NDP52 mediates an antiviral response to hepatitis B virus infection through Rab9-dependent lysosomal degradation pathway

Abstract

Autophagy receptor NDP52 triggers bacterial autophagy against infection. However, the ability of NDP52 to protect against viral infection has not been established. We show that NDP52 binds to envelope proteins of hepatitis B virus (HBV) and triggers a degradation process that promotes HBV clearance. Inactivating NDP52 in hepatocytes results in decreased targeting of viral envelopes in the lysosome and increased levels of viral replication. NDP52 inhibits HBV at both viral entry and late replication stages. In contrast to NDP52-mediated bacterial autophagy, lysosomal degradation of HBV envelopes is independent of galectin 8 and ATG5. NDP52 forms complex with Rab9 and viral envelope proteins and links HBV to Rab9-dependent lysosomal degradation pathway. These findings reveal that NDP52 acts as a sensor for HBV infection, which mediates a unique antiviral response to eliminate the virus. This work also suggests direct roles for autophagy receptors in other lysosomal degradation pathways than canonical autophagy.

Fig. 1. NDP52 interacts with the preS2 region of HBV envelope proteins.

a Schematic presentation of HBV three envelope proteins. The preS1, preS2 and S regions are indicated. Glycosylation within S and preS2 is indicated. b GST pull-down assays using cell extract of HEK293T cells transfected with plasmids encoding Flag-tagged S, M or L and GST or GST fusion of NDP52. Coprecipitated proteins were detected with indicated antibodies. c Coimmunoprecipitation assays with anti-NDP52 antibody in either Huh7 cells transfected with Flag-tagged vector (V), S, M or L (upper panel), or HepAD38 cells without doxycycline (Dox) (lower panel). Immunoprecipitates were detected by indicated antibodies. d Immunofluorescence with anti-NDP52 and anti-preS2 antibodies shows colocalization of NDP52 with M and L in HepAD38 cells without Dox and with Dox. The scale bar is 10 µm for full cell images. e Representative fluorescence micrographs of Huh7 cells transfected with plasmids coding for GFP fusion of L, M or S proteins and immunostained for NDP52. The scale bar is 10 µm for full cell images. f Pearson’s correlation coefficients for colocalizations in e (n = 7 biological replicates). ns non significance. g Representative fluorescence micrographs of HepAD38 cells replicating HBV stained with anti-NDP52 and anti-HBs antibodies and reagents labeling ER (ER-Tracker), Golgi (anti-TGN46 antibody) and multivesicular body (anti-CD63). The scale bar is 10 µm for full cell images, 2.5 µm for zoomed images. Data are means ± SD. Statistical significance in f is determined by a two-sided unpaired t-test. Source data for b, c and f are provided as a Source Data file.

Fig. 2. NDP52 restricts HBV infection.

a Immunoblot analysis of indicated proteins from cellular extract of CRISPR control NDP52^HepG2-NTCPWT (Ctrl), NDP52^{HepG2- NTCPKO} (KO), NDP52^HepG2-NTCPKO restored with mCherry tag vector or mCherry-tagged NDP52 (KO + vector, KO + NDP52), or NDP52^HepG2-NTCPWT overexpressing mCherry tag vector or mCherry-tagged NDP52 (WT + vector, WT + NDP52) cells. b Indicated cells were infected with HBV. Cytoplasmic core DNA was purified fourteen days post infection and detected by Southern blot analysis using an HBV DNA probe. M: DNA marker. Depending on infection, variations in detected viral replication species were observed. Therefore, ensemble replicative species are designated as replication forms. Quantitative PCR of HBV DNA from infected cells (c) and extracellular virions (d) seven days post infection (n = 3 biological replicates). e Medium levels of viral HBeAg released from infected cells (n = 3 biological replicates). S:CO signal-to-cutoff ratio, Dpi day post infection. f NDP52^HepG2-NTCPWT and NDP52^HepG2-NTCPKO cells were infected with HBV. Cells were collected at indicated time. Viral DNA, cccDNA and pregenomic RNA (pgRNA) were analyzed by qPCR. Expression in Ctrl cells was arbitrarily set as 1 (n = 3 biological replicates). Data are means ± SD. Statistical significance in c–f is determined by a two-sided unpaired t-test. Source data for a–f are provided as a Source Data file.

