Neuraminidase of Influenza A Virus Binds Lysosome-Associated Membrane Proteins Directly and Induces Lysosome Rupture

Abstract

As a recycling center, lysosomes are filled with numerous acid hydrolase enzymes that break down waste materials and invading pathogens. Recently, lysosomal cell death has been defined as "lysosomal membrane permeabilization and the consequent leakage of lysosome contents into cytosol." Here, we show that the neuraminidase (NA) of H5N1 influenza A virus markedly deglycosylates and degrades lysosome-associated membrane proteins (LAMPs; the most abundant membrane proteins of lysosome), which induces lysosomal rupture, and finally leads to cell death of alveolar epithelial carcinoma A549 cells and human tracheal epithelial cells. The NA inhibitors peramivir and zanamivir could effectively block the deglycosylation of LAMPs, inhibit the virus cell entry, and prevent cell death induced by the H5N1 influenza virus. The NA of seasonal H1N1 virus, however, does not share these characteristics. Our findings not only reveal a novel role of NA in the early stage of the H5N1 influenza virus life cycle but also elucidate the molecular mechanism of lysosomal rupture crucial for influenza virus induced cell death.

Importance: The integrity of lysosomes is vital for maintaining cell homeostasis, cellular defense and clearance of invading pathogens. This study shows that the H5N1 influenza virus could induce lysosomal rupture through deglycosylating lysosome-associated membrane proteins (LAMPs) mediated by the neuraminidase activity of NA protein. NA inhibitors such as peramivir and zanamivir could inhibit the deglycosylation of LAMPs and protect lysosomes, which also further interferes with the H5N1 influenza virus infection at early stage of life cycle. This work is significant because it presents new concepts for NA's function, as well as for influenza inhibitors' mechanism of action, and could partially explain the high mortality and high viral load after H5N1 virus infection in human beings and why NA inhibitors have more potent therapeutic effects for lethal avian influenza virus infections at early stage.

FIG 1

Lysosomes were decreased significantly upon H5N1 influenza virus infection. (A) MTS assay of cell viability of A549 cells and primary HTEpiC treated with vehicle or H1N1 or H5N1 virus at 48 h postinfection. (B) q-PCR analysis of influenza viral gene M in A549 cells infected with H1N1 or H5N1 virus. (C) A549 cells were infected with either seasonal H1N1 or H5N1 influenza virus or treated with the vehicle as a control. At 24 h postinfection, the cells were stained by LysoTracker or immunolabeled for peroxisomal membrane protein PMP70 and α-tubulin, respectively. Images were obtained using confocal microscopy, and the fluorescence signal of each cell shown on the right was estimated by examining more than 100 cells. (D) Confocal microscopy determination of changes of lysosomes in primary HTEpiC that were infected with either seasonal H1N1 or H5N1 influenza virus or treated with the vehicle allantoic fluid (AF) as a control at 24 h postinfection (h.p.i.). (E) FACS analysis of A549 cells and HTEpiC after staining with LysoTracker at 24 h after seasonal H1N1 or H5N1 infection or treatment with vehicle, with unstained cells as the negative control. (F) Pearson correlation coefficients between cell viability at 48 h postinfection with H5N1 virus and the relative lysosome numbers measured by FACS using LysoTracker before infection in A549, HeLa, MRC-5, H1650, and HEK293T cells. All scale bars indicate 50 μm, and data represent means ± the SEM of three independent experiments. *, P < 0.05; **, P < 0.01.

FIG 2

Lysosomes were ruptured at a late stage of H5N1 influenza virus infection. (A) Representative images of LAMP1 (green) and LAMP2 (red) in A549 cells infected with the seasonal H1N1 virus or H5N1 virus or treated with AF at 24 h postinfection. The fluorescence signal of each cell analyzed from 50 to 100 cells is shown on the right. (B) Confocal microscopy of the cathepsin D and LAMP1 after immunolabeling in A549 cells infected with seasonal H1N1 virus or H5N1 virus or treated with vehicle at 24 h postinfection. The colocalization rate of these two proteins was analyzed for 50 to 100 cells using Fluoview software. All scale bars indicate 50 μm, and the data represent means ± the SEM of three independent experiments. *, P < 0.05; **, P < 0.01.

