Monitoring Taiwanese bovine arboviruses and non‐arboviruses using a vector‐based approach

Abstract

In current sampling approaches, there exists a divergence between the surveillance of arthropod‐borne and that of non‐arthropod‐borne viruses. It is commonly held that the collection of vector specimens applies only to arbovirus surveillance and that the surveillance of non‐arboviruses must rely on traditional methods that involve the sampling of blood, faeces or saliva, or other examinations. The vector‐based approach is a sampling method that has the ability to survey both arboviruses and non‐arboviruses by distinguishing engorged vector specimens from entire vector samples. Accordingly, five arboviruses and three non‐arboviruses were detected in a study using a vector‐based approach conducted during 2012–2015. Hence, this report provides the first description of the Taiwanese vector species for the bovine arboviruses detected. The present investigations demonstrate that the vector‐based approach applies not only to the surveillance of arboviruses, but also has potential as a possible tool for monitoring non‐arboviruses on livestock farms in the future.

• Vector‐based sampling has potential for the simultaneous monitoring of arboviruses and non‐arboviruses on livestock farms.

• The vector or potential vector species of five bovine arboviruses detected in Taiwan were identified for the first time.

• Aino virus and Peaton virus were detected in Taiwan for the first time.

Introduction

Infectious viral diseases pose continual and substantial threats to the dairy industry and have serious implications for animal welfare, food security and even human health. In the interests of farm animal welfare, rapid diagnosis and treatment are required to avoid pain, injury and disease in animals. Furthermore, outbreaks of disease directly reduce the production or quality of dairy products, resulting in huge economic losses and threatening food security. In 2001, an outbreak of foot‐and‐mouth virus infected around one million cattle in the U.K. and the resulting economic losses were estimated to amount to US$12.3–13.8 bn (Thompson et al., 2002). However, zoonotic viral diseases of cattle, such as Rift Valley fever and Crimean–Congo haemorrhagic fever, remain constant threats to human populations (Ganter, 2015; Schuster et al., 2016). Therefore, a sustainable system for the surveillance of infectious disease on livestock farms is essential to the monitoring and control of such diseases.

In Taiwan, several bovine arboviruses are reported, including bovine fever ephemerovirus (BEFV) (Ephemerovirus: Rhabdoviridae), a Simbu serogroup virus, Akabane orthobunyavirus (AKAV) (Orthobunyavirus: Peribunyaviridae), a Palyam subgroup virus, Chuzan virus (CHUV) (Orbivirus: Reoviridae), a strain of epizootic haemorrhagic disease virus serotype 2, Ibaraki virus (IBAV) (Orbivirus: Reoviridae), and bluetongue virus (Orbivirus: Reoviridae) (Liao et al., 1994, 1996a, 1996b; Lee et al., 2011; Ting et al., 2016). However, entomological investigations into the relationships between vectors and bovine arboviruses are still meagre in Taiwan. In addition, several non‐arboviruses have been reported, including bovine leukaemia virus (BLV) (Deltaretrovirus: Retroviridae), foot‐and‐mouth disease virus (Aphthovirus: Picornaviridae), bovine respiratory syncytial virus (BRSV) (Pneumovirus: Paramyxoviridae), bovine coronavirus (BCoV) (Betacoronavirus: Coronaviridae), bovine viral diarrhoea virus (BVDV) (Pestivirus: Flaviviridae), rotavirus A (RVA) (Rotavirus: Reoviridae) and bovine enterovirus (BEV) (Enterovirus: Picornaviridae) (Wang, 1991; Huang et al., 2000; Tsai, 2001; Lee, 2004; Pan, 2013). However, a comprehensive surveillance system that is effective in detecting a broad range of targets and is also affordable is still lacking in Taiwan.

Surveillance of arboviruses through the monitoring of vector insects, mosquitoes and midges is an intuitive concept and is commonly applied on livestock farms. Recent studies have indicated that surveillance of haemoparasites or human viruses has been carried out in engorged mosquitoes, but has not yet been applied to cattle (Boothe et al., 2015; Grubaugh et al., 2015). The present study monitored five bovine arboviruses [BEFV, AKAV, Aino virus (AINOV), CHUV and IBAV] and six bovine non‐arboviruses (BLV, BVDV, BEV, RVA, BRSV and BCoV) in vector insects, and examined the feasibility of a vector‐based approach as a possible surveillance tool in dairy farms.

