Identification of Host Blood Meals of Mosquitoes (Diptera: Culicidae) Collected at the Aripo Savannas Scientific Reserve in Trinidad, West Indies

Abstract

Surveillance for blood-fed female mosquitoes was performed between August 2015 and February 2016 at sites along the periphery of the Aripo Savannas Environmentally Reserve (ASSR) located in northeastern Trinidad, West Indies. We collected engorged female mosquitoes representing 13 species. DNA extractions from dissected abdomens were subjected to PCR amplification with three primer pairs targeting the mitochondrial cytochrome oxidase I and cytochrome b gene sequences. High-quality sequence information and host identification were obtained for 42 specimens representing eight mosquito species with at least one primer combination. A broad range of vertebrates including humans were identified, but the majority were nonhuman mammals, both domestic and wild. Domestic dogs were the most common host and may represent potential sentinel species for monitoring local enzootic arbovirus activity in Trinidad. Culex declarator Dyer and Knab and Culex nigripalpus Theobald were the most common blood-fed mosquito species comprising 79.1% of the total number identified. These species obtained blood meals from birds, nonhuman mammals, and human hosts, and therefore pose significant risks as potential bridge vectors for epizootic arbovirus transmission in the ASSR area as well as other sylvan areas in Trinidad. These data represent the first such results for Trinidad.

Key knowledge for informing on the potential for zoonotic arboviral threats can be obtained from identification of and surveillance for enzootic mosquito vector species at likely geographic foci for human, animal, and mosquito contact. Trinidad, West Indies, has a rich history of pioneering arbovirus and mosquito research dating back to the 1950s with collaboration between the Trinidad Regional Virus Laboratory (TVRL) and the Rockefeller Foundation. These efforts (Downs et al. 1955; Aitken et al. 1964, 1968, 1969; Jonkers et al. 1968; Worth et al. 1968) and subsequent research by the Caribbean Epidemiology Centre (CAREC) and others (Price 1978, Tikasingh et al. 1983, Auguste et al. 2010, Thompson et al. 2012) identified and characterized a large number of viruses endemic to forested areas both well-known and obscure or novel, their mosquito vectors, and vertebrate hosts. However, little effort has been directed toward determining bloodmeal host specificity of individual mosquito species.

Here we explored the potential for host bloodmeal identification at locations around the Aripo Savannas Scientific Reserve (ASSR) in northeast Trinidad where local human populations would be at risk for epizootic arbovirus transmission. Our goal was to obtain information on the host blood-feeding patterns of mosquito species in this area, particularly the identification of both the vertebrate and mosquito species demonstrating the greatest potential for promoting epizootic transmission and associated arbovirus disease outbreaks. Determination of the bloodmeal host range for individual mosquito species in such areas can facilitate identification of key species for targeted monitoring and subsequent prediction of arboviral outbreaks in human populations. Host identification can be determined based on sequence analysis of DNA extracted from host blood dissected from engorged mosquitoes. The most commonly sequenced targets have been mitochondrial genes due to their high-copy number and frequent taxon DNA sequence specificity (Crabtree et al. 2013, Börstler et al. 2016, Brugman et al. 2017). This study represents the first attempt to characterize associations between individual vertebrate hosts and the mosquito species that blood feed on them in Trinidad.

Materials and Methods

Study Site

Mosquitoes were collected at three locations along the periphery of the ASSR of Trinidad (Fig. 1) between August 2015 and February 2016. ASSR covers about 18 km², is one of the few remaining natural savanna ecosystems in Trinidad, and was declared a protected area by the Trinidad and Tobago government (Environmental Management Authority 2007). There are three main habitats at the ASSR: open savanna, marsh forest, and palm marsh. ASSR has been shown to support a diverse range of flora (Baksh-Comeau et al. 2016), as well as insect and vertebrate fauna (Schwab 1988). Although illegal, human intervention such as hunting, hiking, and bird watching within the reserve has been common, and human settlements are located around the periphery as well (Fig. 1).

Fig. 1.

