Human MARCH1, 2, and 8 block Ebola virus envelope glycoprotein cleavage via targeting furin P domain

Membrane‐associated RING‐CH (MARCH) family proteins were recently reported to inhibit viral replication through multiple modes. Previous work showed that human MARCH8 blocked Ebola virus (EBOV) glycoprotein (GP) maturation. Our study here demonstrates that human MARCH1 and MARCH2 share a similar pattern to MARCH8 in restricting EBOV GP‐pseudotyped viral infection. Human MARCH1 and MARCH2 retain EBOV GP at the trans‐Golgi network, reduce its cell surface display, and impair EBOV GP‐pseudotyped virions infectivity. Furthermore, we uncover that the host proprotein convertase furin could interact with human MARCH1/2 and EBOV GP intracellularly. Importantly, the furin P domain is verified to be recognized by MARCH1/2/8, which is critical for their blocking activities. Besides, bovine MARCH2 and murine MARCH1 also impair EBOV GP proteolytic processing. Altogether, our findings confirm that MARCH1/2 proteins of different mammalian origins showed a relatively conserved feature in blocking EBOV GP cleavage, which could provide clues for subsequent MARCHs antiviral studies and may facilitate the development of novel strategies to antagonize enveloped virus infection.

Originally, MARCH proteins were identified owing to homology with the E3 ubiquitin ligase of K3 and K5 of the Kaposi's sarcomaassociated herpesvirus (KSHV). 2,3The MARCH family proteins now comprise 11 members, most of which share a similar molecular structure, including a cytoplasmic N-terminal C4HC3 RING-finger (RING-CH finger) domain and two or more transmembrane domains, except for MARCH7 and MARCH10, which contain no predicted transmembrane domains.4][15] MARCH1 and MARCH2 were originally found to downregulate transferrin receptor (TfR) and CD86 (B7-2). 3MARCH1 was involved in MHC II ubiquitination and thus promoted dendritic cell selection of natural regulatory T cells. 16In addition, MARCH1 inhibited type I IFN signaling pathways. 17MARCH2 recognized syntaxin-6, 18 regulated secretory proteins trafficking, 19 and negatively regulated cell autophagy 20 and NF-κB essential modulator (NEMO) signaling. 21cently, MARCH proteins were found to show antiviral activities. 22,23Tada et al. first reported that MARCH8 inhibited HIV-1 and VSV-G pseudotyped virus infection by targeting their envelope GPs. 246][27][28] The viral proteins targeted by MARCH proteins now are extended to include envelope GPs of EBOV, 27,29,30 severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), 27,28,30 murine leukemia virus (MLV), 24,27 influenza virus (IAV), 27,29,31 Nipah virus (NiV), 27 spring viremia of carp virus (SVCV) 32 and rabies virus (RABV), 28 M2 protein of IAV, 33 and N protein of porcine epidemic diarrhea virus (PEDV). 15Interestingly, in contrast to the negative role on viral infection, MARCH8 also critically supported hepatitis C virus (HCV) replication. 34nerally, two antiviral modes are employed by MARCH proteins to restrict viral infection.On one side, MARCH proteins downregulated viral envelope GPs expression by inducing their intracellular degradation, which was the viral protein cytoplasmic tail dependence (CTD), including envelope GPs of VSV and RABV, and M2 protein of IAV. 24,26,28,33On the other side, MARCH proteins retained viral envelope GPs at intracellular compartments without degradation, which was the viral protein cytoplasmic tail independence (CTI), including envelope GPs of HIV-1, EBOV, and SARS-CoV-2. 24,26,29,30,35 is critical for EBOV infection of target cells.EBOV GP 0 precursor experiences initial N-glycosylation modification in the endoplasmic reticulum (ER) and is then transported to the trans-Golgi network (TGN), where it finishes mature N-and O-glycosylation. 36e mature N, O-glycosylated GP 0 is cleaved into two subunits (GP 1 and GP 2 ) by furin protease at the highly conserved polybasic sites (R-X-A/R-R). 36The GP 1 and GP 2 subunits are re-linked via disulfide bonds and form GP 1,2 , 37 assembled into self-trimers, and transported to the cell membrane.GP 1 contains a mucin-like domain (MLD), which is both N-glycosylated and highly O-glycosylated and thus dramatically increases its molecular weight.Therefore, molecular weight of the mature glycosylated GP 1 is usually larger than that of the full-length GP 0 of immature glycosylation.
It was previously reported that human MARCH8 retained EBOV GP at the TGN and blocked its proteolytic cleavage via targeting furin under the CTI antiviral mode. 29Here, we extend to investigate the anti-EBOV GP potency of human MARCH1/2, bovine MARCH1/2, and murine MARCH1/2.Our results show that human MARCH1 and MARCH2, bovine MARCH2, and murine MARCH1 inhibit EBOV GP intracellular cleavage.Critically, human MARCH1 and MARCH2 are also found to hijack furin, as MARCH8 did, to trap EBOV GP at the TGN and block its maturation, demonstrating a conserved CTI antiviral mode among MARCH molecules and thus shedding light on subsequent antiviral studies.

