Origins and Evolution of tetherin, an Orphan Antiviral Gene

Abstract

Tetherin encodes an interferon-inducible antiviral protein that traps a broad spectrum of enveloped viruses at infected cell surfaces. Despite the absence of any clearly related gene or activity, we describe possible scenarios by which tetherin arose that exemplify how protein modularity, evolvability, and robustness can create and preserve new functions. We find that tetherin genes in various organisms exhibit no sequence similarity and share only a common architecture and location in modern genomes. Moreover, tetherin is part of a cluster of three potential sister genes encoding proteins of similar architecture, some variants of which exhibit antiviral activity while others can be endowed with antiviral activity by a simple modification. Only in slowly evolving species (e.g., coelacanths) does tetherin exhibit sequence similarity to one potential sister gene. Neofunctionalization, drift, and genetic conflict appear to have driven a near complete loss of sequence similarity among modern tetherin genes and their sister genes.

Figure 1. Location and architecture of TM-CC gene products proximal to tetherin in human and mouse genomes

(A) Diagram of genes surrounding human and mouse tetherin. (B) Organization of TM-CC genes and their protein products for human Tetherin (GenBank: NP_004326.1), mouse TM-CC(aT) (GenBank: XP_003945491.1), and human PV1 (GenBank: NP_112600.1) proteins. Glycosylation and cysteine residues are indicated as brown Y symbols and stars respectively. Numbers indicate amino acid positions. Structural features of TM-CC(aT) and PV1 are based on predictions using TMHMM, COILS, Pred-GPI and GlycoEP. See also Figure S1.

Figure 2. Antiviral activity of GPI-modified TM-CC(aT) and PV1 proteins

(A) Infectious virion yield measured using HeLa TZM-bl indicator cells following transfection with a Vpu-deficient HIV-1 proviral plasmid along with increasing amounts of the indicated unmodified (WT) or GPI-modified (+GPI) Tetherin, TM-CC(aT) or PV1 proteins (RLU= relative light units, mean ± SD, n=3). (B) Western blot analyses (anti-CA) of cell lysates and virions corresponding to (A). Numbers at the bottom represent virion CA protein levels relative to those obtained in the absence of an inhibitor. See also Figure S2.

Figure 3. Antiviral activity of divergent Tetherin/TM-CC-GPI proteins

(A) Tetherin/TM-CC-GPI Protein sequences from human (GenBank: NP_004326.1), mouse (GenBank: NP_932763.1), opossum (GenBank: XP_007489270.1 and XP_007489271.1), Tasmanian devil (GenBank: XP_012399618.1), turtle (GenBank: XP_008169758.1, XP_005279001.1 and XP_005279003.1), turkey (inferred from GenBank: XP_010723307.1), falcon (inferred from GenBank: XP_005444407.1 and Gnomon prediction: 2189215010.p), alligator (GenBank: XP_006017475.1 and XP_006017476.1), coelacanth (Gnomon prediction: 16424589.p), elephant shark (GenBank: XP_007897024.1) Tetherin/TM-CC-GPI proteins (see also Supplemental Information). The TM, CC domains and GPI anchor are indicated. Conserved residues are highlighted and predicted omega sites (GPI modification) are indicated in grey. (B) Infectious virion yield measured using HeLa TZM-bl indicator cells following transfection of Vpu-deficient HIV-1 proviral plasmids along with plasmids expressing Tetherin/TM-CC-GPI proteins. (RLU= relative light units, Mean ± SD, n=3). (C) Western blot analyses (anti-CA) of cell lysates and virions corresponding to (B). Numbers at the bottom represent virion CA protein levels relative to those obtained in the absence of an inhibitor. See also Figure S3.

Figure 4. Organization of pv1-proximal genes in vertebrates

Diagrams were generated using NCBI, UCSC, Ensembl Genome Browsers and sequence similarity approaches. Branches in gray indicate incomplete genome assemblies. Inclined figures indicate genes in incompletely assembled scaffolds. White and orange asterisks indicate genes that were active or inactive respectively in virion release-inhibition assays. Potential alternatively spliced versions of alligator and turtle tetherin are indicated by dotted lines. The duplicated loci in zebrafish are shown. Phylogeny and speciation dates were based on (Inoue et al., 2010; Janvier, 2006; Venkatesh et al., 2014). See also Figure S4.

Figure 5. Sequence similarity between PV1, TM-CC(aT) and Tetherin/TM-CC-GPI proteins

(A) Heat map showing e-values of all combinations of reciprocal BLASTp analyses using the PV1, TM-CC(aT) and Tetherin/TM-CC-GPI proteins in this study. (B, C and D) Phylogenetic trees of divergent mammalian Tetherin (B) Fish Tetherin/TM-CC-GPI (C) and other vertebrate Tetherin/TM-CC-GPI and TM-CC(aT) (D) protein sequences. Sequences with significant BLAST hits in (A) were used to construct each tree. Maximum likelihood tree was constructed using RAxML with 1000 bootstrap replicates. Nodes and branches in grey were supported by <80% of the bootstrap replicates. Asterisks indicate bootstrap support for each node. (*) ≥ 80%, (**) ≥ 90%, (***) ≥ 95%. Trees were midpoint rooted (indicated in blue). (E) Alignment of divergent TM-CC(aT) and Tetherin/TM-CC-GPI protein sequences from human (adapted from GenBank: XP_011526778.1), mouse (GenBank: XP_003945491.1), turkey (GenBank: XP_010723297.1), alligator (GenBank: KQL90195.1), turtle (adapted from GenBank: XP_008169839.1) and coelacanth (GenBank: XP_006001674.1 and XP_014347293.1, Gnomon prediction: 16424589.p and TM-CC(aT)_A adapted from RNAseq reads of NW_005819727.1). Sequences spanning the TM, CC domains and GPI anchor are indicated. Residues that comprise the proline-rich domain in human and mouse TM-CC(aT) proteins are indicated in red. See also Figure S5.

Figure 6. Antiviral activity of non-mammalian TM-CC(aT) variants

(A) Transcript structure and C-terminal protein sequences of potential alternatively spliced isoforms of tm-cc(at) in non-mammalian species. The TM, CC and proline-rich domains and hydrophobic patch are indicated in color. The omega site (underlined in blue) and specificity (1 – false positive rate) were predicted using PredGPI. The number of RNAseq reads supporting the occurrence or absence of splicing events are indicated between the exons. (B) Infectious virion yield measured using HeLa TZM-bl indicator cells following transfection of Vpu-deficient HIV-1 proviral plasmids along with plasmids expressing alternatively spliced isoforms of TM-CC(aT) proteins. NH= no hydrophobic (isoforms lacking the hydrophobic patch). (RLU= relative light units, Mean ± SD, n=3). (C) Western blot analyses (anti-CA) of cell lysates and virions corresponding to (B). Numbers at the bottom represent virion CA protein levels relative to those obtained in the absence of an inhibitor. See also Figure S6.

Figure 7. Possible evolutionary scenarios for the emergence of tetherin/tm-cc-gpi gene(s) in the pv1–cilp2 locus

(A) Tetherin/TM-CC-GPI originated once, prior to the division of sharks from other jawed vertebrate lineages via sequential duplications of pv1 and tm-cc(at). (B) Tetherin/TM-CC-GPI originated independently in multiple vertebrate lineages via duplications of pv1 and tm-cc(at). See also Figure S7.