Materials and methods

Vector sampling on dairy farms

The six sampling sites were distributed in Yunlin County, Chiayi County, Tainan City, Kaohsiung City and Pingtung County in southern Taiwan (Fig. 1). A modified blacklight trap (wavelength: 365 nm) was set up on each farm, and insect collections were conducted 2 h after sunset, when Culicoides (Diptera: Ceratopogonidae) biting midges and mosquitoes are more active on dairy farms. Live mosquitoes and midges were anaesthetized using carbon dioxide in the laboratory and then identified to genus or species level and engorged status under an Ez4D microscope (Leica Microsystems GmbH, Wetzlar, Germany). A maximum of 100 mosquitoes and 200 midges per pool were placed in a 2‐mL microcentrifuge tube (Qiagen GmbH, Hilden, Germany) and stored at − 80 °C for further assays.

Figure 1.

Localities of sampling farms in Taiwan. YL, Yunlin; CY, Chiayi; TN, Tainan; KH, Kaohsiung; PT, Pingtung.

Total RNA extraction

Whole insects were homogenized using two 3‐mm diameter steel beads in a TissueLyser II (Qiagen GmbH) at 30 grindings/s for 1 min. Total RNA was then extracted with a PureLink RNA Mini Kit (Thermo Fisher Scientific, Inc., Waltham, MA, U.S.A.). In brief, the vector fragments were mixed with 0.7 mL Lysis solution containing 1% 2‐mercaptoethanol in TissueLyser II to make a homogeneous solution and centrifuged at 16 000g for 10 min. Then, 0.6 mL of supernatants was drawn out and well mixed with 0.6 mL 70% ethanol. The solution was transferred to an RNA binding column and briefly centrifuged at 12 000g. The following wash procedures were performed in line with the manufacturer's instructions for the standard wash protocol. For the RNA elution, 50 µL of nuclease‐free water from the mixture was added to the RNA for 3 min at room temperature and then centrifuged at 12 000g for 3 min for elution. The total RNA was stored at − 80 °C for further assays.

Reverse transcription

The reverse transcription (RT) procedure was performed using an iScript™ cDNA synthesis kit (Bio‐Rad Laboratories, Inc., Hercules, CA, U.S.A.). The concentrations of total RNA samples were measured with a NanoDrop™ 2000 spectrometer (Thermo Fisher Scientific, Inc.). The total RNA was drawn out at between 1 ng and 1 µg for cDNA synthesis, according to the measuring results. Each RNA sample was mixed with 1 µL of reverse transcriptase, 4 µL of 5× reaction buffer containing random hexamer and oligod(T) primer, and then increased to 20 µL with nuclease‐free water. The synthesis of cDNA was conducted at 25 °C for 5 min for primer annealing, at 42 °C for 1 h for synthesis and at 85 °C for 5 min to stop the activity of reverse transcriptase in an Applied Biosystems^® 2720 Thermal Cycler (Thermo Fisher Scientific, Inc.). When the RT reaction was completed, cDNA was diluted to 40 µL using nuclease‐free water and stored at − 80 °C.

Detection of arboviruses

Five arboviruses (BEFV, AKAV, AINOV, CHUV and IBAV) were detected by nested RT‐polymerase chain reaction (PCR). The primer sets are listed in Table S1 following Shin et al. (2009). All viruses were detected using an Ex‐Taq DNA Polymerase Kit (Takara Bio, Inc., Shiga, Japan). The PCR mix was composed of 0.2 µL of Ex‐Taq DNA polymerase, 1 µL of DNA template (cDNA or first PCR product), 1 µL of each primer at 10 µM, 2 µL of 10× reaction buffer and 1 µL of 2.5 mM dNTP mix, and increased to a final volume of 20 µL by the addition of nuclease‐free water. The PCR thermal conditions are listed in Table S2. The secondary PCR products were analysed by electrophoresis using 1.5% agarose gel.

Detection of non‐arboviruses

Three non‐arboviruses (BLV, BVDV and BRSV) were detected by nested RT‐PCR, and three other viruses (BEV, BCoV and RVA) were detected by RT‐PCR in 192 pools of engorged vector specimens in 2014 and 2015, respectively. The primer sets and PCR thermal conditions are listed in Tables S1 and S2. The PCR mix and electrophoresis analysis were identical to those used in arbovirus detections.