| Cell lines
Human embryonic kidney 293T cells, cervical cancer HeLa cells, human liver Huh-7 cells, and African green monkey kidney Vero cells were cultured at 37°C and 5% CO 2 in DMEM (Gibco) supplemented with 8% fetal bovine serum (FBS) (Gibco), streptomycin, and penicillin (100 µg/mL).All cell lines were obtained from the American Type Culture Collection (ATCC).

| Plasmids
The bovine MARCH1/2 and murine MARCH1/2 were acquired from bovine/murine peripheral blood mononuclear cells (PBMC) through reverse transcription polymerase chain reaction (PCR) amplification and cloned into the pCAGGS vector with a Cterminal hemagglutinin (HA) tag by EcoRI/XhoI restriction sites.

| Confocal microscopy assay
HeLa cells (2 × 10 5 ) were seeded on glass slides 10 h before transfection and transfected with Lipofectamine 3000 reagent (Thermo Fisher).Generally, 200-500 ng of each plasmid was transfected in each slide in the confocal assay.Thirty-six hours later, the cells were washed, fixed with 4% paraformaldehyde, and permeabilized with 0.1% Triton X-100 for 5 min.After washing, the cells were blocked with 5% bovine serum albumin (BSA).Cells were then incubated with an anti-HA primary antibody (at a dilution of 1:1000) for 1 h at RT.After three washes, cells were stained with an Alexa Fluor 647-conjugated secondary antibody (at a dilution of 1:2000) for 1 h at room temperature (RT).Following three washes with PBS, cells were stained with 4',6-diamidino-2-phenylindole (DAPI) for 30 min, washed for 1 h at RT, and used for confocal microscopy assay (ZEISS LSM700).At least 100 random cells per slide were analyzed, and the most representative images from each slide were selected for presentation.

| Flow cytometry assay
The bimolecular fluorescence complementation (BiFC) assay was applied to verify MARCH-furin intracellular interaction.The MARCH1-VN/MARCH2-VN/MARCH8-VN and Furin-VC expression vectors were cotransfected into 293T cells.After 36 h, cells were collected, and 1 × 10 4 cells were selected for flow cytometry analysis (BD Biosciences company).The MARCH-VN/Furin-VC protein pairs produced green fluorescence (GF) when approached closely, which indicated their cellular interaction.Transfected cells containing the GF were sorted and labeled in a ratio accounting for the total cells.

| Western blot (WB) assay
Moreover, 293 T cells (5 × 10 5 or 2 × 10 6 ) were seeded on six-well plates or 10 cm dishes 12 h before transfection.Thirty-six hours posttransfection, cells were lysed with radioimmunoprecipitation assay (RIPA) buffer (Sigma) and cleared by centrifugation.The supernatants were collected, mixed gently with loading dye, and boiled in a water bath for 5 min.Subsequently, the samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene difluoride (PVDF) membrane, which was blocked with 4% nonfat milk for 1 h at RT.After washing, the membrane was incubated with the primary antibody (generally with a dilution of 1:5000) and horseradish peroxidase (HRP)-conjugated secondary antibody (with a dilution of 1:10 000).Alternatively, the membrane was directly incubated with an HRP-conjugated primary antibody (with a dilution of 1:10 000).
Finally, the PVDF membrane was exposed to enhanced chemiluminescence (ECL).

