Molecular basis of cell membrane adaptation in daptomycin-resistant Enterococcus faecalis

Abstract

Daptomycin is a last-resort lipopeptide antibiotic that disrupts cell membrane (CM) and peptidoglycan homeostasis. Enterococcus faecalis has developed a sophisticated mechanism to avoid daptomycin killing by redistributing CM anionic phospholipids away from the septum. The CM changes are orchestrated by a 3-component regulatory system, designated LiaFSR, with a possible contribution of cardiolipin synthase (Cls). However, the mechanism by which LiaFSR controls the CM response and the role of Cls are unknown. Here, we show that cardiolipin synthase activity is essential for anionic phospholipid redistribution and daptomycin resistance since deletion of the 2 genes (cls1 and cls2) encoding Cls abolished CM remodeling. We identified LiaY, a transmembrane protein regulated by LiaFSR, and Cls1 as important mediators of CM remodeling required for redistribution of anionic phospholipid microdomains. Together, our insights provide a mechanistic framework on the enterococcal response to cell envelope antibiotics that could be exploited therapeutically.

Keywords: Bacterial infections; Infectious disease; Lipid rafts; Peptides.

Figure 1. The cls genes are upregulated in association with DAP and can compensate for each other.

(A) qPCR results evaluating cls1 and cls2 expression in DAP-R Efs OG117ΔliaX relative to DAP-S Efs OG117. ****P < 0.0001, n = 3–4 replicates, individual t test. (B) qPCR results evaluating cls1 and cls2 expression in DAP-S Efs OG117 with DAP 1 μg/mL relative to DAP-S Efs OG117. n > 3 replicates via unpaired 2-tailed t test. (C) cls1 or cls2 gene expression in Efs OG117Δcls1 and Efs OG117Δcls2 relative to Efs OG117. n = 3, unpaired 2-tailed t test. (D) cls1 or cls2 gene expression in Efs OG117ΔliaXΔcls1 or Efs OG117ΔliaXΔcls2 relative to Efs OG117ΔliaX; n = 3, unpaired 2-tailed t test.

Figure 2. Both cls genes are required for DAP-R anionic phospholipid redistribution phenotype.

(A) Representative images of NAO staining (top row), bright-field (middle row), and overlay (bottom row) in cls mutant derivatives of Efs OG117ΔliaX and DAP MICs (under images). Scale bar: 2μm. White arrows represent anionic phospholipid microdomains at mid-cell or non–mid-cell locations. (B) Quantification of septal localization of anionic phospholipid microdomains with NAO in cls mutant derivatives of Efs OG117ΔliaX by counting a minimum of 50 cells per replicate (n = 3–5 replicates, ****P < 0.0001, 1-way ANOVA with multiple comparisons). (C) Representative images of NAO staining (top row), bright-field (middle row), and overlay (bottom row) in cls mutant derivatives of Efs OG117ΔliaX complemented with vector pMSP3535 and derivatives with DAP MICs (under images). White arrows represent anionic phospholipid microdomains at mid-cell or non–mid-cell locations. Scale bar: 2 µm. (D) Quantification of septal localization of anionic phospholipid microdomains with NAO in cls mutants of Efs OG117ΔliaX complemented with vector pMSP3535 and derivatives by counting a minimum of 50 cells per replicate. n = 3–4, **P < 0.01, ***P < 0.001, 1 way ANOVA with multiple comparisons. Whole images were adjusted for “Black Balance” per BZ-X800 Image Analysis Software with individual representative selected.

Figure 3. Cell membrane phospholipid changes associated with DAP-R and deletion of cls genes.

(A) Quantification of lipid classes in DAP-susceptible E. faecalis OG117 and DAP-R E. faecalis OG117ΔliaX. (B) Quantification of lipid classes in DAP-susceptible E. faecalis OG117 (parental) and cls mutant derivatives. (C) Quantification of cardiolipin (CL) content in DAP-susceptible E. faecalis OG117 and cls mutant derivatives. (D) Quantification of lipid classes in DAP-R E. faecalis OG117ΔliaX and cls mutant derivatives. (E) Quantification of cardiolipin (CL) content in DAP-R E. faecalis OG117ΔliaX and cls mutant derivatives. DG, diacylglycerol; DGDG, diglycosyldiacylglycerol; PG, phosphatidylglycerol; LysylPG, lysyl-phosphatidylglycerol; CL, cardiolipin. n = 4, *P < 0.05, **P < 0.01, ****P < 0.0001, 1- or 2-way ANOVA with multiple comparisons

Figure 4. Cls1 localizes in nonseptal anionic phospholipid microdomains in DAP-R E. faecalis.

(A) Left 2 panels: Representative images of NAO staining (top panel), tetracysteine-tagged Cls1 with ReAsH reagent (red fluorescence, middle panel), and overlay of both images (bottom panel). White arrows represent anionic phospholipid microdomains at mid-cell or non–mid-cell locations. Both anionic phospholipid microdomains and Cls1 are shown to relocalize and overlap away from the septum in DAP-R Efs (OG1RFΔliaX) (Pearson Correlation Coefficient, 0.91 [95% CI, 0.74–0.97]) compared with their septal pattern in DAP-S E. faecalis OG1RF (Pearson Correlation Coefficient, 0.94 [95% CI, 0.83–0.98]). Right 2 panels: NAO staining for anionic phospholipid microdomains (green) and tetracysteine tagged-Cls1 localization (red) in E. faecalis OG117ΔliaXΔcls1Δcls2 harboring pMSP3535 carrying the gene expressing the tetracysteine tagged-Cls1 with or without nisin induction at 50 ng/mL, overlap between NAO and ReAsH (Pearson Correlation Coefficient, 0.91 [95% CI, 0.76–0.96]). White arrows represent anionic phospholipid microdomains at mid-cell or non–mid-cell locations. Whole images were adjusted for “Black Balance” per BZ-X800 Image Analysis Software with individual representative selected. (B) Quantification of Cls septal localization for strains, minimum 50 cells counted/strain, n = 3.

