Type IV Pili of Streptococcus sanguinis Contribute to Pathogenesis in Experimental Infective Endocarditis

Abstract

Streptococcus sanguinis is a common cause of infective endocarditis (IE). Efforts by research groups are aimed at identifying and characterizing virulence factors that contribute to the ability of this organism to cause IE. This Gram-positive pathogen causes heart infection by gaining access to the bloodstream, adhering to host extracellular matrix protein and/or platelets, colonizing the aortic endothelium, and incorporating itself into the aortic vegetation. While many virulence factors have been reported to contribute to the ability of S. sanguinis to cause IE, it is noteworthy that type IV pili (T4P) have not been described to be a virulence factor in this organism, although S. sanguinis strains typically encode these pili. Type IV pili are molecular machines that are capable of mediating diverse virulence functions and surface motility. T4P have been shown to mediate twitching motility in some strains of S. sanguinis, although in most strains it has been difficult to detect twitching motility. While we found that T4P are dispensable for direct in vitro platelet binding and aggregation phenotypes, we show that they are critical to the development of platelet-dependent biofilms representative of the cardiac vegetation. We also observed that T4P are required for in vitro invasion of S. sanguinis into human aortic endothelial cells, which indicates that S. sanguinis may use T4P to take advantage of an intracellular niche during infection. Importantly, we show that T4P of S. sanguinis are critical to disease progression (vegetation development) in a native valve IE rabbit model. The results presented here expand our understanding of IE caused by S. sanguinis and identify T4P as an important virulence factor for this pathogen. IMPORTANCE This work provides evidence that type IV pili produced by Streptococcus sanguinis SK36 are critical to the ability of these bacteria to attach to and colonize the aortic heart valve (endocarditis). We found that an S. sanguinis type IV pili mutant strain was defective in causing platelet-dependent aggregation in a 24-h infection assay but not in a 1-h platelet aggregation assay, suggesting that the type IV pili act at later stages of vegetation development. In a rabbit model of disease, a T4P mutant strain does not develop mature vegetations that form on the heart, indicating that this virulence factor is critical to disease and could be a target for IE therapy.

Keywords: Streptococcus sanguinis; bacterial pathogenesis; infective endocarditis; motility; platelet adherence; platelet aggregation; rabbit model; type IV pili.

FIG 1

Type IV pili in S. sanguinis SK36. (A) Model of T4P filament assembly in S. sanguinis, with PilB localized to the end of the filament based on structural modeling by Raynaud et al. (66). PilACIJK are not shown due to unknown function or localization. Image created with BioRender. (B) Genes involved in T4P biogenesis in monoderm bacteria, adapted from Pelicic (55). Known or predicted functions are indicated below the operon. Genes of the same color are predicted to encode proteins of similar function, while those colored in white have no identified homologous counterparts in other species. An asterisk indicates genes dispensable for filament assembly but essential for functional motility.

FIG 2

S. sanguinis SK36 produces T4P regulated by growth phase and detectable by immunoblotting. (A) R250 Coomassie staining for sheared T4P from S. sanguinis strains SK36 and 2908 and their isogenic ΔpilF mutants grown for 24 h in TH. (B) Colloidal G250 Coomassie staining of sheared SK36 and 2908 pili at different time points (red box). The 8-h time point denotes an OD₆₀₀ of ∼1.0 (generally 8 h), while 16- and 24-h time points were normalized to an OD₆₀₀ of 1.0 at the specified time postinoculation and then pelleted to ensure a comparable number of bacteria were processed. (C) Immunoblot of sheared pili (PilE) from SK36 and the ΔpilF mutant at 24 h. (D) Immunoblot of sheared pili from ΔpilB, MIDAS, and complementation strains. Results shown are representative of at least two independent experiments. Molecular weights (kDa) are indicated to the left of each image.

FIG 3

Twitching motility of SK36 when cultured on blood agar in the presence of excess (100 μM FeCl₂), standard (0.3 mM dipyridyl + 100 μM FeCl₂), or limited iron (3 mM dipyridyl). Plates were made with 5% sheep blood and 80% Levinthal’s medium base. Passaging of cultures was performed by scraping the outer edge of the colony with a sterile toothpick and restreaking organisms for isolation. Each passage was incubated for 4 days at 37°C in a candle jar containing sterile water at the bottom to provide humidity. Results shown are representative of at least three independent experiments.