| Antibodies
Mouse anti-actin, mouse anti-HA, and mouse anti-FLAG monoclonal antibodies were purchased from Sigma.HRP-conjugated anti-mouse immunoglobulin G secondary antibody was purchased from Pierce.Alexa Fluor 647-conjugated secondary antibody was purchased from Thermo Fisher.HRP-conjugated anti-HA and anti-FLAG primary antibodies were purchased from Roche and Sigma, respectively.The ECL detection kit was purchased from Thermo Fisher.

| Co-immunoprecipitation (Co-IP) assay
In addition, 2 × 10 6 293 T cells were seeded in 10 cm culture dishes and transfected 12 h later.Forty-eight hours posttransfection, cells were lysed with RIPA buffer and cleared by centrifugation.The supernatants were collected and incubated with anti-FLAG antibodyconjugated magnetic beads (Sigma) overnight at 4°C.Then, the beads were collected with a magnetic rack, repeatedly washed, and used for WB assay.

| Viral infectivity assay
Moreover, 5 × 10 5 293T cells were seeded on six-well plates 12 h before transfection.In addition, 293T cells were cotransfected with FLAG-GP-ΔMLD/VSV-G, pNL-Luc-ΔEnv, and MARCH expression vectors.Forty-eight hours posttransfection, virion-containing supernatants were collected to infect the indicator Vero cells.Before infection, supernatants were subjected to ultracentrifugation at 40 000 × g for 1 h, and pellets were used for WB or P24 Gag enzyme-Linked Immunosorbent Assay quantity assay.Vero cells were lysed, and the supernatants were collected and used for luciferase activity assay.

| Luciferase activity assay
One day before viral infection, 1 × 10 4 Vero cells were seeded on 96well plates.Virion-containing supernatants were then added to Vero cells.After 48 h, the cell culture was terminated.Cells were washed with PBS three times and lysed with 200 μL of RIPA buffer.In addition, 100 μL of cell lysate was taken, mixed with an equal amount of substrate (Bright-Glo luciferase assay system; Promega) for 5 min at RT (avoid light exposure), and the luciferase activity was measured using an EnSpire (PerkinElmer) microplate reader.

| Human MARCH1 and MARCH2 inhibited EBOV GP proteolytic processing
Human MARCH1 and MARCH2 showed antiviral activities, [26][27][28] and human MARCH8 suppressed EBOV GP proteolytic processing. 29,30quence alignment indicated that human MARCH1 shared a high (64%), whereas human MARCH2 was a low (20.5%)homology to human MRAH8 (Figure 1).However, they all shared a similar membraneassociated structure. 23Thus, we aim to test whether mammalian MARCH1 and MARCH2 could block EBOV GP cleavage.We cloned MARCH1 and MARCH2 from humans, cattle, and mice.We first trace these mammalian MARCH proteins' intracellular distribution.We inserted the green fluorescent protein (GFP) in MARCH proteins' C-terminus as our previous work. 29Confocal analysis indicated that these MARCH proteins showed membrane-associated and scattered punctate distribution (Figure 2), similar to the intracellular sublocalization of MARCH8, as previously reported. 29Next, we sought to verify the anti-EBOV GP activities of these MARCH proteins.As shown in Figure 3A, in the presence of human MARCH1/2, no or less GP 1 was produced, while the GP 0 was less affected, as previously described for human MARCH8, 29 indicating that human MARCH1 and MARCH2 also suppressed GP processing.Similarly, bovine MARCH2 and murine MARCH1 blocked GP cleavage (Figure 3A).Intriguingly, murine MARCH2 nearly lost such inhibition effects.Notably, the expression of GP 0 and GP 1 was significantly impaired in the presence of bovine MARCH1 (Figure 3A), different from the other MARCH proteins.To further confirm human MARCH1 and MARCH2 blocking effects, we employed two mutants, human MARCH1-W114A and MARCH2-W97A, which proved to lose their antiviral activities, 26 to verify the above phenotype.As indicated in Figure 3B, the MARCH1-W114A and MARCH2-W97A lost such inhibition effects as compared to their wild-type, which suppressed GP cleavage in a dose-dependent manner (Figure 3C).In addition, knockdown of human MARCH1 and MARCH2 restored GP cellular cleavage (Figure S1).These lines of evidence indicated that human MARCH1 and MARCH2 inhibited EBOV GP intracellular cleavage.
We then applied the GP-ΔMLD-mediated pseudovirus infection system to check whether the MARCH proteins could block viral infection. 29Compared to the control, viral infectivity was significantly suppressed in the presence of the MARCH proteins except for murine MARCH2 (Figure 3D), which matched their blocking effects on GP cleavage.Importantly, human MARCH1 and MARCH2 also inhibited VSV-G pseudotyped virion infectivity in a dose-dependent manner (Figure 3E), similar to a previous report. 26EBOV could infect nearly all cell types in vivo, and its primary infection targets mainly included monocyte-derived macrophages (MDMs), dendritic cells, and liver cells.Therefore, in this study, we used Huh-7 cells to evaluate the antiviral role of MARCHs, and we demonstrated that human MARCH1 and MARCH2 suppressed EBOV GP cleavage in human liver Huh-7 cells (Figure 3F).

