A streptococcal lipid toxin induces membrane permeabilization and pyroptosis leading to fetal injury

Group B streptococci (GBS) are Gram-positive bacteria that cause infections in utero and in newborns. We recently showed that the GBS pigment is hemolytic and increased pigment production promotes bacterial penetration of human placenta. However, mechanisms utilized by the hemolytic pigment to induce host cell lysis and the consequence on fetal injury are not known. Here, we show that the GBS pigment induces membrane permeability in artificial lipid bilayers and host cells. Membrane defects induced by the GBS pigment trigger K+ efflux leading to osmotic lysis of red blood cells or pyroptosis in human macrophages. Macrophages lacking the NLRP3 inflammasome recovered from pigment-induced cell damage. In a murine model of in utero infection, hyperpigmented GBS strains induced fetal injury in both an NLRP3 inflammasome-dependent and NLRP3 inflammasome-independent manner. These results demonstrate that the dual mechanism of action of the bacterial pigment/lipid toxin leading to hemolysis or pyroptosis exacerbates fetal injury and suggest that preventing both activities of the hemolytic lipid is likely critical to reduce GBS fetal injury and preterm birth.


SUPPLEMENTARY INFORMATION
. Kinetics of K + and Hb efflux due to hemolysis induced by Staphylococcus aureus αtoxin and Triton X 100.
3) Figure S3. Disruption of artificial lipid bilayers by GBS pigment (75nM), pore forming porin MspA of Mycobacterium smegmatis and detergent SDS. 4) Figure S4. The GBS pigment induces secretion of IL18 but not IL6, TNF-α and IFN-γ from THP-1 derived macrophages. 5) Figure S5. Western blots of THP-1 shRNA knockdown cell lines. 6) Figure S6. Osmoprotectants and the caspase 3/7 inhibitor do not provide protection from macrophage cell death observed with hyperpigmented GBS strains. 7) Figure S7. Increasing amounts of the caspase inhibitor z-YVAD-FMK provided significant protection from GBS pigment mediated cell death in macrophages. 8) Figure S8. Nucleic acid or proteins are absent from purified GBS pigment. 9) Figure S9. Inactive GBS pigment is not hemolytic and does not induce cell death or IL-1β secretion in THP-1 macrophages.

10) Reference Cited
Figure S1. Kinetics of K + and Hb efflux due to hemolysis induced by Staphylococcus aureus αtoxin and Triton X 100.
(A) Release of K + and Hb release was measured from RBC treated with 0.47µM S. aureus α-toxin. Efflux of K + occurred faster than Hb, as measured by time to 50% release (9.6min vs 13.3min, p < 0.001, extra sum-of-squares F test); this lag is similar to the lag observed for GBS pigment-mediated hemolysis (see Fig. 1B). Data are the average ± SEM of two independent experiments. (B) K + and Hb release was measured from RBC treated with 0.22µM Triton-X 100 and 100% release of K + and Hb occurred instantly. Osmoprotectants provide minimal protection from hemolysis caused by SDS (a direct lysis mechanism). Notably, the osmoprotectants PEG1500 and PEG3000 provided complete protection to lysis induced by the GBS pigment, suggesting that the pigment induces hemolysis not via a direct lysis mechanism, but rather a colloidal-osmotic mechanism (see Fig. 1C). Data are the average ± SEM of two independent experiments.

Figure S3. Disruption of artificial lipid bilayers by GBS pigment (75nM), pore forming porin MspA of Mycobacterium smegmatis and detergent SDS. (A, B) GBS pigment (75nM) induces membrane permeability in black lipid membranes (BLMs), see A & for detail t = 320-345s, see B.
Equivalent volume of ΔcylE extract does not disrupt BLMs (see C). As controls, BLMs were incubated with MspA (0.51nM; D) or SDS (350µM; E). For MspA, protein was added at t = 0s, and mixed by pipetting from t =10-20s. Stepwise increase in current indicative of multiple pore formation is observed. For SDS, detergent was added at t=0s and mixed by pipetting from t =17-22s. A rapid and large increase in conductance is observed around 23s which is sustained for a few seconds before bilayer disruption, indicative of bilayer solubilization. Levels of IL-6, TNF-α, and IFN-γ or IL-18 in the supernatant of pigment or ΔcylE extract treated THP-1 cells were measured. While secretion of IL-18 is significantly higher in pigment treated cells, there is no significant increase in IL-6, TNF-α, or IFN-in response to purified GBS pigment. This suggests that the purified pigment induces activation of the NLRP3 inflammasome, it does not induce secretion of TLRmediated cytokines (n=3, ***p = 0.0002, *p = 0.017 for 0.5 µM pigment, *p = 0.025; for 0.25µM pigment, Bonferroni's multiple comparison test following ANOVA; error bars ± SEM).

Figure S6. Osmoprotectants and the caspase 3/7 inhibitor do not provide protection from macrophage cell death observed with hyperpigmented GBS. THP-1 cells transfected with empty
vector, scrambled control, shASC or shNLRP3 were incubated with GBSΔcovR (MOI=1) for 4 hours in the presence or absence of 30mM PEG1500 (A) or 100µM caspase 3/7 inhibitor (Z-DEVD-FMK, shown below as DEVD or with control DMSO (B). Percent cell death was then measured by LDH release. The addition of the osmoprotectant (PEG1500) or DEVD did not reduce the amount of cell death in THP-1 derived macrophages including shASC and shNLRP3 cells. (Left Panel) Purified GBS pigment (1.6nmol) and equivalent amount of control ΔcylE extract were resolved by agarose gel electrophoresis and stained with ethidium bromide for detection of nucleic acids (DNA or RNA). (Middle Panel) Purified GBS pigment (9.3nmol) and equivalent amount of control ΔcylE extract were resolved by SDS-PAGE followed by Sypro Ruby staining for detection of proteins.
(Right Panel) Three independently purified batches of GBS pigment (0.6nmol in 3µL) and ΔcylE extract were tested for the presence of GBS RNA using RT-PCR (Qiagen); reverse transcription and PCR for the housekeeping gene rpsL was performed as previously described (Lembo et al, 2010). As a control, 500ng of RNA isolated from the GBS strain A909 was tested in the presence and absence of DMSO:0.1% TFA.