Methods

Study locations
Fieldwork was undertaken at two locations on the Philippine island of Luzon: Bulacan Province (13–26 April, 2010) and Subic Bay Freeport Zone (20 Nov-7 Dec, 2010) (Fig. 1). Bulacan Province was the focus of RESTV detections in pigs and associated pig workers, and the focus of our initial surveillance. The primary field locations in Bulacan Province were Biak na Bato National Park in the municipality of San Miguel (N 15° 06’ 33.9” E 121° 05’ 44.6”) and Puning Cave in the municipality of Doña Remedios Trinidad (N 14° 57’ 29.7” E 121° 05’ 27.4”). Biak-na-Bato National Park is an extensive protected area comprising forested riverine gorges and cave networks. Puning Cave is a riverine limestone cave complex within remnant forest habitat, surrounded by farmland. Both locations have diverse and abundant bat populations. A known flying-fox roost in the Cubi area of Subic Bay Freeport Zone (N 14° 47’ 16.63” E 120° 16’ 22.02”) was the focus of the later surveillance [19]. The roost is in a peri-urban forest remnant within an urban and farmland mosaic adjacent to an extensive area of largely intact forest.

Bat capture and sampling
In Bulacan Province, sampling targeted insectivorous bats and small fruit bats, including taxa (or related taxa) previously associated with filoviruses in Africa [1]. We strategically deployed mist-nets and harp traps [20] after dusk to capture bats as they exited caves or foraged through the evening. Nets were continuously monitored, and bats removed on capture and placed in individual cotton bags; harp traps were monitored either continuously or periodically, the latter typically at hourly intervals, and bats removed from the holding bag of the trap and placed in individual calico bags. Bags were carried to the processing station (maximum 10 min) and hung from horizontal lines at minimum of 150 mm apart to ensure adequate ventilation. Following sample and data collection, bats were offered fruit juice for hydration and energy, and immediately released.
At Subic Bay, the sampling targeted pteropodid fruit bats (flying foxes), which were captured by mist-net [20] pre-dawn and post-dusk in the immediate vicinity of the known roost. Captured bats were held individually in cotton pillowcases and transported 3.4 km by vehicle to the Wildlife in Need Rescue Centre for processing. Bats were sequentially anaesthetised using the inhalation agent isofluorane [21]. Following data and sample collection, bats were recovered from anaesthesia, offered fruit juice (for hydration and energy), and released at their capture location within four hours of capture [19].

Sample and data collection
Biological samples were collected from bats using non-lethal, minimally invasive techniques by multiple teams, each including a veterinarian and wildlife biologist with experience in handling bats. In turn, each bat was removed from its bag, and species, sex, pregnancy/lactation status, forearm length, weight and body condition score recorded [20]. A venous blood sample was collected as described by Smith et al. [22] for small species (<100 g), and by Epstein and Field [20] for flying-foxes (>500 g), and stored at 4C on wet ice in the field. Blood sample volume did not exceed 1 % bodyweight in accordance with animal ethics guidelines [23]. Duplicate oropharyngeal, urine/urogenital and rectal swabs were collected where possible, placed in lysis buffer (Nuclisens, Biomerieux, USA), and temporarily stored in a dry shipper at −70C. Samples were transported to the Bureau of Animal Industries (BAI) Philippine Animal Health Centre (PAHC) laboratory in Quezon City daily. Serum was yielded from the blood samples by centrifugation, and all samples collated into two sets where possible, one of which was archived at the PAHC laboratory, and the other subsequently forwarded to the CSIRO Australian Animal Health Laboratory (AAHL) in Geelong, Australia. Cross-contamination in the field was avoided by holding each bat in a separate clean bag from capture, processing individual bats sequentially, and adopting appropriate biosecurity protocols such as changing or disinfecting gloves between bats, disinfecting the immediate worksite and non-disposable equipment between bats, and using disposable and sterile consumables.

Laboratory analyses
Serology and molecular assays were undertaken at AAHL. Samples were handled at BSL 4 until inactivated. An indirect ELISA (using mixed recombinant RESTV and EBOV (formally ZEBOV) N antigens) was used to screen sera, with Western blot (using individually run recombinant RESTV and EBOV N antigens) performed on ELISA-positive sera as described in Olival et al. [10]. Cut-off values to determine ELISA-positive sera were determined using a statistical approach as described in Pourrut et al. [1] and Olival and Hayman [5]. Confirmatory Western blot analysis was conducted as described in Hayman et al. [24]. Molecular assays comprised quantitative (q) and conventional RT-PCR in series. Swabs of the same sample type were pooled (five per pool) and RNA extraction undertaken using a QIAamp viral mini kit according to manufacturer’s instructions. Eluates were tested using a US CDC qPCR which targeted the RESTV NP gene (P. Rollin, 2010, pers. comm.). A sample yielding a repeatable Ct value of less than 40 was regarded as positive, and the authenticity of the amplified products corroborated by melt curve analysis; a sample yielding a repeatable Ct value of 40–45, or a non-repeatable Ct value of less than 40 was regarded as potentially positive. All other samples were regarded as ‘not detected’. Positive and potentially positive pools were re-tested in the same assay, as were the component individual samples. Where adequate sample remained, positive or potentially positive individual samples were tested by a PCR targeting the NP gene [25] adapted to a hemi-nested PCR with a second forward primer (FiloNP-hnFe – TGATGGTAATCTTYAGATTGATGAGG) in an attempt to gain adequate product for direct sequencing. Purified PCR products were sequenced at the AAHL sequencing facility using a BigDye Terminator v1.0 Kit (Applied Biosystems) and an ABI PRISM 377 DNA Sequencer (Applied Biosystems). Every nucleotide was sequenced with a minimum of threefold redundancy to ensure a consensus and repeatable sequence data. The Clone Manager and Align Plus programs in the Sci Ed Central software package (Scientific and Educational Software) were used for sequence management and analysis. Phylogenetic analysis based on the 519 bp fragment of the NP genes from different Ebola virus sequences was conducted using the MEGA5 program [26]. The phylogenetic tree was constructed using the maximum likelihood algorithm with bootstrap values determined by 1,000 replicates.

Statistics and data analysis
The study employed a cross-sectional design to investigate the ‘presence or absence’ of RESTV infection in Philippine bats. A target sample size of 120 individuals per taxa was set, based on the averaged ZEBOV and MARV seroprevalence reported in Pourrut et al. [1], to allow negative findings to be interpreted as providing 95 % statistical confidence of absence of infection in the taxa at a design prevalence of 2.5 %. The findings are presented as descriptive summary statistics.

Animal ethics
Fieldwork was carried out under the Philippine Government Protected Areas and Wildlife Bureau (now Biodiversity Management Bureau) permit no. 2010–197, issued following approval from the Philippines National Wildlife Management Committee. The latter is responsible for assessing applications for the use of wildlife for scientific research in accordance with the Philippines Rules and Regulations on the Conduct of Scientific Procedures Using Animals. The permit approved the defined study procedures for capture, handling, sample collection, and release of bats, including Acerodon jubatus, listed as ‘Endangered’ by the International Union for Conservation of Nature (IUCN) and the Philippine National List of Threatened Wild Fauna. All procedures reflect current best practice, minimising stress and discomfort, and reducing the risk of injury, mortality, and interference to natural behaviour. Local approvals were obtained from the Protected Area Management Board of Biak Na Bato National Park, the Municipality of Doña Remedios Trinidad, and Subic Bay Metropolitan Authority.