg DNA and competence stimulating peptide (CSP) (EIRQTHNIFFNFFKRR; GenScript) at a final concentration of 250 nM. The cultures were then incubated under aerobic conditions for 4 h at 37°C, 5% CO₂ before selection on blood agar plates supplemented with the appropriate antibiotics.

Bacterial strains were grown in THY medium until OD₅₈₀ 0.6. Cells were then pelleted by centrifugation for 20 min at 3,000 rpm, washed twice in PBS, before the cell pellet was sonicated on ice using three 30 s bursts at 200–300 W. Cell lysates were then centrifuged to pellet all cellular debris, and protein concentrations measured in the supernatant using the BioRad DC^TMProtein Assay. 6.25 μl of LDS Sample Buffer (4X) and 2.5 μl of Sample Reducing Agent (10X) was added to a known lysate concentration, making up to total volume of 25 μl using water, denatured at 70°C for 10 min prior to analyses using SDS-PAGE and Western blotting using pre-prepared 4–12% Bis-Tris Plus SDS-PAGE gels. Electrophoresis was carried out using 1X MES running buffer at a constant 200 V for 30 min, and the protein gels were transferred to PVDF membrane using a dry transfer iBlot system. Membranes were blocked in 5% milk in 1X TBS-Tween (TBS-T) for 2 h at room temperature, then incubated in primary antibody (1 h at room temperature or overnight at 4°C). Membranes were washed 3 times in TBS-T for 45 min and then incubated in secondary antibody for 1 h at room temperature. After washing again 3 times, membranes were developed using Luminata^TMCrescendo HRP substrate (Merck Millipore) and imaged using the ImageQuant LAS 4000.

CD1 (female, 8–12 weeks) or C57BL/6 C3- (female, 14–15 weeks) mice were injected intraperitoneally with varying CFU suspended in 100 μl PBS. Blood and spleen were obtained from mice sacrificed at set time points (1, 4, 24 h), and serial dilutions of blood and spleen homogenates plated onto blood agar plates to enumerate bacterial CFU. In select experiments, Ly6G + neutrophils were depleted using 600 μg anti-ly6G monoclonal antibody (1A8, BioXCell) administered intraperitoneally in a 200 μl volume 24 h prior to bacterial challenge; we have previously shown this treatment reduces neutrophils recruited to lavage fluid by 94.8% 24 h post-infection (Wilson et al., 2015, 2017).

The relative virulence of the strains is presented as a competitive index (CI), defined as the ratio of the test strain (single or double mutant strain) to the reference strain (wild-type or a single mutant strain) recovered from the mice divided by the ratio of the test strain to the reference strain in the inoculum. A CI of < 1.0 indicates that the test strain is reduced in virulence compared with the reference strain. The lower the CI the greater the reduction in virulence.

Statistical analysis was performed using GraphPad Prism 8.0. Data were presented as group means ± the standard error of the mean (SEM). Results expressed as means were compared using either one way ANOVAs with post hoc tests for multiple groups or Students T-test when comparing the mean of two groups only. In vitro data are representative of results obtained from at three or more repeated experiments.

All in vivo experiments were carried out at the biological services unit at UCL according to United Kingdom national guidelines for animal use and care, and were approved by the UCL Biological Services Ethical Committee and the UK Home Office (Project Licence PPL70/6510).

To assess whether differences in sensitivity to complement independent of capsule expression may underpin the relatively low virulence of S. mitis strains we investigated C3b/iC3b deposition on four S. pneumoniae capsular serotypes (4, 19C, 36, and 45) and four S. mitis strains, three of which express S. pneumoniae capsular serotypes (19C, 36, and 45)(Figure 1A). The last S. mitis strain (SK142) expresses a capsule with an unknown chemical structure and serotype. C3b/iC3b deposition on the S. mitis and S. pneumoniae strains was measured using an established flow cytometry assay (Brown et al., 2002). With the exception of S. mitis serotype 36 (SK1126), there was significantly greater C3b/iC3b deposition on the S. mitis strains compared to the S. pneumoniae strains. Comparing between S. mitis strains, S. mitis SK142 was the most susceptible to complement deposition, with S. mitis serotype 36 being the least sensitive (Figure 1A). Amongst the S. pneumoniae strains, serotype 36 (1095/39) was again the most resistant to opsonisation with C3b/iC3b. These data indicate that S. mitis strains are more readily recognised by the complement system than S. pneumoniae strains even when they express a capsule, including capsules chemically identical to S. pneumoniae serotypes.

