key: cord-0930926-zsui1s62
authors: Qian, Jin; Le Duff, Yann; Wang, Yimeng; Pan, Qinghua; Ding, Shilei; Zheng, Yi-Min; Liu, Shan-Lu; Liang, Chen
title: Primate lentiviruses are differentially inhibited by interferon-induced transmembrane proteins
date: 2015-01-01
journal: Virology
DOI: 10.1016/j.virol.2014.10.015
sha: e4c0846438059d270c635e3aeadb06828b4dfa3b
doc_id: 930926
cord_uid: zsui1s62

Interferon-induced transmembrane (IFITM) proteins inhibit the entry of a large number of viruses. Not surprisingly, many viruses are refractory to this inhibition. In this study, we report that different strains of HIV and SIV are inhibited by human IFITM proteins to various degrees, with SIV of African green monkeys (SIV(AGM)) being mostly restricted by human IFITM2. Interestingly, SIV(AGM) is as much inhibited by human IFITM2 as by IFITM3 of its own host African green monkeys. Our data further demonstrate that the entry of SIV(AGM) is impaired by human IFITM2 and that this inhibition is overcome by the cholesterol-binding compound amphotericin B that also overcomes IFITM inhibition of influenza A viruses. These results suggest that IFITM proteins exploit similar mechanisms to inhibit the entry of both pH-independent primate lentiviruses and the pH-dependent influenza A viruses.

Interferon-induced transmembrane (IFITM) proteins inhibit a wide range of viruses (reviewed in (Diamond and Farzan, 2013; Perreira et al., 2013) ). Humans have five IFITMs including IFITM1, 2, 3, 5 and 10 among which IFITM1, 2 and 3 exhibit antiviral activities (Hickford et al., 2012) . These three IFITMs are ubiquitously expressed in different tissues, and respond to type I interferon stimulation. IFITM5 is strictly expressed in osteoblasts and has a role in bone mineralization (Moffatt et al., 2008) . The function of IFITM10 remains unknown. Many important human pathogenic viruses are sensitive to IFITM restriction. These include influenza A virus, flaviviruses, Ebola virus, SARS coronavirus, Rift Valley fever virus, reovirus, human immunodeficiency virus type 1 (HIV-1), etc. (Anafu et al., 2013; Brass et al., 2009; Huang et al., 2011; Jiang et al., 2010; Lu et al., 2011; Mudhasani et al., 2013) . The importance of IFITM proteins in host antiviral defense is demonstrated by the high mortality of ifitm3-knockout mice infected with influenza A virus and by the possible association of an SNP in the human ifitm3 gene with the disease severity caused by influenza virus infection (Bailey et al., 2012; Everitt et al., 2012; Wakim et al., 2013; Zhang et al., 2013) .

IFITMs are small transmembrane proteins containing 120 to 135 amino acids (Siegrist et al., 2011) . IFITM2 and 3 share higher homology as compared to IFITM1, which has a relatively shorter N-terminal region and a longer C-terminal region. IFITM proteins have two predicted transmembrane (TM) domains. However, recent data suggest that only the C-terminal TM domain of IFITM3 crosses the membrane, whereas the N-terminal one serves as an intramembrane domain (IMD) (Bailey et al., 2013; Jia et al., 2012; Yount et al., 2012) . This IMD likely associates with the cytoplasmic leaflet of the lipid bilayer with the help of palmitoylated cysteine residues (Yount et al., 2012) . This type of membrane topology allows the cytoplasmic exposure of a large portion of IFITM sequences that may interact with cellular factors and machineries that collectively modulate the functions of IFITMs. One example is the 20-YEML-23 motif of IFITM3 that interacts with the adaptor protein AP-2 and regulates IFITM3 endocytosis from the plasma membrane en route to late endosomes (Chesarino et al., 2014; Jia et al., 2012) . The K24 residue of IFITM3 is a major site of ubiquitination. This modification affects IFITM3 subcellular localization and antiviral activity (Yount et al., 2012) . The IFITM proteins inhibit virus infection by impairing virus entry (Feeley et al., 2011) . Two models of inhibition have been proposed. One model suggests that IFITM proteins interfere with membrane hemifusion . This model is supported by the results that IFITM proteins suppressed cell membrane hemifusion that was created by low pH at cold temperature. This inhibition was rescued by oleic acid that promotes membrane hemifusion . The second model proposes that IFITM proteins impede the formation of viral fusion pore (Desai et al., 2014) . In support of this latter model, no effect was detected on lipid mixing between viral membrane and endosomal membrane upon IFITM3 overexpression. Yet, the release of viral genome into the cytoplasm was blocked by IFITM3 (Desai et al., 2014) . Both models are consistent with the notion that IFITM proteins can modulate the biophysical property of lipid bilayer, such as membrane fluidity and curvature, through mechanisms that possibly involve the interaction of IFITMs with VAPA and the disruption of intracellular cholesterol homeostasis (Amini-Bavil-Olyaee et al., 2013) . In support of this mechanism, a cholesterol-binding agent amphotericin B overcomes the inhibition of influenza A virus infection by IFITM3 (Lin et al., 2013) .