FIG 3

LAMP1 and LAMP2 were significantly deglycosylated after H5N1 influenza virus infection. (A) Immunoblot analysis of LAMP1 and LAMP2 in A549 cells infected with seasonal H1N1 or H5N1 virus at the indicated times (in hours) postinfection. β-Actin was used as a control. (B) Immunoblot analysis of LAMP1 and LAMP2 in 293T cells infected with H1N1 or H5N1 virus or treated with AF at 48 h postinfection. β-Actin was used as a loading control. (C) Immunoblot analysis of LAMP1 and LAMP2 in A549 cells infected with seasonal H1N1 or H5N1 virus in the presence of actinomycin D (ACD) or cycloheximide (CHX) at 24 h postinfection. (D) Immunoblot analysis of highly glycosylated LAMP1 and those that were digested with the glycosidase PNGase F in A549 cells infected with H1N1 or H5N1 virus at the indicated time points. The asterisk (*) indicates an unspecific band. (E) Immunoblot analysis of TFEB, M6PR, cathepsin B, cathepsin D, cathepsin K, cathepsin L, ATPase, H⁺ transporting, lysosomal 70-kDa, and V1 subunit A (ATP6V1A) in A549 cells at the indicated time points in hours postinfection (h.p.i.) with H1N1 or H5N1 virus.

FIG 4

Influenza virus NA could bind to the LAMPs directly and induce the deglycosylation of LAMPs. (A) Plasmids encoding H5N1 influenza viral proteins (PB1, PB2, PA, NP, HA, NA, M1, M2, NS1, and NS2) (for the abbreviation definitions, see Materials and Methods) were separately transfected in 293T cells, and LAMP1 and LAMP2 were detected by immunoblot analysis. The expression of each viral protein was detected using flag antibody. (B) Immunoblot analysis of LAMP1 or LAMP2 in A549 cells transfected with control vector or the NA-flag plasmid. (C) Immunoblot analysis of glycosylated LAMP1 and those digested with glycosidase PNGase F in 293T cells transfected with plasmids encoding NA of H5N1 or vector. An asterisk (*) indicates an unspecific band. (D) FACS analysis of lysosomes stained with LysoTracker in 293T cells transfected with NA of the H5N1 virus or vector. (E) MTS assay of cell viability of A549 cells transfected with control vector or NA(H5N1) gene at 48 h posttransfection. (F) Immunoblot analysis of LAMP1 and LAMP2 in A549 cells treated with NA protein of H1N1 or H5N1 virus which was expressed and purified from 293T cells at the indicated amounts. β-Actin was used as a control. (G) Confocal microscopy analysis showing the interaction between viral NA and the lysosomal membrane marker LAMP1 in A549 cells at 3 h after H1N1 or H5N1 virus infection or treated with the vehicle as a control. Scale bars, 50 μm. (H) 293T cells were transfected with vector, NS1-flag, or NA-flag encoding plasmids, and an immunoprecipitation (IP) assay was performed with anti-flag or control IgG. Cell lysates (input) and immunoprecipitated complexes were analyzed by immunoblotting with anti-LAMP1, anti-LAMP2, anti-NS1-flag, anti-NA-flag, or anti-β-actin antibodies. (I) MTS assay of the cell viability of A549 cells, which were transfected with nontarget control siRNA (NC) or LAMP1- or LAMP2-specific siRNA and then treated with vehicle, H1N1 virus, or H5N1 virus. The knockdown efficiency is shown on the right. (J) q-PCR analysis of influenza viral M gene in A549 cells after transfection of LAMP1, LAMP2, or nontarget control (NC) siRNA and infected with H5N1 virus. (K) MTS assay of the viability of A549 cells transfected with vector control, LAMP1, or LAMP2 and then infected with H5N1 virus. The overexpression efficiency is shown on the right. All scale bars indicate 50 μm, and data represent means ± the SEM of three independent experiments. *, P < 0.05; **, P < 0.01.

