Fig 1.
MARCH2 gene expression is unaffected by type I IFN and HIV-1 infection.
MARCH2 fold expression change relative to mock normalized to GAPDH in (A) PMA-differentiated THP-1, H9, and 293T cells, (B) primary CD4+ T cells from three donors and (C) primary Monocyte Derived Macrophages (MDMs) from three donors treated with human IFN-β (500 units/ml) for 4 h, 8 h, 16 h, and 24 h. (D) MARCH2 fold expression relative to mock normalized to GAPDH in PMA-differentiated THP-1 cells infected with HIV-1JR-CSF and H9 cells infected with HIV-1NL4-3 at 4 h, 8 h, 16 h, and 24 h post infection (hpi). Mock indicates mock-treated (PBS). Graphs represent mean ± SD from n = 3 independent experiments.
Fig 2.
G61 residue on the N- terminal cytoplasmic tail of human MARCH2 is critical for its antiretroviral function.
(A) Sequence alignment of human MARCH2 (hM2: NCBI Ref. NP_001356705.1) and mouse MARCH2 (mM2: NCBI Ref. NP_001239409.1). Differing amino acid residues are shown in red and their positions are indicated. In green are conservative changes in the amino acid sequence. (B and C) Western blots of cells and culture media from co-transfections of MLV (B) or HIV-1 (NL4-3) (C) in the presence of either wild type hM2 or mM2, cytoplasmic tails swapped M2 chimeras (hM2N-mM2C or mM2N-hM2C) or empty vector (E.V.). (D and E) Substituting any 4 N-terminal cytoplasmic tail residues (S18, Q58, C61, and N75) of mM2 with those from hM2 renders mM2 antiviral. In D, western blots of cell and culture media of co-transfections of MLV in the presence of either wild type hM2 or mM2, mM2 mutants (S18G, Q58P, C61G, or N75C) or E.V. In E, infectivity assays of MLV containing a firefly luciferase reporter genome generated in the presence or absence of wild type hM2 or mM2 or the aforementioned mM2 mutants. Luciferase levels were measured 48 hpi and normalized to MLV p30 levels of the input virus. (F and G) G61 residue at the N-terminal cytoplasmic tail of hM2 is critical for its anti-HIV-1 function. In F, western blots of HIV-1 (NL4-3) co-transfections in the presence of either wild type hM2 or mM2, hM2 mutants (G18S, P58Q, G61C, or C75N) or E.V. In G, infectivity assays of NL4-3 Env pseudotyped luciferase reporter viruses generated in the presence of wild type hM2 or mM2, hM2 mutants (G18S, P58Q, G61C, or C75N) or E.V. Luciferase levels were measured 48 hpi and normalized to HIV-1 p24 levels of input virus. For B, C, D and F, representative immunoblotting results from n = 3 independent experiments are shown. Western blots were probed with anti-Myc, anti-gp70, anti-p15E, anti-Pr65Gag, anti-p24, anti-gp120, anti-gp41 and anti-β-Actin antibodies. For E and G, the percentage (%) of relative infectivity was determined with respect to virus produced in the presence of E.V. and all graphs represent mean ± SD from 3 independent experiments. Statistical significance was determined using one-sample t-test (two-tailed) when comparisons were performed with E.V. and unpaired t-test (two-tailed) was used for non-E.V. comparisons. ns, non-significant; *, P ≤ 0.05; **, P ≤ 0.01.
Fig 3.
The RING-CH domain and the TM2 domain of human MARCH2 are important for the degradation of HIV-1 envelope glycoprotein.
293T cells were co-transfected with an HIV-1 infectious clone (NL4-3) along with either empty vector (E.V.), wild type (WT) MARCH2 (M2) or the following M2 mutants: (A) M2 RING-CH domain mutants (M2 W97A and M2 C64/67S), (B) M2 with deletions of N-terminal amino acid residues 2–30 (M2 Δ2–30) or 31–56 (M2 Δ31–56), (C) M2 with mutations in the PDZ binding motif (M2 PDZmut) and (D) M2 with either N-terminal (TM1) or C-terminal (TM2) swapped with those of human MARCH4 (M2 TM1M4) and human transferrin receptor (M2 TM2TR) respectively. Cells were harvested 48 h post transfection, and proteins were analyzed by immunoblotting using anti-Myc, anti-p24, anti-gp120, anti-gp41 and anti-β-Actin antibodies. For A- D, representative immunoblotting results from n = 3 independent experiments are shown.
Fig 4.
Only the canonical long isoform of MARCH2 has anti-HIV-1 function.