Not all enveloped viruses are inhibited by IFITM proteins. Examples are lymphocytic choriomeningitis virus (LCMV), Lassa virus (LASV), Machupo virus (MACH), human papillomavirus, cytomegalovirus and adenovirus that are all resistant to IFITMs (Brass et al., 2009; Warren et al., 2014) . Among retroviruses, murine leukemia virus (MLV) is relatively refractory to IFITMs, the HIV-1 strain BH10 is inhibited, whereas another HIV-1 strain IIIB exhibits resistance (Brass et al., 2009; Lu et al., 2011) . In this study, we examined a panel of HIV and SIV strains for their sensitivity to IFITM inhibition. The results revealed various degrees of inhibitions ranging from no inhibition for HIV-1 A/G to 10-fold inhibition for SIV of African green monkeys.

In order to evaluate the susceptibility of different primate lentiviruses to inhibition by IFITM1, 2 and 3, we selected the following viruses for study, including three HIV-1 strains (the laboratory adapted strain NL4-3, primary isolate YU-2 and the circulating recombinant form A/G), one HIV-2 strain (HIV-2 Rod ), five SIV strains from chimpanzees (SIV CPZ1.9 ), African green monkeys Chlorocebus sabaeus (SIV AGM-sab ) and Chlorocebus tantalus (SIV AGM-tan ), rhesus macaques (SIV MAC-1A11 ) and sooty mangabeys (SIV SMM ). Since these viruses either use CXCR4 or CCR5 as the coreceptor, we chose to infect the HIV indicator cell line TZM-bl that expresses both CXCR4 and CCR5 and are thus susceptible to infection by all these viruses. We first transduced TZM-bl cells with retroviral vectors expressing human IFITM1, 2 or 3 and selected the stably transduced cell lines with puromycin. Ectopic expression of IFITM1, 2 and 3 was confirmed by western blotting (Fig. 1A) . We then challenged these TZM-bl cell lines with different doses of HIV or SIV. Virus infection was monitored by measuring luciferase activity that was expressed under the control of HIV-1 LTR promoter in TZM-bl cells. The data report the effect of IFITM proteins on the early phase of HIV/SIV infection until viral Tat protein is produced. Fig. 1B shows the luciferase activities of one representative infection experiment that was performed with different doses of viruses. The averages of three independent experiments are summarized in Fig. 1C . The results showed that SIV AGM-tan was inhibited the most, whereas infection of HIV-1, SIV CPZ1.9 and SIV MAC were not profoundly affected by the three human IFITM proteins. On the basis of the degrees of inhibition, these primate lentiviruses are ranked as SIV AGM-tan 4SIV AGM-sab , SIV SMM , HIV-2 Rod 4HIV-1 NL4-3 , HIV-1 YU-2 , HIV-1 A/G 4SIV CPZ1.9 and SIV MAC . The results also revealed that IFITM2 was the most inhibitory, followed by IFITM3 and IFITM1.

