key: cord-0258633-eksi9yia
authors: Hayward, Joshua A.; Tachedjian, Mary; Johnson, Adam; Irving, Aaron T.; Gordon, Tamsin B.; Cui, Jie; Nicolas, Alexis; Smith, Ina; Boyd, Vicky; Marsh, Glenn A.; Baker, Michelle L.; Wang, Lin-Fa; Tachedjian, Gilda
title: Unique Evolution of Antiviral Tetherin in Bats
date: 2020-11-12
journal: bioRxiv
DOI: 10.1101/2020.04.08.031203
sha: 3d0588de1fd7be113b77957bf6becc045bbd2694
doc_id: 258633
cord_uid: eksi9yia

Bats are recognised as important reservoirs of viruses deadly to other mammals, including humans. These infections are typically nonpathogenic in bats raising questions about host response differences that might exist between bats and other mammals. Tetherin is a restriction factor which inhibits the release of a diverse range of viruses from host cells, including retroviruses, coronaviruses, filoviruses, and paramyxoviruses, some of which are deadly to humans and transmitted by bats. Here we characterise the tetherin genes from 27 species of bats, revealing that they have evolved under strong selective pressure, and that fruit bats and vesper bats express unique structural variants of the tetherin protein. Tetherin was widely and variably expressed across fruit bat tissue-types and upregulated in spleen tissue when stimulated with Toll-like receptor agonists. The expression of two computationally predicted splice isoforms of fruit bat tetherin was verified. We identified an additional third unique splice isoform which includes a C-terminal region that is not homologous to known mammalian tetherin variants but was functionally capable of restricting the release of filoviral particles. We also report that vesper bats possess and express at least five tetherin genes, including structural variants, a greater number than any other mammal reported to date. These findings support the hypothesis of differential antiviral gene evolution in bats relative to other mammals.

Bats are reservoirs of viruses that are highly pathogenic to other mammals, including humans. Viruses 18 such as Hendra, Nipah, Ebola, Marburg, severe acute respiratory syndrome coronaviruses (SARS-CoV -19 1 and likely also SARS-CoV-2), and the Sosuga virus have crossed species barriers from bats into 20 humans (1-5). Laboratory studies demonstrate that specific bat species can be infected with Ebola, 21

Marburg, SARS-CoV-1, Hendra, and Nipah viruses without showing clinical signs of disease (6-10). 22 Introduction 3 in immune signalling events by triggering the NF-κB signalling pathway, leading to stimulation of the 1 antiviral interferon response (30) (31) (32) (33) . 2

In humans, tetherin is expressed across most cell types, including its BST-2 namesake bone marrow 3 stromal cells, and its expression is upregulated by stimulation with type I interferons (34) (35) (36) . Tetherin 4 is a dimeric dual-anchor type II membrane protein that contains one protein anchor, a transmembrane 5 domain near its N-terminus, an extracellular coiled-coil domain, and a glycophosphatidylinositol (GPI) 6 lipid anchor, which is attached to its C-terminus as a post-translational modification (37) (38) (39) . Tetherin 7 contains a number of conserved cytosine and asparagine motifs within the extracellular domain, with 8 respective roles in dimerisation and glycosylation, and a dual-tyrosine motif (Y·x·Y) in its cytoplasmic 9 region, which has a role in viral particle endocytosis and immune signalling cascades (31, 40) . Tetherin 10 is located in lipid rafts at the plasma membrane where many viral particles bud during acquisition of 11 their host membrane-derived viral envelopes (39) . During the viral budding process one anchor 12 remains embedded in the nascent viral envelope while the other remains attached to the plasma 13 membrane, tethering the virion to the cell and preventing its release into the extracellular 14 environment (28, 38) . 15 Tetherin is common to all mammals, and orthologs which share structural and functional, but not 16 sequence similarity, exist in other vertebrates (29) . Most mammals carry only a single tetherin gene; 17 however gene duplication has been observed in sheep, cattle, opossums and wallabies (29, 41) . 18 Furthermore, human tetherin is expressed in two alternative isoforms (30) , the long (l-tetherin) and 19

short (s-tetherin) isoforms, which differ through the truncation of 12 amino acid (AA) residues at the 20 N-terminus of l-tetherin. Computational analysis of the genomes of megabats from the Pteropus genus 21 predict two additional isoforms, X1 and X2, with an internal, rather than N-terminal difference in 22 amino acid sequences, although whether these isoforms are expressed is unknown (12) . The predicted 23 isoform X1 is homologous to human l-tetherin, and the X2 isoform is a splice variant which contains a 24 7 AA exclusion within the extracellular coiled-coil domain relative to isoform X1. A recent analysis of 25 tetherin in the fruit bats Hypsignathus monstrosus and Epomops buettikoferi revealed that these 26 species expressed a homolog of Pteropus isoform X1 and that it is a functional restriction factor 27 capable of inhibiting the release of the bat-hosted Nipah virus, as well as virus-like particles (VLPs) 28 produced from the Ebola virus and the human immunodeficiency virus (23) . 29 Introduction retroviral VLPs, in addition to the two computational predictions, tetherin isoforms X1 and X2, which 1 were generated through the automated NCBI annotation pipeline from the published P. alecto 2 genome [NCBI: PRJNA171993] (12) . We report that microbats of the suborder Yangochiroptera, 3 genus Myotis, possess at least five tetherin genes, two of which contain large structural differences 4 in the extracellular coiled-coil domain. An analysis of the tetherin genes of 27 species of bats 5 revealed that bat tetherin has been subjected to strong positive selection, indicating that these 6 flying mammals have experienced evolutionary pressure on this antiviral gene. Collectively our 7 findings indicate that bats have undergone tetherin gene expansion and diversification relative to 8 other mammals, supporting the hypothesis of differential antiviral gene evolution in bats. 9