(A) Both isoforms of human MARCH2 (M2), M2 long isoform (M2LI) and M2 short isoform (M2SI) are expressed in H9 and THP-1 cells. M2 isoform specific products were amplified by PCR from H9 and THP-1 cDNAs followed by agarose gel analysis. GAPDH was used as a loading control and M2 plasmid DNA was used as positive control. (B) M2LI reduces HIV-1 infectivity. NL4-3 Env pseudotyped luciferase reporter viruses were generated in the presence of either M2LI, M2SI or empty vector (E.V.). Pseudoviruses were used to infect U373-MAGI-CXCR4 cells, luciferase levels were measured 48 hpi and normalized to HIV-1 p24 levels of the input virus. The percentage (%) of relative infectivity was determined with respect to virus produced in the presence of E.V. In A, representative gel images from n = 2 independent experiments are shown. In B, graphs represent mean ± SD from 3 independent experiments. Statistical significance was determined using one-sample t-test (two-tailed) when comparisons were performed with E.V. and unpaired t-test (two-tailed) was used for the comparison between M2LI and M2SI. ns, non-significant; **, P ≤ 0.01; ***, P ≤ 0.001. (no template control, NTC; + Control, positive control).
Fig 5.
MARCH2 and HIV-1 gp41 interaction is mediated by their TM domains.
(A) Endogenous MARCH2 (M2) interacts with HIV-1 gp41. Cell lysates from HIV-1NL4-3 infected H9 cells were immunoprecipitated (IPed) with an anti-HIV-1 gp41 antibody followed by western blots probing with anti-M2, anti-HIV-1 gp41 and anti-GAPDH antibodies. (B) The C- terminal TM domain (TM2) of M2 is critical for its interaction with HIV-1 gp41. CoIPs of cell lysates of 293T cells co-transfected with an HIV-1 infectious clone along with either empty vector (E.V.), M2 WT or M2 TM domain mutants (M2 TM1M4, M2 TM2TR). Cells were harvested, lysed and immunoprecipitated with an anti-HIV-1 gp41 antibody followed by western blots probing with anti-Myc (M2), anti-HIV-1 gp41 and anti-β-Actin antibodies. (C) HIV-1 gp41 TM domain is required for M2-mediated restriction. 293T cells were co-transfected with plasmids encoding either NL4-3 Env or mutant NL4-3 Env, in which the TM domain has been swapped with that of transferrin receptor, (NL4-3 TMTR Env) along with either E.V. or M2. Cells were lysed followed by western blot analysis with anti-HIV-1 gp41, anti-Myc (M2) and anti-β-Actin antibodies. (D) The TM domain of HIV-1 gp41 is required for its interaction with M2. Cells were transfected as in C, lysed and immunoprecipitations were performed with an anti-HIV-1 gp41 antibody followed by western blots probing with anti-Myc (M2), anti-HIV-1 gp41 and anti-β-Actin antibodies. For A- D, representative immunoblotting results from n = 3 independent experiments are shown.
Fig 6.
MARCH2 is incorporated in HIV-1 virions in a transmembrane (TM)-dependent manner.
(A-C) C-terminal TM domain (TM2) of MARCH2 (M2) is critical for M2 incorporation in nascent HIV-1 particles. HIV-1 particles produced in the presence of either wild type (WT) M2 or MARCH2 TM domain mutants (M2 TM1M4, M2 TM2TR). Culture supernatants were harvested 48 h post transfection, concentrated through 30% sucrose cushion, and subjected to OptiPrep (5 to 20%) density gradient fractionation. Proteins were recovered from the fractions by TCA precipitation and analyzed by western blots probing with anti-Myc (M2) and anti-p24 antibodies. (D) M2 co-localizes with HIV-1 Gag at the cell surface. AD293 cells co-transfected with plasmids expressing HIV Gag-iGFP and M2 WT were subjected to immunostaining. The table at the bottom shows Pearson’s correlation coefficient (mean ± SD) for co-localization between M2 and HIV-1 Gag (Gag) calculated from 5 independent experiments. (E) The M2 short isoform (M2SI) that lacks the TM domain is not incorporated in nascent HIV-1 virions. Virus from the culture media and cell lysates from 293T cells transfected with an HIV-1 infectious clone and either empty vector (E.V.), M2SI or the M2 long isoform (M2LI) were analyzed by western blots probing with anti-Myc (M2), anti-p24 and anti-β-Actin. For A-C and E representative of n = 3 independent experiments are shown. In D, representative deconvolved single Z-section images are shown. (Fraction 1, F1; Fraction 15, F15).
Fig 7.