One mechanism by which S. pneumoniae evades opsonisation with C3b is by PspC-mediated binding of FH (Dave et al., 2001). Homologues of pspC have not been found in the sequenced S. mitis genomes to date suggesting lack of PspC might be one reason why there are differences in complement sensitivity between S. pneumoniae and S. mitis strains expressing the same capsular serotype. We therefore assessed FH recruitment to the S. mitis strains after incubation in 10% human serum using flow cytometry. This demonstrated no evidence of significant factor H binding to any of the S. mitis strains whereas, as previously described, S. pneumoniae strains showed a high level of binding to factor H (Figures 1B,C).

The above data suggested increased sensitivity of S. mitis strains to complement independent of capsular serotype could be related to their inability to bind FH via PspC. To investigate this possibility, the S. pneumoniae TIGR4 strain pspC was expressed in S. mitis SK142. Immunoblots confirmed that the S. mitis pspC⁺ expressed PspC protein (Figure 2A). Expression of pspC resulted in a reduction in growth levels over 8 h (Supplementary Figure 2) but demonstrated significant levels of FH binding to S. mitis (Figures 2B,C). The functional consequences of expression of pspC by S. mitis was assessed using in vitro assays of C3b/iC3b deposition, growth in blood, and neutrophil phagocytosis (Figure 3). Flow cytometry confirmed that the S. mitis pspC⁺ strain was indeed more resistant to opsonisation with complement, showing similar levels of C3b/iC3b deposition as the S. pneumoniae TIGR4 strain (Figures 3A,B). The S. mitis pspC⁺ strain showed increased survival in whole human blood compared to the parental wild type S. mitis strain, with an increase in CFU compared to the wild type of almost 150% after 4 h (Figure 3C). The S. mitis pspC⁺ strain also demonstrated a reduced bacterial association with neutrophils when incubated in whole human serum (NHS), but not when incubated in PBS or in heat-inactivated human serum (Figure 3D), suggesting this effect was due to the reduced opsonisation of S. mitis pspC⁺ with complement. These data demonstrate that functional PspC was successfully expressed by the S. mitis pspC⁺ strain and, as predicted, increased the strain’s resistance to complement-mediated immunity.

A mouse model of sepsis was used to investigate whether differences in complement sensitivity might contribute toward the relatively low virulence of S. mitis. In this model, the S. mitis SK142 strain is very rapidly cleared with a 99% reduction in bacterial CFU after 30 min (Rukke et al., 2014). As neutrophils mediate complement dependent immunity against streptococcal species we first assessed the importance of neutrophils for the rapid clearance of S. mitis in the mouse model of infection. Mice were depleted of neutrophils using an anti-Ly6G monoclonal antibody administered 24 h prior to infection. This causes an eightfold reduction in neutrophil numbers (Wilson et al., 2015). At 1 h post-intraperitoneal inoculation of 5 × 10⁷ CFU S. mitis SK142, blood CFU were approximately 10-fold higher in neutrophil depleted mice (6.3 log₁₀ SD 0.32) compared to untreated mice (5.4 log₁₀ SD 0.18, P-value 0.04 Student’s t-test), but no differences were seen in CFU recovered from the spleen (6.47 log10 SD 0.41, vs. 6.50 log10 SD 0.14, respectively; Figure 4A). These data suggest some contribution from neutrophils toward the rapid clearance of S. mitis from the blood during systemic infection of mice, and is compatible with a significant role for complement in this process. To investigate the potential role for complement further, infection experiments were repeated in mice deficient in complement factor C3. One hour after intraperitoneal inoculation with 5 × 10⁶ CFU of the S. mitis SK142, blood and spleen CFU were approaching 2 log₁₀ higher in C3^–/– mice compared to wild-type mice, although the low n-number prevented these differences being statistically significant (Figure 4B). However, 4 h after inoculation, two out of three C3 deficient mice were still able to clear bacteria completely from the blood and spleen (Figure 4C).