| Human MARCH1 and MARCH2 retained EBOV GP at the TGN apparatus
Generally, mature GP 1,2 is presented on the cell surface, enveloped into virions, and responsible for binding to target cell receptors in EBOV new round of infection.Human MARCH1 and MARCH2 reduced both GP 1 expression and GP 1,2 -mediated viral infectivity.We thus deduced that human MARCH1 and MARCH2 downregulated GP 1 cell surface display.
To confirm this hypothesis, the GP-ΔMLD-GFP and human MARCH1 or MARCH2 were cotransfected in HeLa cells and their protein localization was analyzed by confocal microscopy assay.GP-ΔMLD-GFP displayed a plasma membrane (PM) localization, and the human MARCH1 and MARCH2 proteins showed scattered punctate distribution (Figure 4A).
In the presence of human MARCH1 or MARCH2, the GP-ΔMLD-GFP was downregulated from the PM and was retained intracellularly (Figure 4A).These results demonstrated that human MARCH1 and MARCH2 trapped GP-ΔMLD-GFP at an intracellular compartment, where it was unable to be transported to the PM.To further verify this case, we introduced the BiFC system to determine where the GP was  S2).In the presence of MARCH1/2/8, EBOV GP-ΔMLD was trapped intracellularly, consistent with the former results (Figure 4B).Further detection indicated MARCH1-VN/MARCH2-VN/MARCH8-VN and GP-ΔMLD-VC pair groups showed minimal overlap with CNX-mCherry but displayed a strong co-localization with TGN-mCherry (Figure 4B).Therefore, we concluded human MARCH1 and MARCH2 retained GP at the TGN, MARCH8 did. 29

