Discovery and Genomic Characterization of a Novel Henipavirus, Angavokely Virus, from Fruit Bats in Madagascar

Henipaviruses (HNVs) belong to a genus of bat-, rodent-, and shrew-borne viruses within the family Paramyxoviridae with demonstrated zoonotic potential. HNVs can manifest extreme virulence in human hosts, as exemplified by the prototypical HNVs, Hendra virus (HeV), and Nipah virus (NiV), which cause severe acute respiratory distress and/or encephalitis in humans, yielding case fatality rates that can exceed 90% (1 to 3). This high pathogenicity and the lack of approved HNV therapeutics or vaccines for humans have garnered HeV and NiV classification as Biological Safety Level 4 (BSL4) agents and WHO priority diseases. Since their discovery in the 1990s, HeV and NiV have periodically spilled over to humans from their reservoir hosts, pteropodid bats. HeV zoonosis is mediated by spillover to intermediate horse hosts, from which humans acquire infection (4). NiV can spillover to humans via intermediate transmission through pig hosts, or directly from bat to human, resulting in near-annual outbreaks of fatal encephalitis in South Asia, where subsequent human-to-human transmission also occurs (2, 5 to 7).

Novel HNVs continue to emerge from wildlife hosts and represent ongoing threats to human health. Initially, the Henipavirus genus comprised only HeV and NiV; however, the last 2 decades have witnessed the discovery of five new HNVs: bat-borne Cedar virus (CedV) and Ghanaian bat virus (GhV), rodent-borne Mojiang virus (MojV), and shrew-borne Gamak (GAKV) and Daeryong viruses (DARV) (8 to 11). News of an eighth putative henipavirus—and the fifth recognized zoonotic HNV—came to light while this article was in press: Langya henipavirus (LayV), a close relative of MojV, was recently identified in febrile patients in eastern China (12). Of these novel HNVs, at least three show evidence of zoonotic potential: serological data suggest prior human exposure to GhV or to an antigenically related virus in West Africa (13), while MojV was first identified following a human outbreak of severe pneumonia in Chinese mine workers, all of whom died after infection (9). Additionally, LayV was first identified in part with sentinel surveillance of human febrile patients with a history of animal contact, then later found to circulate at high seroprevalence in shrews (12). Thus, in addition to their high potential for pathogenicity, HNVs possess a broad host range that spans at least seven mammalian orders, including bats (10, 14).

Cross-species viral spillover necessitates effective interspecies transmission, which first requires a virus to successfully enter the cells of diverse host species. In general, HNVs use the highly conserved ephrin family of proteins, both type A and type B, as cell entry receptors (1, 8, 15 to 17). A notable exception to this pattern is MojV, which does not use ephrin proteins—or the sialic acid and CD150 receptors common to non-HNV paramyxoviruses—to gain cell entry (15, 18). Indeed, as of yet, the viral entry receptor for MojV—and the closely related GAKV, DARV, and LayV—remain unknown. In general, viruses in the genus Henipavirus have broad host ranges and cause high case fatality rates following human spillover, making the characterization of new HNVs a high public health priority.

The HNV genome consists of six structural proteins: nucleocapsid (N), phosphoprotein (P), matrix (M), fusion (F), glycoprotein (G), and polymerase (L). In comparison with other members of the family Paramyxoviridae, HNVs have relatively larger genomes (approximately 18 kb versus 16 kb). This extended length is largely due to several, long 3′ untranslated regions (UTRs) of the N, P, F and G mRNAs (19, 20). The genome length of HNVs, like all paramyxoviruses, adheres to the so-called “rule of six,” whereby viral genomes consistently demonstrate polyhexameric length (21). The “rule of six” is believed to be a requirement for efficient genome replication under the unique mRNA editing features of the paramyxovirus genome (21). The paramyxovirus P locus exhibits notable transcription properties that are shared across most members of the Paramyxoviridae family. The P gene permits the translation of additional accessory proteins from either gene editing events within the locus (prior to translation) or an overlapping open reading frame (ORF) in the P gene itself. All HNVs, with the exception of CedV, harbor a highly conserved mRNA editing site at which the insertion of additional guanine residues can result in the translation of accessory proteins, V and W, involved in viral antagonism and evasion of the host immune system (1). The HNV P gene also contains an overlapping ORF that allows for the synthesis of a third accessory protein, C, which is also involved in viral host immune evasion (1).

