Restriction of Influenza A Virus by SERINC5

Abstract

Serine incorporator 5 (Ser5), a transmembrane protein, has recently been identified as a host antiviral factor against human immunodeficiency virus (HIV)-1 and gammaretroviruses like murine leukemia viruses (MLVs). It is counteracted by HIV-1 Nef and MLV glycogag. We have investigated whether it has antiviral activity against influenza A virus (IAV), as well as retroviruses. Here, we demonstrated that Ser5 inhibited HIV-1-based pseudovirions bearing IAV hemagglutinin (HA); as expected, the Ser5 effect on this glycoprotein was antagonized by HIV-1 Nef protein. We found that Ser5 inhibited the virus-cell and cell-cell fusion of IAV, apparently by interacting with HA proteins. Most importantly, overexpressed and endogenous Ser5 inhibited infection by authentic IAV. Single-molecular fluorescent resonance energy transfer (smFRET) analysis further revealed that Ser5 both destabilized the pre-fusion conformation of IAV HA and inhibited the coiled-coil formation during membrane fusion. Ser5 is expressed in cultured small airway epithelial cells, as well as in immortal human cell lines. In summary, Ser5 is a host antiviral factor against IAV which acts by blocking HA-induced membrane fusion. IMPORTANCE SERINC5 (Ser5) is a cellular protein which has been found to interfere with the infectivity of HIV-1 and a number of other retroviruses. Virus particles produced in the presence of Ser5 are impaired in their ability to enter new host cells, but the mechanism of Ser5 action is not well understood. We now report that Ser5 also inhibits infectivity of Influenza A virus (IAV) and that it interferes with the conformational changes in IAV hemagglutinin protein involved in membrane fusion and virus entry. These findings indicate that the antiviral function of Ser5 extends to other viruses as well as retroviruses, and also provide some information on the molecular mechanism of its antiviral activity.

Keywords: SERINC5; hemagglutinin; influenza A virus; restriction; virus entry.

FIG 1

Ser5 reduces the infectivity of pseudovirions with IAV HA proteins. (A) Determination of the antiviral effect of Ser2, Ser3 and Ser5 on pseudovirions with IAV H5HA, HIV-1 (NL4-3) Env, and VSV-G glycoproteins. Viruses were prepared in cells transfected with 1 μg of Ser2, Ser3 or Ser5 plasmid, and then MDCK or TZM-bl cells were infected with the viral supernatant. After 72 h pi, luciferase activity was measured and finally the specific infectivity was determined. The results are presented as the mean of 3 independent experiments ± SEM and analyzed by two-way ANOVA (**, P < 0.01; ****, P < 0.001; ns: no significance). (B) Western blot analysis of the virus pellet and cell lysate of IAV H5HA and HIV-1 (NL4-3) Env from (A). Arrows indicate the corresponding proteins. (C) Ser5 inhibited IAV group 1 and 2 HA proteins in a dose-dependent manner. Pseudovirions with the indicated glycoproteins were prepared in the presence of different amounts of Ser5 plasmid. The results are presented as the mean of 3 independent experiments ± SEM and analyzed by two-way ANOVA (*, P < 0.05; ****, P < 0.001; ns: no significance).

FIG 2

Nef antagonizes Ser5 effect on the pseudovirions with IAV HA. The antiviral activity of Ser5 to indicated glycoproteins is antagonized with Nef_97ZA012 protein. Pseudovirions with different glycoproteins were prepared in the presence of a fixed amount of Ser5 (1 μg for IAV; 0.5 μg for HIV-1 [NL4-3]) and varied Nef amounts (0, 0.25, 0.5 and 1 μg), and finally the relative specific infectivity was determined by comparison with the specific infectivity of virions without Ser5 and Nef proteins. The results are presented as the mean of 4 independent experiments ± SEM and analyzed by two-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.005; ****, P < 0.001).