Bats possess structural homologs of human tetherin 11 To advance our understanding of bat tetherin homologs, we mined all available publicly accessible 12 sequence read archives representing 26 species of bats through a BLASTn analysis (Table 1) 1A & 1C) from a multiple sequence alignment (SI Data 1). The primary sequence lengths range from 20 177 to 221 AA, compared to 180 AA for human l-tetherin, and of these only 23 residues were 21 conserved in 100% of the bat tetherin sequences analysed ( Figure 1B ). The assembled tetherin 22 sequence for all bats was a homolog of human l-tetherin, except for the Black-bearded Tomb bat 23 (Taphozous melanopogon), which was predicted to possess an equivalent of the human s-tetherin. 24

The predicted T. melanopogon tetherin nucleotide sequence contains two key mutations upstream of 25 its short isoform start site, an ATG to CTG change in the long isoform start codon, and a TAG stop 26 codon in between the two start codon sites, either of which would be sufficient to prevent the 27 production of the long isoform of tetherin. No nucleotide conflicts exist between the 17 reads 28 mapping to this region of the T. melanopogon tetherin sequence, indicating that these mutations are 29 unlikely to represent sequencing or consensus alignment errors. 30

The dual-tyrosine Y·x·Y motif, which is critical for mediating viral-particle endocytosis and is involved Results Y|C·x·Y|H ( Figure 1A and 1C). All bat species possess at least one tyrosine residue within this motif 1 ( Figure 1C ). Conservation of the protein domain organization and key structural motifs of bat tetherin, 2 despite significant amino acid sequence diversity, supports the present understanding of tetherin as 3 a protein whose functions are mediated through structural rather than sequence-mediated 4 interactions (28). 5

Bat tetherin genes are under significant positive selection 6 To assess the evolutionary pressure applied during the speciation of bats, a selection test was 7 performed to analyse bat tetherin genes spanning to ~65 million years ago (mya) of Chiropteran 8 history. The analysis revealed that overall, bat tetherin genes have been subjected to positive 9 selection, with an average ratio of non-synonymous (dN) to synonymous (dS) mutations, dN/dS, of 10 1.125 over the Chiropteran phylogeny, with numerous specific positions subjected to a large degree 11 of positive selection, with dN/dS values of 2.626 -2.668 (Table 2 ). These sites are predominantly 12 located in the transmembrane domain and the regions of the cytoplasmic and extracellular domains 13 immediately adjacent to the transmembrane domain ( Figure 1A ). This analysis suggests the presence 14 of evolutionary pressure exerted on tetherin, by viral antagonists or countermeasures that target 15 residues within tetherin, following the speciation of the Chiropteran order. The apparent ubiquity of 16 tetherin among bats, in concert with significant sequence diversity and evidence of strong positive 17 selection pressures operating on each species, supports the notion that tetherin plays a major role in 18 the antiviral repertoire of bats . 19 Fruit bats possess unique structural isoforms of the tetherin protein generated 20 through alternative splicing of a single tetherin gene 21 The expression of the fruit bat tetherin gene was initially assessed by a BLASTn search within the 22 transcriptome of an Australian fruit bat, the Black flying fox, Pteropus alecto, using the human tetherin 23 amino acid sequence as the search query. We identified 30 contigs matching the query (lowest E-value 24 = 9.21 x 10 -23 ). A single P. alecto homolog of tetherin, herein described as isoform A, homologous to 25 human l-tetherin and the predicted X1 isoform (GenBank: XM_006904279), was identified. P. alecto 26 tetherin isoform A has low primary amino acid sequence conservation (37%) compared to human l-27 tetherin ( Figure 2A ) and other mammalian (28 -47%) tetherin sequences homologous to human l-28 tetherin ( Figure 2B ). Despite the low amino acid sequence identity, the predicted secondary structures 29 and protein domains are largely conserved between bats and other mammals ( Figure 2B ). All tetherin 30 proteins containing the cytoplasmic (CD), transmembrane (TM), extracellular domains (ED), and the To verify the expression of bat tetherin isoform A, a cDNA library was prepared from P. alecto spleen 3 tissue, comprising several immune cell types, and assessed for the presence of transcripts matching 4 that identified in the P. alecto transcriptome. Primers were designed to flank the open reading frame 5 of tetherin isoform A, and amplicons were generated by PCR. Amplicons were then cloned into 6 plasmid vectors for DNA sequencing. This analysis revealed two additional isoforms of tetherin, 7

isoforms B & C. Isoform B was homologous to the computationally predicted isoform X2 (GenBank: 8 XM_015587122). Both isoforms B and C were predicted to contain structural differences relative to 9 isoform A. They both possess a 7 AA exclusion in the middle of the extracellular domain while isoform 10 C was additionally predicted to contain an alternative C-terminus, without the GSP domain ( Figure  11 2B). The absence of a GSP domain suggests that it is unlikely that parallel homodimers of isoform C 12 would possess the capacity to tether viral particles to the cell. 13

Mapping the tetherin cDNA sequences against the P. alecto genome identified a single tetherin gene 14 To further investigate the diversity of tetherin in non-fruit bat species, we amplified the tetherin 25 nucleotide sequences for the vesper bat species Myotis ricketti and Miniopterus schreibersii using 26 primers designed on the basis of the transcriptome sequences (Table 3A) . These primers were used 27 for PCR of cDNA samples generated from M. ricketti spleen tissue and a M. schreibersii kidney cell line 28 to amplify the tetherin coding domain sequences (Table 3A) . Primers for M. ricketti were also used to 29 amplify the tetherin coding region from cDNA generated from a Myotis macropus kidney cell line as it 30

was reasoned that since these species belong to the same genus, primers designed for one might be 31 capable of amplifying tetherin from the other. Following PCR, amplified DNA from M. ricketti, These microbats were found to express various homologs of tetherin that include encoding of unique 1 structural variations ( Figure 4A and 4B). The tetherin homolog of human l-tetherin, predicted for M. 2 schreibersii, was detected ( Figure 4A ). Five unique tetherin variants, differing in their encoded amino 3 acid sequences (tetherin A -E), were identified for M. macropus, two of which, tetherin A and B, were 4 also detected in M. ricketti ( Figure 4A ). These tetherin variants encode three distinct homologs 5 (tetherin A, C, D) of the human l-tetherin, sharing the same protein domains but differing in their 6 amino acid sequence ( Figure 4B ). Pairwise comparisons ( Figure 5) domain, transmembrane domain, and less than half of the extracellular domain ( Figure 4B ). 15

