A novel receptor - ligand pathway for entry of Francisella tularensis in monocyte-like THP-1 cells: interaction between surface nucleolin and bacterial elongation factor Tu

Abstract

Background: Francisella tularensis, the causative agent of tularemia, is one of the most infectious human bacterial pathogens. It is phagocytosed by immune cells, such as monocytes and macrophages. The precise mechanisms that initiate bacterial uptake have not yet been elucidated. Participation of C3, CR3, class A scavenger receptors and mannose receptor in bacterial uptake have been already reported. However, contribution of an additional, as-yet-unidentified receptor for F. tularensis internalization has been suggested.

Results: We show here that cell-surface expressed nucleolin is a receptor for Francisella tularensis Live Vaccine Strain (LVS) and promotes LVS binding and infection of human monocyte-like THP-1 cells. The HB-19 pseudopeptide that binds specifically carboxy-terminal RGG domain of nucleolin inhibits LVS binding and infection of monocyte-like THP-1 cells. In a pull-down assay, elongation factor Tu (EF-Tu), a GTP-binding protein involved in protein translation, usually found in cytoplasm, was recovered among LVS bacterial membrane proteins bound on RGG domain of nucleolin. A specific polyclonal murine antibody was raised against recombinant LVS EF-Tu. By fluorescence and electron microscopy experiments, we found that a fraction of EF-Tu could be detected at the bacterial surface. Anti-EF-Tu antibodies reduced LVS binding to monocyte-like THP-1 cells and impaired infection, even in absence of complement and complement receptors. Interaction between EF-Tu and nucleolin was illustrated by two different pull-down assays using recombinant EF-Tu proteins and either RGG domain of nucleolin or cell solubilized nucleolin.

Discussion: Altogether, our results demonstrate that the interaction between surface nucleolin and its bacterial ligand EF-Tu plays an important role in Francisella tularensis adhesion and entry process and may therefore facilitate invasion of host tissues. Since phagosomal escape and intra-cytosolic multiplication of LVS in infected monocytes are very similar to those of human pathogenic F. tularensis ssp tularensis, the mechanism of entry into monocyte-like THP-1 cells, involving interaction between EF-Tu and nucleolin, might be similar in the two subspecies. Thus, the use of either nucleolin-specific pseudopeptide HB-19 or recombinant EF-Tu could provide attractive therapeutic approaches for modulating F. tularensis infection.

Figure 1

Down-regulation of cell surface nucleolin on THP-1 monocyte-like cell surface after LVS infection. A: THP-1 cells in suspension were incubated first with anti-nucleolin either polyclonal (part A1) or MAb (part A2), diluted 1/200 and then treated with PFA. In part A3, THP-1 cells were first treated by PFA and then incubated with polyclonal anti-nucleolin Ab. Analysis was performed by fluorescence microscopy at 63× magnification. B: FACS analysis of nucleolin expression on cell surface of uninfected cells (B1) or cells infected for 30 min by LVS (B2). THP-1 cells were incubated either with anti-GAPDH as isotype negative control (dotted line) or with anti-nucleolin (solid line) MAbs. These experiments are representative of five different experiments.

Figure 2

Interaction of nucleolin with LVS is involved in binding and infection of human monocyte-like THP-1 cells. A: THP-1 cells were incubated in RPMI, in absence or in presence of 5 μM HB-19 pseudopeptide or 9R control peptide. Opsonized LVS transformed with GFP expressing plasmid (LVS-GFP) were then added for 30 min. THP-1 cells were analyzed by fluorescence microscopy. Fifteen fields containing average of 100 cells were examined to quantify number of bacteria bound on human cells. Data show mean from three independent experiments ± SD values indicated as error bars. * = p < 0.01, indicates significant difference for HB-19-treated group compared to control group with 9R peptide. B: Intracellular replication of LVS. THP-1 cells were incubated with RPMI, 5 μM HB-19 or 9R before addition of opsonized LVS for 30 min. Cells were washed with RPMI containing gentamicin to kill extracellular bacteria and incubated in RPMI-FCS and gentamicin for indicated times. Quantification of intracellular bacteria was performed as described in Methods. Results show mean from three independent experiments, each performed in triplicate ± SD values indicated as error bars. C: THP-1 cells were incubated either with RPMI, 10 μM HB-19 alone or in presence of anti-CR3 Mab, 10 μM 9R or F3, or anti-CR3 Mab. Opsonized LVS were added for 30 min. Cells were washed with RPMI containing gentamicin and incubated in RPMI-FCS and gentamicin for 22 h. Quantification of intracellular bacteria was performed as described in Methods. Results show mean from five independent experiments, each performed in triplicate ± SD values indicated as error bars.

