We collected pooled urine samples from plastic sheets placed underneath flying fox roosts in southeastern Queensland and mid- to north-coast NSW during December 2016–September 2020 (Figure). We placed sheets in areas of the roost where P. alecto flying foxes were roosting, although other species were often also present. We recorded the number and species of bats immediately above the sheets. We also captured individual bats in mist nests; recorded species, sex, and age class; then collected urine samples directly from each anaesthetised bat or from a urine collection bag attached to its holding bag. Shortly after collection, we placed samples into viral lysis buffer, virus transport media, or an empty cryovial and stored them at −80°C (Appendix).

We used the QIAamp Viral RNA Kit using a QIAcube HT automated system (QIAGEN, https://www.qiagen.com) to extract RNA, then eluted it in 150 μL of TE buffer and first screened it for HeV-g1 using a qRT-PCR assay targeting the P gene (Table 1). We stored extracted RNA at −80°C and then screened it for HeV-g2 using the new multiplexed qRT-PCR assay, targeting the M gene with primers specific for HeV-g1 and HeV-g2 (2,3) (Table 1; Appendix). We used 10-fold dilutions with a known number of genome copies to construct a standard curve, calculate copy numbers/mL, and estimate limit of detection. We amplified the partial cytochrome b gene from all positive samples (10,11) (Table 1) and confirmed host species identity based on sequence identity across 402-bp sequences (Appendix).

We screened 4,539 pooled urine samples collected from 129 underroost sampling sessions and 1,674 urine samples collected from individual bats over 39 catching sessions during July 2017–September 2020 (Appendix Tables 1, 2). Eight pooled urine samples and 2 samples from individual flying foxes tested positive for HeV-g2 (Table 2). Positive samples were from Sunnybank in Queensland and Clunes, Lismore, Dorroughby, Maclean, and Nambucca Heads in NSW.

We detected HeV-g2 in samples collected across all seasons. Prevalence in sessions with positive detections ranged from 2.5% to 6.5% (95% CI 0.1%–22.8%). In pooled samples, HeV-g2 was only detected in sessions when HeV-g1 was also detected (HeV-g1 prevalence range 2.5%–50.1%); however, we found no statistically significant correlation between HeV-g1 and HeV-g2 prevalence (Pearson correlation analysis ρ = 0.09; p = 0.87). Most (8/10) of the HeV-g2–positive samples had low genome copies, but 2, ARSUN015_15_1 and ARLIS002_55_1, had considerably higher copy numbers (Table 2).

Individual flying foxes that tested positive included a P. poliocephalus juvenile female captured in Maclean, NSW, and a P. alecto adult male captured in Clunes, NSW (Appendix Table 3). We detected HeV-g2 in pooled samples from mixed-species roosts containing P. alecto and P. poliocephalus flying foxes. Cytochrome b sequencing identified DNA from P. alecto flying foxes in 6/8 positive underroost samples and from P. poliocephalus flying foxes in 2/8 (Table 2).

Urine is the route of HeV excretion from flying foxes and the source of virus transmission to horses. Detecting the novel Hendra variant HeV-g2 in the urine of flying foxes helped identify its distribution range, associated host species, transmission dynamics, and spillover risk. We show evidence that P. alecto and P. poliocephalus flying foxes excrete HeV-g2 in urine and both are likely competent reservoir hosts. We did not screen urine samples from P. conspicillatus or P. scapulatus flying foxes, so the potential of these species to excrete HeV-g2 in urine remains unconfirmed.

Although HeV-g1 has been detected in flying fox urine samples collected across all seasons, prevalence peaks in winter in subtropical regions (4,12), which is consistent with our preliminary HeV-g2 seasonality findings (5/8 detections in late May–late August) in the study area. The significantly lower prevalence of HeV-g2 than HeV-g1 could indicate actual lower prevalence in the sampled population. Alternatively, repeated freeze-thaw cycles in our samples or the bias toward collecting P. alecto urine in our sampling design might have led to lower detection. Tissue samples from flying foxes submitted for lyssavirus testing after contact with humans or pets showed higher HeV-g2 prevalence than our samples from wild populations (7), which might reflect higher prevalence in sick or stressed bats or geographical differences. HeV-g2 was previously detected in tissue samples from South Australia (3 positives from 4 samples), Victoria (7/64), and Western Australia (1/2) (7). Our findings extend the known distributional range of HeV-g2 to southeastern Queensland and mid- to north-coast NSW, areas proximate to the 2 known cases of HeV-g2 spillover to horses (3,5).

Our findings support expanding the expected geographic risk area for HeV spillover to include the distribution of P. poliocephalus flying foxes. Screening flying fox urine samples from a broader geographic range, including regions where P. alecto flying foxes are absent, should better inform epidemiologic relationships and relative prevalence of HeV variants. Given that data on the true diversity of HeV and related viruses in flying fox populations are incomplete, unbiased or Paramyxoviridae family–level viral surveillance in reservoir and spillover hosts might identify further variants. Developing a panel of diagnostic tools to detect a more comprehensive range of the viruses capable of spillover would substantially advance our ability to forecast spillover risk, manage biosecurity, and provide guidance to horse owners, veterinarians, and other stakeholders.