Nucleoside-Diphosphate-Kinase of P. gingivalis is Secreted from Epithelial Cells In the Absence of a Leader Sequence Through a Pannexin-1 Interactome

Abstract

Nucleoside-diphosphate-kinases (NDKs) are leaderless, multifunctional enzymes. The mode(s) of NDK secretion is currently undefined, while extracellular translocation of bacterial NDKs is critical for avoidance of host pathogen clearance by opportunistic pathogens such as Porphyromonas gingivalis. P. gingivalis-NDK during infection inhibits extracellular-ATP (eATP)/P2X₇-receptor mediated cell death in gingival epithelial cells (GECs) via eATP hydrolysis. Furthermore, depletion of pannexin-1-hemichannel (PNX1) coupled with P2X₇-receptor blocks the infection-induced eATP release in GECs, and P. gingivalis-NDK impacts this pathway. Ultrastructural and confocal microscopy of P. gingivalis-co-cultured GECs or green-fluorescent-protein (GFP)-P. gingivalis-NDK transfected GECs revealed a perinuclear/cytoplasmic localization of NDK. eATP stimulation induced NDK recruitment to the cell periphery. Depletion of PNX1 by siRNA or inhibition by probenecid resulted in significant blocking of extracellular NDK activity and secretion using ATPase and ELISA assays. Co-immunoprecipitation-coupled Mass-spectrometry method revealed association of P. gingivalis-NDK to the myosin-9 motor molecule. Interestingly, inhibition of myosin-9, actin, and lipid-rafts, shown to be involved in PNX1-hemichannel function, resulted in marked intracellular accumulation of NDK and decreased NDK secretion from infected GECs. These results elucidate for the first time PNX1-hemichannels as potentially main extracellular translocation pathway for NDKs from an intracellular pathogen, suggesting that PNX1-hemichannels may represent a therapeutic target for chronic opportunistic infections.

Figure 1. Nucleoside diphosphate kinase (NDK) secretion from P. gingivalis (P.g.) is shown both in the cytoplasm of the infected host cells and on the surface of P. gingivalis bacteria at 12 hours after infection in gingival epithelial cells (GECs).

NDK was visualized by transmission electron microscopy using immunogold labelling and rabbit anti-P. gingivalis NDK antibody. NDK (blue arrows) is seen on the P. gingivalis bacterial surface and in the host cytoplasm independently of the bacteria (A,B). An enlarged image of the boxed area is shown to the right (A). P. gingivalis with no primary antibody incubation (C), ndk-deficient mutant strain, ΔNDK (D) and GECs without infection (E) with both primary and secondary antibody incubations were used as controls. The black arrows point to non-specific background level of gold labelling staining in the control samples. Bar represents 1 μm.

Figure 2. Immunofluorescence analysis of P. gingivalis-NDK cellular localization in GECs.

(A) 12 h of P. gingivalis infection; (B) transfection with GFP-linked P. gingivalis-NDK. (C) 3 mM ATP stimulation of 12 h P. gingivalis-infected GECs; P. gingivalis-NDK in infected cells (A,C) was detected using rabbit anti-P. gingivalis NDK antibody and visualized with anti-rabbit AlexaFluor488 secondary antibody; (D) 3 mM ATP of GECs transfected with GFP-linked P. gingivalis-NDK (E) No-infection, no-transfection control cells showed no unspecific staining of NDK. (F) Transfection with GFP-containing plasmid, without P. gingivalis-NDK insertion (empty vector), showed an unspecific uniform GFP expression throughout the transfected cells as expected. These are representative images of at least three separate experiments performed in duplicates. All images represent maximum image projections of z-sections performed on laser confocal microscope. Bars represent 15 μm.

Figure 3. Detected NDK enzyme ATPase activity in cell culture media of P. gingivalis-infected GECs in the presence or absence of PNX1 or lipid raft inhibition.