Cloning and sequencing of bovine viruses

The virus amplicon was extracted using a Quick Gel™ Extraction Kit (Thermo Fisher Scientific, Inc.). Then, the purified DNA was ligated into a pGEM^®‐T Easy Vector using T4 DNA ligase (Promega Corp., Madison, WI, U.S.A.) at 16 °C for 14 h. The ligated plasmids were transformed into competent cells of Escherichia coli DH5α at 42 °C for 30 s and then incubated on lysogeny broth (LB) agar plates with 50 µm ampicillin (Sigma‐Aldrich Corp., St Louis, MO, U.S.A.), 40 µL of 100 mm IPTG solution (Sigma‐Aldrich Corp.) and 40 µL of 1% X‐gal solution (5‐bromo‐4‐chloro‐3‐indolyl‐β‐D‐galactopyranoside) (Sigma‐Aldrich Corp.) at 37 °C for 16 h. The target inserts in colonies were checked by PCR with specific target primers and electrophoresis run in 1% agarose gel. The checked colonies were sequenced by a local company (Genomic Ltd, Taipei, Taiwan). The sequencing results were transferred into vectornti ^® Version 10 (Thermo Fisher Scientific, Inc.) and trimmed. The sequence identity was verified by blast, and each verified sequence was uploaded to GenBank at the National Center for Biotechnology Information (NCBI) (Bethesda, MD, U.S.A.).

Minimum infection rates

Minimum infection rates (MIRs) for five arboviruses detected in the unengorged vector specimens via species identification were calculated with the following equation, with only one infected individual assumed in each positive pool:

$$\frac{\mathit{\text{number of positive pools}}}{\mathit{\text{total number of unengorged specimens tested}}}\times 1000$$

Results

From September 2012 to December 2015, a total of 58 576 Culicoides biting midges in 520 pools and 11 741 mosquitoes in 569 pools were collected. Overall, 5855 Culicoides biting midges in 123 pools and 8630 mosquitoes in 425 pools were identified to species and the remainder were identified to genus level. Unengorged pools accounted for 39.8% (n = 450), engorged pools for 28.1% (n = 317), and unidentified pools for 32.1% (n = 363) of the total. Five Culicoides species (Culicoides oxystoma, Culicoides nipponensis, Culicoides arakawae, Culicoides actoni and Culicoides bubalus) were identified. Culicoides oxystoma was the dominant species (n = 5484, 93.7%), followed by C. nipponensis (n = 236, 4.0%), C. actoni (n = 108, 1.8%), C. arakawae (n = 16, 0.3%) and C. bubalus (n = 11, 0.2%). Of the 11 mosquito species identified within the genera Anopheles, Aedes, Armigeres and Culex (all: Diptera: Culicidae), six belonged to the genus Culex. Among these, Culex tritaeniorhynchus was the most abundant (n = 7758, 89.9%), followed by Culex quinquefasciatus (n = 354, 4.1%) and Armigeres subalbatus (n = 252, 2.9%). Each of the other eight species accounted for < 1% of the mosquitoes identified to species level (Table 1).

Table 1.

Vector insects collected for the detection of arboviruses and non‐arboviruses.

				Engorged pools, n
Species		Specimens, n	Pools, n	No	Yes	Unidentified
Biting midges	Culicoides spp.	52 721	397	28	71	298
	C. oxystoma	5484	92	65	27	0
	C. nipponensis	236	10	0	10	0
	C. arakawae	16	11	10	1	0
	C. actoni	108	8	7	1	0
	C. bubalus	11	2	2	0	0
Mosquitoes	Anopheles spp.	214	63	26	31	6
	An. sinensis	77	27	14	13	0
	Aedes spp.	4	4	1	2	1
	Ae. albopictus	19	7	5	2	0
	Ae. vexans	11	2	1	1	0
	Armigeres spp.	154	22	9	11	2
	Ar. subalbatus	252	42	22	20	0
	Ar. baisasi	2	1	1	0	0
	Culex spp.	2739	109	49	36	24
	Cx. tritaeniorhynchus	7758	208	135	73	0
	Cx. quinquefasciatus	354	50	41	8	1
	Cx. pipiens molestus	73	19	17	2	0
	Cx. fuscocephala	57	8	6	2	0
	Cx. annulus	26	6	4	2	0
	Cx. sinensis	1	1	1	0	0

