Research ArticleMICROBIOLOGY

All major cholesterol-dependent cytolysins use glycans as cellular receptors

See allHide authors and affiliations

Science Advances  22 May 2020:
Vol. 6, no. 21, eaaz4926
DOI: 10.1126/sciadv.aaz4926
  • Fig. 1 Glycan array analyses of glycan binding by CDCs.

    Red indicates binding above background. Blue indicates binding observed here and previously (9). White indicates that the structure was not bound. Classes of glycans arrayed are indicated to the left of the figure. A complete list of glycan structures on the array and the mean fold fluorescence increase above background of three replicate glycan array experiments are shown in data S1. GAGs, glycosylaminoglycans.

  • Fig. 2 Analysis of the hemolytic activity of CDCs in the presence of candidate glycan cellular receptors.

    (A to F) The hemolytic activity of the CDCs at the hemolytic dose 100 (HD100) or HD50 against 1% (v/v) human group O RBCs was determined after preincubation with phosphate-buffered saline, or with the indicated concentration of candidate glycan receptor, or with the same concentration of cellobiose, which served as a common, nonbinding glycan control. For the analysis of the impact of sTF and TF on ILY binding to RBCs (G), RBCs were fixed to the wells of a 96-well plate, and the ability of 2 mM sTF, 2 mM TF, or 2 mM cellobiose to block the binding of 25 ng of ILY was assessed. The HD100 of SLO was also assessed in the presence of blood group B type IV pentasaccharide and pregnenolone sulfate (PS) (sodium salt) (H), individually and in combination (see fig. S7G for determination of HD100 of SLO and fig. S7, I and J for hemolytic assays in the presence of a range of concentrations of blood group B type IV pentasaccharide and PS). Results presented are a representative assay conducted with a single blood donor and are shown as the mean of triplicate samples, with error bars showing ±1 SD from the mean. Multiple independent assays (minimum of three) were performed all showing the same trend. The percentage values for each data point used to generate bar graphs are shown in table S4A. The structure of the glycan receptor and the KD of the CDC for the glycan as determined by SPR are shown above the appropriate bar in each histogram. Statistical significance was determined using a two-tailed unpaired Student’s t test. **P < 0.005, ***P < 0.0005, and ****P < 0.0001. P values for t tests comparing hemolytic activity of CDCs without and with glycan/PS are shown in table S4B. B pentaose, blood group B type IV pentasaccharide; A pentaose, blood group A type IV pentasaccharide; H tetraose, blood group H antigen type IV tetraose; sLeC, sialyl-Lewis C; α2-6SLN, Neu5Acα2-6 LacNAc; P1, P1 antigen; Xeno, Xeno antigen/Galili epitope; A-tri, blood group A trisaccharide; sTF, sialyl-TF.

  • Fig. 3 Analysis of the interaction of Ply with sLeX-glycosylated Mac-1 and cytotoxicity for THP-1 cells with and without CD11b.

    (A) SPR measurements for the affinity of Ply for Mac-1, Mac-1 with the N-linked sLeX glycan removed using peptide N-glycosidase (PNGase) (Mac-1 + PNGase), and the CD11b I-domain showing sLeX enhanced binding to Mac-1 (P = 0.0163) and that glycan-independent binding of Ply to Mac-1 occurs via the I-domain as no difference in binding was observed between Mac-1 lacking sLeX and recombinant human I-domain (P = 0.4031). NCDI, no concentration dependent interaction detected at the concentrations tested. A graphical representation of the Mac-1 complex is shown under the SPR KD values: CD18 = yellow, CD11b = blue, and I-domain = orange, with sLeX glycosylation of I-domain structure shown. (B) Cytotoxicity of S. pneumoniae D39 expressing wild-type Ply (D39) or a nontoxic version of Ply (Ply460D and L460D) for THP-1 cells with control shRNA (shRNA control) or THP-1 CD11b shRNA knockdown cells (shRNA CD11b). A multiplicity of infection of 2.5 of S. pneumoniae cells was used. Results are shown as the mean of duplicate, independent assays (each assay consisting of triplicate samples), with error bars showing ±1 SD from the mean. Statistical significance was determined using a two-tailed unpaired Student’s t test. *P < 0.05.

  • Fig. 4 Analysis of the impact, on ILY binding, of removing the sialic acids and O-linked glycans from RBC hCD59.

    Solubilized RBC membrane proteins were treated with α2-3 neuraminidase (N) to remove the sialic acids from hCD59, or with both α2-3 neuraminidase and O-glycosidase (N + O) to remove O-linked glycans from hCD59, or with enzyme buffer only under the same conditions (buffer); then were probed with biotin labeled ILY (ILY). Western blotting with anti-CD59 polyclonal antibody confirmed that ILY was binding to hCD59 (anti-CD59). M, molecular weight marker. See fig. S4B for complete far Western/Western blot images and Coomassie-stained gel.

  • Fig. 5 NMR, molecular dynamics and site-directed mutagenesis, confirmed structures of SLY D4 engaged with two distinct glycan receptors.