The Aripo Savannas Scientific Reserve (ASSR) location within Trinidad. (A) Locations (stars) of sentinel traps around the periphery; (B) ASSR location in Trinidad.

Mosquito Collection and Species Identification

Mosquitoes were collected using three trap types distributed in five habitat types: residential, palm marsh, open savanna, marsh forest, and palm marsh/savanna ecotone at ASSR (Supp Table 1 [online only]). These included 1) CDC miniature light traps baited with CO₂; 2) Trinidad #17 Traps baited with CO₂ (Davies 1971); and 3) BG-Sentinel traps with BG-Lure. Trap numbers H1, H2, C2–C8 were located at the southwest trap site and included four CDC traps and five #17 traps; V2–V7 were located at the north trap site and included three CDC traps and four #17 traps; V1 was located at the southeast trap site with a CDC trap (Fig. 1). Collections were performed for ~24 h periods weekly beginning in August 2015 and ending in February 2016, with the exception of December 2015 (no collections made) as indicated in Supp Table S1 [online only]. Single BG-Sentinel traps were placed at the H2 and V5 sites for the 2016 collections only.

Individual collections from each trap were placed in ~500-ml plastic cups with mesh covering and placed on dry ice for transport to the laboratory in the Department of Life Sciences at the University of the West Indies. Engorged females were separated, and species identification was based on microscopic examination of morphological characteristics with standard taxonomic keys (Lane 1953, Bram 1967, CAREC 2002, WRBU 2018). Thereafter, individual specimens were placed in 2.0-ml cryovials and stored at −80°C until shipment on dry ice to the University of Notre Dame for bloodmeal analyses, where they were again stored at −80°C.

DNA Extraction and Bloodmeal Identification

Abdomens were removed from individual females and subjected to standard alkaline DNA extraction to a final volume of 50 µl with 0.06 M NaOH and 0.11 M Tris–HCl, pH 8.0 (Rudbeck and Dissing 1998). PCR amplification was first attempted using a cytochrome oxidase I (COI) multiplexed primer set as described elsewhere (Ivanova et al. 2007, Crabtree et al. 2013). If samples failed to amplify with COI, amplification was attempted with two cytochrome b (Cytb) primer sets following Börstler et al. (2016). Samples were visualized via 1.5% gel electrophoresis, and successfully amplified products were purified using the EZNA Cycle Pure PCR kit following the manufacturer’s instructions (Omega Bio-Tek, Norcross, GA). Samples were sequenced on a 3730 xl DNA Analyzer (Applied Biosystems, Foster City, CA) using the same PCR primers with BigDye (ThermoFisher Scientific, Waltham, MA) by the Genomics and Bioinformatics Core Facility (http://genomics.nd.edu/genomics-bioinformatics-core-facility) at the University of Notre Dame. Sequences were analyzed utilizing SequenceScanner (Applied Biosystems) and Geneious (Biomatters) software. COI gene sequences were identified using the Barcode of Life Database (http://barcodeoflife.org). Cytb sequences were similarly identified using the NCBI GenBank database (http://blast.ncbi.nlm.nih.gov/Blast.cgi). The Fisher’s exact probability test was used to determine whether any bias in amplification success was evident between COI and Cytb primers with respect to mammal versus non-mammal hosts (http://www.vASSRrstats.net).

