Identification of lipopolysaccharide-binding peptide regions within HMGB1 and their effects on subclinical endotoxemia in a mouse model

Abstract

Lipopolysaccharide (LPS) triggers deleterious systemic inflammatory responses when released into the circulation. LPS-binding protein (LBP) in the serum plays an important role in modifying LPS toxicity by facilitating its interaction with LPS signaling receptors, which are expressed on the surface of LPS-responsive cells. We have previously demonstrated that high mobility group box 1 (HMGB1) can bind to and transfer LPS, consequently increasing LPS-induced TNF-α production in human peripheral blood mononuclear cells (PBMCs). We report here on the identification of two LPS-binding domains within HMGB1. Furthermore, using 12 synthetic HMGB1 peptides, we define the LPS-binding regions within each domain. Among them, synthetic peptides HPep1 and HPep6, which are located in the A and B box domains of HMGB1, bind to the polysaccharide and lipid A moieties of LPS respectively. Both HPep1 and HPep6 peptides inhibited binding of LPS to LBP and HMGB1, LBP-mediated LPS transfer to CD14, and cellular uptake of LPS in RAW264.7 cells. These peptides also inhibited LPS-induced TNF-α release in human PBMCs and induced lower levels of TNF-α in the serum in a subclinical endotoxemia mouse model. These results indicate that HMGB1 has two LPS-binding peptide regions that can be utilized to design anti-sepsis or LPS-neutralizing therapeutics.

Figure 1

LPS-binding specificity of HMGB1 domains. (A) Biotin-labeled E. coli LPS was incubated with 6× His-tagged HMGB1 A and B box proteins and pull-down assays were performed. The beads were subjected to 12% SDS-PAGE and Western blot analysis was performed using anti-His Ab. (B, C) An aliquot of 5 μg/mL of biotin–LPS was incubated with 5 μg/mL of His-tagged A box or B box HMGB1 protein that had been preincubated with various amounts of E. coli delipidated LPS, S. minnesota lipid A, S. minnesota Re595 LPS, or WT S. minnesota LPS as inhibitors. Biotin–LPS was precipitated and analyzed using Western blotting with an anti-His Ab. (C) The line indicates the cutline of the same blot membrane. Data shown are representative of two independent experiments.

Figure 2

Mapping of the LPS-binding regions of HMGB1. (A) Twelve synthetic biotin-labeled HMGB1 peptides were prepared for the LPS binding study. Boxes A and B and the acidic tail domain are underlined. (B) Briefly, 10 μg/mL of each biotin-labeled HMGB1 peptide was incubated with 10 μg/mL of LPS. Pull-down assays were performed with streptavidin agarose beads and analyzed by Western blotting. The membrane was probed with an anti-LPS Ab. WT HMGB1 was used as a positive control (left). This assay was repeated using four selected peptides (right). (C) Microtiter plates were coated with 10 μg/mL of LPS in PBS and washed with 0.05% Tween-20 PBS. Various concentrations of each biotin-labeled HMGB1 peptide were added to the wells followed by the addition of HRP-conjugated streptavidin. TMB solution was used as a substrate for color development. Data shown are representative of two independent experiments.

Figure 3

Binding of HMGB1 peptides to LPS. (A) Competitive binding analysis of the interaction between the HMGB1 peptides and LPS. Microtiter plates were coated with LPS, and the same amount of peptide No. 1 (HPep1) or No. 6 (HPep6) was added to the wells in the presence of various amounts of S. minnesota Re595 LPS, S. minnetosa lipid A, or E. coli delipidated LPS. WT S. minnetosa LPS was used as a positive control inhibitor. The binding of the HMGB1 peptides to LPS was probed by HRP-conjugated streptavidin. (B, C) Inhibition of LPS binding to LBP and HMGB1 by HMGB1 peptides. (B) Briefly, 100 ng/mL of LBP was added to LPS-coated wells in the presence of various amounts of HMGB1 peptides. The binding of LBP to LPS was probed using an anti-LBP Ab. Polymyxin B was used as a positive control inhibitor. (C) Biotin-labeled LPS was incubated with 5 μg/mL of HMGB1 in the presence of a mixture of HPep1 and HPep6 peptides, and pull-down assays were performed using streptavidin agarose beads. Western blot analysis was performed using anti-HMGB1 Ab. Data shown are representative of two independent experiments.