    Sections of 1H-15N HSQC NMR CSP spectra of SLY D4 in the presence of αGal/Galili antigen (A) and P1 antigen (B) at a protein:ligand ratio of 1:10. Signals of apo-SLY D4 are shown in blue and SLY D4 in complex with glycans are shown in red. Full 1H-15N HSQC NMR CSP spectra are shown in fig. S6 (B and C). Amino acids that showed an intensity change or CSP are highlighted in yellow in the SLY D4 structure. Unbiased docking experiments of SLY D4 with αGal/Galili antigen (C) and P1 antigen (D) represent an energetically favored bound conformation. Both modeled structures are in excellent agreement with 1H-15N HSQC NMR titration experiments (see Fig. 5, A and B). Binding residues identified in molecular docking experiments only are highlighted in cyan. Coordinating amino acids are shaded in green. Dotted yellow bars represent hydrogen bonds. Green bars represent strong hydrophobic interactions, whereas orange bars represent weaker hydrophobic interactions. For an enlarged version of (C) and (D) showing the key coordinating residues, see fig. S6D. See table S3 for SPR analysis of site-directed mutants of key coordinating residues highlighted in (A) and (B). The whole SLY toxin crystal structure is shown in blue (D1), purple (D2), yellow (D3), and gray (D4) in (A) (Protein Data Bank code 3HVN). ppm, parts per million.

  • Table 1 A subset of SPR results for CDCs against glycans plus cholesterol and PS.

    Glycans, cholesterol, and PS were tested at concentrations listed in data S2 and the Materials and Methods. Data shown are the mean KDs from at least three independently run SPR analyses. NCDI, no concentration dependent binding observed up to the maximum concentration tested; n.t., not tested, sample was not run against that toxin. α1-3Gal tri, Galili (αGa1) antigen trisaccharide; sTF, Neu5Acα2-3Ga1β1-3Ga1NAc; sLeC, Neu5Acα2-3Ga1β1-3G1cNAc; α2-6sLN, Neu5Acα2-6LacNAc; BGA tri, blood group A trisaccharide; BGA type 4, blood group A type IV pentasaccharide; BGB type 4, blood group B type IV pentasaccharide; BGH di, blood group H antigen disaccharide; BGH type 4, blood group H antigen type IV tetraose; PS, pregnenolone sulfate (sodium salt). For the full range of glycans tested, error, and more than two significant figures, see table S1.

    GT2n.t.NCDIn.t.n.t.n.t.n.t.6.7 nM10 nMn.t.n.t.n.t.n.t.
    α1-3Gal trin.t.n.t.n.t.n.t.n.t.n.t.n.t.n.t.258 nM547 nMNCDINCDI
    P1 Antigenn.t.n.t.n.t.n.t.n.t.n.t.n.t.n.t.359 nM342 nMNCDINCDI
    LNnTn.t.0.93 nM2.8 μMNCDINCDIn.t.NCDINCDIn.t.n.t.4.9 μMn.t.
    TFn.t.n.t.46 μMn.t.NCDIn.t.n.t.n.t.n.t.n.t.60 μMn.t.
    sTFn.t.n.t.13 μMn.t.NCDIn.t.n.t.n.t.NCDINCDI599 nM568 nM
    sLeCn.t.1.4 μM1.6 μMNCDI16 μMn.t.n.t.n.t.NCDINCDI16 μM4.9 μM
    α2-6sLNn.t.n.t.23 μM633 nM1.6 μM1.8 μMn.t.n.t.n.t.n.t.NCDINCDI
    sLeX1.3 μM3.1 μM3.9 μMn.t.n.t.n.t.n.t.n.t.NCDINCDI2.8 μM4.9 μM
    BGA trin.t.1.3 μM5.1 μMn.t.NCDIn.t.n.t.n.t.8.7 nM4.4 nM9.9 μM4.9 μM
    BGA type 4n.t.105 nM150 μMn.t.NCDIn.t.n.t.n.t.5.7 μM9.2 μMNCDINCDI
    BGB type 4n.t.0.12 nM32 μMn.t.NCDIn.t.n.t.n.t.8.4 μM6.4 μMNCDIn.t.
    BGH din.t.603 nMNCDIn.t.NCDIn.t.n.t.n.t.19 μM26 μMNCDIn.t.
    BGH type 4n.t.2.6 μMNCDIn.t.NCDIn.t.n.t.n.t.n.t.n.t.NCDIn.t.
    PS0.4 nM52 nMn.t.n.t.n.t.997 nMn.t.214 nM1.8 μMn.t.588 nMNCDI

Supplementary Materials

  • Supplementary Materials

    All major cholesterol-dependent cytolysins use glycans as cellular receptors

    Lucy K. Shewell, Christopher J. Day, Freda E.-C. Jen, Thomas Haselhorst, John M. Atack, Josephine F. Reijneveld, Arun Everest-Dass, David B. A. James, Kristina M. Boguslawski, Stephan Brouwer, Christine M. Gillen, Zhenyao Luo, Bostjan Kobe, Victor Nizet, Mark von Itzstein, Mark J. Walker, Adrienne W. Paton, James C. Paton, Victor J. Torres, Michael P. Jennings

    Download Supplement

    The PDF file includes:

    • Methodological Details
    • Figs. S1 to S8
    • Tables S1 to S4
    • Legends for datasets S1 to S2
    • References

    Other Supplementary Material for this manuscript includes the following:

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article