Figure 5. LiaY, a member of the LiaFSR system, mediates changes in phospholipid architecture associated with DAP resistance.

(A) Representative images of anionic phospholipid localization by NAO staining (top row), bright-field (middle row), and overlay (bottom row) in DAP-R Efs OG1RFΔliaX*289 and liaYZ mutants. DAP MICs (under images). White arrows represent anionic phospholipid microdomains at mid-cell or non–mid-cell locations. Scale bar: 2 μM. (B) Quantification of septal localization of anionic phospholipid microdomains by NAO staining in Efs OG1RF and liaXYZ mutants from A (>50 cells counted per strain per replicate, n = 3–4 replicates). ***P < 0.001 via 1-way ANOVA with multiple comparisons. (C) Representative images of anionic phospholipid localization by NAO staining (top row), bright-field (middle row), and overlay (bottom row) in Efs OG1RF liaX*289ΔliaYZ mutant transformed with pAT392 and derivatives. DAP MICs (under images). White arrows represent anionic phospholipid microdomains at mid-cell or non–mid-cell locations. Scale bar: 2 μM. (D) Quantification of septal localization of anionic phospholipid microdomains by NAO staining of Efs OG1RF liaX*289ΔliaYZ mutant complemented with pAT392 and derivatives from C (>50 cells counted per strain per replicate, n = 6–9 replicates). *P < 0.05; **P < 0.001, 1-way ANOVA with multiple comparisons. Whole images were adjusted for “Black Balance” per BZ-X800 Image Analysis Software with individual representative selected.

Figure 6. LiaY bridges the LiaFSR response with changes in membrane architecture via Cls.

(A) Protein-protein interactions between LiaY and Cls1 using the bacterial 2-hybrid system. Proteins were tagged at either the N- or C-terminus, cotransformed into E. coli BTH101, and activity recorded via a β-galactosidase assay. A leucine zipper interaction was used as the positive control (T18-zip/T25-zip, red bar) with 2 nontagged empty vectors used as negative controls (T18/T25). n = 3–8, ****P < 0.001, 1-way ANOVA with multiple comparisons with (-) against LiaY/Cls1 combinations. (B) Protein-to-protein interactions between LiaZ and Cls1 using the bacterial 2-hybrid system and following the same methodology as in A. n = 4–7. (C) Representative images of NAO staining (top row), mCherry (second row), bright-field (third row), and overlay (bottom row) in DAP-S (Efs OG1RFΔliaY) (Pearson Correlation Coefficient, 0.67 [95% CI, 0.24–0.88]) and DAP-R (Efs OG117ΔliaX) (Pearson Correlation Coefficient, 0.91 [95% CI, 0.75–0.97]). Both strains were transformed with pMSP3535::liaY-mCherry. White arrows represent anionic phospholipid microdomains at mid-cell or non–mid-cell locations. Scale bar: 2 μM. (D) Representative images of Cls1-GFP (top row), LiaY-mCherry (second row), bright-field (third row), and overlay (bottom row) of LiaY-mCherry and Cls1-GFP via fluorescence microscopy in Efs OG117 (Pearson Correlation Coefficient, 0.69 [95% CI, 0.28–0.89]) and Efs OG117ΔliaX GFP-Cls1 (Pearson Correlation Coefficient, 0.86 [95% CI, 0.63–0.95]) (with gfp-cls1 introduced in the chromosome) transformed with pMSP3535::liaY-mCherry. White arrows represent anionic phospholipid microdomains at mid-cell or non–mid-cell locations. Scale bar: 2 μM. Whole images were adjusted for “Black Balance” per BZ-X800 Image Analysis Software with individual representative selected. (E and F) Quantification of septal localization, minimum 50 cells counted per strain. n = 3, ***P < 0.001, 2-tailed t test.

Figure 7. Mechanistic model of cell membrane adaptation in daptomycin-resistant E. faecalis.

In the absence of daptomycin (top panel), anionic phospholipid microdomains are located at the bacterial division septum where other critical membrane-associated proteins involved in cell division, cell wall or lipid biosynthesis (such as Cls1) are also found. Members of the LiaFSR system (namely liaXYZ) are produced at basal levels and localize in septal areas. LiaX plays an inhibitory role through its C-terminal domain likely via interactions with members of the LiaFSR system. Upon activation of LiaFSR in the presence of daptomycin (bottom panel), the N-terminal domain of LiaX is released to the milieu and interacts with DAP. Concomitant changes in the C-terminus of LiaX cause “activation” of LiaY, generating interactions with Cls1. The LiaY-Cls1 complexes localize in areas away from the septum, presumably in regions where the membrane is being damaged by the antibiotic. Cls1 then produces high amounts of cardiolipin in these nonseptal areas, attracting DAP molecules to these membrane regions, resulting in alteration of the interaction of DAP with its transmembrane target (likely lipid II intermediates), preventing further DAP-mediated damage.