In this study, four arboviruses (BEFV, AKAV, AINOV and CHUV) were detected. The finding of Shamonda orthobunyavirus [Peaton virus (PEAV)] was unexpected. Samples containing PEAV were distinguished from AINOV samples after sequencing and blast analysis. Taiwanese and Japanese AINOV exhibited 98.8–99.7% NP gene homology and PEAV exhibited 98.8–99.1% NP gene homology (Yanase et al., 2010). The detected arboviruses were deposited in GenBank under accession numbers KF676987–KF676999 and MF083618–MF083684. In the detected arboviruses, the structure genes on the M segments of AKAV, AINOV and PEAV and segments 2 and 6 of CHUV were sequenced for further molecular epidemiological studies (data not shown). In 2012, BEFV, AINOV and PEAV were detected in Yunlin County site A (YL‐A), and AKAV was detected in YL‐A, Yunlin County site B (YL‐B) and Pingtung County (PT) (Fig. 1). In 2013, positive detections of AKAV, AINOV and PEAV were observed from October to December. Akabane orthobunyavirus was detected in Kaohsiung City (KH), AINOV in YL‐A, YL‐B and PT, and PEAV in YL‐A. Chuzan virus was detected from July to September in YL‐A and YL‐B. In 2014, there were 13 detections of BEFV from April to September and 18 detections of AINOV from January to June. Five viruses, BEFV, AKAV, AINOV, PEAV and CHUV, were detected in 2015; among them, CHUV accounted for 45.5% of overall detections (Table 2).

Table 2.

Polymerase chain reaction (PCR) detection of bovine arboviruses on dairy farms in southern Taiwan.

		BEFV						AKAV						AINOV						PEAV						CHUV
Date		YL‐A	YL‐B	CY	TN	KH	PT	YL‐A	YL‐B	CY	TN	KH	PT	YL‐A	YL‐B	CY	TN	KH	PT	YL‐A	YL‐B	CY	TN	KH	PT	YL‐A	YL‐B	CY	TN	KH	PT
2012	September	•						•	•				•
	October–December							•						•						•
2013	January–March
	April–June
	July–September																									•	•
	October–December											•		•	•				•	•
2014	January–March							•						•	•	•		•
	April–June	•			•										•		•
	July–September	•	•	•	•							•
	October–December
2015	January–March				•			•
	April–June							•									•										•			•
	July–September	•	•					•			•	•															•	•	•
	October–December				•						•									•									•

•, positive by PCR detection.

BEFV, bovine fever ephemerovirus; AKAV, Akabane orthobunyavirus; AINOV, Aino virus (Shuni orthobunyavirus); PEAV, Peaton virus (Shamonda orthobunyavirus); CHUV, Chuzan virus (Palyam virus). YL‐A and YL‐B are located in Yunlin County, CY in Chiayi County, TN in Tainan City, KH in Kaohsiung City, and PT in Pingtung County.

Two species of Culicoides biting midge and six species of mosquito were found to carry target arboviruses that were determined in unengorged specimens. Culicoides oxystoma was found to carry AKAV, AINOV, CHUV and BEFV, and C. arakawae carried BEFV. Anopheles sinensis carried BEFV and AKAV, and the other mosquito species carried various viruses (Table 3).

Table 3.

Polymerase chain reaction (PCR) detection of bovine arboviruses in unengorged vector specimens.

		Positive detections in unengorged specimens
Species		BEFV	AKAV	AINOV	PEAV	CHUV
Biting midges	Culicoides spp.	•		•	•
	C. oxystoma	•	•	•	•	•
	C. nipponensis
	C. arakawae	•
	C. actoni
	C. bubalus
Mosquitoes	Anopheles spp.		•	•
	An. sinensis	•	•
	Aedes spp.
	Ae. albopictus
	Ae. vexans
	Armigeres spp.
	Ar. subalbatus			•		•
	Ar. baisasi
	Culex spp.	•		•
	Cx. tritaeniorhynchus	•		•		•
	Cx. quinquefasciatus		•	•
	Cx. pipiens molestus					•
	Cx. fuscocephala					•
	Cx. annulus

•, positive by PCR detection.

BEFV, bovine fever ephemerovirus; AKAV, Akabane orthobunyavirus; AINOV, Aino virus (Shuni orthobunyavirus); PEAV, Peaton virus (Shamonda orthobunyavirus); CHUV, Chuzan virus (Palyam virus).

The MIRs in C. oxystoma for infection by the five arboviruses detected ranged from 0.191 to 0.572, whereas MIRs in Cx. tritaeniorhynchus were 0.644, 0.805 and 0.161 for infection with BEFV, AINOV and CHUV, respectively. The MIRs in the other six vector species are shown in Table 4.

Table 4.

Minimum infection rates (MIRs) in the vector species.