Figure 3

Expression of CR3 is not affected by HB-19. A. FACS analysis of CR3 expression on surface of THP-1 cells preincubated with 10 μM HB-19 (A1) or 10 μM 9R (A2). THP-1 cells were incubated either with anti-GAPDH as negative control (dotted line) or with anti-CR3 (solid line) MAbs. These experiments are representative of three different experiments. B. Binding of E-C3bi on THP-1 cells preincubated either with RPMI, 10 μM HB-19, 10 μM 9R or 20 μg anti-CR3 was determined by counting 100–200 cells by phase-contrast microscopy. Data show mean from three independent experiments ± SD values indicated as error bars.

Figure 4

Identification of bacterial ligands for nucleolin. Part A. Samples were prepared as outlined in Experimental Procedures, normalized to 50 μg proteins per lane before separation by SDS-PAGE. Immunoblotting (I.B.) was performed either with anti-EF-Tu Ab diluted 1/100,000 (part A1) or anti-IglA Ab diluted 1/5,000 (part A2). Part B. 500 μg LVS membrane proteins were incubated with different concentrations of biotinylated p63 peptide (p63*). Complexes, formed between bacteria proteins and biotinylated p63 peptide, were isolated by purification using avidin-agarose. Purified proteins were analyzed by SDS-PAGE. Immunoblotting (I.B.) was performed with anti-EF-Tu Ab diluted 1/100,000. Mobility of marker proteins is indicated on left side. Part C. Specific binding of EF-Tu present in LVS membranes with RGG domain of nucleolin. 500 μg LVS membrane proteins were preincubated without (lane -) or with 50 μM unlabeled p63 peptide, before incubation with 5 μM biotinylated p63 peptide (p63*). Complex formed between EF-Tu present in bacteria membrane proteins and biotinylated p63 peptide was analyzed by immunoblotting with anti-EF-Tu Ab diluted 1/100,000. In control, 50 μg LVS membrane proteins were run directly.

Figure 5

Recombinant EF-Tu specifically interacts with RGG carboxy-terminal domain of nucleolin. A. P63 peptide, corresponding to RGG domain of nucleolin binds in dose-dependent manner to recombinant EF-Tu. Aliquots containing 1 μg recombinant His-EF-Tu were incubated with different concentrations of biotin-labeled p63 peptide. Complex formed between His-EF-Tu and biotin-labeled p63 peptide was isolated by purification using avidin-agarose. Purified proteins were analyzed by SDS-PAGE. Lane S shows purified preparation of 0.2 μg His-EF-Tu. Presence of EF-Tu was detected by immunoblotting using anti-EF-Tu Ab diluted 1/100,000. B. Specific binding of recombinant EF-Tu with RGG domain of nucleolin. Aliquots containing 1 μg of His-EF-Tu were preincubated without (lane -) or with 40 or 80 μM unlabeled p63 peptide before further incubation with 4 μM biotin-labeled p63 peptide. Biotin-labeled p63 peptide (4 μM) was also preincubated with HB-19 (40 or 80 μM), 80 μM F3 or 80 μM 9Arg (9R) peptides before further incubation with 1 μg His-EF-Tu. Complexes formed between His-EF-Tu and biotin-labeled p63 peptide were isolated and analyzed by immunoblotting as in section A. On left is position of molecular weight protein markers. On right is position of 43 kDa His-EF-Tu. C: 5 μg GST (lane 2) or GST-EF-Tu (lane 3) bound on glutathione-Sepharose beads were incubated with THP-1 solubilized membrane proteins (500 μg). After extensive washes, bound proteins were run on SDS-PAGE with, in control, 50 μg proteins solubilized from THP-1 membranes, run directly without incubation with beads (lane 1). Immunoblotting was performed with anti-nucleolin MAb, diluted 1/10,000. Estimated molecular mass (in kDa) of nucleolin is indicated on left of black line.