Since IFITM proteins are known to inhibit virus entry (Feeley et al., 2011; Lu et al., 2011) , we asked whether the strong inhibition of SIV AGM by IFITM2 is a result of impaired virus entry. To this aim, we prepared the BlaM-Vpr-containing HIV and SIV particles, and used these virions to infect IFITM-expressing TZM-bl cells. The efficiency of virus entry was determined by measuring the cleavage of CCF2 by BlaM-Vpr that enters the cytoplasm together with viral cores. The results showed that the entry of HIV-1 NL4-3 and SIV MAC into TZM-bl was marginally affected by IFITM1, 2 or 3 ( Fig. 2) . In contrast, the entry of SIV AGM-tan and SIV AGM-sab , to a lesser extent SIV SMM , was strongly impaired by IFITM2 and IFITM3 (Fig. 2) . This similar reduction in the entry of SIV AGM-tan and SIV AGM-sab contrasts with a moderately stronger inhibition of SIV AGM-tan infection by IFITM2 and 3 as shown in Fig. 1 . This difference suggests that SIV AGM-tan may be inhibited not only at the entry step, but also at a downstream step until viral Tat is produced, which is measured in the assays shown in Fig. 1 .

We next asked whether the endogenous IFITM2 and 3 are able to inhibit the entry of SIV AGM . We first used shRNA to knock down IFITM2 and 3 in TZM-bl cells (Fig. 3A ). Both SIV AGM-tan and SIV AGM-sab showed significantly higher infection in the IFTIM2/3-knockdown cells (Fig. 3B ). Since the BlaM-Vpr containing SIV AGM-sab particles generated much stronger signals in the entry assay than SIV AGM-tan (Fig. 2) , we further measured the effect of IFITM2/3-knockdown on the entry of SIV AGM-sab . We also treated TZM-bl cells with IFNα2b to increase the expression of endogenous IFITM2 and 3. The results showed that IFNα2b reduced the entry of SIV AGM-sab by 2-fold and this diminution was completely lost when IFITM2 and 3 were depleted with shRNA ( Fig. 3C and D). When the endogenous IFITM2 and 3 were knocked down in a human T cell line called C8166 that constitutively express relatively high level of IFITM2, the entry of SIV AGM-sab increased by approximately 50% (Fig. 3E -G). IFNα2b treatment increased the expression of IFITM2 and 3, and results in a 40% reduction in SIV AGM-sab entry. Depletion of IFITM2 and 3 under IFNα2b treatment restored SIV AGM-sab entry to the control level ( Fig. 3E and F). We also observed that shRNA3 depleted IFITM2 much more efficiently compared to shRNA1 and shRNA2 ( Fig. 3E ), which correlates with a moderately greater entry of SIV AGM-sab in shRNA3-transdued C8166 cells than in shRNA1-or shRNA2transduced cells, albeit that this increase does not reach statistical significance ( Fig. 3F and G). Together, these data indicate that endogenous IFITM2 and 3 inhibit the entry of SIV AGM-sab .

It has been reported that amphotericin B prevents IFITM3 from inhibiting influenza A virus through modulating membrane fluidity (Lin et al., 2013) . We suspected that, if IFITM2 and 3 use the same mechanism to inhibit SIV AGM and influenza A virus, then amphotericin B should also rescue the infection of SIV AGM in IFITM2/3expressing cells. Indeed, when amphotericin B was added with increasing doses, the infection of both SIV AGM-sab and SIV AGM-tan in IFITM2 or IFITM3-expressing TZM-bl cells were restored to the level of infection in control cells ( Fig. 4A and B ). HIV and SIV are known as pH-independent viruses (McClure et al., 1988 ). Yet, the high sensitivity of SIV AGM to IFITM2 and 3 inhibition raises the possibility that this SIV may have become pH-dependent similar to the influenza A virus. Contrary to this speculation, SIV AGM showed resistance to the treatment of chloroquine or bafilomycin A1 both of which neutralize the pH in late endosomes and thus inhibit VSV G protein-mediated pH-dependent virus entry (Fig. 4C) . In further support of this observation, the dynamin inhibitor dynasore moderately stimulated the infection of SIV AGM (Fig. 4D) , which suggests that, in contrast to VSV that enters the cell via endocytosis and is thus sensitive to dynasore treatment, the entry of SIV AGM is independent of endocytosis. Together, these data indicate that IFITM2 and 3 exploit a similar mechanism to inhibit viruses that enter cells via different pathways.