| MARCH1 and MARCH2 inhibited furin-cleavage of EBOV GP
Human MARCH8 suppressed furin-cleavage of EBOV GP. 29 Therefore, based on the above results (Figure 2B), we further verified whether human MARCH1 and MARCH2 could block furincleavage of EBOV GP.We first applied the BiFC system to detect human MARCH1/MARCH2/MARCH8 and furin intracellular interaction.MARCH1-VN/MARCH2-VN/MARCH8-VN and furin-VC (FR-VC) were cotransfected into 293T cells.Flow cytometry analysis showed MARCH1-VN/MARCH2-VN/MARCH8-VN and FR-VC had strong intracellular interaction signals (Figure 5A).Then, MARCH1-VN/MARCH2-VN/MARCH8-VN and FR-VC were cotransfected into HeLa cells.The results displayed that human MARCH1/2/8 and furin had a clear intracellular colocalization (Figure S3).Next, the Co-IP assay was performed to confirm their intracellular interactions in 293T cells.The furin expression vector was cotransfected with human MARCH1/MARCH2 and EBOV GP-ΔMLD expression vectors to verify furin's binding to MARCH1/2 and EBOV GP.As shown in Figure 5B, human MARCH1/2 and GP-ΔMLD could be co-immunoprecipitated by furin, indicating they formed an intracellular complex, similar to the results obtained in human MARCH8 as previously identified. 29Relying on these results, we identify whether these human MARCH proteins blocked furincleavage of GP.Generally, mature GP 0 was cleaved into GP 1 (~130 kDa) and GP 2 (~26 kDa) by endogenous furin.However, in the F I G U R E 5 MARCH proteins suppressed EBOV GP intracellular cleavage by furin.(A) MARCH1/2/8-VN/furin-VC BiFC pairs were cotransfected into 293T cells.After 36 h, cell samples were collected and pretreated for flow cytometry assay (Sony MA900 sorter instrument).Cells contained the green fluorescence were sorted and labeled a ratio accounting for the total cells.(B) FLAG-tagged furin was cotransfected with HA-tagged MARCH1/2 and HA-tagged GPΔMLD expression vectors in 293T cells.Samples were immunoprecipitated with anti-FLAG and used for WB assay.Furin was detected with anti-FLAG, and MARCH1/2 and GPΔMLD were verified by anti-HA.(C) GP was coexpressed with furin and MARCH1/2/8 in 293T cells.GP and furin were detected by anti-FLAG, and MARCH1/2/8 were detected with anti-HA.BiFC, bimolecular fluorescence complementation; EBOV, Ebola virus; FR, furin; GP, glycoprotein; HA, hemagglutinin; MARCH, membrane-associated RING-CH; M1, human MARCH1; M2, human MARCH2; M8, human MARCH8; SP, signal peptide.presence of exogenous furin, mature GP 0 was cleaved into three parts, that is, the GP 1 , GP 1 *, and GP 2 (Figure 5C, Lane 2), as previously demonstrated. 29Full-length GP was coexpressed with furin in the absence or presence of these human MARCH proteins.
Like human MARCH8, human MARCH1 and MARCH2 retained EBOV GP at the TGN (Figure 4B) and blocked its transportation to the PM, thus reducing its cell surface presentation (Figure 4A).MARCH1 but not MARCH2 displayed a relatively high sequence homology to MARCH8 (Figure 1).However, they all inhibited EBOV GP cleavage, implying protein conformation was critical to these MARCH proteins' CTI antiviral effects.Interestingly, both EBOV GP 0 and GP 1 were devoid of expression in the presence of bovine MARCH1 (Figure 3A).It is uncertain in which step bovine MARCH1 blocked EBOV GP expression, for example, at the transcription or expression processes, and further work is needed to clarify this accurate mechanism.In contrast to its human and bovine paralogue, murine MARCH2 lost blocking effects on EBOV GP cleavage (Figure 3A), similar to a previous report of MARCH2 inefficiency in inducing MLV envelope GP degradation. 27Given the high sequence similarity between human/bovine and murine MARCH2, it would be expected to identify which domains are critical for their gain or loss of function in the CTI antiviral mode.
Furin cleaves a group of host intracellular proproteins, which regulate many critical biological processes. 38It also recognizes many viral envelope GPs, including HIV-1 Env, EBOV GP, IAV H5N1 HA, and SARS-CoV-2 S. 38,39 Viral GPs cleavage by furin or other proprotein convertases (PCs) is important for their target cell infections.This study demonstrated that human MARCH1, MARCH2, and MARCH8 29 specifically interacted with furin (Figures 5A,B and 6), thus blocking EBOV GP cleavage (Figure 5C).antiviral mode among these human MARCH proteins (the antiviral mode is summarized in Figure 7).Notably, furin is an autocleavage protein, and mature furin is cleaved by itself at the TGN or the cell surface and then secreted outside the cells.Our results showed that intracellular furin expression was enhanced in the presence of human MARCH proteins (Figure 5C, Lane 3/4/5), because these human MARCH proteins blocked furin autocleavage activity and retained it at the TGN, thus leading to its intracellular accumulation.Future work is needed to elucidate this mechanism.
Besides MARCH8, 29 many host antiviral proteins also target furin to block virus replication. 40In this report, we demonstrated that human MARCH1 and MARCH2, and maybe the bovine MARCH2 and murine MARCH1, also inhibited furin-cleavage of EBOV GP.Thus, targeting cellular furin or other PCs to restrict enveloped viruses' infections may be a conserved defense mode employed by host antiviral factors.
Only recently, Liu et al. reported that M2 protein of IAV H1N1 evaded MARCH8-induced degradation through its cytoplasmic tail K78/K79 variation. 33Senecavirus A (SVA) 2AB was found to induce MARCH8 autophagic degradation. 45Human MARCH1, MARCH2, and MARCH8 are highly expressed in MDMs, 24,26 which are critical infection targets of EBOV.Thus, to effectively replicate in MDMs, whether EBOV needs to overcome these human MARCH proteins' restriction deserves further investigation. 46