Accordingly, it is predicted to lack the GPI signal peptide required for the post-translational addition 16 of a GPI-anchor, also found for P. alecto tetherin isoform C ( Figure 2B ). The absence of a GPI anchor 17

indicates that it is unlikely that homodimers of M. macropus tetherin C isoform B would possess the 18 capacity to tether viral particles to the cell. M. macropus tetherin B and E were found to be structurally 19 unique, possessing a large deletion of ~60 amino acids in the extracellular coiled-coil domain ( Figure  20 4B), compared to the other tetherin homologs, which has been shown to be critical for viral particle 21 restriction (28, 42) . This deletion results in the exclusion of the conserved disulphide bond-forming 22 cysteine residues, and the conserved asparagine-linked glycosylation sites, indicating that this form of 23 tetherin is unlikely to form dimers or be glycosylated in the manner of human tetherin (28, 43 comprising seven potential M. lucifugus tetherin genes ( Figure 6A ) and indicating the presence of a 3 tetherin gene locus within this scaffold. Further analysis of the mapped exons, as described below, 4 enabled confirmation of the M. lucifugus tetherin regions. The M. lucifugus tetherin gene locus was 5 located between the MVB12A and PLVAP genes, as observed for the single P. alecto tetherin gene 6 ( Figures 3 and 6) . 7

The genome assembly of M. davidii also contained significant matches to M. macropus tetherin 8 sequences, however these were spread across seven different gene scaffolds against which no M. ENSMLUG00000023562 (18) . Herein these genes are labelled Tetherin 1, 2, 6, and 7, respectively 24 ( Figure 6A ) . 25 Tetherin expression pattern in P. alecto tissue samples 26 Tetherin is widely and variably expressed across human tissues and is upregulated in response to 27 interferon (35, 36) . To assess the relative expression of bat tetherin across different tissue types, tissue 28 samples were obtained from three individual P. alecto bats: a male, a pregnant female, and a juvenile 29 male. Using primers that amplify all three isoforms of P. alecto tetherin, expression was analysed by All tetherin coding sequences were modified with the addition of a haemagglutinin (HA) tag sequence 28 (N-SGYPYDVPDYAGS-C) to enable antibody-based detection of protein expression. In all cases the HA 29 tag was inserted at the position immediately following the end of the coiled-coil region of the 30 extracellular domain ( Figure 1A ) which is the equivalent location to that previously utilised in the 31 tagging of human tetherin (28) .

To evaluate tetherin protein expression, mammalian HEK293T cells, which do not express tetherin in 1 the absence of type I interferon stimulation (21) , were transfected with bat tetherin expression 2 constructs. Cell lysates were extracted using a method for enhanced extraction of GPI-anchored 3 proteins (48) and tetherin expression was observed by non-reducing SDS-PAGE and Western blot 4 analysis ( Figure 10 ). 5 P. alecto tetherin isoforms A and B were detected primarily as broad bands at the expected positions 6 ~56-75 kDa while isoform C was present as both dimers and predominantly, monomers (~32-35 kDa) 7

( Figure 10A ). The presence of broad/multiple bands in close proximity to each other likely reflects 8 variable levels of tetherin glycosylation as previously reported for human tetherin (28) . M. macropus 9 tetherin A was also detected as a dimer at the expected size of ~55 kDa while tetherin B was only 10 detected as a monomer with a size of ~26 kDa ( Figure 10B ). Bat tetherin that are homologous to 11 human l-tetherin dimerised as expected, while the structurally unique tetherin proteins behaved 12 differently. In this regard, P. alecto tetherin isoform C formed both dimers and monomers, while M. 13 macropus tetherin B was detected exclusively as a monomer. These data show that HA-tagged 14 tetherin can be expressed and that these proteins differ in their ability to dimerise. 15

Human tetherin localises to the plasma membrane at sites of virus budding, in addition to membranes 16 of the trans-Golgi network and recycling compartments (21, 49) . To determine localisation of P. alecto 17 tetherin isoforms, HEK293T cells were transfected with constructs expressing HA-tagged tetherin that 18 was visualised by immunofluorescence. P. alecto tetherin isoform A localised to the plasma 19 membrane, displaying a similar cellular localisation pattern as human tetherin ( Figure 11A and 11C), 20 while isoform C was predominantly localised in the cytoplasm ( Figure 11E ). P. alecto tetherin isoform 21 B demonstrated a fluorescence pattern with features shared by both isoforms A and C ( Figure 11D ). 22

These data show that bat tetherin isoforms can be expressed with isoforms A and B, but not C 23 predominately localising at the plasma membrane. 24 P. alecto and M. macropus tetherin proteins display distinct ability to restrict 25 HIVΔVpu virus-like particles 26 Tetherin from humans and other mammals including cats, sheep, and other bat species (Hypsignathus 27 monstrosus and Epomops buettikoferi) restrict the release of HIV-1 VLPs from cells (23, 50, 51) . To 28 examine if tetherin from P. alecto and M. macropus function to similarly block viral particle release, 29

we assessed their ability to restrict HIV-1 VLPs. HEK293T cells were co-transfected with constructs 30 expressing bat tetherin and HIVΔVpu VLPs, that do not express Vpu, an antagonist of human tetherin 31 (21) . P. alecto tetherin isoforms A and B inhibited the release of HIVΔVpu VLPs in contrast to tetherin HIVΔVpu VLP release from HEK293T cells; however tetherin B was unable to inhibit HIVΔVpu VLP 1 egress ( Figure 12B ). These data indicate that, with the exception of P. alecto tetherin isoform C and 2 M. macropus tetherin B, tagged bat tetherin proteins are functionally capable of restricting the release 3 of HIV-1 particles from mammalian cells in the absence of Vpu. 4