We next asked whether SIV AGM is inhibited by IFITM proteins of its natural host, African green monkeys. To answer this question, we first cloned IFITM1 and IFITM3 from the African green monkey kidney cell lines called COS-7 and Vero. The agmIFITM1 exhibits a much higher homology to macaque IFITM1 than to IFITM1 of humans and chimpanzees (Fig. 5A ). Similar to human IFITM1, agmIFITM1 exhibited no or only weak inhibition of HIV or SIV (Fig. 5B ). Four agmIFITM3 sequences were identified that differ at amino acid positions 22 and 38 (Fig. 5A ), no distinct agmIFITM2 sequence was found. All four agmIFITM3 proteins inhibited HIV and SIV, with SIV AGM-tan and SIV AGM-sab being the most inhibited (Fig. 5B) . The agmIFITM3 appears to elicit a much stronger inhibition than human IFITM2 or IFITM3 (Fig. 1) , which might have enabled agmIFITM3 to suppress the infection of SIV AGM-tan and SIV AGM-sab to similar low levels. The differences at amino acid positions 22 and 38 affected the antiviral activity of IFITM3, agmIFITM3(I22H38) elicited the strongest inhibition (Fig. 5B) . Notably, the I22 residue is located within the 20-YEML-23 motif that directs the endocytosis of human IFITM3 (Chesarino et al., 2014; Jia et al., 2012; Jia et al., 2014) . This may explain the effect of M22I change on the antiviral function of agmIFITM3. Since SIV AGM is potently inhibited by both human IFITM2 and agmIFITM3, it is likely that SIV AGM is much more susceptible to IFITM restriction than other HIV and SIV strains regardless of the species origin of the IFITM proteins. 

Viruses of different families are affected by IFITM proteins to various degrees. Some viruses are strongly inhibited, examples are influenza A viruses, flaviviruses, Ebola viruses, and SARS coronaviruses (Brass et al., 2009; Huang et al., 2011) . Some are moderately inhibited, such as vesicular stomatitis virus . Some are resistant to IFITM, including LCMV, MACH, LASV, MLV, papillomavirus, cytomegalovirus and adenovirus (Brass et al., 2009; Warren et al., 2014) . Infection of human coronavirus OC43 is even stimulated by IFITM2 or 3 (Zhao et al., 2014) . Results of our study showed that members of the same virus family such as lentiviruses can be affected by IFITM proteins to markedly different degrees that range from virtually no inhibition of SIV CPZ1.9 and SIV MAC to a 10-fold restriction of SIV AGM by human IFITM2.

This observation appears to mirror the species-specific antiviral activity that has been described for some host restriction factors such as tetherin (Sauter et al., 2009) . Human tetherin does not inhibit HIV-1 because of the countering action of viral antagonist Vpu, but it can block the release of SIV AGM . In a similar vein, African green monkey tetherin potently inhibits HIV-1 owing to the inability of HIV-1 Vpu to antagonize this monkey tetherin. However, our results showed that agmIFITM3 inhibits SIV AGM more potently than human IFITM2 does, which suggests that SIV AGM may have not been sufficiently pressured to escape from the agmIFITM3 restriction. It is known that, in contrast to HIV and SIV infections that cause AIDS, SIV AGM or SIV SMM infection of their natural hosts is not pathogenic, which is at least partially attributable to the rapid resolution of innate immune responses at the acute infection stage and the lack of chronic immune activation (Harris et al., 2010; Jacquelin et al., 2014) . Therefore, after establishing the infection, SIV AGM or SIV SMM does not face the inhibition by interferon and ISGs including IFITM proteins. In the absence of selection by interferon, SIV AGM or SIV SMM may, in theory, develop any level of sensitivity to IFITM proteins. It is therefore not a surprise that SIV AGM is more restricted by IFITM3 compared to SIV SMM .

Human IFITM1, 2 and 3 often show different antiviral activities as a result of their sequence divergence (Lu et al., 2011) . Given the closer homology between IFITM2 and 3 than to IFITM1, the antiviral spectra of IFITM2 and 3 are better aligned than with IFITM1. This trend is also seen in this study. IFITM1 is a weaker inhibitor of HIV and SIV as compared to IFITM2 and 3.

We successfully cloned agmIFITM1 from COS-7 and Vero cells. We also cloned four agmIFITM3 variants that differ at amino acid positions 22 and 38. No distinct agmIFITM2 was found. Although we cannot completely rule out the possibility that African green monkeys do not have an ifitm2 gene, no ifitm2 gene has been described in the macaque genome except for two copies of ifitm3 genes. In other words, neither of the two copies of macaque ifitm3 genes has changed significantly enough to become a distinct ifitm2 gene as seen in humans and chimpanzees. It is possible that in different primate hosts, the ifitm3 duplicate have undergone different levels of selection and as a result, exhibit different levels of divergence. In some primates, like humans and chimpanzees, one ifitm3 copy has accumulated enough changes to become a distinct ifitm2 gene.