Fig. 3. NDP52 interaction with preS2 is required for inhibition of HBV.

a A schematic map of NDP52 showing the domain organization and the deletion constructs used to identify the interaction domain with the preS2 region. SKICH skeletal muscle and kidney enriched inositol phosphatase carboxyl homology domain, CLIR LC3C interacting region, CC coiled-coil, GIR galectin 8 interacting region, Zn zinc finger. b Coimmunoprecipitation assays using cell extract of HEK293T cells transfected with plasmids encoding Flag-tagged NDP52 full length (FL) or truncations and HA-tagged M envelope protein. Flag-coprecipitated proteins were detected with indicated antibodies. c Immunoblot analysis of indicated proteins from cellular extract of NDP52^HepG2-NTCPKO cells restored with mCherry tag vector, mCherry-tagged wild-type NDP52 (WT) or mCherry-tagged NDP52 lacking GIR domain (ΔGIR). d Indicated cells were infected with HBV. Core DNA was purified fourteen days post infection and detected by Southern blot analysis using an HBV DNA probe. M: DNA marker. e Quantitative PCR of HBV DNA from infected cells seven days post infection (n = 3 biological replicates). f Medium levels of HBeAg from infected cells. S:CO signal-to-cutoff ratio, Dpi day post infection (n = 3 biological replicates). g GST pull-down assays using cell extract of HEK293T cells transfected with plasmids encoding Flag-tagged S and M constructs and GST fusion of NDP52. GST pulldown proteins were detected with indicated antibodies. In schematic presentation of S and M constructs, glycosylation sites are presented. mGly: mutant with the two glycosylation sites Asn4 and Thr37 at the preS2 region changed to Ala4 and Ala37. h Representative fluorescence micrographs of Huh7 cells transfected with plasmids coding for mCherry-NDP52 and Flag-M and immunostained with anti-Flag antibody. The scale bar is 10 µm for full cell images. i, j Pearson’s correlation coefficients for colocalizations in h (n = 6 biological replicates). k Quantitative PCR of HBV DNA from NDP52^HepG2-NTCPWT (Ctrl) and NDP52^HepG2-NTCPKO (KO) cells transfected with pPres2Δ19-26 or pPreS2WT with pHBVcore(-) four days post transfection (n = 3 biological replicates). Data are means ± SD. Statistical significance in e, f, i–k is determined by a two-sided unpaired t-test. Source data for b–g, i–k are provided as a Source Data file.

Fig. 4. NDP52 targets viral envelopes to lysosomal degradation.

a Colocalization of NDP52, viral envelope proteins (HBs) and Lamp2 in HepAD38 cells (upper panel). The scale bar is 10 µm for full cell images, 2.5 µm for zoomed images. Immunoblot analysis of indicated proteins from either whole cell lysate (WCL) or lysosome lysate (LL) of HepAD38 cells in the presence of Dox (lower panel). The major band close to 35 kd was also detected by anti-preS2 antibody and is shown in the other figures of WB with lysosome extract. b Short hairpin RNAs targeting NDP52 (shNDP52) or non-targeting control (shCtrl) were expressed in HepAD38 cells. NDP52-knockdown HepAD38 or control cells were immunostained for viral envelope proteins (HBs) and Lamp2. The scale bar is 10 µm for full cell images, 2.5 µm for zoomed images. Pearson’s correlation coefficients for colocalization are presented (n = 5 biological replicates). c Immunoblot analysis of indicated proteins from cellular extract of HepAD38 cells mock-treated or treated with bafilomycin A1 (BafA1). The ratio of the envelope protein normalized to GAPDH in BafA1-treated cells versus mock-treated cells is presented. d Immunoblot analysis of indicated proteins from either whole cell lysate (WCL) or lysosome lysate (LL) of HepAD38 cells in the presence of Dox. The ratio of the envelope protein normalized to GAPDH or Lamp2 in shNDP52 cells versus shCtrl cells is presented. e NDP52^HepG2-NTCPWT (Ctrl) and NDP52^HepG2-NTCPKO (KO) cells were infected with HBV and treated with NHCl₄. Cytoplasmic core DNA was purified fourteen days post infection and detected by Southern blot analysis using an HBV DNA probe. M: DNA marker. f NDP52^HepG2-NTCPWT and NDP52^HepG2-NTCPKO cells were infected with HBV and treated with NHCl₄. Quantitative PCR of HBV DNA from infected cells seven days post infection (n = 3 biological replicates). Data are means ± SD. Statistical significance in b and f is determined by a two-sided unpaired t-test. Source data for a–f are provided as a Source Data file.