The structurally unique P. alecto tetherin isoform C restricts the release of filoviral 5 virus-like particles 6 We next investigated whether P. alecto tetherin isomers are able to restrict filovirus VLPs, including 7 isoform C, which contains a unique C-terminal domain compared to isomers A and B ( Figure 2B ). In 8 contrast to experiments with HIV-1 VLP, tetherin isoform C was able to restrict the release of VLPs 9 composed of Ebola and Marburg virus VP40 matrix proteins (P = 0.008 for both; Figure 13A & 13B). 10 Ebola VLPs were restricted by isoforms A, B, and C to similar extents ( Figure 13A ), while Marburg VLPs 11

were restricted by isoform C to a lesser extent than isoforms A and B ( Figure 13B ). These data 12 demonstrate that all P. alecto isomers were able to restrict filoviral VLPs including isomer C which 13 lacks the GSP domain. 14

Bats are increasingly being recognised as hosts of viruses with zoonotic potential, driving efforts to 16 better understand these host-virus relationships and the evolutionary features of bats that 17 differentiate them from other mammals (14, 17, 52, 53) . It has been hypothesised that the 18 evolutionary adaptation to flight, including changes to the DNA damage response, increased 19 metabolic rates, and higher body temperatures, have influenced the immune system of bats in such a 20 way as to make them ideal hosts for viruses (12, 54, 55) . To determine the differences between the 21 innate antiviral defences of bats relative to other mammals, we analysed the genes and expressed 22 transcripts of tetherin from diverse bat genera and species within the order Chiroptera. We found 23 that in all but one species, bats possess genes that express transcripts encoding a tetherin protein 24 homologous to the long isoform of human tetherin (l-tetherin). In addition, we found that P. alecto 25 expresses three isoforms from a single tetherin gene, and that vesper bats (genus Myotis) encode five, 26 and possibly as many as seven, distinct tetherin genes. 27

The one bat species that lacked the human l-tetherin homolog, T. melanopogon, possessed an 28 equivalent of the short isoform of human tetherin (s-tetherin), an observation also reported for cats 29 (56) . In a study of feline tetherin, short isoform-exclusivity was found to improve viral restriction and 30 decrease sensitivity to tetherin antagonism by HIV-1 Vpu when compared to an engineered long adapted to the feline host, is not restricted by either long or short isoforms of feline tetherin (56) . In 1 humans, l-tetherin acts as a virus sensor which induces NF-κB signalling in contrast to s-tetherin, which 2 lacks the dual tyrosine motif required for eliciting this innate immune response (30, 31) . The findings 3 presented here suggest that if bat tetherin proteins are confirmed to similarly mediate cytoplasmic 4 domain-mediated immune signalling, then T. melanopogon, which only encodes a homolog of 5 s-tetherin, would be predicted not to mediate this effect. 6

Bat tetherin amino acid sequences were found to be highly variable. The predicted protein lengths 7 range from 177 to 221 AA (human l-tetherin is 180 AA) and of these, only 23 amino acid residues were 8 conserved in 100% of the 27 bat tetherin sequences analysed ( Figure 1B ). Among these are the 9 structurally important cysteine and asparagine residues which are responsible for tetherin 10 dimerisation and glycosylation, respectively (57, 58) . The dual-tyrosine motif, responsible for 11 mediating viral particle endocytosis and immune signalling (31, 40) , was found to exist as a variable 12

Y|C·x·Y|H motif across the bat tetherin variants analysed. All bats maintained at least one of the two 13 tyrosine residues. This observation is significant because mutational studies of human tetherin have 14 demonstrated that the dual tyrosines provide redundancy for both the endocytic and signalling 15 activities, which are maintained as long as either tyrosine is present (31, 40) . 16 To understand the evolutionary selective pressures on tetherin genes across bat species, a selection 17 test was performed that revealed that bat tetherin genes are under strong positive selection with 18 amino acid positions in and around the transmembrane domain being the region under the strongest 19 selective pressure (Table 2 and Figure 1A ). This is in agreement with previous reports that primate 20 tetherin possess multiple sites under positive selection in this same region (28) . The driver of positive 21 selection in the case of primate tetherin is antagonism by viral counter-measures including HIV-1 Vpu 22 (59) . Venkatesh et al. (60) demonstrated that the configuration that tetherin dimers adopt during 23 viral particle retention primarily consists of the GPI-anchor being embedded in the viral envelope and 24 the transmembrane domain remaining attached to the cellular membrane. If this paradigm holds true 25 for bat tetherin, then it follows that the tetherin-cell membrane interface is the major site of tetherin 26 antagonism in bats and it would be reasonable to speculate that the drivers of this selection are viral 27

antagonists analogous in the mode of interaction, if not structure or function, to lentiviral tetherin 28

antagonists. 29 We amplified tetherin from spleen-derived cDNA of an Australian fruit bat, P. alecto, and confirmed 30 the expression of the two computationally predicted splice variants (isoforms A [X1] and B [X2]), and 31 additionally identified the expression of a third isoform of tetherin, isoform C. Mapped against the P.

( Figure 3B ). P. alecto tetherin isoform B possesses a 7 AA exclusion within the extracellular coiled-coil 1 domain relative to isoform A, while isoform C is predicted to harbour the same 7 AA exclusion and an 2 alternative C-terminus, that lacks the GSP domain predicted to be present in tetherin isoforms A and 3 B which is necessary for the post-translational addition of the GPI-anchor. This is important because 4 studies of human tetherin have demonstrated that the presence of a GPI-anchor is essential for 5 restricting the release of viral particles (28, 42) . Sheep and cows possess a duplication of the tetherin 6 gene (41, 51) . In sheep, the duplicate tetherin, named tetherin B, similarly does not encode a GSP, and 7 studies of sheep tetherin B function reveal that while it is capable of limited restriction of VLPs, it is 8 significantly less potent than sheep tetherin A (51) . The mechanism of its function is unknown and its 9 C-terminal amino acid sequence is entirely dissimilar from that of P. alecto tetherin isoform C. One Tetherin was expressed widely and variably across the tissues of P. alecto, which is consistent with 29 previous observations of human tetherin (36) . We observed the highest levels of P. alecto tetherin 30 expression in the thymus, possibly reflecting the role of the thymus in expressing a large proportion 31 of the proteome due to its role in central tolerance. High levels of expression were also observed in 32 lung tissue, particularly in the lung of an individual female bat. This may reflect a frontline defensive Tetherin was upregulated in spleen tissue treated with TLR agonists LPS and PIC. Interestingly, we 1 observed that the expression of isoforms A, B, and C was highly variable among individual spleen 2 samples. In several spleens, only isoform B was expressed, while in one spleen treated with PIC the 3 vast majority of expression was attributed to isoform A. The remainder expressed variable levels of 4 each isoform A and B, with the majority of expression trending toward isoform B. Surprisingly, isoform 5 C was only expressed, at a low level, in a single spleen sample that had been treated with PIC. The 6 biological importance of this observation is not presently known. Ongoing assessments of bat tetherin 7 expression should determine if this strong bias in alternative isoform expression is present in other 8 tissues, such as the thymus and lung. 9