HIV and SIV are generally regarded as pH-independent viruses as opposed to pH-dependent viruses such as influenza A viruses that require low pH to accomplish viral membrane fusion (Earp et al., 2005) . We further confirmed the pH-independent nature of SIV AGM by showing its resistance to chloroquine and BafA1 both of which neutralize the pH of late endosomes. We also observed that SIV AGM infection is refractory to dynamin inhibitor dynasore, which suggests that SIV AGM does not need endocytosis to enter cells. Since both SIV AGM and influenza A viruses are potently inhibited by IFITM proteins, it is evident that the routes of virus entry do not necessarily determine the susceptibility of viruses to IFITM restriction. Since the cholesterol-binding compound amphotericin B counters the inhibition of SIV AGM and influenza A virus by IFITMs, IFITM proteins likely utilize the same mechanism to change the membrane fluidity at the virus entry portals and thus block viral membrane fusion.

In conclusion, our study demonstrates that IFITM proteins are able to potently restrict the pH-independent virus such as SIV AGM that does not require endocytosis for its entry. This observation suggests that IFITM proteins do not strictly depend on its localization to late endosomes for its antiviral action. Given the relative closeness of SIV AGM to other IFITM-resistant SIV strains such as SIV MAC , there exists an opportunity to identify the viral determinant behind this IFITM resistance phenotype.

The HEK293T, TZM-bl, COS-7 and Vero cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (Invitrogen) and penicillin/streptomycin (Invitrogen). TZM-bl is an HeLa-derived cell line that expresses CD4, CXCR4 and CCR5 and has a luciferase gene driven by HIV-1 LTR (Wei et al., 2002) . C8166 were grown in RPMI1640 medium supplemented with 10% fetal bovine serum and penicillin/streptomycin. Transfection of HEK293T cells was performed with Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.

HIV-1 YU-2 (GenBank M93258), SIV CPZ1.910 (GenBank EF394356), SIV AGM-tan (GenBank U58991), SIV AGM-sab (GenBank U04005), and SIV MAC-1A11 (GeneBank M76764) proviral clones were obtained through NIH AIDS Reagent Program. HIV-1 A/G and HIV-2 ROD clones were generously provided by Mark Wainberg. SIV SMM-E543 (GenBank U72748) clone was a gift from James Whitney. The human IFITM1, 2 and 3 cDNA sequences were cloned into the retroviral vector pQCXIP as described previously (Lu et al., 2011) . shRNA plasmids targeting human ifitm2 and 3 genes were purchased from Sigma and were used to produce lentiviral particles for transduction.

To produce HIV-1 (NL4-3, YU2, A/G), HIV-2 (ROD) and SIV (MAC, CPZ, AGM-sab, AGM-tan, SMM), HEK293T cells were transfected with 6 μg of each viral DNA. For the production of BlaM-Vpr-containing viruses, HEK293T cells were transfected with 6 μg of viral plasmid DNA and 2 μg of plasmid DNA encoding the BlaM-Vpr fusion protein. For the production of NL4-3 virus pseudotyped by the VSV G protein, HEK293T cells were transfected with 6 μg of NL4-3(env-) DNA and 0.2 μg of plasmid DNA encoding the G protein of VSV. Forty-eight hours after transfection, culture supernatants were collected and clarified by centrifugation in the CS-6R centrifuge (Beckman Coulter) at 3000 rpm for 30 min at 4 1C. Viruses were aliquoted and stocked at À80 1C. The amounts of viruses were determined by measuring the levels of viral reverse transcriptase activity.

TZM-bl cells were infected with retroviral particles expressing IFITM1, IFITM2 or IFITM3. Cells that stably express IFITMs were selected with puromycin (2 μg/ml). For assessment of IFITM1, IFITM2 and IFITM3 expression by western blot, TZM-bl cells were lysed in the cytobuster buffer (Novagen #71009-4) containing protease inhibitors (Roche #11 836 153 001) for 20 min on ice. Cell lysate was resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), blotted onto PVDF (polyvinylidene fluoride) membrane (Roche). IFITMs proteins, containing a Flag tag at their N-termini, were detected using monoclonal antibody against the Flag tag (Sigma #F1804) and horseradish peroxidase conjugated secondary antibodies (GE Healthcare).