Figure 4

Ability of HMGB1 A- and B-box domain-containing peptides to transfer BODIPY FL-LPS to CD14. (A) A mixture of BODIPY FL-LPS and sCD14 was incubated in the presence of HMGB1 A box, B box, and GST-acidic tail proteins, and fluorescence levels were measured after 10 h at 25°C. LBP was used as a positive control, and 2% SDS was used to completely solubilize the disaggregated state of LPS for maximum fluorescence. (B) Interaction of HMGB1 with CD14. Either 2 or 10 μg of GST-HMGB1 protein was incubated with whole cell lysate from RAW264.7 cells (top) or recombinant CD14 protein (middle) and then precipitated with glutathione-Sepharose 4B beads. Western blot analysis was performed using an anti-CD14 Ab. Whole-cell lysate was loaded as a positive control. Binding of HMGB1 A and B box peptides to recombinant CD14 protein was measured by ELISA (bottom). A titration of 6-His-tagged A and B box proteins were added to the CD14-coated wells and anti-His Ab was used as the primary Ab. (C) HMGB1 peptide-mediated transfer of LPS to sCD14. In all, 1 μg/mL of BODIPY FL-LPS and 5 μg/mL of CD14 protein were incubated in the presence of 2.5 μg/mL of each HMGB1 peptide. Fluorescence levels were measured at 525 nm with a 488 nm excitation after 10 h at 25°C. LBP and HMGB1 proteins were used as positive controls. Heat-treated (Hx) HMGB1 was used as a control. Data shown are representative of three independent experiments. Error bars: standard deviation.

Figure 5

Inhibition of LPS transfer by HPep1 and HPep6. (A) A mixture of BODIPY FL-LPS, CD14 protein, and LBP was incubated in the presence of various amounts of HPep1 and HPep6, and changes in fluorescence levels were measured. (B) RAW264.7 cells were incubated with a mixture of BODIPY FL-LPS and LBP in the presence of each HMGB1 peptide and the fluorescence levels of RAW264.7 cells were measured after washing. (C) RAW264.7 cells were incubated with a preincubated mixture of 100 μg/mL of FITC-conjugated LPS and 100 μg/mL of HMGB1 peptide HPep1 or HPep6 in 10% FBS-DMEM. The binding of FITC-LPS was analyzed by flow cytometry after washing. Data shown are representative of two or three independent experiments.

Figure 6

Inhibition of TNF-α production in human PBMCs by HMGB1 peptides. Human PBMCs (5×10⁶ cells/mL) were stimulated with a preincubated mixture of 1 ng/mL of LPS and 200 ng/mL of LBP in the presence of 2.5 μg/mL of each HMGB1 peptide in serum-free Opti-MEM® medium. The cultures were incubated 16 h at 37°C and the concentration of TNF-α in the culture supernatants was determined using sandwich ELISA. Data shown are mean and standard deviation of three independent experiments.

Figure 7

Neutralizing effect of the HMGB1 peptides HPep1 and HPep6 in a subclinical endotoxemia mouse model. BALB/c mice were injected intravenously with a subclinical dose of 100 ng of ultrapure E. coli LPS and 100 μg of HPep1 or HPep6. HPep3 (HMGB1 peptide No. 3), which does not bind to LPS, was used as a negative control. Six mice per group were evaluated. Serum samples were collected 2 h after injection and the serum concentration of TNF-α was determined using sandwich ELISA. Error bars of standard deviation are represented. Data shown are representative of two independent experiments. Dunn's test (nonparametric) was used for calculation of the p-values.