The expression of the P. alecto tetherin isoforms A, B, and C, and M. macropus tetherin A and B, reveals 10 differences in their relative capacities to form homodimers under our assay conditions ( Figure 10 ). 11

This observation is notable because the ability of tetherin to form cysteine-linked dimers is required 12

for the restriction of HIV-1 viral particles (58) which we chose as the VLPs against which bat tetherin 13 proteins were functionally validated for inhibiting viral particle release. In contrast to restriction of 14 HIV-1, tetherin dimerisation is not required for the restriction of arenaviruses and filoviruses (61), 15 suggesting that the need for dimerisation is virus-dependent. 16 The P. alecto tetherin isoforms A and B, and M. macropus tetherin A, that were most similar to human 17 l-tetherin, were all found to predominantly form dimers. This was determined by non-reducing SDS-18 PAGE and Western blot analysis where bands were observed almost exclusively at the expected size 19 of tetherin dimers ( Figure 10 ). In contrast, P. alecto tetherin isoform C was present as both dimeric 20 and monomeric forms ( Figure 10A ). The presence of monomeric forms is surprising because P. alecto 21 tetherin isoform C contains all of the conserved cysteine residues and the same extracellular coiled-22 coil domain as isoforms A and B. The distinct C-terminal region in isoform C is the result of an 23 alternative splicing event that also causes the loss of the final three amino acid residues of the coiled-24 coil region possessed by isoforms A and B. Additionally, the inserted HA tag is located in the position 25 immediately following the coiled-coil domain. It is possible that the absence of the three terminal 26 amino acid residues in the coiled-coil domain of isoform C relative to isoform A and B accounts for the 27 difference in relative extents of dimerisation. Alternatively, the observed extent of dimerisation of 28 isoform C may be an artefact of the method used to process the samples for analysis and may not be 29 reflective of the extent of dimerisation in situ. It is also important to note that we cannot currently 30 rule out the possibility that the insertion of the HA tag in between the coiled-coil domain and the 31 alternative C-terminal sequence might be responsible for the observed reduction in dimerisation. M. 32 macropus tetherin B was observed exclusively as a monomer ( Figure 10B ). The monomeric exclusivity for dimerisation within the 60 AA deletion in the extracellular coiled-coil domain of tetherin B (Figure  1 4). 2

Given the differing extent of homodimer formation for each bat tetherin and reports that dimerisation 3 is important for the restriction of some viruses and not required for others (58, 61) , it is likely that 4 these differences might affect the extent by which each tetherin is capable of restricting various VLPs. 5

Tetherin function was validated through an assessment of the capacity of P. alecto tetherin isoforms 6 A, B, and C, and M. macropus tetherins A and B to restrict the release of HIVΔVpu VLPs. While primate 7 lentiviruses are not known to natively infect bats, bats have recently been discovered to host extant 8 gammaretroviruses and deltaretroviruses, and bat genomes contain diverse endogenous retroviral 9

sequences (27, (62) (63) (64) (65) . P. alecto tetherin isoforms A and B inhibited the release of HIVΔVpu VLPs, in 10 contrast to isoform C which lacked restriction activity ( Figure 7A ). M. macropus tetherin A restricted 11

HIVΔVpu VLP release, while tetherin B did not ( Figure 7B ). These findings are consistent with the 12 expected effects of tetherin dimerisation on the restriction of HIVΔVpu VLPs (58) . 13

Previous reports on the necessity of the tetherin GPI anchor for inhibition of viral particle release (28, 14 42) indicates that the lack of a GPI anchor on isoform C would likely result in an inability to inhibit viral 15 egress from the host cell. However, sheep tetherin B, which does not possess a GPI-signal peptide 16 ( Figure 2B ), is capable of limited restriction of betaretroviral VLPs through an unknown mechanism 17 (51) . Because of the unique sequence of the C-terminus and lack of GSP in P. alecto tetherin isoform 18 C, we extended our evaluation of its restrictive capacity to include filoviral VLPs derived from Ebola 19 and Marburg virus VP40 matrix proteins. Surprisingly, P. alecto tetherin isoform C was capable of 20 restricting the release of Ebola VLPs to an extent similar to isoforms A and B, although it was marginally 21 less restrictive of Marburg VLPs compared to isoforms A and B ( Figure 13 ). 22

How isoform C could be capable of restricting the release of VLPs without a C-terminal GPI anchor is 23 not presently known, although there are at least two possible explanations. The first is that the 24 alternative C-terminal sequence of isoform C is involved in envelope or cellular membrane binding 25 through an unknown mechanism. The second is that isoform C forms dimers in a manner distinct from 26 that of isoforms A and B. One possibility is that it can form an antiparallel homodimer configuration 27 in such a way that the N-terminus of one monomer aligns with the C-terminus of another monomer, costs of existence and extension. Matching reads were downloaded and assembled in the same manner, and the assembled consensus sequence was used in a third SRA BLAST using the same parameters as the 1 second. This process was iteratively repeated for the assembled consensus sequence of each bat tetherin 2 until it extended through the tetherin coding domain in both directions into the 5' and 3' untranslated 3 regions, respectively demarcated by the locations of the start methionine and stop codon. 4