| CONCLUSION
Funding information Natural Science Funding Key Program of Yibin Vocational and Technical College, Grant/Award Number: ZR22ZD-01; Science and Technology Planning Project of Yibin Vocational and Technical College, Grant/Award Number: YBZYSC22BK07; Nanchong Vocational and Technical College for Basic Scientific Research, Grant/Award Number: ZRA1904; Science and Technology Planning Project of Nanchong, Grant/Award Number: 22XCZX0003 origins showed a relatively conserved feature in blocking EBOV GP cleavage, which could provide clues for subsequent MARCHs antiviral studies and may facilitate the development of novel strategies to antagonize enveloped virus infection.K E Y W O R D S cleavage, Ebola virus, envelope glycoprotein, furin, MARCH 1 | INTRODUCTION Currently, a group of host cell membrane-associated proteins has been discovered to target viral envelope glycoproteins (GPs), including interferon (IFN)-induced transmembrane (IFITM) proteins, serine incorporator (SERINC) and MARCH proteins. 1These host innate immunity factors exerted their antiviral activities at multiple stages of viral replication.

1
Sequence alignment of human MARCH1, MARCH2, and MARCH8.The human MARCH8 amino acid (AA) sequence was aligned with that of human MARCH1 and MARCH2.MARCH1 and MARCH2 AA Sequence identical to MARCH8 were marked in black region and dots.Identical AA sequences of MARCH1/MARCH8, MARCH2/MRACH8, and MARCH1/MARCH2 were indicated in the dark regions and dots.Dashes represented sequence deletion.M1, human MARCH1; M2, human MARCH2; M8, human MARCH8.F I G U R E 2 Subcellular localization of mammalian MARCH proteins.Human/bovine/murine MARCH1/2-GFP and human MARCH8-GFP vectors were transfected into HeLa cells.Nuclei were stained with DAPI, and fluorescent signals were detected by confocal microscopy.DAPI, 4',6-diamidino-2-phenylindole; GFP, green fluorescent protein; MARCH, membrane-associated RING-CH.F I G U R E 3 MARCH proteins blocked EBOV GP cleavage.(A) GP (full-length) was cotransfected with human/bovine/murine MARCH1/2 or human MARCH8 in 293T cells.GP was detected with anti-FLAG, and MARCH proteins were detected with anti-HA.(B) GP and human MARCH1/2/8 or the human MARCH1-W114A/MARCH2-W97A expression vectors were cotransfected into 293T cells.GP was detected with anti-FLAG, and MARCH proteins were detected with anti-HA.Compared to the control (Lane 1), wild-type MARCH8, MARCH1, and MARCH2 potently restricted EBOV GP proteolytic cleavage (Lane 2/3/5).Under the expression of two human MARCH mutants, EBOV GP restored their intracellular cleavage (Lane 4/6).(C) Human MARCH1 and MARCH2 inhibited EBOV GP cleavage in a dose-dependent manner.GP and human MARCH1/2/8 expression vectors were cotransfected into 293T cells.GP was detected with anti-FLAG, and MARCH proteins were detected with anti-HA.As the expression of human MARCH1 and MARCH2 elevated, EBOV GP cleavage gradually declined.(D) MARCH proteins reduced EBOV GP-mediated pseudovirion infectivity.GPΔMLD, pNL-Luc-ΔEnv, and MARCH expression vectors were cotransfected into 293T cells.Supernatants containing the pseudovirions were collected to infect Vero cells.Infectivity was indicated as luciferase values.(E) Human MARCH1 and MARCH2 reduced VSV-G-mediated pseudovirion infectivity in a dose-dependent manner.VSV-G, pNL-Luc-ΔEnv, and MARCH expression vectors were cotransfected into 293T cells.Supernatants containing the pseudovirions were collected to infect Vero cells.Infectivity was indicated as luciferase values.(F) GP (full-length) was coexpressed with human MARCH1/2 in human Huh-7 cells.GP was detected with anti-FLAG, and MARCH proteins were detected with anti-HA.The statistical analysis was performed via one-way analysis of variance and paired t test.BM, bovine MARCH; EBOV, Ebola virus; GP, glycoprotein; HM, human MARCH; MARCH, membrane-associated RING-CH; MM, murine MARCH; ns, not significant.*p < 0.05; **p < 0.01; ***p < 0.001.