FIG 5

Influenza virus NA induced the LAMP deglycosylation depending on the its NA activity and low-pH environment. (A) Immunofluorescent colocalization of MAL I-labeled sialic acid and lysosomal membrane marker LAMP2 in A549 cells infected with H1N1 or H5N1 virus or treated with vehicle as a control. The sialic acid fluorescence signal colocalized with LAMP2 per cell (on the right) was estimated by examining 50 to 100 cells. Scale bars, 50 μm. (B) Immunoblot analysis of LAMP1 and LAMP2 deglycosylation in 293T cells transfected with plasmids encoding vector, wild-type (WT) NA, or mutant NAs using NA-flag as an overexpression control and β-actin as a loading control. (C) 293T cells were transfected with NS1-flag, wild-type NA-flag, or mutant NA (E258G) encoding plasmids, respectively, and the immunoprecipitation (IP) assay was performed with anti-flag antibody. Cell lysates (input) and immunoprecipitated complexes were analyzed by immunoblotting with anti-LAMP2 and anti-flag antibodies. (D) 293T cells were transfected with control plasmid or NA(H5N1)-encoding plasmids. After 6 h, peramivir (30 μg/ml), zanamivir (30 μg/ml), or PBS was separately added into the media of these cultured cells. The deglycosylation of LAMP1 and LAMP2 was analyzed by immunoblotting, and flag was used for NA detection and β-actin as a loading control. (E) Immunoblot analysis of LAMP1 and LAMP2 deglycosylation in A549 cells infected with H5N1 virus and treated with PBS, peramivir (30 μg/ml), or zanamivir (30 μg/ml), respectively, at the indicated time points. (F) FACS analysis of the lysosome numbers in A549 cells treated with AF and in cells infected with H5N1 virus and treated with peramivir (30 μg/ml), zanamivir (30 μg/ml), or PBS. (G) Assay of NA activity of H1N1 or H5N1 influenza virus or vehicle control AF in different buffers (pH 4.0, 5.0, 6.0, or 7.4) at 37°C. RFU, relative fluorescence units. (H) Activity assay evaluating 0.1 μg of H1N1 or H5N1 influenza virus NA protein in different buffers (pH 4.0, 5.0, 6.0, or 7.4) at 37°C. The amounts of NA protein were also determined by immunoblot analysis (shown on the right). (I) The extracellular region of LAMP2 fused to the human IgG-Fc fragment (LAMP2-Fc) and the control protein human IgG-Fc (Fc) were conjugated to protein A-beads. After incubation with vehicle or H1N1 or H5N1 virus in different buffers (pH 4.0, 5.0, 6.0, or 7.4) at 37°C for 6 h, protein changes were detected by immunoblotting with anti-LAMP2 antibody or Fc antibody as a control. (J) 293T cells were transfected with NA(H5N1) gene and treated with control PBS, chloroquine (CQ), bafilomycin A1 (BfaA1), chlorpromazine (CPZ), methyl-β-cyclodextrin (MβCD), or NH₄Cl, and then, 48 h later, LAMP1 and LAMP2 were detected by immunoblotting with NA-flag as an overexpression control and β-actin as a loading control. All data represent means ± the SEM of three independent experiments. *, P < 0.05; **, P < 0.01.

FIG 6

Neuraminidase inhibitors could decrease influenza virus cell entry. (A) A549 cells were treated with PBS or the NA inhibitors peramivir (30 μg/ml) or zanamivir (30 μg/ml) 1 h before or at the indicated times after H5N1 virus infection, and an MTS assay was performed to determine the viability of these cells at 48 h postinfection. (B) A549 cells were treated with PBS or the NA inhibitors peramivir (30 μg/ml) and zanamivir (30 μg/ml) 1 h before H1N1 or H5N1 virus infection and 4 h postinfection; the NP-positive nucleus was analyzed by confocal microscopy, and the percentages of NP-positive nuclei were determined using ImageJ software. (C and D) q-PCR analysis of the influenza viral gene M at the indicated time points in A549 cells that infected with H1N1 virus (C) or H5N1 virus (D) and treated with PBS, peramivir (30 μg/ml), or zanamivir (30 μg/ml) 1 h before the infection. All data represent the means ± the SEM of three independent experiments. *, P < 0.05; **, P < 0.01.