Evolutionary selection test 5 To determine if evolutionary selective pressures were being applied to bat tetherin, a selection test was 6 performed. To detect the positively selected sites among bat tetherin sequences, a maximum likelihood 7 (ML) phylogeny was generated with CODEML implemented in PAML4 software (67) . The input tree was 8 generated based on a pre-existing bat species tree (68) as well as cytb phylogeny derived from MEGA6 9 To amplify tetherin nucleotide sequences from bat cDNA, polymerase chain reaction (PCR) assays were 13 performed using cDNA generated from P. alecto and M. ricketti spleen tissue, and M. macropus and M. 14 schreibersii kidney cell lines, using various combinations of forward and reverse primers (Table 3A) . 15

Primers were designed using the tetherin predictions identified in the contig and SRA analyses. Primers 16 were designed to bind to the 5' and 3' untranslated regions of bat cDNA such that the full CDS could be 17 amplified. Bat capture, tissue collection and RNA extraction was conducted as previously reported (73) 18 with the exception that RNAlater (Ambion, USA) preserved spleen tissue from four male adult bats was 19 pooled before tissue homogenisation and followed with total RNA extraction with the Qiagen RNeasy Mini 20 kit with on-column extraction of genomic DNA with DNase I. Total RNA was reverse transcribed into cDNA 21 with the Qiagen Omniscript reverse transcriptase according to the manufacturer's protocol with the 22 exception that the reaction contained 100 ng/μl total RNA, 1 μM oligo-dT18 (Qiagen) and 10 μM random 23 hexamers (Promega). 24

All PCR amplification assays were performed using the Roche FastStart High Fidelity PCR system (Cat # 25 04738292001) with an annealing temperature gradient of 54°C to 64°C in 2°C increments. Each reaction 26

was made up to a total of 20 μl, containing 1 unit of polymerase, 2 ng of total cDNA, and 8 pmol of each 27 primer. All other parameters for PCR amplification were performed according to the manufacturer's 28

Amplicons from PCR reactions were analysed by agarose gel electrophoresis (74) . Individual DNA bands 1 were physically excised from the gel. DNA was purified from the gel fragments using the Wizard SV Gel 2 and PCR Clean Up kit (Promega, Fitchburg, USA) according to the manufacturer's protocol. 3

To further analyse the PCR amplified DNA, each DNA fragment was blunt-end ligated into the pCR2.1-4 TOPO-TA or pCR-Blunt-II-TOPO plasmid vector (Invitrogen, Waltham, USA). Ligation was performed using 5 the TOPO TA or Zero Blunt TOPO PCR Cloning Kit (Invitrogen) according to the manufacturer's instructions 6 and plasmids were transformed into Top10 E. coli using a standard heat-shock method (74) . Plasmids were 7 purified from these cultures using a Wizard Plus SV Miniprep DNA Purification kit (Promega). All inserted 8 sequences were confirmed by Sanger sequencing using M13 forward and reverse primers (M13F and 9 M13R; Table 3B ). Generation of tagged tetherin constructs for expression in mammalian cells 28 To enable detection of tetherin expression, the P. alecto tetherin isoforms A, B, and C, and M. macropus 29 tetherin A and B were genetically modified with the insertion of nucleotide sequences encoding the 30 haemagglutinin (HA) antibody-binding epitope. Tetherin sequences were modified through a 2-step PCR 1 process. PCR reactions were performed using the Roche FastStart HighFidelity PCR kit according to the 2 manufacturer's recommendations. 3

To express tetherin proteins in a mammalian cell culture system the tagged tetherin inserts were sub-4 cloned from the pCR2.1-TOPO (P. alecto tetherin isoforms A, B, and C) and pCR-Blunt-II-TOPO (M. 5 macropus tetherin A and B) vectors into pcDNA3.1 mammalian expression vectors. The HA-tagged 6 tetherin constructs contained terminal enzyme restriction sites. P. alecto tetherin constructs contained 7

XhoI and XbaI sites at their 5' and 3' ends, respectively. M. macropus tetherin constructs contained an 8

EcoRI site at each end. These sites, which are also present in the pcDNA3.1 vector, were used for digestion-9 ligation transfer of the HA-tagged tetherin sequences into pcDNA3.1. The ligation reaction products were 10 transformed into Top10 E. coli and plasmid clones were purified from E. coli colonies using the method 11 described under 'cDNA analysis'. All inserted sequences were verified by Sanger sequencing using the 12 pcDNA3.1 vector sequencing primers, T7F forward and BGHR reverse (Table 3B) . 13 Expression, extraction, and detection of tetherin in a mammalian cell culture system (DMEM-10). Cells were incubated at 37°C with 5% CO2. 22

The expression was performed in 6-well plates. Each well was seeded with 3.0x10 5 cells/well in 2 ml of 23 DMEM-10. Cells were transfected when the monolayer had reached 50-60% confluency. The tetherin 24 constructs analysed are listed in Table 4 . Each plasmid construct was transfected into cells in duplicate 25 wells at 2 μg/well using the transfection reagent Lipofectamine 2000 (Thermo Fisher) according to the 26 manufacturer's protocol. Tetherin was extracted using a previously published GPI-anchored protein 27 extraction protocol (48) . 28

Protein samples were analysed by size-based separation through SDS-PAGE and visualisation using 29

Western blot analysis (74) . For the Western blot analysis, the primary antibody solution contained a 1/1000 dilution of a monoclonal rabbit anti-HA antibody (C29F4, Cell Signaling Technology, Danvers, USA) 1 in TBS containing 0.1% Tween-20, and the secondary antibody solution contained a 1/10,000 dilution of a 2 polyclonal goat anti-rabbit IRD800 fluorophore-conjugate secondary antibody (Rockland, USA). The 3 primary antibody solution was incubated overnight at 4°C, and the secondary antibody solution was 4 incubated at room temperature for 1 h. To visualise fluorescent antibody-bound tetherin proteins, 5 membranes were scanned using the Odyssey Imaging System (LI-COR Biosciences, Lincoln, USA) at 6 wavelengths of 600 and 800 nm, using the default software settings. 7