(A) A representative image of the expression of the 45 kDa PNX1 protein in PNX1 knock-down GECs at 48 h post transfection, as detected by Western blot assay; β-tubulin (51 kDa) was used as loading control; Full-length blots are presented in Supplementary Figure 2; (B) GECs were infected with P. gingivalis and treated with inhibitors of PNX1 (probenecid), Myosin-9 (ML9), an actin cytoskeleton inhibitor (cytochalasin D), or a lipid raft inhibitor (MβCD), or were transfected with PNX1 siRNA and infected with P. gingivalis in combination with treatment with ML9 or cytochalasin D. ndk-deficient mutant strain of P. gingivalis was used as control. All data represent an average of at least three separate experiments. P-values were calculated using a two-tail student t-test. ** Represent P-values < 0.001.

Figure 4. Confocal Fluorescence analysis of NDK and Myosin-9 co-localization in P. gingivalis-infected primary GECs (Top) at 6 h post infection, or (Middle) at 24 h post infection with P. gingivalis.

Uninfected GECs were used as staining controls (Bottom). Myosin-9 was labelled in red by rabbit anti-human Myosin-9 monoclonal antibody, visualized with anti-rabbit AlexaFluor594 secondary antibody. P. gingivalis-NDK is labelled in green, detected by monoclonal P. gingivalis-NDK specific antibody, visualized with anti-mouse AlexaFluor488 secondary antibody. Cell nuclei were visualized with DAPI. Mander’s co-localization (yellow) coefficients for the two molecules ranged from 0.465 at 6 h to 0.791 at 24 h. Bars represent 5 μm.

Figure 5. NDK secretion detected by ELISA in cell culture media of GECs infected with P. gingivalis.

GECs were infected with P. gingivalis and treated with inhibitors of PNX1 (probenecid), Myosin-9 (ML9), an actin cytoskeleton inhibitor (cytochalasin D), or a lipid raft inhibitor (MβCD), or were transfected with PNX1 siRNA and infected with P. gingivalis in combination with treatment with ML9 or cytochalasin D. All data represent an average of at least three separate experiments. P-values were calculated using a two-tail Student t-test. ** and * represent P-values < 0.001 and P < 0.05 respectively.

Figure 6. Immunofluorescence analysis of NDK intracellular accumulation.

(A) P. gingivalis-infected primary GECs after treatment with probenecid, MβCD, ML9 or cytochalasin D. Uninfected, untreated cells were used as a control; (B) PNX1 siRNA-transfected P. gingivalis-infected primary GECs with or without MβCD, ML9 or cytochalasin D treatment. Non-target siRNA-transfected P. gingivalis-infected GECs were used as a control. P. gingivalis-NDK is labelled in green, detected by monoclonal P. gingivalis-NDK specific antibody, visualized with anti-rabbit AlexaFluor488 secondary antibody. Bar represents 1 μm. (C) NIH ImageJ analysis was performed for measuring cell fluorescent intensity. Cell boundaries were determined by actin labelling with phalloidin-TRITC. Corrected total cell fluorescence was calculated and measurements were normalized to the mean intensity of the uninfected, untreated cells. All data represent an average of at least three separate experiments. ** Denote P-values < 0.001. Select exact P-values are also shown.

Figure 7. A proposed pathway of NDK secretion from infected host cells.

Upon infection with P. gingivalis, an initial activation of ATP release from the cell arises, and the P2X₇-receptor/Pannexin-1 (PNX1)-hemichannel is activated. P. gingivalis-NDK is accumulated in the cytoplasm and mostly in the perinuclear area. The accumulated P. gingivalis-NDK is activated upon ATP release, which can act as an autocrine danger signal to the host by stimulating the P2X₇-receptor/PNX1-hemichannel. Following activation, P. gingivalis-NDK interacts Myosin-9 motor molecule and is trafficked along the Myosin-9 and actin filaments to the cell periphery. NDK translocates to the extracellular space through the forming P2X₇-receptor/PNX1 channel. Secreted NDK hydrolyzes the danger-signal eATP, thus attenuating the stimulation of the P2X₇-receptor/PNX1-hemichannel complex, and the downregulating the associated downstream events, such as reactive oxygen species production and intracellular bacterial killing.