Cells were lysed in the cytobuster buffer (Novagen) on ice for 20 min. After a brief spin, cell lysate was collected and mixed with 2 Â Laemmli buffer. Proteins were resolved by PAGE and transferred onto the PVDF membrane (Roche). Membranes were blocked for 1 h at room temperature in 4% milk and incubated with primary antibodies overnight at 4 1C. Anti-Flag antibody was purchased from Sigma, anti-IFITM2 and anti-IFITM3 antibodies from ProteinTech. After several washes in 1 Â phosphate buffer saline (PBS), membranes were further incubated with horseradish peroxidase-conjugated secondary antibodies (GE Healthcare). Signals were detected by electrochemiluminescence (ECL) and visualized by exposure to a Kodak film.

shRNA silencing of IFITM proteins in TZM-bl and C8166 cells TZM-bl or C8166 cells were transduced by lentiviral particles allowing the expression of shRNAs against ifitm2 or 3 mRNAs. Three shRNAs were able to deplete both IFITM2 and 3 proteins (TRCN0000057502, TRCN0000118022, TRCN0000118025 from Sigma) and were used in this study. Their sequences are:

shRNA1:CCGGCCTGTTCAACACCCTCTTCTTCTCGAGAAGAAGA-GGGTGTTGAACAGGTTTTTG;

shRNA2:CCGGCCTGTTCAACACCCTCTTCATCTCGAGATGAAGAG-GGTGTTGAACAGGTTTTTG;

shRNA3:CCGGGCTTCATAGCATTCGCCTACTCTCGAGAGTAGGCG-AATGCTATGAAGCTTTTTG.

Twenty-four hours after transduction, cells were treated with or without interferon α2b (IFNα2b) at 1000 U/ml for 24 h, followed by infection with BlaM-Vpr-containing viruses to assess virus entry efficacy.

Viral fusion assay was performed by infecting TZM-bl or C8166 cells as described previously (Lu et al., 2011) . In brief, cells were exposed to BlaM-Vpr containing viruses and polybrene (5 μg/ml), and spun for 45 min, 1800 rpm (CS-6R, Beckman Coulter) at 4 1C, then incubated at 37 1C for 2 h. Cells were washed once with CO 2 independent medium (Invitrogen), incubated with CCF2-AM containing medium for 1 h at room temperature, washed once and incubated with developing medium at room temperature for 16 h. The following day, cells were washed 2 times in phosphate buffered saline (containing 2% FBS), fixed with 3.7% formaldehyde and analyzed by flow cytometry (BD, Fortessa) to monitor the cleavage of CCF2-AM substrate by BlaM.

Cloning IFITM cDNA from African green monkey cells COS-7 and Vero cells are both African green monkey kidney cells. They were treated with IFNα2b for 16 h to induce IFITM expression, followed by extraction of total cellular RNA with Trizol (Invitrogen). Purified RNA was reverse transcribed and amplified using a one-step RT-PCR system (Roche) with following primers: AGM_IFITM1_F1: 5 0 -CAACAGGGGAAAGCAGGGCTC-3 0 AGM_IFITM1_F2: 5 0 -CAACACTTCTTTCCCCAAAGCCAG-3 0 AGM_IFITM1_R1: 5 0 -GTCATTGTGGACAGGTGTGTGGG-3 0 AGM_IFITM1_R2: 5 0 -CTGTATCTAGGGGCAGGACCAAG-3 0 AGM_IFITM2/3_F1: 5 0 -GGGAAAGGGAGGGCCCACTGAG-3 0 AGM_IFITM2/3_F2: 5 0 -CCCACTAACCCGACCACCGCTG-3 0 AGM_IFITM2/3_R1: 5 0 -GTGTGTGAGGATAAAGGGCTG-3 0 AGM_IFITM2/3_R2: 5 0 -GGGCAGAGCTCCTGGCCTGAATG-3 0 These primers were designed as follows. To clone African green monkey IFITM1 (agmIFITM1), we aligned the 5 0 and 3 0 untranslated regions (UTRs) of the ifitm1 genes of human, chimpanzee, pongo, and macaque that are available in Genbank. Two forward and two reverse primers are designed on the basis of the conserved sequences. Primers for agmIFITM2/3 cloning were similarly designed on the basis of the conserved UTRs of human ifitm2 and 3, chimpanzee ifitm2 and 3, and two macaque ifitm3 genes. To amplify IFITM1 or IFITM2/3 cDNA, four RT-PCR reactions were performed using the combinations of the 2 forward and 2 reverse primers. The amplified IFITM DNA fragments were cloned into the pQCXIP retroviral vector (Clontech) between the BamHI and EcoRI restriction sites and sequenced. A Flag tag was added to the N-termini of the agmIFITM proteins and the agmI-FITM cDNAs were cloned into pQCXIP for expression.