Fig. 5. Rab9 is involved in NDP52-mediated viral degradation.

a Colocalization of HBs, NDP52 and Rab9 or HBs, Rab9 and Lamp2 in HepAD38 cells. The scale bar is 10 µm for full cell images, 2.5 µm for zoomed images. b NDP52 knockdown HepAD38 (shNDP52) or knockdown control cells (shCtrl) were immunostained for HBs and Rab9 (n = 10 biological replicates). The scale bar is 10 µm for full cell images. Pearson’s correlation coefficients for colocalization are presented. ns: non significance. c Coimmunoprecipitation assays using HepAD38 cells. Rab9 is coprecipitated with NDP52. d Coimmunoprecipitation assays with anti-GFP antibody in Huh7 cells transfected with GFP-tagged Rab9 and Flag-tagged M. e Coimmunoprecipitation assays with anti-GFP antibody in Huh7 cells transfected with GFP-tagged Rab9, GST-tagged NDP52 and Flag-tagged M. f HepAD38 cells were treated with bafilomycin A1 (BafA1), Earl’s Balanced Salts (EBSS), rapamycin (Rapa) or etoposide (Etop). Proteins from either whole cell lysate (WCL) or lysosome lysate (LL) were analyzed by immunoblot assay. The ratio of the envelope protein expression in WCL and LL normalized to GAPDH or Lamp2 respectively in reagent-treated cells versus mock-treated cells is presented. g Immunoblot analysis of indicated proteins from either WCL or LL of Rab9-knockdown (shRab9) or control (shCtrl) HepAD38 cells. The ratio of protein expression in shRab9 cells versus shCtrl cells in WCL and LL normalized to GAPDH or Lamp2 respectively is presented below each graph. h Immunoblot analysis of indicated proteins from cellular extract of Rab9 knockdown control Rab9^HepG2-NTCPWT (Ctrl), Rab9^HepG2-NTCPKD (KD), Rab9^HepG2-NTCPKD overexpressing mCherry-tagged NDP52 (KD + NDP52) or mCherry tag vector cells (KD + vector). i Rab9^HepG2-NTCPWT (Ctrl), Rab9^HepG2-NTCPKD (KD) cells were infected with HBV. Core DNA was purified fourteen days post infection and analyzed by Southern blot using an HBV DNA probe. M: DNA marker. j Quantitative PCR of HBV DNA from infected cells (n = 3 biological replicates). k Medium levels of HBeAg from infected cells (n = 3 biological replicates). S:CO signal-to-cutoff ratio, Dpi day post infection. l Rab9^HepG2-NTCPKD overexpressing NDP52 (KD + NDP52) or control cells (KD + vector) cells were infected with HBV. Core DNA was purified and analyzed by Southern blot using an HBV DNA probe. m Quantitative PCR of HBV DNA from infected cells seven days post infection (n = 3 biological replicates). ns non significance. n Medium levels of HBeAg from infected cells (n = 3 biological replicates). S:CO signal-to-cutoff ratio, Dpi day post infection. Data are means ± SD. Statistical significance in b, j, k, m, and n is determined by a two-sided unpaired t-test. Source data for b–n are provided as a Source Data file.

Fig. 6. NDP52 suppresses HBV in vivo.

a–c Mice were transduced with AAVHBV to establish chronic hepatitis B. Four weeks later (week 0), mice were injected either with adenovirus vector (AdCtrl) or expressing human NDP52 (AdNDP52). Serum levels of HBsAg (a), HBeAg (b) and alanine aminotransferase (ALT) (c) were measured. The dashed lines indicate the levels in naïve mice (n = 9 animals at weeks 0–2. n = 6 animals at week 3. n = 3 animals at weeks 4–7). d Quantitative PCR of HBV DNA and HBV pregenomic RNA (pgRNA) in mouse liver 3 weeks post adenovirus injection (n = 8 animals). Data are means ± SD. Statistical significance is determined by a two-sided unpaired t-test. e Immunohistochemical staining of livers with antibodies against S protein (α-HBs) or core protein (α-HBc). Source data for a–d are provided as a Source Data file.