FIG 4

T4P are not required for in vitro platelet binding or in vitro platelet aggregation. (A) Binding of immobilized bacteria to human platelets. Bacteria were first adhered to the tissue culture well surface, wells were blocked with BSA, and human platelets were added to the wells to allow binding to the bacteria. Adherence shown is relative to WT and is correlated with acid phosphatase released from attached platelets in an acid phosphatase assay, with pNPP as substrate. (B) Aggregation of human platelets was measured by light transmission aggregometry using a modified microtiter plate assay. The ΔpbrA strain was included in both sets of experiments as a control for both decreased platelet adherence and platelet aggregation. Two independent experiments with two different donors were performed for each panel. Platelet adherence was analyzed by one-way ANOVA (A) and was analyzed by two-way ANOVA (B), both corrected for multiple comparisons by the method of Dunnett. *, P < 0.05. Error bars represent the standard deviation.

FIG 5

T4P are required for invasion of iHAECs but are dispensable for adherence to iHAECs. (A) Adherence of S. sanguinis SK36 or ΔpilF mutant to iHAECs. Bacteria were mixed in a ratio of 50:1 (bacteria/iHAECs) and allowed to bind to the iHAEC surface for 3 h. Wells were washed to remove planktonic bacteria before lysing and plating dilutions of cell-associated bacteria to quantitate adherence levels. (B) Invasion of S. sanguinis SK36 or ΔpilF mutant into iHAECs. Bacteria were mixed in a ratio of 50:1 (bacteria/iHAECs) and allowed to interact with the cells. At the end of 3 h, wells were washed three times, and the M200 medium was replaced with fresh M200 medium supplemented with 100 μg/ml of gentamicin for 1 h. Subsequently, the cells were washed three times, lysed, and dilutions were plated to quantitate intracellular CFU. (C) Representative invasion of iHAEC cells by complementation strains and the ΔpilB/MIDAS mutants relative to WT invasion, performed as in panel B. Percent adherence and invasion were determined by contrasting the CFU observed following the assay to the CFU observed from plating the inoculum. At least two independent experiments were performed for panel A, and three independent experiments were performed for panel B. At least five independent experiments were performed for panels C and D. Statistical analysis was performed using two-tailed, unpaired t tests for panels A and B. Panels C and D were analyzed via one-way Brown-Forsythe and Welch ANOVA corrected for multiple comparisons by the method of Dunnett’s T3 test. *, P < 0.05. Variance of the WT was determined by normalizing within-experiment replicates to one relative value for each independent experiment. Error bars represent the 95% confidence interval.

FIG 6

T4P contribute to platelet-dependent biofilm formation and function independently of platelet aggregation. Overnight bacterial cultures were diluted into a 1:1 mixture of TH/PRP and grown for 24 h in an anaerobic chamber at 37°C. Total biomass representing both bacterial and platelet factors were washed, heat fixed, and stained with crystal violet. (A) Representative biomass development in both PRP and PPP from two independent experiments. (B) Relative platelet-dependent biofilm formation in PRP. Values are from seven independent experiments. (C) Effects of ADP and aspirin on platelet-dependent biofilm formation. At least three independent experiments were performed. Statistical analysis in panel A was performed using one-way ANOVA corrected for multiple comparisons by the method of Sidak. Data in panels B and C were analyzed via one-way ANOVA corrected for multiple comparisons by the method of Dunnett. *, P < 0.05. Error bars represent the 95% confidence interval.

FIG 7

Disease pathology of infective endocarditis is significantly attenuated in rabbits infected with the ΔpilF mutant. (A) Representative images of cardiac vegetations from WT (left) and ΔpilF mutant (right) rabbits with lesions encircled. (B) Vegetation sizes from experimental rabbits. All except one rabbit infected with WT developed vegetations greater than 20 mg; no vegetations larger than 10 mg were observed in rabbits infected with the ΔpilF mutant, with the average being 3.5 mg. (C) CFU recovered from cardiac lesions or the valve surface are significantly different between WT and mutant strains. (D) Rabbits infected with the mutant strain experienced significantly less splenomegaly as measured by spleen weight. Means for vegetation size and CFU are geometric. Statistical analysis was performed using Mann-Whitney. *, P < 0.05. Error bars represent the 95% confidence interval. Two additional data points for the WT strain are present due to the hearts of two rabbits being sent for histology; vegetations from these rabbits were thus not able to be quantified during necropsy. (E) Pathology of distal organs is not significantly different between WT and the ΔpilF mutant. Gross pathology of the kidney, liver, and lungs from infected rabbits scored by a previously described scale (38). Organs were scored by three blinded investigators. The SK36 ΔpilF mutant did not exhibit distinct organ pathology compared to WT. Kidneys and liver exhibited moderate pathology on average, characterized by rare, small lesions evident on the surface. By contrast, lungs generally exhibited more severe pathology characterized by multifocal hemorrhaging and necrosis evident on one or both lobes. Statistical analysis was performed using unpaired, two-tailed t tests. Error bars represent the 95% confidence interval.