3. 4 |
Human MARCH1, MARCH2, and MARCH8 recognized furin P domainAs the Co-IP assay and flow cytometry analysis indicated (Figure5A,B), human MARCH1, MARCH2, and MARCH8 interacted with furin.Furin CD and P domains are critically important for substrate cleavage.To identify which domain was necessary for recognition by human MARCH proteins, we utilized two previously created furin mutants,29 that is, the furin CD deletion mutant (FRΔCD) and the furin P deletion mutant (FRΔP), to perform the Co-IP assay.As shown in Figure6, human MARCH1, MARCH2, and MARCH8 could bind to both full-length furin and the FRΔCD but rarely recognize the FRΔP, implying these MARCH proteins recognize furin P domain.

4 |
DISCUSSION MARCH proteins have recently been shown to inhibit viral infections by targeting their envelope GPs.Generally, MARCH proteins exerted their antiviral functions through two different modes, that is, the CTD and CTI.Human MARCH8 blocked EBOV GP and SARS-CoV-2 S GP cleavage and glycosylation maturation through the CTI antiviral mode.
In summary, we revealed a conserved CTI antiviral mode across mammalian species.MARCH1/2 of different mammalian origins inhibited EBOV GP proteolytic cleavage, which provides new insights for current antiviral studies and may facilitate the development of novel antiviral therapeutics.AUTHOR CONTRIBUTIONS Changqing Yu, Yuanzhe Bai, Wenbo Tan, and Yu Bai contributed to the concept and design of this work, acquisition, analysis, interpretation of data, and manuscript preparation.Xuemei Li, Yulong Zhou, Jingbo Zhai, and Mengzhou Xue contributed to analysis, discussion, or acquisition of data; Qiang Liu, Yan-Dong Tang, and Chunfu Zheng contributed to the concept of this work, acquisition of data, images acquisition, analysis and interpretation of F I G U R E 7 Antiviral model of human MARCH1/2/8 in restricting EBOV GP.MARCH proteins were located at membrane-associated compartments, including the ER, TGN, and PM.EBOV GP experiences mature glycosylation modification at the TGN and is cleaved by furin into GP 1 and GP 2 subunits.GP 1 and GP 2 are relinked and transported from the TGN to the PM.However, in the presence of MARCH proteins, furin interacted with both MARCHs and EBOV GP, forming a MARCH-furin-EBOV GP complex at the TGN.Resultantly, EBOV GP cleavage/ glycosylation maturation was blocked and was trapped by MARCH proteins at the TGN, which blocked its translocation from the TGN to the PM.EBOV, Ebola virus; ER, endoplasmic reticulum; GP, glycoprotein; MARCH, membrane-associated RING-CH; M1, human MARCH1; M2, human MARCH2; M8, human MARCH8; PM, plasma membrane; SP, signal peptide; TGN, trans-Golgi network.data, and manuscript revision.All authors approved the submitted version.