FIG 3

The antiviral activity of Ser5 is associated with its plasma membrane-bound property and interaction with HA. (A) The antiviral activity of Ser5 and its incorporation into virus particles are associated with plasma membrane localization. Pseudovirions with IAV H5 HA protein were prepared in cells transfected with the wild-type Ser5 or Ser5K130A (1 μg), and finally the specific infectivity was determined. The results are presented as the mean of four independent experiments ± SEM and analyzed by one-way ANOVA (***, P < 0.005; ns: not significant). Western blot analysis of the virus pellet and cell lysate of indicated preparations. The analysis was performed thrice and one representative example is shown. Arrows indicate the corresponding proteins. (B) Representative confocal images of the localization of the wild-type Ser5 and Ser5K130A in cells producing IAV pseudovirions. The HA epitope tag was inserted between residues 290 and 291 of Ser5. Immunofluorescence staining of HA (green) indicates that it is mainly found on the plasma membrane. In addition, the staining of Ser5 (red) reveals that the wild-type Ser5 is mainly found on the plasma membrane whereas the mutant Ser5K130A is only found intracellularly. The merged images further show that HA is colocalized with the wild-type Ser5 (white arrows). Scale bar is 2 μm. The Pearson colocalization coefficient between HA and Ser5 signal of each treatment was determined accordingly (n = 5 cells). The results are presented as the mean of experiments ± SEM and analyzed by Student's t test (****, P < 0.001). (C) Co-immunoprecipitation assay of FLAG-tagged Ser5 (Ser5-FLAG) or Ser5K130A (Ser5K130A-FLAG), and HA. 293T cells were co-transfected with Ser5-FLAG, Ser5K130A-FLAG or a filler plasmid, and plasmid encoding IAV H5 HA with or without plasmid pNL4-3.LucR-E-. Co-immunoprecipitation tests were performed on extracts of the transfected cells. Representative immunoblots are shown. Arrows indicate the corresponding proteins. The fourth panel shows both HA2 and p24 bands because the same membrane was sequentially probed with both anti-HA and anti-p24 antisera. The ratio of HA2 protein to Ser5 signal in IP fractions was determined accordingly (n = 9 samples). The results are presented as the mean of experiments ± SEM and analyzed by Student's t test (****, P < 0.001).

FIG 4

Ser5 inhibits the membrane fusion induced by IAV HA. (A) Cells were infected with pseudovirions containing BlaM-Vpr and the indicated glycoproteins for 3 to 4 h with or without Ser5 (Ser5 to glycoprotein plasmid ratio, 2:1), and then loaded with CCF2. After incubation with the indicated pseudovirions, the cells were fixed and analyzed for the cleavage of CCF2 by measuring fluorescence emission at 520 nm (uncleaved CCF2) and 447 nm (cleaved CCF2) using flow cytometry. These results are the representative of 3 independent experiments. The graph in the right panel depicts the relative percentage of virus-cell fusion. The results are presented as the mean of three independent experiments ± SEM and analyzed by two-way ANOVA (****, P < 0.001). (B) 293T cells serving as “producer” cells were transfected with indicated glycoproteins; a filler plasmid or Ser5 (2 μg); and DSP1-7. Corresponding target cells as mentioned in the Methods session were transfected with DSP8-11. About 24 h posttransfection, target cells were detached by trypsin, and were transferred to a 96-well plate in sextuplicate. At 48 h posttransfection, target cells of HIV-1 (NL4-3) were incubated with 30 μM Enduren (Promega) for 1 h, and then the producer cells were added to the target cells in an equal volume, resulting in a 15 μM final concentration of Enduren. Luciferase activity was measured 2 h after the addition of producer target cells. For the cell-cell fusion induced by IAV HA, producer cells with or without Ser5 were added to the target cells at 48 h posttransfection. After 16 h additional incubation, cells were incubated in acidic DMEM (pH 5) at 37°C for 10 min and then neutralized with complete culture medium. After that, medium was replenished with complete culture medium in the presence of 15 μM Enduren. Luciferase-induced luminescence was measured 2 h after the acidic treatment. These results are the representative of 4 independent experiments. The graph in the right panel depicts the relative percentage of cell-cell fusion. The results are presented as the mean of 4 independent experiments ± SEM and analyzed by two-way ANOVA (****, P < 0.001).