Our lab has previously presented evidence of exposure to henipa-like viruses in serum collected from three endemic Madagascar fruit bat species (E. dupreanum, Pteropus rufus, and Rousettus madagascariensis) using a Luminex serological assay that identified cross-reactivity to CedV/NiV/HeV-G and -F proteins (22). The most significant antibody binding previously detected corresponded to the NiV-G antigen for E. dupreanum serum and the HeV-F antigen for P. rufus and R. madagascariensis serum, suggesting the potential circulation of multiple HNVs in the Madagascar fruit bat system (22). Fruit bats, including E. dupreanum, are consumed widely in Madagascar as a source of human food, presenting opportunities for cross-species zoonotic emergence. This underscores the importance of further characterization of the pathogenic and zoonotic potential of potential HNVs circulating in the Madagascar fruit bat system. Here, we describe and characterize a novel bat HNV genome, Angavokely virus (AngV), recovered from urine samples collected from the Madagascar fruit bat, E. dupreanum. Our work suggests AngV is part of an ancestral group of HNVs that may rely on a novel, non-ephrin-mediated viral entry pathway.

We describe and characterize a novel HNV, AngV, from a urine sample collected from an E. dupreanum Malagasy fruit bat. In this study, urine samples from 206 unique fruit bats were assessed by metagenomic sequencing, yielding an overall positive HNV detection rate of 4.9% (10/206) for all bats studied and an HNV prevalence of 9.4% (10/106) for the E. dupreanum hosts. Of all the HNV positive samples, only one sample yielded sufficient reads for assembly of a complete coding sequence and subsequent genomic analysis. In a 6-year collection period spanning multiple wet/dry seasons, HNV-positive samples were only recovered from E. dupreanum bats in the Angavokely roosting site, despite prior serological evidence of HNV infection in P. rufus and R. madagascariensis bats, as well (22). HNV RNA was recovered from E. dupreanum in both wet and dry seasons, though higher sampling intensity throughout the wet season precludes any conclusions regarding underlying seasonal patterns in these data. Previous work in this system has suggested a seasonal increase in fruit bat seroprevalence across the winter low-nutrient season, which also overlaps the gestation period for these synchronously breeding fruit bats (22). In fruit bat systems elsewhere, HNVs are also shed in urine at higher rates during the nutrient-poor dry seasons for the localities in question (26 to 28); in the case of NiV and HeV, these seasonal viral shedding pulses have been linked to zoonotic spillover.

The recovered genome of AngV exhibits a structural organization characteristic of the Henipavirus genus and a nucleotide and amino acid identity to HeV and NiV that is comparable to those shared with the more distantly related HNVs, MojV, GhV, and CedV. A limited quantity of available original sample precluded full genome recovery for AngV (as evidenced by the lack of the 5′ UTR region of the N ORF), which prevented analysis of the extent to which the full AngV genome may abide by the “rule of six,” observed by all other members of the Orthoparamyxovirinae subfamily (29). Phylogenetic analyses of AngV support classification of this virus as a distinct novel bat-borne Henipavirus (L gene amino acid distance <0.82 distance for the subfamily Orthoparamyxovirinae), in accordance with the International Committee on Taxonomy of Viruses (ICTV) criteria (20). This novel HNV is estimated to have diverged approximately 9,800 years ago, prior to the currently known African and Asian bat-borne HNV lineages but considerably more recently than the estimated mid- to late-Miocene divergence of E. dupreanum from its sister species, E. helvum, on the African continent (30). Recent characterization of Betacoronaviruses in Madagascar fruit bats demonstrates surprising identity to lineages circulating in West Africa (31), suggesting that, though resident only in Madagascar, Malagasy fruit bats likely experience some form of contact with the African continent. Of the 49 bat species that inhabit the island nation of Madagascar, nine species are widely distributed across Africa, Asia, and/or Europe, presenting opportunities for interspecies viral transmission via island-hopping. Intensified viral sampling of Madagascar’s insectivorous bat populations for HNVs thus represents an important future research priority.

