Streptococcus agalactiae glycolipids promote virulence by thwarting immune cell clearance

Abstract

Streptococcus agalactiae [group B Streptococcus (GBS)] is a leading cause of neonatal meningitis, with late-onset disease (LOD) occurring after gastrointestinal tract colonization in infants. Bacterial membrane lipids are essential for host-pathogen interactions, and the functions of glycolipids are yet to be fully elucidated. GBS synthesizes three major glycolipids: glucosyl-diacylglycerol (Glc-DAG), diglucosyl-DAG (Glc₂-DAG), and lysyl-Glc-DAG (Lys-Glc-DAG). Here, we identify the enzyme, IagB, as responsible for biosynthesis of Glc-DAG, the precursor for the two other glycolipids in GBS. To examine the collective role of glycolipids to GBS virulence, we adapted a murine model of neonatal meningitis to simulate LOD. The GBS∆iagB mutant traversed the gut-epithelial barrier comparable to wild type but was severely attenuated in bloodstream survival, resulting in decreased bacterial loads in the brain. The GBS∆iagB mutant was more susceptible to neutrophil killing and membrane targeting by host antimicrobial peptides. This work reveals an unexplored function of GBS glycolipids with their ability to protect the bacterial cell from host antimicrobial killing.

Fig. 1.. IagB is required for Glc-DAG biosynthesis in GBS.

(A) Glycolipid biosynthesis pathways in GBS (created in BioRender). (B) Major GBS glycolipids (Glc-DAG, Glc₂-DAG, and Lys-Glc-DAG) are present in COH1 WT. COH1∆iagA lacks Glc₂-DAG, while COH1∆mprF lacks Lys-Glc-DAG. As expected, COH1∆iagB lacks all three glycolipids, consistent with the fact that IagB is required for the synthesis of Glc-DAG and its derived Glc₂-DAG and Lys-Glc-DAG. Shown are the positive extracted ion chromatograms. (C) ELISA for GBS type I LTA on the bacterial whole cell and culture supernatant. COH1iagB synthesizes significantly more LTA and sheds it into the culture supernatant compared to COH1 WT and COH1∆iagA [biological and technical triplicate, mean and SEM, ordinary one-way analysis of variance (ANOVA) with Fisher’s least significant difference (LSD)]. P values indicated; ns, not significant (P value > 0.05). (D) Representative TEM images (n = 6) of COH1 WT (top) and COH1∆iagB (bottom) indicate a thickening of the cell envelope in ∆iagB mutant cells. Magnification: left panels, 23k; right panels, 68k. Scale bars: 200 nm (left) and 50 nm (right).

Fig. 2.. GBS∆iagB is attenuated in murine hematogenous meningitis model.

(A) Schematic of murine hematogenous meningitis model (created in BioRender). (B) Kaplan-Meier survival curve of groups of CD-1 mice injected intravenously (iv) with 10⁹ to 10¹⁰ COH1 WT or COH1ΔiagB strains; bacterial counts were assessed in the (C) brain, heart, and lungs and (D) blood at moribund state or 72 hours (median indicated; WT, n = 20; ΔiagB, n = 19). (E) KC chemokine production measured by ELISA. In vitro assays for (F) adherence (total cell associated), (G) invasion, and (H) normalized hCMEC invasion to total cell associate (WT set to 100%) indicate that iagB contributes to adherence but not invasion to brain endothelium (mean of three replicate experiments with four technical replicates, mean and SEM). Transcript levels of (I) Snai1 and (J) Cldn5 in hCMEC cells after 5-hour infection at a multiplicity of infection (MOI) of ~20 (mean of three replicate experiments with three technical replicates, mean and SEM). No difference in transcript levels between WT and ∆iagB-infected wells, indicating similar transcriptional changes resulting in BBB permeability. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. Statistical analyses: (B) log-rank test, [(C) to (E)] Mann-Whitney U test, and [(F) to (J)] ordinary one-way ANOVA with Fisher’s LSD test. UI, uninfected. P values indicated; ns, not significant (P value > 0.05).

Fig. 3.. Murine neonatal GBS meningitis model via GI colonization.