FIG 5

Ser5 inhibits IAV (PR8/H1N1) infection. (A) Representative Western blot analysis of the virus pellets produced from indicated cells and the lysates of virus-producing cells in the presence or absence of doxycycline. MDCK-Ctrl and MDCK-Ser5 cells were pretreated with doxycycline for 16 h, and then infected with IAV (PR8/H1N1) at an MOI of 0.01 for 72 h. The virus supernatant was ultracentrifuged, and cell lysate was harvested at 72 h pi. (B) The virus titer of the concentrated virus from the indicated cells was determined by the plaque assay, in which the same vRNA copy of the virus supernatant was serially diluted in a 10-fold manner. The virus titer (PFU/mL) from each preparation was determined accordingly. Finally, the relative virus titer to the one produced from MDCK-Ctrl cells without doxycycline treatment was calculated. The results are presented as the mean of 4 independent experiments ± SEM and analyzed by two-way ANOVA (**, P < 0.01; ns: no significance). (C) The infectivity test of IAV produced from one of the independent experiments as in (A). Equal vRNA copy amounts of IAV with or without Ser5 infected MDCK cells. The percentage infection by the concentrated IAV with or without Ser5 was finally analyzed by flow cytometry at 12 h pi. Representative plots are shown. (D) Graph depicts normalized infection percentage measurements from four independent experiments as in (C). The results are presented as the mean of 4 independent experiments ± SEM and analyzed by two-way ANOVA (***, P < 0.005; ns: no significance).

FIG 6

Endogenously expressed Ser5 effect interferes with IAV infectivity. (A) Total RNA of IAV-susceptible and primary cells was extracted, and the copy numbers of mRNA of Ser5 and GAPDH were quantified by qPCR in comparison with their corresponding standard curves. The results are presented as the mean of 3 independent experiments ± SEM (***, P < 0.005; ****, P < 0.001; ND: not detectable). (B) Endogenous Ser5 in human cells possesses antiviral activity against IAV. The Ser5 knockdown (Ser5KD) and control (Ctrl) CaCo2 and A549 cells were infected with influenza A/PR/8/34 (H1N1) virus at MOI of 0.1 and 1 respectively. Infected cells were harvested at 8, 24, and 32 h pi. The vRNA and cellular RNA was extracted from infected cells, and then the vRNA and GAPDH mRNA were quantified by qPCR. Finally, the virus infectivity in infected cells was determined by normalizing IAV vRNA of M gene with cellular GAPDH level. The results are presented as the mean of 3 independent experiments ± SEM and analyzed by one-way ANOVA (*, P < 0.05; **, P < 0.01).

FIG 7

Ser5 affects HA dynamics as determined by smFRET. IAV pseudovirions with or without Ser5 were prepared and labeled with fluorophores as described in Materials and Methods. (A) Experimental setup of viral particles immobilized on a quartz substrate for imaging with TIRF microscopy. (B) As previously described, fluorophores were attached to the HA2 domain. The indicated inter-fluorophore distances in the pre- and post-fusion HA conformations were estimated through molecular dynamic simulation (33). (C) Ser5 destabilizes the pre-fusion conformation of HA as indicated by the reduction in high-FRET occupancy, and the increase in low-FRET occupancy across the range of pHs tested. High- and low-FRET state occupancies were determined through HMM analysis. (D) Ser5 impedes the transition to the coiled-coil conformation of HA. Fluorescently labeled pseudovirions with or without Ser5 were incubated at pH 4.8 for the indicated times to trigger HA conformational changes. The particles were then returned to neutral pH and evaluated by smFRET. The presence of Ser5 increased the return to the high-FRET pre-fusion conformation after 10- and 20-minute incubations at acidic pH. In (C) and (D), bars indicate the average occupancy over the population of traces, with error bars representing the SEM. The corresponding FRET histograms are shown in Fig. S5. Statistical significance of differences in FRET state occupancies was evaluated using a t test (*, P < 0.01; **, P < 0.001; ***, P < 0.0001; ns: no significance).