Virion-incorporated MARCH2 blocks HIV-1 particle entry by targeting virus-cell fusion in a RING-CH-independent manner.
(A and B) Low levels of MARCH2 (M2) have no effect on envelope (Env) glycoprotein levels yet still reduce HIV-1 particle infectivity. HIV-1 particles produced in the presence of either low amounts of M2 or empty vector (E.V.) were analyzed by western blots. In B, HIV-1 from A was normalized for gp120 and used to infect TZM-bl cells; luciferase levels were measured 48 hpi and normalized to HIV-1 p24 levels of the input virus. (C and D) M2 RING-CH mutant reduces HIV-1 particle infectivity. NL4-3 Env pseudotyped luciferase reporter viruses were produced in either increasing concentrations of RING-CH domain mutant of MARCH2 (M2 RING-CHmut) or eGFP-N1 (serving as E.V.). Culture media were harvested and either analyzed by (C) western blots or (D) used to infect U373-MAGI-CXCR4 cells, luciferase levels were measured 48 hpi and normalized to HIV-1 p24 levels of the input virus. (E and F) M2 TM2TR has no effect on gp120 and gp41 levels and is not incorporated in nascent virions. NL4-3 Env pseudotyped luciferase reporter viruses were produced in either increasing concentrations of M2 TM2TR, low levels of wild type M2 or E.V. (E) western blots and (F) infectivity assays as described in C and D. (G) MARCH2 inhibits HIV-1 Env-mediated fusion. H9 cells were infected with p24- normalized HIV-1 produced in the presence of either M2 RING-CHmut or E.V. Fusion inhibitor enfuvirtide (T20) served as a negative control. Cells were stained with CCF2-AM and then subjected to FACS. Fusion efficiency was quantified as the percentage of cleaved CCF2. Representative FACS plots and graphs representing mean ± SD of % cleaved CCF2 positive cells from n = 3 independent experiments are shown. (H and I) M2-mediated inhibition of envelope incorporation in nascent HIV-1 virions and inhibition of virus entry by M2 inhibit HIV-1 particle infectivity in an additive manner. NL4-3 Env pseudotyped luciferase reporter viruses were generated in the presence of E.V., wild type M2 or M2 RING-CHmut. In H, western blots and in I, infectivity assays as described in (C) and (D). Luciferase levels were measured 48 hpi and normalized to HIV-1 p24 levels of the input virus. In B, D, F and I, the percentage (%) of relative infectivity was determined with respect to virus produced in the presence of E.V. For A, C, E and H, representative western blots from n = 3 independent experiments are shown probing with anti-gp120, anti-gp41, anti-p24, anti-Myc (M2) and anti-β-Actin antibodies. Graphs in B, D, F and I represent mean ± SD from 3 independent experiments. Statistical significance was determined using one-sample t-test (two-tailed) when comparisons were performed with E.V. in B, D, F and I and unpaired t-test (two-tailed) between non E.V. conditions in D and I and for any comparisons in G. ns, non-significant; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001.
Fig 8.
MARCH2 restricts HIV-1 replication in a T cell specific manner.
(A) MARCH2 (M2) knockdown verification in H9 cells knocked down with a M2 specific (shM2) or a scramble control (shCtrl) shRNA. Fold expression change of human M2 and MARCH3 (M3) transcripts in M2 depleted H9 cells relative to shCtrl treated cells and normalized to GAPDH is shown. Bottom panels show western blots of endogenous M2 proteins in shCtrl- or shM2- treated H9 cells. GAPDH serves as loading control. (B) Endogenous M2 reduces HIV-1 replication in H9 cells. Wild type (shCtrl) or M2 depleted (shM2) H9 cells were infected with HIV-1NL4-3 and harvested at the indicated time points. HIV-1 nef and env DNA levels were determined by RT-PCR normalized to GAPDH. (C) HIV-1 virions produced in H9 cells knocked down for endogenous M2 have higher Env levels. Western blots of virions produced in wild type (shCtrl) or M2 depleted (shM2) H9 cells probing with anti-gp120, anti-gp41 and anti-p24 antibodies. Densitometry analysis of gp120 and gp41 band intensity relative to shCtrl and normalized to p24 are shown below (n = 3). (D) Endogenous M2 is incorporated into nascent HIV-1 virions. Virus from C was used to detect endogenous M2 by western blots probing with anti-M2 and anti-p24 antibodies. (E) M2 and MARCH8 (M8) knockdown verification in THP-1 cells depleted for either M2, M8 or both M2 and M8. In the left panel, fold expression changes of human M2 and M3 transcripts in shM2 expressing THP-1 cells relative to shCtrl expressing cells and normalized to GAPDH. In the right panel, fold expression