IFITM3:n restriktiokyky perustuu N-terminaalisiin S-palmitylaatioihin

https://www.life-science-alliance.org/content/3/1/e201900542

Open Access

Bat IFITM3 restriction depends on S-palmitoylation and a polymorphic site within the CD225 domain

View ORCID ProfileCamilla TO Benfield  Correspondence email, Farrell MacKenzie, Markus Ritzefeld, View ORCID ProfileMichela Mazzon, Stuart Weston, View ORCID ProfileEdward Tate, Boon Han Teo, Sarah E Smith, View ORCID ProfilePaul Kellam, Edward C Holmes, View ORCID ProfileMark Marsh Abstract

Host interferon-induced transmembrane proteins (IFITMs) are broad-spectrum antiviral restriction factors. Of these, IFITM3 potently inhibits viruses that enter cells through acidic endosomes, many of which are zoonotic and emerging viruses with bats (order Chiroptera) as their natural hosts. We previously demonstrated that microbat IFITM3 is antiviral. Here, we show that bat IFITMs are characterized by strong adaptive evolution and identify a highly variable and functionally important site—codon 70—within the conserved CD225 domain of IFITMs. Mutation of this residue in microbat IFITM3 impairs restriction of representatives of four different virus families that enter cells via endosomes.

 This mutant shows altered subcellular localization and reduced S-palmitoylation, a phenotype copied by mutation of conserved cysteine residues in microbat IFITM3. Furthermore, we show that microbat IFITM3 is S-palmitoylated on cysteine residues C71, C72, and C105, mutation of each cysteine individually impairs virus restriction, and a triple C71A-C72A-C105A mutant loses all restriction activity, concomitant with subcellular re-localization of microbat IFITM3 to Golgi-associated sites. Thus, we propose that S-palmitoylation is critical for Chiropteran IFITM3 function and identify a key molecular determinant of IFITM3 S-palmitoylation.

Introduction

Interferon-induced transmembrane proteins (IFITMs) are antiviral factors that act uniquely and early in viral replication cycles to restrict the entry of a diverse range of primarily enveloped viruses into cells (1). Humans possess three IFN-inducible IFITM genes—IFITM1, IFITM2, and IFITM3—encoding proteins with antiviral functions and two IFITM family members that lack antiviral function—IFITM5 and IFITM10. Mice have orthologs of all these IFITMs as well as two additional genes, Ifitm6 and Ifitm7. Phylogenetic analysis of vertebrate IFITMs indicates that IFITM1, IFITM2, and IFITM3 group with murine ifitm6 and ifitm7 in a clade of immunity-related IFITMs (IR-IFITMs), with IFITM5 and IFITM10 falling as separate lineages (2). IFITMs belong to the CD225/pfam04505 or “dispanin” protein superfamily (http://pfam.xfam.org/family/PF04505) (3) that contains more than 2,000 members, including both prokaryotic and eukaryotic proteins, all of which encode a conserved CD225 protein domain.

As their name suggests, IFITMs are membrane proteins, allowing them to police the cell surface and endocytic membranes that viruses must cross to invade cells. Studies of IFITM topology suggest a type II transmembrane configuration with a cytosolic N terminus, cytosolic conserved intracellular loop (CIL) domain, transmembrane domain, and extracellular (or intraluminal) C terminus (4, 5), although there is evidence that other IFITM topologies exist (6, 7, 8). The results of spectroscopic topological studies agree with the type II transmembrane configuration, as do bioinformatic predictions of IFITM3 secondary structure that reveal three alpha helices, with the C-terminal helix forming a single transmembrane domain (9, 10). The CD225 domain is highly conserved among IFITMs and comprises an intramembrane domain (IMD) and CIL domain. The hydrophobic IMD contains a 10-residue amphipathic helix (amino acid residues 59–68 of human IFITM3) that is required for the antiviral activity of both IFITM3 and IFITM1 (9). The subcellular localization of IFITMs is a key determinant of their antiviral profile. When expressed singly, IFITM3 and IFITM2 preferentially localize to early and late endosomes and lysosomes, restricting viruses that enter via these endolysosomal compartments. In contrast, IFITM1 primarily localizes at the cell surface and can restrict viruses that enter through the plasma membrane (11, 12, 13, 14). Indeed, mutants of IFITM3 that lack an N-terminal endocytic sorting motif ²⁰YEML²³ localize to the plasma membrane and lose their ability to inhibit influenza A virus (IAV), alphavirus, and coronavirus infection by endosomal routes (14, 15, 16, 17, 18).

Studies focusing on IFITM3 restriction of IAV and Semliki Forest virus (SFV) indicate that virus internalization is unaffected by IFITM3 expression and, for SFV at least, the viral envelope glycoprotein undergoes low pH-induced conformational changes (14). However, for both viruses, the viral core components are not delivered to the cytoplasm, suggesting that membrane fusion fails. Experiments with IAV indicate that hemifusion (i.e., lipid-mixing between viral and cellular membranes) can occur in the presence of IFITM3, but the subsequent formation of a fusion pore is inhibited (13, 19). Recent work has shown that IFITM3-positive vesicles fuse with incoming virus-bearing vesicles before hemifusion and that IFITM3 enhances the rate of virus trafficking to lysosomes (20). The co-localization of viral cargo with IFITM3-positive endosomes is specific to restricted viruses, suggesting that IFITM-insensitive viruses such as Lassa virus enter via different endosomal compartments and thereby escape IFITM engagement and restriction (13, 20). Further examples of virus-specific IFITM action include the ability of murine IFITM6 to restrict filoviruses, but not IAV (21), and amino acids within the IFITM3 CIL domain that are preferentially needed for IAV but not dengue virus restriction (22). Other post-entry mechanisms for IFITM3 restriction have also been proposed (23, 24, 25).