Figure 6

Elongation factor Tu is localized on external side of F. tularensis LVS membranes. Part A. Fluorescence microscopy analysis of LVS pellet incubated either with NIS (from rabbit or from mice), anti-EF-Tu or anti-LVS Abs all diluted 1/10,000. Bright fields (part A1) or fluorescence microscopy (part A2) of 15 fields for each group were observed at 63× magnification. Images are representative of 3 different experiments. Part B. Electron micrographs showing attachment of gold spheres to surface of LVS incubated either with anti-EF-Tu or rabbit anti-LVS Abs and examined by nc-IEM. Control micrograph of LVS incubated with murine NIS shows no beads on bacteria cell surface. Electron microscopy was observed at 20,000× magnification. Size bars equal 500 nm.

Figure 7

Cell surface nucleolin co-localizes with LVS elongation factor Tu. THP-1 cells were incubated for 30 min with opsonized LVS and their interaction was observed by confocal microscopy at 63× magnification. Human cell surface was labeled with rabbit anti-nucleolin Ab diluted 1/200. Bacteria were labeled with murine anti-EF-Tu Ab diluted 1/2,000. Merging was observed with 3 × Zoom either with fluorescence light (right panel) or as bright field (left panel). Red arrows indicate colocalization of LVS with nucleolin present on cell surface. White arrows indicate LVS bound on cell surface in absence of nucleolin. These two photos are representative of five different experiments.

Figure 8

Elongation factor Tu of LVS participates in binding to human monocyte-like THP-1 cells and in their infection. A: THP-1 cells were infected for 30 min by opsonized LVS-GFP that had been pre-incubated either with RPMI, NIS or anti-EF-Tu Ab (diluted 1/2,000). THP-1 cells were also pre-incubated with 50 μg His-EF-Tu and then infected for 30 min by opsonized LVS-GFP pre-incubated with NIS. Fluorescence microscopy was analyzed at 63× magnification. Fifteen fields containing average of 100 cells were examined to quantify number of bound bacteria. Results shown are means ± SD from three independent experiments. *, p < 0.01 for group of bacteria treated with anti-EF-Tu Ab compared to group incubated with RPMI. B: Intracellular replication of LVS. THP-1 cells were infected for 30 min with opsonized LVS that were first incubated either with RPMI (■), mouse NIS (◆) or anti-EF-Tu Ab (●), both diluted 1/2,000. Cells were then washed with RPMI containing gentamicin and further incubated in RPMI-FCS and gentamicin for indicated times. Quantification of intracellular bacteria was performed as described in Methods. Results shown are means from five independent experiments, each in triplicate ± SD indicated by error bars. C: LVS were incubated either with Human serum, H.I. serum or with No Human serum. LVS in these different conditions were further incubated either with RPMI, mouse NIS or anti-EF-Tu Ab, both diluted 1/2,000. THP-1 cells were then infected by the different LVS preparations. THP-1 cells, pre-incubated with anti-CR3 MAb were infected with opsonized LVS that had been pretreated or not by anti-EF-Tu Ab. THP-1 cells were also preincubated with 50 μg His-EF-Tu before infection by opsonized LVS. Cells were washed in RPMI with gentamicin and incubated in RPMI-FCS and gentamicin for 22 h. Quantification of intracellular bacteria was performed as described in Methods. N.E. (not examined): anti-CR3 MAb, anti-CR3 MAb and anti-EF-Tu Ab and His-EF-Tu samples were not tested in H.I serum and No human serum conditions. Results shown are means from five independent experiments, each in triplicate ± SD indicated by error bars.