Fluorescence microscopy 8 To visualise tetherin localisation within cells, 500 ng of plasmids encoding HA-tagged human l-tetherin 9 and P. alecto tetherin isoforms A, B, and C were transfected into HEK293T cells seeded on glass coverslips 10 as described above. At 48 h post transfection, cells were fixed with 4% paraformaldehyde (Sigma) in PBS 11 for 10 min at room temperature, and then permeabilised in 0.2% Triton X-100 in PBS for 5 min at room 12 temperature. Tetherin localisation in cells was detected by staining cells with anti-HA-tag rabbit 13 monoclonal IgG (Thermo Fisher) diluted in 0.2% Triton X-100, 3% BSA in PBS for 1 h at room temperature. qPCR analysis of tetherin expression across multiple bat tissues 20 To assess tetherin expression across various P. alecto tissues, tetherin mRNA levels were measured by 21 qPCR analysis. The primers and probes were designed using the program Primer Express (Perkin-Elmer, 22

Applied Biosystems, USA). The tetherin primers amplify all three known isoforms of P. alecto tetherin. P. 23 alecto bats were trapped in Queensland, Australia, and transported alive by air to the ACDP in Victoria, 24

where they were euthanised for dissection using methods approved by the ACDP animal ethics committee 25 (AEC1389). Tissues were stored at −80 •C in RNAlater (Ambion). Total RNA was extracted from frozen P. Transcriptome analysis of isoform expression under immune-stimulating treatments 8 To compare the expression of alternative isoforms of P. alecto tetherin and the impact of treatment with 9 immune-stimulating compounds, bats were treated with PBS, LPS (Invivogen, #tlrl-pb5lps) or poly(I:C) 10 (Invivogen, #vac-pic) as published previously (46) . Briefly, 5 h post-intraperitoneal injection, bats were 11 anaesthetised, culled and organs were processed for RNA, DNA, protein and cell suspensions as described 12

previously (78) . RNA libraries were prepared using RiboZero Plus rRNA-depletion kits (Illumina, USA) and 13 cDNA was generated using a mix of oligo-dT/random hexamer primers, prior to sequencing in 2x150PE on 14 the Illumina HiSeq platform. 15

Sequencing read libraries were quality controlled by analysis using FastQC (79) . Illumina sequence 16 adapters were removed and reads were trimmed or discarded on the basis of quality using the Trim 17 Sequences tool in CLC. Overlapping paired reads were merged using the Merge Overlapping Pairs tool in 18 CLC. Using the RNA-Seq Analysis tool in CLC, sequence reads were mapped against the P. alecto gene 19 scaffold containing the tetherin gene (GenBank accession: KB030270.1), which was manually annotated 20 with the tetherin gene and mRNA sequences for isoforms A, B, and C. The following parameters were used 21 for the RNA-Seq analysis: mismatch = 3, insertion = 3, deletion = 3, length fraction = 0.6, similarity fraction 22 = 0.95; default parameters were used otherwise. Isoform expression levels were normalised for post-23 trimming read library size and reported as counts-per-million reads (CPM). Non-parametric one-tailed 24

Mann-Whitney tests were performed to calculate statistical significance between treatments. The P. 25 alecto tetherin gene scaffold, annotation files, and read maps are provided in SI Data 2. 26 Functional validation of tetherin activity The consensus bat tetherin amino acid sequence structures and motifs generated through the multiple sequence alignment of tetherin from 27 bat species (SI Data 1) and compared against human tetherin. Strong positive selection of bat tetherin is revealed by the ratio of non-synonymous to synonymous mutations (Table 2) , and is represented as evolutionary hotspots. The YxY dual tyrosine motif, the alternative start site for the short isoform of tetherin, and the GPI anchor attachment site positions are indicated by arrows. The red asterisk (*) indicates the location of the inserted molecular tags in the P. alecto and M. macropus tetherin expression constructs. CD, cytoplasmic domain; TM, transmembrane domain; ED, extracellular domain; GPI, glycophosphatidylinositol; GSP, GPI signal peptide. B. Significant sequence diversity exists among bat tetherin proteins, indicated by the percentage of amino acid sequence conservation at each site of the consensus bat tetherin. Amino acids conserved in all 27 sequences are represented by their letters. Amino acids represented by X are variable and are included to indicate the sequence distance between closely positioned conserved residues. C. The consensus bat tetherin sequence. Amino acid residues represented in > 50% of bat tetherin sequences are indicated with their letter. Positions at which no amino acid residue is represented in > 50% of bat tetherin sequences are indicated with 'X'. The red line indicates the position of the dual tyrosine motif. M. macropus tetherin B and E analysed separately to A, C and D to account for the inclusion of 60 AA in M. macropus tetherin A, C, and D that is otherwise removed following gap treatment of the alignment of M. macropus B and E. Protein sequences of human and bat tetherins were aligned and gaps were removed. The horizontal axis represents percentage identity between pairs of amino acid sequences, coloured with a white to red gradient indicating increasing similarity as red intensity increases. The vertical axis represents the number of amino acid residue identity differences between sequence pairs, coloured with a white to blue gradient indicating increasing difference as blue intensity increases. A Primers used to amplify tetherin homologues from cDNA generated from Myotis macropus , M. ricketti , and Miniopterus schreibersii . Combinations of forward and reverse primers identified as capable of amplifying tetherin are indicated alongside the highest annealing temperature at which amplification was successful. B The sequencing primers used to confirm the sequence of tetherin clones and modified tetherin constructs inserted into plasmid vectors are listed. C The qPCR primers and probes used to analyse tetherin expression in bat tissues are listed. Probe 5' reporters and 3' quenchers are indicated with italics. UTR, untranslated region; PCR, polymerase chain reaction; TA°C, primer annealing temperature. Table 4 . Primers used for the generation of haemagglutinin-tagged Pteropus alecto and Myotis macropus tetherin expression constructs through a 2-Step PCR process.