changes of human M2 and M8 transcripts in shM2, shM8 and shM2/M8 expressing THP-1 cells relative to shCtrl expressing cells normalized to GAPDH. Bottom panels show western blots of endogenous M2 proteins in shCtrl- or shM2, shM8, shM2/M8- expressing THP-1 cells. β-Actin serves as loading control. (F) Endogenous M2 has no effect on HIV-1 replication in THP-1 cells. PMA-differentiated THP-1 cells expressing shCtrl or shM2, shM8, both shM2 and shM8 were infected with HIV-1AD8 and infected cells were harvested at the indicated time points. HIV-1 nef and env DNA levels were determined by qPCR and normalized to GAPDH. (G and H) Only HIV-1 virions produced in THP-1 cells depleted for endogenous M8 have elevated Env levels. PMA-differentiated THP-1 cells expressing shCtrl, shM2, shM8, or shM2/M8 were infected with HIV-1AD8, cell lysates and culture supernatants were harvested at 48 hpi. In G, culture supernatants were concentrated by ultracentrifugation and p24 normalized purified virions fractions were used for western blot analyses probing for gp120, gp41, p24. The graphs on the right show quantification analysis of virion associated gp120 or gp41 band intensities normalized to p24 band intensities relative to shCtrl conditions. In H, p24 normalized purified virions fractions were used to infect TZM-bl cells; luciferase levels were measured 48 hpi, and the percentage (%) of relative infectivity was determined with respect to virus produced in the presence of shCtrl conditions. (I) Primary CD4+ T cells were purified from 3 donors, activated, and transduced with lentiviruses expressing either shCtrl or shM2 followed by infection with HIV-1NL4-3. Infected cells were harvested at the indicated time points. HIV-1 nef and env DNA levels were determined by RT-PCR and normalized to GAPDH. (J) Primary CD14+ cells were purified from 2 donors, differentiated, and transduced with lentiviruses expressing either shCtrl or shM2, shM8 or both shM2/M8 followed by infection with HIV-1AD8. Infected cells were harvested at the indicated time points and analyzed as in I. In A, C, D, E and G, representative blots from 3 independent experiments are shown. Graphs in A, B, E, F, G, H represent mean ± SD from 3 independent experiments. Graphs in I and J represent mean ± SD from 3 and 2 human donors respectively. Statistical analysis in A, B, E, F, I and J performed using unpaired t-test (two-tailed). In H, statistical significance was determined using one-sample t-test (two-tailed) when comparisons were performed with shCtrl and unpaired t-test (two-tailed) between non-Ctrl conditions. ns, non-significant; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001. (days post infection, dpi).
Fig 9.
MARCH2 inhibits HIV-1 cell-to-cell transmission and is recruited at virological synapse.
(A) MARCH2 (M2) overexpression reduces HIV-1 cell-to-cell transfer. CellTrace Far Red (CTFR)-labelled H9 cells were cocultured with 293T donor cells co-transfected with plasmids for HIV Gag-iGFP along with either M2 or empty vector (E.V.). At 16 h of co-culture, cells were harvested and subjected to FACS. (B) Endogenous M2 reduces HIV-1 cell-to-cell transfer. CTFR-labelled TZM-bl cells were cocultured with H9 cells stably expressing shRNA-M2 or shRNA-Ctrl and infected with HIV Gag-iGFP. Cells were harvested 24 h post co-culture followed by FACS analysis. (C) M2 localizes predominantly in intracellular compartments. AD293 cells transfected with an M2 expressing plasmid were subjected to immunostaining. (D) M2 is recruited at the virological synapse during HIV-1 cell-to-cell transmission. CellTracker Voilet (CVT)-BMQC-labelled H9 cells were co-cultured with AD293 cells co-transfected with plasmids expressing HIV Gag-iGFP and M2 followed by immunostaining. Insets represent 55.6 × zoomed images from the indicated boxed regions. The graph at the right shows representative line scanning analyses for co-localization between MARCH2 (M2), HIV-1 Gag (Gag) and ICAM-1. The pixel intensity in each channel is measured along a line drawn on the image (white line in the inset) and relative pixel intensity is plotted versus distance along the line. In A and B, representative FACS plots and cumulative graphs representing mean ± SD of % GFP+ positive H9 or TZM-bl cells from n = 3 independent experiments are shown. Statistical analysis was performed using unpaired t-test (two-tailed). **, P ≤ 0.01. For C and D, images were acquired from multiple fields from 3 independent experiments and representative deconvolved single Z-section images are shown.