Fusobacterium nucleatum adhesin FadA binds vascular endothelial cadherin and alters endothelial integrity

Abstract

Fusobacterium nucleatum is a Gram-negative oral anaerobe, capable of systemic dissemination causing infections and abscesses, often in mixed-species, at different body sites. We have shown previously that F. nucleatum adheres to and invades host epithelial and endothelial cells via a novel FadA adhesin. In this study, vascular endothelial (VE)-cadherin, a member of the cadherin family and a cell-cell junction molecule, was identified as the endothelial receptor for FadA, required for F. nucleatum binding to the cells. FadA colocalized with VE-cadherin on endothelial cells, causing relocation of VE-cadherin away from the cell-cell junctions. As a result, the endothelial permeability was increased, allowing the bacteria to cross the endothelium through loosened junctions. This crossing mechanism may explain why the organism is able to disseminate systemically to colonize in different body sites and even overcome the placental and blood-brain barriers. Co-incubation of F. nucleatum and Escherichia coli enhanced penetration of the endothelial cells by the latter in the transwell assays, suggesting F. nucleatum may serve as an 'enabler' for other microorganisms to spread systemically. This may explain why F. nucleatum is often found in mixed infections. This study reveals a possible novel dissemination mechanism utilized by pathogens.

© 2011 Blackwell Publishing Ltd.

Figures

Figure 1. Diagram of VE-cadherin structure and main interacting proteins

VE-cadherin is a cell surface exposed protein composed of 5 extracellular domains (EC), a unique transmembrane domain, and an intracellular domain interacting with catenins among which are p120 and β-catenin (β-cat). The cadherin-catenin complex is involved in the maintenance of cell-cell adherent junction and barrier integrity of the endothelium. The region bound by FadA identified by Y2H is a sequence of 119 residues composed of the 2nd half of EC4 and the 1st half of EC5.

Figure 2. Binding of FadAc to VE-cadherin415–534

(A) Purified FadAc mixed with GST or GST-VE-cadherin415–534 proteins were loaded onto the glutathione column. Following elution, the samples (2 μg/lane) were loaded onto 12% SDS-PAGE, followed by Coomassie blue staining (top panel) or Western blot using anti-FadA antibodies (bottom panel). (B) Co-precipitation assay was performed by mixing GST or GST-VE-cadherin415–534 with E. coli lysate expressing FadAc, mFadA, or the FadA mutant G4, respectively. The GST complexes were captured with equilibrated GST-bind resin, and eluted with 4x Laemmli buffer. One-fifth of the elution was subjected to 12% SDS-PAGE followed by Coomassie blue staining (top panel) and Western blot using anti-FadA antibodies (bottom panel). (C) Representative fluorescence emission spectra of dansylated-FadAc (4 μM) alone, or in the presence of 20 μM GST or 20 μM GST-VE-cadherin415–534. (D) Increased fluorescence intensity due to the presence of VE-cadherin415–534 (F) relative to that of dansylated-FadAc alone (F0) was plotted against the concentration of VE-cadherin415–534 added, where F = Fv − Fg, with Fv being the fluorescent intensity emitted in the presence of GST-VE-cadherin415–534 and Fg being that emitted in the presence of GST. The data were fit to the Boltzmann sigmoidal equation and the dissociation constant was calculated as described in the experimental procedures.

Figure 3. Co-localization of FadAc and VE-cadherin by double immunofluorescent staining and confocal microscopy

HUVEC monolayers were blocked in growth medium supplemented with 1% BSA and incubated for 1 hr at 37°C, followed by 2 hr incubation with BSA (A), Alexa-Fluor 488-conjugated cytochrome c (B), or Alexa-Fluor 488-conjugated FadAc (green) (C). Cells were washed and permeabilized, followed by incubation with mouse monoclonal antibody anti-VE-cadherin (red) and Alexa-Fluor-594-conjugated chicken-anti-rabbit polyclonal antibodies and counterstained with DAPI (blue). The cells were observed using a laser scanning confocal microscope. The yellow color results from the merge of red and green, indicating co-localization of VE-cadherin and FadAc.

Figure 4. VE-cadherin is required for F. nucleatum to attach to HUVEC

(A) Western blot of HUVEC cells transfected with 100 nM anti-CDH5 siRNA pool or non-targeting siRNA Pool#2 from Dharmacon using rabbit polyclonal anti-VE-cadherin antibodies. Approximately 5×103 cells were loaded onto each lane. The blot shown is a representative of three independent experiments. (B) Attachment of F. nucleatum 12230 to HUVEC cells treated in (A). Expressed are relative attachment levels with that exhibited by F. nucleatum 12230 to untreated cells designated as 100%. Data are the means ± SD from three independent experiments performed at least in triplicate. *** p<0.001 based on t tests.

Figure 5. Dose-dependent inhibition of F. nucleatum attachment and invasion by FadAc and VE-cadherin

The proteins pre-incubated with HUVEC were indicated below the x axis. The level of F. nucleatum 12230 attachment (A) or invasion (B) was expressed as the percentage of bacteria recovered following cell lysis relative to the total number of bacteria initially added. Data are the means ± SD from two independent experiments each performed in triplicates. *** p<0.001 based on t tests.

Figure 6. FadA increases the permeability of the endothelial barrier

A–C. Confluent HUVEC monolayers in transwells were incubated with 50 μg/ml FadAc (■), 0.5 UI/mL Thrombin (◇), or untreated (×) (A); or with increasing concentrations of FadAc (B); or with 50 μg/ml FadAc in combination with varying concentrations of GST or GST-VE-cadherin415–534, or untreated (C). Cell permeability was assessed by adding Texas Red-conjugated dextran to the upper chamber (final concentration = 1 mg/ml). At indicated times (A), or after 30 min (B and C), aliquots from the lower chamber were taken and the fluorescence was measured. D. Cells were incubated with F. nucleatum 12230 (■), fadA-deletion mutant US1 (▲), fadA-complementing strain USF81(×), or E coli DH5α (□), each at an MOI=100, followed by the addition of fluorescent dextran. The permeability assay was performed as described above. E. E. coli was added to the transwell inserts with or without HUVEC, or in combination with F. nucleatum 12230, US1, or USF81. At the indicated times, aliquots were taken from the lower chamber and plated on LB agar plates followed by incubation in air. The rate of E. coli penetration was calculated as percent of bacteria recovered from the lower chamber over the total bacteria added to the upper chamber. Data are the means ± SD from at least three independent experiments performed in duplicates or triplicates. * p<0.05 and *** p<0.001 based on t test (A to D) and ANOVA test followed by Bonferroni’s post test (E).

Figure 7. HUVEC viability assay

HUVEC were incubated with 4% formaldehyde (A), untreated (B), incubated with 1 mg/ml of FadAc (C) or F. nucleatum 12230 at an MOI of 1000:1 (D) for 2.5 hours, and stained with Trypan Blue. Cells were observed immediately under a light microscope immediately. A representative field for each condition is shown. The dead cells were stained blue (A) while the viable cells remained unstained (B–D). FadA adhesin (green) from Fusobacteirum nucleatum co-localizes with VE-cadherin (red) on the endothelial cells, causing VE-cadherin to relocate from cell-cell junctions (top two rows) to intracellular compartments (bottom row).