		MIR
Vector species		BEFV	AKAV	AINOV	PEAV	CHUV
Biting midges	Culicoides oxystoma	0.191	0.191	0.191	0.191	0.572
	Culicoides arakawae	66.667	−	−	−	−
Mosquitoes	Anopheles sinensis	17.241	17.241	−	−	−
	Armigeres subalbatus	−	−	5.291	−	10.582
	Culex tritaeniorhynchus	0.644	−	0.805	−	0.161
	Culex quinquefasciatus	−	2.959	5.917	−	−
	Culex pipiens molestus	−	−	−	−	14.085
	Culex fuscocephala	−	−	−	−	18.519

−, virus not identified in the vector species.

The MIRs represent numbers of infected individuals per 1000 vector individuals.

BEFV, bovine fever ephemerovirus; AKAV, Akabane orthobunyavirus; AINOV, Shuni orthobunyavirus; PEAV, Peaton virus (Shamonda orthobunyavirus); CHUV, Chuzan virus (Palyam virus).

In surveys of non‐arboviruses, BLV, BVDV and RVA were detected in engorged Culicoides biting midges and mosquitoes. The results of blast analyses indicated that 160 bp of 5′UTR sequences in BVDV belonged to BVDV type 1 and 384 bp of the VP6 gene sequence in RVA belonged to rotavirus A. In 2014, BLV was detected in YL‐A, and BLV and RVA were detected in Tainan City (TN). In 2015, BLV was detected in YL‐A, YL‐B, TN and KH, whereas BVDV was detected in YL‐A (Table 5).

Table 5.

Polymerase chain reaction (PCR) detections of three positive bovine non‐arboviruses.

		BLV						BVDV						RVA
Date		YL‐A	YL‐B	CY	TN	KH	PT	YL‐A	YL‐B	CY	TN	KH	PT	YL‐A	YL‐B	CY	TN	KH	PT
2014	January–March	•
	April–June	•			•												•
	July–September
	October–December
2015	January–March				•	•		•
	April–June	•	•		•
	July–September	•	•		•	•
	October–December				•

•, positive by PCR detection.

BLV, bovine leukaemia virus; BVDV, bovine viral diarrhoea virus; RVA, rotavirus A.

YL‐A and YL‐B are located in Yunlin County, CY in Chiayi County, TN in Tainan City, KH in Kaohsiung City, and PT in Pingtung County.

Discussion

Investigations into arbovirus vectors and their ecological characteristics are important for controlling arboviruses in the field, but the vector species of bovine arboviruses in Taiwan are still unknown. A comprehensive investigation conducted in dairy farms in Taiwan found 17 species of Culicoides biting midge (Chien, 2015). However, the present study identified only five species, although this may reflect the short duration of the period of collection of live specimens for virus detections. Culicoides oxystoma is a dominant species on dairy farms, and is also associated with paddy rice fields (Yanase et al., 2013). This species is identified as a vector of BEFV, AKAV, AINOV, PEAV and CHUV in Taiwan and Japan (Yanase et al., 2005). The present findings revealed an ornithophilic species, C. arakawae, which may also be a potential vector of BEFV, and its occurrence may be related to the dynamic of wild birds or to distance from a poultry farm (Yanase et al., 2005). Although target bovine arboviruses are considered to be mainly Culicoides‐borne viruses, mosquitoes may also play an important role in bovine arbovirus transmission. Culex tritaeniorhynchus may be a potential vector of BEFV, AINOV and CHUV. In a previous study, AKAV was isolated from Cx. tritaeniorhynchus in Japan (Yamakawa et al., 2006), but was not in the present study; however, AKAV was detected in Cx. quinquefasciatus in Taiwan. Thus, the susceptibility of Cx. quinquefasciatus and Cx. tritaeniorhynchus to Taiwanese AKAV strains requires further examination. Culex quinquefasciatus may be a potential vector of AINOV, whereas Cx. tritaeniorhynchus, Culex pipiens molestus and Culex fuscocephala are possible vectors of CHUV in Taiwan. The present authors recorded the first finding of Ar. subalbatus as a potential vector of AINOV and CHUV and propose that An. sinensis is a potential vector of BEFV and AKAV. The habitats of these mosquito vectors are associated with rice paddy fields, human residues and cattle dung (Ohba et al., 2013). Dairy farms in Taiwan are commonly located near other livestock farms or close to rice paddies, which provides these vector species with various ideal habitats.