The antiviral effector IFITM3 disrupts intracellular cholesterol homeostasis to block viral entry

Interferon inducible transmembrane protein 3 (IFITM3) restricts reovirus cell entry

Ifitm3 limits the severity of acute influenza in mice

Interferon-induced transmembrane protein 3 is a type II transmembrane protein

The IFITM proteins mediate cellular resistance to influenza A H1N1 virus, West Nile virus, and dengue virus

Phosphorylation of the antiviral protein interferon-inducible transmembrane protein 3 (IFITM3) dually regulates its endocytosis and ubiquitination

IFITM3 restricts influenza A virus entry by blocking the formation of fusion pores following virus-endosome hemifusion

The broad-spectrum antiviral functions of IFIT and IFITM proteins

The many mechanisms of viral membrane fusion proteins

IFITM3 restricts the morbidity and mortality associated with influenza

IFITM3 inhibits influenza A virus infection by preventing cytosolic entry

Downregulation of robust acute type I interferon responses distinguishes nonpathogenic simian immunodeficiency virus (SIV) infection of natural hosts from pathogenic SIV infection of rhesus macaques

Evolution of vertebrate interferon inducible transmembrane proteins

Distinct patterns of IFITM-mediated restriction of filoviruses, SARS coronavirus, and influenza A virus

Innate immune responses and rapid control of inflammation in African green monkeys treated or not with interferon-alpha during primary SIVagm infection

The N-terminal region of IFITM3 modulates its antiviral activity by regulating IFITM3 cellular localization

Identification of an endocytic signal essential for the antiviral action of IFITM3

Identification of five interferon-induced cellular proteins that inhibit West Nile virus and dengue virus infections

IFITM proteins restrict viral membrane hemifusion

Amphotericin B increases influenza A virus infection by preventing IFITM3-mediated restriction

The IFITM proteins inhibit HIV-1 infection

Human immunodeficiency virus infection of CD4-bearing cells occurs by a pH-independent mechanism

Bril: a novel bone-specific modulator of mineralization

IFITM-2 and IFITM-3 but not IFITM-1 restrict Rift Valley fever virus

IFITMs restrict the replication of multiple pathogenic viruses

Tetherin-driven adaptation of Vpu and Nef function and the evolution of pandemic and nonpandemic HIV-1 strains

The small interferon-induced transmembrane genes and proteins

Enhanced survival of lung tissue-resident memory CD8( þ) T cells during infection with influenza virus due to selective expression of IFITM3

The antiviral restriction factors IFITM1, 2 and 3 do not inhibit infection of human papillomavirus, cytomegalovirus and adenovirus

Emergence of resistant human immunodeficiency virus type 1 in patients receiving fusion inhibitor (T-20) monotherapy

Interferoninduced cell membrane proteins, IFITM3 and tetherin, inhibit vesicular stomatitis virus infection via distinct mechanisms

S-palmitoylation and ubiquitination differentially regulate interferon-induced transmembrane protein 3 (IFITM3)-mediated resistance to influenza virus

Interferon-induced transmembrane protein-3 genetic variant rs12252-C is associated with severe influenza in Chinese individuals

Interferon induction of IFITM proteins promotes infection by human coronavirus OC43

We thank Mark Wainberg, Andres Finzi, and James Whitney for providing agents and valuable discussions. This study was supported by funding from Canadian Institutes of Health Research to CL (MOP-133479 and HVI-98828) and from National Institutes of Health to SLL (R01AI112381, R21AI105584, R21AI109464, and R56AI107095).