Central Nervous System-Infecting Pathogens Escherichia coli and Cryptococcus neoformans Exploit the Host Pdlim2 for Intracellular Traversal and Exocytosis in the Blood-Brain Barrier

Abstract

Microbial penetration of the blood-brain barrier, a prerequisite for the development of central nervous system (CNS) infection, involves microbial invasion, intracellular traversal, and exocytosis. Microbial invasion of the blood-brain barrier has been investigated, but the molecular basis for microbial traversal and exit from the blood-brain barrier remains unknown. We performed transcriptome analysis of human brain microvascular endothelial cells (HBMEC) infected with Escherichia coli and Cryptococcus neoformans, representative bacterial and fungal pathogens common in CNS infections. Among the targets upregulated in response to E. coli and C. neoformans infection, PDLIM2 was knocked down by small hairpin RNA (shRNA) in HBMEC for further investigation. We demonstrated that Pdlim2 specifically regulated microbial traversal and exit from HBMEC by assessing microbial invasion, transcytosis, intracellular multiplication, and egression. Additionally, the defective exocytosis of internalized E. coli cells from the PDLIM2 shRNA knockdown cells was restored by treatment with a calcium ionophore (ionomycin). Moreover, we performed proximity-dependent biotin labeling with the biotin ligase BioID2 and identified 210 potential Pdlim2 interactors. Among the nine Pdlim2 interactors enriched in response to both E. coli and C. neoformans infection, we selected MPRIP and showed that HBMEC with knockdown of MPRIP mimicked the phenotype of PDLIM2 knockdown cells. These results suggest that the CNS-infecting microbes hijack Pdlim2 and Mprip for intracellular traversal and exocytosis in the blood-brain barrier.

Keywords: BioID; Cryptococcus neoformans; Escherichia coli; blood-brain barrier; exocytosis; human brain microvascular endothelial cell.

FIG 1

Upstream regulator analysis of the infection-induced gene expression. Each regulator was predicted by Ingenuity Pathway Analysis (IPA) to be activated (red, Z-score > 1.5) or repressed (blue, Z-score < −1.5) during infection of HBMEC with either E. coli strain RS218 or C. neoformans strain H99. All the predicted pathways achieved a Z-score of ≥|1.5| in at least one time point for both organisms. White indicates no predicted activation or repression.

FIG 2

Pdlim2 regulated microbial exocytosis from HBMEC. (A) Colocalization of Pdlim2-GFP with F-actin, E. coli S242 mCherry, or C. neoformans KN99 mCherry. F-actin was labeled with Alexa Fluor 594 phalloidin (red). DNA was labeled with DAPI (blue). Colocalization is highlighted by the arrowheads. Scale bars, 10 μm. (B) Pdlim2 expression in control shRNA cells and PDLIM2 shRNA knockdown cells. The assay was performed in duplicate. The results from one of the duplicates are presented. Student’s t test was performed to determine statistical significance. P = 0.016. (C) Relative invasion frequencies of E. coli strains S242, RS218, N603, and F32. The results were calculated as the percentages of the initial inoculum and are presented as the relative percentages of invasion compared with the invasion in the control shRNA cells. Student’s t test was performed to determine statistical significance. (D) Rescue experiment of PDLIM2 shRNA knockdown cells. The results were calculated as the percentages of the initial inoculum and are presented as the relative percentages of invasion compared with the invasion in the control shRNA cells. One-way analysis of variance (ANOVA) and multiple-comparison tests were performed to determine statistical significance. (E) Relative transcytosis frequencies of E. coli S242 and C. neoformans. The results were calculated as the percentages of the initial inoculum and are presented as the relative percentages of transcytosis compared with the transcytosis in the control shRNA cells. Student’s t test was performed to determine statistical significance. (F) Intracellular multiplication of E. coli S242 in control shRNA cells. ANOVA and multiple-comparison tests were performed to determine the statistical significance. (G) Intracellular multiplication of E. coli S242 in PDLIM2 shRNA knockdown cells. ANOVA and multiple-comparison tests were performed to determine statistical significance. (H) Egression of E. coli S242 from HBMEC. 0 h represents the initial internalized E. coli CFU. Student’s t test was performed to determine statistical significance. (H) Egression of E. coli S242 from HBMEC in the presence of ionomycin for 4 h. Student’s t test was performed to determine statistical significance. Each assay was performed in triplicate. Error bars indicate mean values ± standard deviations (SD). *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., no significant difference.

FIG 3

BioID revealed host proteins interacting with Pdlim2 in response to microbial infection. (A) Pdlim2 was fused with BioID2-HA and stably expressed in HBMEC. LTR, long terminal repeat; RRE, Rev response element; cPPT, central polypurine tract; P2A, 2A self-cleaving peptides; WPRE, woodchuck hepatitis virus posttranscriptional regulatory element. (B) The proteins biotinylated by Pdlim2-BioID2 were detected with HRP-conjugated streptavidin after SDS-PAGE separation. The expression of Pdlim2-BioID2 in HBMEC in the presence of biotin led to a drastic increase in biotinylation of the endogenous proteins (top). Fusion protein Pdlim2-BioID2-HA was detected with anti-HA antibody (bottom). (C) Pdlim2-BioID2-HA colocalized with F-actin. Pdlim2-BioID2-HA was labeled with anti-HA antibody (green). F-actin was labeled with Alexa Fluor 594 phalloidin (red). DNA was labeled with DAPI (blue). Colocalization is highlighted by the arrowhead. Scale bars, 10 μm. (D) Biotinylated proteins enriched in HBMEC infected with E. coli. Gray lines, BioID interactions; black lines, known protein interactions from the String networks; node sizes, normalized log₂ fold enrichment. Proteins enriched in response to both E. coli and C. neoformans are highlighted in red. (E) Biotinylated proteins enriched in HBMEC infected with C. neoformans. Gray lines, BioID interactions; black lines, known protein interactions from the String networks; node sizes, normalized log₂ fold enrichment. Proteins enriched in response to both E. coli and C. neoformans are highlighted in red.

FIG 4

Mprip was involved in intracellular traversal and exocytosis in HBMEC. (A) Colocalization of Mprip with F-actin, Pdlim2, E. coli S242 mCherry, or C. neoformans KN99 mCherry. Mprip-HA was labeled with anti-HA antibody (green). F-actin was labeled with Alexa Fluor 594 phalloidin (red). Mprip was labeled with anti-MPRIP polyclonal antibody (red). DNA was labeled with DAPI (blue). Colocalization is highlighted by the arrowheads. Scale bars, 10 μm. (B) Mprip expression in control shRNA cells and MPRIP shRNA knockdown cells. The assay was performed in duplicate. The results for one of the duplicates are presented. Student’s t test was performed to determine statistical significance. P = 0.048. (C) Relative invasion frequencies in MPRIP shRNA knockdown and MPRIP rescue cells. The results were calculated as the percentages of the initial inoculum and are presented as the relative percentages of invasion compared with the invasion in the control shRNA cells. ANOVA and multiple-comparison tests were performed to determine statistical significance. (D) Relative transcytosis frequencies in MPRIP shRNA knockdown cells. The results were calculated as the percentages of the initial inoculum and are presented as the relative percentages of transcytosis compared with the transcytosis in the control shRNA cells. Student’s t test was performed to determine statistical significance. Error bars indicate mean values ± SD. Each assay was performed in triplicate. *, P < 0.05.