To further investigate the role of complement during actual S. mitis infection, mice were inoculated intraperitoneally with a high dose (5 × 10⁷ CFU) of S. mitis SK142 and the relatively complement resistant S. mitis serotype 36 strain. Bacterial CFU were obtained from blood and spleen after 1 h to identify differences between strains. Mice inoculated with S. mitis serotype 36 showed outward signs of mild sickness (ruffled fur, hunched posture, and reduced mobility) that were not observed in mice infected with the SK142 strain, and also had approximately 100-fold more CFU in the blood after 1 h (around 10⁷ per ml) (Figure 4D). There was no significant difference seen between strains in the CFU recovered from the spleen. Repeat experiments showed persistence of serotype 36 S. mitis CFU 4 h after inoculation (Figure 4E), a timepoint when SK142 has been almost completely cleared from both blood and spleen (Figure 4C). Hence, the complement resistant serotype 36 S. mitis strain showed increased resistance to rapid bacterial clearance in mice compared to the more complement sensitive S. mitis SK142 strain, suggesting a role for complement in early clearance of S. mitis from the blood. However, by 24 h there was complete clearance of the serotype 36 S. mitis strain from both the blood and spleen (data not shown) demonstrating that even this relatively complement resistant S. mitis strain was unable to establish sustained infection.

To further assess the role of complement for controlling S. mitis CFU in the mouse model of sepsis we investigated whether the increased complement resistance of the SK142 S. mitis pspC⁺ strain resulted in a reduction in the usual rapid clearance of the SK142 strain in the sepsis model. Previously it has been shown that S. pneumoniae PspC (Dave et al., 2001) does not bind mouse FH, and compatible with that observation no mouse FH binding was identified to the S. mitis pspC⁺strain incubated in mouse serum (Figure 3E). Hence to investigate whether human factor H could protect S. mitis pspC⁺ from complement during mouse infection, bacteria were pre-incubated in purified human FH before inoculation into mice. In vitro experiments demonstrated that pre-incubation in human FH bound by S. mitis pspC⁺ inhibited C3b/iC3b deposition when the bacteria were subsequently incubated in baby rabbit complement (Figure 3F). For the in vivo experiments, we used competitive infection of the S. mitis pspC⁺ strain with the S. mitis wild-type (complement sensitive) strain as this is highly sensitive at identifying differences in virulence between bacterial strains (Yuste et al., 2005). However, after pre-incubation in human FH the S. mitis pspC⁺ strain had no competitive advantage with a mean CI of 0.9 (SD 0.23) 1 h post intraperitoneal injection. These data indicate that increased resistance to complement does not overcome the rapid clearance of S. mitis in the mouse model.

As replication under conditions encountered during infection is essential for bacterial virulence independent of the ability to evade innate immunity, we compared the growth in blood and serum of the S. pneumoniae and S. mitis strains. Three of the four S. mitis strains studied had poor survival in blood, with < 50% of the inoculum CFU recovered after 4 h incubation (Figure 5A). The exception was S. mitis serotype 36 which increased CFU by approximately 100% compared to the inoculum. However, this increase in CFU was markedly lower than that seen for the S. pneumoniae strains, which all demonstrated a > 10-fold increase in CFU. The serotype 19C and serotype 36 had particularly large increases in CFU after 4 h incubation in blood with a > 40-fold increase. These data indicate that the S. mitis strains have an impaired ability to replicate in blood compared to S. pneumoniae (Figure 5B), and that this effect is independent of capsular serotype as two of the S. mitis strains expressing S. pneumoniae capsular serotypes were still unable to survive in blood. To determine whether the low survival rates of S. mitis in blood was due to the action of persisting cellular and soluble immune effectors within the blood (e.g., complement and neutrophils), or due to poor growth of S. mitis under physiological conditions, growth of all strains in heat-inactivated (HI, to deplete complement activity) human serum was assessed (Figure 5C). Survival of all four S. mitis strains significantly increased in HI-sera compared to blood, with all strains demonstrating at least maintenance of inoculum CFU. Again, higher levels of S. mitis ST36 were recovered in comparison to the other three strains. However, the S. pneumoniae strains still grew considerably better in serum than the S. mitis strains, with only a marginally smaller difference in percentage survival between the pooled data for all four S. pneumoniae strains compared to the S. mitis strains in serum compared to blood (Figure 5D). These results suggest that although the presence of cells and active complement in the human blood could be exacerbating the differences in survival between S. mitis and S. pneumoniae, differences in growth rates in physiological fluid had a more dominant effect on differences in replication rates between S. mitis and S. pneumoniae.