Group A Streptococcus exploits human plasminogen for bacterial translocation across epithelial barrier via tricellular tight junctions

Abstract

Group A Streptococcus (GAS) is a human-specific pathogen responsible for local suppurative and life-threatening invasive systemic diseases. Interaction of GAS with human plasminogen (PLG) is a salient characteristic for promoting their systemic dissemination. In the present study, a serotype M28 strain was found predominantly localized in tricellular tight junctions of epithelial cells cultured in the presence of PLG. Several lines of evidence indicated that interaction of PLG with tricellulin, a major component of tricellular tight junctions, is crucial for bacterial localization. A site-directed mutagenesis approach revealed that lysine residues at positions 217 and 252 within the extracellular loop of tricellulin play important roles in PLG-binding activity. Additionally, we demonstrated that PLG functions as a molecular bridge between tricellulin and streptococcal surface enolase (SEN). The wild type strain efficiently translocated across the epithelial monolayer, accompanied by cleavage of transmembrane junctional proteins. In contrast, amino acid substitutions in the PLG-binding motif of SEN markedly compromised those activities. Notably, the interaction of PLG with SEN was dependent on PLG species specificity, which influenced the efficiency of bacterial penetration. Our findings provide insight into the mechanism by which GAS exploits host PLG for acceleration of bacterial invasion into deeper tissues via tricellular tight junctions.

Figure 1. PLG is associated with bacterial localization in tTJs.

(a) Caco-2 cells were infected with an EGFP-expressing GAS NIH35 strain as green images at an MOI of 10 for 2 h in the presence or absence of 2 μM human PLG. Tricellulin was labeled with anti-tricellulin and Alexa Fluor 594-conjugated antibodies as red images, whereas ZO-1 was labeled with anti-ZO-1 and Alexa Fluor 647-conjugated antibodies as blue images. GAS-infected cells were analyzed using a confocal laser-microscope. White arrows indicate bacterial association with tricellulin. Data shown are representatives of at least three separate experiments. (b) Bacterial localization in tTJs was assessed as described in the Methods section. Data obtained from ten fields of view (x630) are presented as the mean ± S.D. *P < 0.01.

Figure 2. Tricellulin is a determinant for bacterial localization in tTJs.

(a) Caco-2 cells were transfected with tricellulin-targeted siRNA. A scrambled siRNA was utilized as a negative control. At 72 h after transfection, whole cell lysates were subjected to Western blot analysis using anti-tricellulin antibody. β-actin served as a loading control. (b) Tricellulin knockdown or control cells were infected with EGFP-expressing GAS strain at an MOI of 10 for 2 h in the presence of 2 μM human PLG. ZO-1 and tricellulin were immunostained and are shown as blue and red images, respectively. GAS-infected cells were analyzed using a confocal laser-microscope. Data shown are representatives of at least three separate experiments. (c) Bacterial localization in tTJs, bTJs, or other regions was assessed as described in the Methods section. Data obtained from ten fields of view (x630) are presented as the mean ± S.D. *P < 0.01.

Figure 3. PLG functions as molecular bridge between SEN and tricellulin.

TRIC proteins (a) or TRIC-EC1 variants (b) were immobilized on microtiter plates, then increasing amounts of SEN were added in the absence or presence of 1 μM human PLG. Bound SEN was detected using an anti-SEN antibody. All experiments were performed in sextuplet with three technical repeats. Data are shown as the mean ± S.D. of six wells from a representative experiment. *P < 0.01.

Figure 4. Introduction of sorting signal into SEN affects subcellular localization and bacterilal PLG binding.

(a) Schematic diagram of SEN protein. Internal and C-terminus PLG-binding motifs are shown as green and pink images, respectively. The amino acid sequences of the PLG-binding motif in each strain are described below. (b) NIH35 and isogenic sen mutants were grown in THY broth. The culture densities were measured at 37 °C. (c) The strains were grown to an OD₆₀₀ of 0.8 in THY broth, then each fraction was prepared, as described in Methods. (d) NIH35 and the isogenic sen mutants (OD₆₀₀ = 0.6) were bound to microtiter plates, and bound cells were incubated with 1 μM human PLG. Cell-bound PLG was detected by ELISA using an anti-PLG antibody. Data are shown as the mean ± S.D. of six samples from a representative experiment. *P < 0.05; **P < 0.01.

Figure 5. Bacterial translocation is mediated by interaction of SEN with human PLG.

(a) Caco-2 cells were grown using a Millicell filter system, and then infected with the NIH35 strain or sen mutants at an MOI of 10 for 2 h. After removing non-adherent bacteria, the ability of the GAS strains to translocate across epithelial cells at 8 h after infection in the presence of bovine PLG was assessed by examining medium samples obtained from the lower chambers. Data are shown as the mean ± S.D. of six wells from a representative experiment. *P < 0.01. (b) Human and bovine PLG were immobilized on a nitrocellulose membrane after serial two-fold dilutions. After blocking of the membrane, 40 μg of SEN was overlaid and the binding signal was detected with an anti-SEN antibody. (c) TRIC-EC1 protein was immobilized on microtiter plates, and increasing amounts of SEN were reacted in the presence of 1 μM human PLG or 1 μM bovine PLG. Bound SEN was detected using an anti-SEN antibody. Data are shown as the mean ± S.D. from three independent experiments. *P < 0.01. (d) The effect of human PLG on bacterial translocation was analyzed as described in (a). Data are shown as the mean ± S.D. of six wells from a representative experiment. *P < 0.01.

Figure 6. SEN is involved in GAS-induced disruption of intercellular junctions.

(a) Effects of bovine or human PLG on reduction of TER in cells infected with GAS strains for 8 h. The TER value of the non-infected cells was set to 100%. All experiments were performed in sextuplet with three technical repeats. Data are shown as the mean ± S.D. of six wells from a representative experiment. *P < 0.01. (b) Caco-2 cells (left panel) or Detroit 562 cells (right panel) were infected with NIH35 or sen mutants at an MOI of 10 for 7 h in the presence of 1 μM human PLG. Cleavage of ZO-1, occludin, tricellulin, and E-cadherin was detected in whole cell lysates by Western blot analysis. β-actin served as a loading control.