IFITMs are heavily regulated by posttranslational modifications (PTMs).
(PTM esimerkisi Ihmisen IFITM3 sekvenssi lähteestä
https://www.ncbi.nlm.nih.gov/protein/NP_066362.2. kopioin sekvenssin:

ORIGIN      
        1 mnhtvqtffs pvnsgqppny emlkeeheva vlgaphnpap ptstvihirs etsvpdhvvw
       61 slfntlfmnp cclgfiafay svksrdrkmv gdvtgaqaya stakclniwa lilgilmtil
      121 livipvlifq ayg
//

 One major modification is S-palmitoylation, a reversible 16-carbon lipid PTM that increases protein hydrophobicity and influences the behavior of proteins in membrane environments (26). For human and murine IFITM3, S-palmitoylation can occur on cysteine residues 71, 72, and 105 and enhances IFITM3 antiviral activity (27, 28). Recent live-cell imaging showed that abrogating C72 palmitoylation slowed IFITM3 trafficking to membrane compartments containing IAV particles (20). Multiple zinc finger DHHC (Asp-His-His-Cys) domain–containing palmitoyltransferases (ZDHHCs) can palmitoylate IFITM3 with marked functional redundancy, although ZDHHC20 may be particularly important (29). For human IFITM3, C72 is also the dominant site for acylation (30). Three other PTMs have also been reported, all of which negatively regulate IFITM3 antiviral activity: ubiquitination on one or more of four lysine residues (27), methylation on K88 (31), and phosphorylation on Y20 (15, 16). IFITM3 also forms homo- and hetero-oligomers. Although these are thought to require amino acids F75 and F78 (22), a recent study reported that these residues are required for antiviral activity but not for IFITM3-IFITM3 interactions (32). In tissue culture systems at least, IFITM3, IFITM2, and IFITM1 restrict viruses in a cooperative manner, with the extent of cooperativity and redundancy between IFITMs varying for different viruses (20, 22). However, the biochemical mechanisms and molecular determinants by which IFITM proteins restrict virus infection are still far from clear.

In mice and humans, there is evidence that IFITM3 influences host antiviral resistance. Ifitm3 knockout mice show enhanced morbidity after infection with IAV, alphaviruses, and flaviviruses (33, 34, 35, 36), and in humans, single nucleotide polymorphisms in IFITM3, which may act by altering IFITM3 expression or subcellular distribution, have been associated with an increase in morbidity in IAV and HIV-1 infections (recently reviewed by references 37 and 38). Antiviral function has also been reported for IFITMs from diverse mammalian and avian species (39, 40, 41, 42, 43, 44), although little is known about the role of IFITMs in antiviral defense in these species and in shaping host range. Phylogenetic studies have shown that the IFITM gene family is evolutionarily conserved and characterized by gene duplication, copy number variation, and the presence of pseudogenes (2, 45). Whereas humans possess a single IFITM3 gene, a remarkable multiplicity of IFITM3-like genes exists in the genomes of some other primates; for example, there are 25 IFITM3-like variants in the marmoset and eight in the African green monkey (2, 46). Interestingly, several of these variants contain N-terminal polymorphisms which, when inserted into human IFITM3, prevent ubiquitination and endocytosis; however, the function of these duplicated IFITM3-like genes has not been tested (46).

The antiviral responses of bats (mammalian order Chiroptera) are of particular interest because these animals have been increasingly recognized as reservoir hosts from which numerous viruses have ultimately emerged, with severe pathogenic and socioeconomic consequences in humans and livestock. Indeed, a recent analysis showed that bats host a significantly higher proportion of zoonotic viruses than other mammalian orders (47). Moreover, Chiroptera are the only mammalian order that harbor significantly more zoonotic viruses than predicted from reporting effort and host traits such as geographic range and mammal sympatry. It is possible that this propensity to be reservoirs for a large number of viruses, many of which may remain asymptomatic, in part reflects aspects of bat immunology (48, 49, 50, 51, 52). Hence, the study of Chiropteran antiviral effectors, such as IFITMs, can potentially reveal mechanisms of viral tolerance as well as the evolutionary signatures of virus–host co-evolution. We previously showed that microbat IFITM3 retains sequence motifs for endocytosis and PTM, traffics to the plasma membrane before endocytic uptake, co-localizes with endosomal markers, and at normal expression levels in primary microbat cells, inhibits infection by pH-dependent enveloped viruses (39).

Here, we performed evolutionary analyses of mammalian IFITMs, including those identified from bats, to shed light on IFITM function and the nature of past selection pressures, and to identify key amino acids for experimental studies of IFITM function.

Results

Phylogenetic analysis of Chiropteran IFITM genes

Results

Phylogenetic analysis of Chiropteran IFITM genes

Viral restriction activity is conserved in microbat IFITM3 (39). However, Chiropteran IFITMs have not been included in previous phylogenetic analyses, although bats comprise ∼20% of all mammalian species. Accordingly, we first analyzed the phylogenetic relationships of IR-IFITM gene sequences from 31 eutherian mammal species, including 13 species of bat. The genes analyzed were identified either via translated BLAST (tBLAST) or from our cDNA analyses (rapid amplification of cDNA ends [RACE] on cells from Myotis myotis, Eptesicus serotinus, and Sus scrofa and proteomics informed by transcriptomics (53) for Pteropus alecto). The resulting phylogeny revealed relatively high levels of sequence divergence (nucleotide substitutions per site) between the IFITM genes (Fig 1). Notably, Chiropteran IFITMs formed a monophyletic group separated from other taxa by a relatively long branch.