Bats: Important reservoir hosts of emerging viruses

Bats and their virome: an important source of emerging viruses capable of infecting humans

A recently discovered pathogenic paramyxovirus, Sosuga virus, is present in Rousettus aegyptiacus fruit bats at multiple locations in Uganda

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Bat-borne virus diversity, spillover and emergence

Experimental inoculation of plants and animals with Ebola virus

Transmission studies of Hendra virus (equine morbilli-virus) in fruit bats, horses and cats

Experimental Nipah virus infection in pteropid bats (Pteropus poliocephalus)

Bat coronaviruses and experimental infection of bats, the Philippines

Experimental inoculation of Egyptian rousette bats (Rousettus aegyptiacus) with viruses of the Ebolavirus and Marburgvirus genera

Antiviral immune responses of bats: A review

Comparative analysis of bat genomes provides insight into the evolution of flight and immunity

Contraction of the type I IFN locus and unusual constitutive expression of IFN-α in bats

Differential evolution of antiretroviral restriction factors in pteropid bats as revealed by APOBEC3 gene complexity

The Egyptian rousette genome reveals unexpected features of bat antiviral immunity

Going to bat(s) for studies of disease tolerance

A metaanalysis of bat phylogenetics and positive selection based on genomes and transcriptomes from 18 species

Virus-and interferon alpha-induced transcriptomes of cells from the microbat Myotis daubentonii

A potent postentry restriction to primate lentiviruses in a yinpterochiropteran bat

Six reference-quality genomes reveal evolution of bat adaptations

Tetherin inhibits retrovirus release and is antagonized by HIV-1 Vpu

Inhibition of Lassa and Marburg virus production by tetherin

Tetherin inhibits Nipah virus but not Ebola virus replication in fruit bat cells

Bats and viruses: Friend or foe?

Severe acute respiratory syndrome coronavirus spike protein counteracts BST2-mediated restriction of virus-like particle release

Viruses in bats and potential spillover to animals and humans

Infectious KoRV-related retroviruses circulating in Australian bats

Tetherin inhibits HIV-1 release by directly tethering virions to cells

Origins and evolution of tetherin, an orphan antiviral gene

Identification of alternatively translated tetherin isoforms with differing antiviral and signaling activities

Innate sensing of HIV-1 assembly by Tetherin induces NFκB-dependent proinflammatory responses

Stimulation of NF-κB activity by the HIV restriction factor BST2

Large-scale identification and characterization of human genes that activate NF-κB and MAPK signaling pathways

Molecular cloning and chromosomal mapping of a bone marrow stromal cell surface gene, BST2, that may be involved in pre-B-cell growth

Tetherin is a key effector of the antiretroviral activity of type I interferon in vitro and in vivo

In vivo expression profile of the antiviral restriction factor and tumor-targeting antigen CD317/BST-2/HM1. 24/tetherin in humans

Structural insight into the mechanisms of enveloped virus tethering by tetherin

Structural basis of HIV-1 tethering to membranes by the BST-2/tetherin ectodomain

Bst-2/HM1. 24 is a raft-associated apical membrane protein with an unusual topology

Clathrin-mediated endocytosis of a lipid-raftassociated protein is mediated through a dual tyrosine motif

Identification and functional analysis of three isoforms of bovine BST-2

HIV-1 accessory protein Vpu internalizes cell-surface BST-2/tetherin through transmembrane interactions leading to lysosomes

Cloning and characterization of the antiviral activity of feline Tetherin/BST-2

Molecular phylogenetic reconstructions identify East Asia as the cradle for the evolution of the cosmopolitan genus Myotis (Mammalia, Chiroptera)

Paleontological evidence to date the tree of life

Studies on B cells in the fruit-eating black flying fox (Pteropus alecto)

Proteomic analysis of Pteropus alecto kidney cells in response to the viral mimic, Poly I:C

Detection of glycophospholipid anchors on proteins

Suppression of Tetherin-restricting activity upon human immunodeficiency virus type 1 particle release correlates with localization of Vpu in the trans-Golgi network

Feline tetherin efficiently restricts release of feline immunodeficiency virus but not spreading of infection

Interplay between ovine bone marrow stromal cell antigen 2/tetherin and endogenous retroviruses

Viruses in bats and potential spillover to animals and humans

Bat biology, genomes, and the Bat1K project: to generate chromosome-level genomes for all living bat species

Bat flight and zoonotic viruses

A potent anti-inflammatory response in bat macrophages may be linked to extended longevity and viral tolerance

Feline tetherin is characterized by a short N-terminal region and is counteracted by the feline immunodeficiency virus envelope glycoprotein

Structural and functional studies on the extracellular domain of BST2/tetherin in reduced and oxidized conformations

The formation of cysteine-linked dimers of BST-2/tetherin is important for inhibition of HIV-1 virus release but not for sensitivity to Vpu

Species-specific activity of HIV-1 Vpu and positive selection of tetherin transmembrane domain variants

Mechanism of HIV-1 virion entrapment by tetherin

Dimerization of tetherin is not essential for its antiviral activity against Lassa and Marburg viruses

Identification of diverse full-length endogenous betaretroviruses in megabats and microbats

The potential role of endogenous viral elements in the evolution of bats as reservoirs for zoonotic viruses

Identification of diverse groups of endogenous gammaretroviruses in mega and microbats

North American big brown bats (Eptesicus fuscus) harbor an exogenous deltaretrovirus

Structural and biophysical analysis of BST-2/tetherin ectodomains reveals an evolutionary conserved design to inhibit virus release

PAML 4: phylogenetic analysis by maximum likelihood

A molecular phylogeny for bats illuminates biogeography and the fossil record

MEGA6: molecular evolutionary genetics analysis version 6.0

Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes

PredGPI: a GPI-anchor predictor

MUSCLE: multiple sequence alignment with high accuracy and high throughput

Molecular characterisation of Toll-like receptors in the black flying fox Pteropus alecto

Molecular cloning

Synthesis and processing of RNA" in Genomes

BLAT-the BLAST-like alignment tool

Analysis of relative gene expression data using real-time quantitative PCR and the 2− ΔΔCT method

Optimizing dissection, sample collection and cell isolation protocols for frugivorous bats

FastQC: a quality control tool for high throughput sequence data

Broad-spectrum inhibition of retroviral and filoviral particle release by tetherin