The present MIR data indicate higher estimations in mosquitoes than in Culicoides biting midges (Table 4). There are no previous reports of MIRs of the five detected arboviruses, but the MIRs of Schmallenberg virus (Orthobunyavirus) and bluetongue virus (Orbivirus) were found to fall in the interval between 0.1 and 100 (Braks et al., 2017). However, MIRs are vulnerable to interference by intrinsic and extrinsic factors, such as the extrinsic incubation period in a vector, sampling population size, and the sensitivity of the detection method (Bustamante & Lord, 2010). Thus, the present group proposes to examine the vector competence of the potential vector species identified for detected arboviruses because most of these can be bred in the laboratory.

A nationwide programme of surveillance of AKAV, CHUV and IBAV showed their seroprevalences to be 83.95%, 67.78% and 80.77%, respectively (Ting et al., 2004), which suggests that these arboviruses might have become endemic after the first isolation in the 1990s. The current arbovirus detections demonstrate that both AKAV and CHUV, but not IBAV, were epidemic during 2012–2015. The present study used a different primer set designed for the Taiwanese IBAV strain (Table S1), but the result was the same as in the previous survey. It is possible that the vectors in the current collections may not carry IBAV in the field or that IBAV was not epidemic in 2012–2015; however, its occurrence needs further investigation.

Both AINOV and PEAV were first detected in Taiwan, although they have since been isolated or detected in Australia and Japan (St George et al., 1980; Yanase et al., 2010). In the present study, the first detections of AINOV were in YL‐A in October 2012, but the main detections were recorded from the end of 2013 to the middle of 2014 in five sampling farms. However, PEAV was detected only in YL‐A in November 2012, and in 2013 and 2015, and this temporal and geographic profile of PEAV requires further study. Some reassortment between AINOV and PEAV may occur in nature (Yanase et al., 2010), although this phenomenon has not yet been observed by cloning their structure genes (data not shown). In 2012–2013, BEFV broke out in both the Yaeyama Islands, Japan, and Taiwan; BEFV may have been transferred from the Yaeyama Islands to eastern Taiwan through an airstream (Ting et al., 2014; Hayama et al., 2016). The initial outbreak of BEFV occurred in September 2012 in eastern Taiwan, but the virus soon spread to western Taiwan via Pingtung County; this temporospatial pattern of occurrence is in line with predictions made by airstream simulation (Hayama et al., 2016). Another viewpoint indicated that BEFV may spread in a seasonal direction from southern China to southwestern Taiwan (Walker & Klement, 2015). Therefore, phylogenetic and biogeographic analyses of vectors and pathogens are essential for understanding temporospatial patterns of occurrence of bovine arboviral diseases in terms of the influences of geographic distributions and seasonal atmospheric conditions in countries of eastern and southeastern Asia.

Bovine leukaemia virus is a blood‐borne pathogen that is responsible for longterm viraemia in cattle and has been diagnosed earlier in Taiwan (Wang, 1991; Barez et al., 2015). Both characteristics are useful for elucidating the feasibility of vector‐based approaches in dairy farms. The present results showed a positive rate of 18.75%, but the conversion of data between regular serum investigations and vector‐based approaches requires further study. The vector‐based approach may also be applicable for the surveillance of other pathogens in cattle blood, such as Anaplasma spp., Babesia spp. and Trypanosoma spp. The detection of BVDV and RVA demonstrated that a vector‐based approach can also be applied to transient viraemia viruses. Rotavirus A is abundant in cattle faeces, and although the contamination by faeces of vector specimens could not be completely excluded in this study, RT‐PCR analyses in three other pools collected at the same time from the same trap showed negative results.

The present study is the first to report the use of engorged vectors (i.e. mosquitoes and biting midges) as tools for the surveillance of non‐arboviruses in dairy farms. The vector‐based approach provides rapid results by RT‐PCR detections within 24 h and allows for a high frequency of sampling. It also reduces the pain and possible distress associated with blood sampling in cattle. Furthermore, users of this approach do not need professional expertise to identify vector species if the targets of analysis are non‐arboviruses, and the pooling of different vector species will not affect the detection of non‐arboviruses. However, the current vector‐based approach requires further improvement if it is to be applied in the field because the efficiency of this approach must be comparable with that of other sampling methods, such as blood or milk sampling, and other details of this approach, such as the parameters of trap setting, limitation of collection time and the effects of blood digested in the midguts of vectors, also require further investigation. Therefore, the present authors suggest that the vector‐based approach has the potential to be used as a surveillance tool to increase understanding of epidemics of both arboviruses and non‐arboviruses on livestock farms.

Supporting information

Table S1. Primer sets used in this study.

Table S2. Polymerase chain reaction conditions in this study.