As an ancestral bat-borne HNV, AngV may provide important insight into HNV evolution and pathogenesis. Similar to other paramyxoviruses, the encoded AngV P gene is able to produce multiple immunomodulatory protein products (32). One such protein product is the V protein, thought to be involved in immune evasion and considered a significant determinant of viral pathogenicity and lethality (33, 34). AngV harbors the highly conserved mRNA editing site and a predicted ORF that encodes a V protein with a conserved cysteine-rich C terminus, suggesting that AngV has the capacity to produce a functional V protein. With the exception of the newly discovered HNVs in shrews, GAKV and DARV, all HNVs harboring a V protein have previously demonstrated evidence of human infection, highlighting the potential for AngV to cause productive infection in humans (1, 9, 13). Further studies are needed to ascertain the virulence potential and host breadth of this novel virus.

Characterization of the AngV glycoprotein (G) through AlphaFold modeling and structure-based alignments revealed a similar structural organization to other HNV glycoproteins. Notably, the AngV glycoprotein surpasses that of GhV as the longest glycoprotein of the Henipavirus genus. Like that of GhV, the AngV glycoprotein harbors a long C-terminal extension (Fig. S4). It is unclear if the C-terminal extension of the AngV glycoprotein has a functional role, though the C-terminal extension of the glycoprotein in GhV is known to play a functional role in receptor-mediated fusion (16).

Henipavirus host tropism and virulence rely on a myriad of factors, one of which is the HNV glycoprotein. The previously characterized HNV glycoproteins of NiV, HeV, CedV, and GhV utilize members of the ephrinA and ephrinB class family as host-cell receptors for viral entry into human cells (16 to 18, 24, 35). However, like MojV, the AngV glycoprotein lacks these well-conserved ephrin binding residues. Structure-based alignments can shed light on potential receptor binding residues when characterizing novel viruses. For instance, sequence-based comparisons of the GhV and NiV glycoproteins were used to predict GhV ephrin binding (13), which was later confirmed by crystallography (16). Structure-based alignment of the AngV glycoprotein shows a lack of highly conserved ephrin binding residues, including NiV E533—a seminal residue for ephrinB2 binding that is conserved across all ephrin binding HNVs. This suggests that, like MojV—and probably DARV, GAKV, and LayV—the AngV glycoprotein may not bind ephrins, pointing to the possible use of an ancestral viral entry pathway. The growing number of novel HNVs that appear not to rely on ephrin binding for cellular entry could warrant reevaluation of the existing HNV genus to better reflect conserved function and pathobiology.

This work presents a novel bat-HNV, AngV, identified from a Malagasy fruit bat. AngV joins a growing group of ancestral HNVs with unknown cell-entry receptors. Discovery of the cell surface receptor for AngV represents an important future research priority that will shed light on the breadth of host range for this virus, including its zoonotic potential.

We thank Anecia Gentles and Kimberly Rivera for help in the field and the lab and acknowledge the Virology Unit at the Institut Pasteur de Madagascar and Maira Phelps of the Chan Zuckerberg Biohub (CZB) for logistical support. We additionally thank Angela Detweiler, Michelle Tan, and Norma Neff of the CZB genomics platform for mNGS support. Molecular graphics and analyses were performed with UCSF ChimeraX, developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, with support from the National Institutes of Health R01-GM129325 and the Office of Cyber Infrastructure and Computational Biology, National Institute of Allergy and Infectious Diseases.

This work was supported by the National Institutes of Health (1R01AI129822-01 grant to J.-M.H., P.D., and C.E.B.; 5T32AI007641-19 to S.M.; and R01AI109022 to H.C.A.), DARPA (PREEMPT Program Cooperative Agreement no. D18AC00031 to C.E.B.), the Bill and Melinda Gates Foundation (GCE/ID OPP1211841 to C.E.B. and J.-M.H.), the Adolph C. and Mary Sprague Miller Institute for Basic Research in Science (postdoctoral fellowship to C.E.B.), AAAS/L’Oréal USA (For Women in Science fellowship to C.E.B.), the Branco Weiss Society in Science (fellowship to C.E.B.), and the Chan Zuckerberg Biohub.

We declare no competing interests.