(A) Schematic of murine neonatal GBS meningitis model (created in BioRender). (B) Kaplan-Meier survival curve of groups of P2 C57Bl/6J mice injected intragastrically with 10⁶ COH1 WT or COH1ΔiagB strains; bacterial counts were assessed in the (C) stomach, intestines, brain, and blood at moribund state or 48 hours (median indicated; WT, n = 14; ΔiagB, n = 15). (D) Brain KC chemokine production measured by ELISA. Statistical analyses: (B) log-rank test, (C) Mann-Whitney U test, and (D) ordinary one-way ANOVA with Fisher’s LSD test. P values indicated; ns, not significant (P value > 0.05).

Fig. 4.. IagB promotes bloodstream survival.

Groups of P2 C57Bl/6J mice were inoculated with 10⁶ COH1 WT or COH1∆iagB and bacterial burdens assessed at (A) 4 hpi and (B) 8 hpi. No difference between tissue bacterial burdens was observed at 4 hpi, indicating that GBS∆iagB is not defective in traversing the GI epithelial barrier; however, at 8 hpi, significant differences in bacterial burden are observed. (C) Blood immune profiling by flow cytometry at 8 hpi of neutrophils. Fewer neutrophils are observed in the blood of WT and ∆iagB-infected mice compared to uninfected controls, suggesting that GBS modulates neutrophil numbers during acute infection. (D) Pooled human whole blood survival of COH1, COH1∆iagB, and complemented strain after 1-hour exposure. COH1∆iagB has a significant decrease in survival compared to WT. Percent survival (% Survival) calculated as the ratio of CFU at 1 hour versus input CFU (mean and SEM, biological triplicate). [(A) to (C)] Median indicated. Statistical analyses: [(A) and (B)] Mann-Whitney U test and [(C) and (D)] ordinary one-way ANOVA with Fisher’s LSD test. P values indicated; ns, not significant (P value > 0.05).

Fig. 5.. GBS∆iagB is more susceptible to neutrophil killing.

(A) HL60-neutrophil killing of COH1 WT and COH1∆iagB at an MOI of 0.01 at 3 hours. % Survival calculated as the ratio of CFU in normal serum versus CFU in heat-killed serum in the presence of HL60 cells. Opsonophagocytic killing of (B) COH1 WT and (C) COH1∆iagB. Cytochalasin D prevents COH1 WT killing but does not completely prevent killing of COH1∆iagB, indicating that neutrophils can kill COH1∆iagB in a different manner. % Survival calculated as the ratio of CFU in normal serum versus CFU in heat-killed serum and HL60 negative control set to 100% (mean and SEM, biological triplicate). (D) Kaplan-Meier curves of neutrophil-depleted (α-Ly6G antibody treated; solid lines) and control (α-IgG2a antibody treated; dashed lines) P2 mice were infected with 10⁶ CFU of COH1 WT (red lines) and COH1∆iagB (blue lines). No significant difference observed between WT (red solid line) and ∆iagB (blue solid line) survival in neutrophil-depleted mice. ∆iagB-infected control mice (blue dashed line) survived significantly better than WT control (red dashed line) and ∆iagB-infected neutrophil-depleted (blue solid line) mice. COH1-depleted, n = 7; ∆iagB-depleted, n = 9; WT control, n = 5; ∆iagB control, n = 6. Statistical analyses: (A) unpaired two-tailed t test, [(B) and (C)] ordinary one-way ANOVA with Fisher’s LSD test, and (D) log-rank test. P values indicated; ns, not significant (P value > 0.05).

Fig. 6.. GBS glycolipids protect the membrane from AMPs.

(A) GBS∆iagB has a significantly more negatively charged membrane compared to WT as measured by FM 4-64 membrane staining. GBS∆iagB has significantly reduced survival compared to GBS WT after 30 min of exposure to (B) mouse mCRAMP (16 μM) and (C) human LL-37 (16 μM) and 1 hour exposure to (D) human HNP-1 (14.5 μM). Statistical analyses: [(A) to (D)] ordinary one-way ANOVA with Fisher’s LSD test compared to WT. Mean and SEM; P values indicated. Biological triplicate with technical replicates performed for each assay.

Fig. 7.. GBS glycolipids promote bloodstream survival.

GBS colonizes the neonatal GI tract where it can disrupt the GI epithelial barrier and spread systemically via the bloodstream. Neonatal neutrophils in the bloodstream use CAMPs to control the infection. GBS∆iagB, which lacks all glycolipids, is unable to protect the cellular membrane from CAMPs and is cleared at a higher rate than WT GBS. GBS attachment and penetration of the BBB results in BBB disruption and inflammation resulting in meningitis, which is reduced in the absence of glycolipids. Created in BioRender.