Results

PCR amplification and sequence data were obtained for at least one of the three mtDNA primer sets for 42 of the 88 blood-engorged females collected (Supp Table 2 [online only]). A significant majority (n = 33, P < 0.000003) of these were with the Cytb primers reflecting a bias in successful amplification of mammal host species, whereas COI primers reflected a bias towards non-mammal host species. Sequence matches of 99–100% sequence identity were considered species-level identification, subject to confirmation that the species had previously been identified as resident in Trinidad (http://sta.uwi.edu/fst/lifesciences/animals-trinidad-tobago). Sequence matches of less than 99% were subject to identification at the most confident taxon level. Detailed information on individual specimens, including sequence identities, is provided in Supp Table S2 (online only). Host bloodmeal identities were determined for eight mosquito species that included 36 to species level, one to genus level, four to order level, and one to class level (Table 1). The most common hosts were domestic dogs (n = 15, 35.7%). Host range was broad and included a diversity of taxa representing humans, amphibians, birds, nonhuman mammals, and reptiles (Table 2). However, the majority of these across all mosquito species were from birds (n = 9, 21.4%) and nonhuman mammals (n = 27, 64.3%). Culex declarator and Culex nigripalpus were the most common engorged specimens collected, comprising 40.5 and 28.6% of the total, respectively. For Cx. declarator and Cx. nigripalpus, nonhuman mammals were the main hosts (76.2 and 50.0%), followed by birds (14.3 and 42.9%). Humans were minor hosts for both as well (4.8 and 7.1%).

Table 1.

Frequency of individual mosquito species with host category and most confident taxon identification

Category	Mosquito species	No. specimens	Taxon level	Identification	Common name
Amphibian	Culex amazonensis Lutz	1	Order	Anura	Frog
Bird	Culex declarator Dyar and Knab	1	Genus	Butorides	Heron
		1	Species	Dendrocygna autumnalis Linnaeus (Anseriformes: Anatidae)	Black-bellied whistling duck
		1	Species	Gallus gallus Linneaus (Galliformis: Phasianidae)	Domestic chicken
	Culex nigripalpus Theobald	1	Species	Glaucidium brasilianum Gmelin (Strigiformes: Strigidae)	Ferruginous pygmy owl
		1	Species	Thamnophilus doliatus Linnaeus (Passeriformes: Thamnophilidae)	Barred antshrike
		4	Species	Dendrocygna autumnalis	Black-bellied whistling duck
Human	Aedes serratus Theobald	1	Species	Homo sapiens Linneaus (Primates: Hominidae)	Human
	Culex declarator	1	Species	Homo sapiens	Human
	Culex nigripalpus	1	Species	Homo sapiens	Human
	Culex pedroi Sirivanakarn and Belkin	1	Species	Homo sapiens	Human
Nonhuman mammal	Culex caudelli Dyer and Knab	1	Species	Caluromys philander Linneaus (Mammalia: Didelphidae)	Bare-tailed woolly opossum
		1	Species	Canis lupus Linneaus (Mammalia: Canidae)	Domestic dog
	Culex declarator	1	Species	Dasyprocta leporina Linnaeus (Mammalia: Dasyproctidae)	Red-rumped agouti
		1	Species	Felis catus Linneaus (Mammalia: Felinae)	Domestic cat
		1	Class	Mammalia	Mammal
		2	Order	Rodentia	Rodent
		11	Species	Canis lupis familiaris	Domestic dog
	Culex nigripalpus	1	Species	Caluromys philander	Bare-tailed woolly opossum
		1	Species	Canis lupis familiaris	Domestic dog
		1	Species	Mazama americana Erxleben (Mammalia: Cervidae)	Red brocket
		1	Order	Rodentia	Rodent
		1	Species	Uroderma bilobatum Peters (Mammalia: Phyllostomidae)	Tent-making bat
		2	Species	Bos taurus Linnaeus (Mammalia: Bovidae)	Cow
	Culex spissipes Theobald	1	Species	Canis lupis familiaris	Domestic dog
	Psorophora albipes Theobald	1	Species	Felis catus	Domestic cat
Reptile	Culex declarator	1	Species	Rhinoclemmys punctularia Daudin (Reptilia: Geoemydidae)	Spotted-legged turtle

Table 2.

Number of blood-fed specimens collected and successful host DNA identification groups based on mtDNA COI or cytochrome b gene analysis

Mosquito species	No. specimens collected	No. blood meals identified	Birds	Human	Nonhuman mammals	Amphibians	Reptiles
Culex declarator	34	21	3	1	16		1
Culex nigripalpus	24	14	6	1	7
Aedes serratus	6	1		1
Culex caudelli	6	2			2
Culex pedroi	5	1		1
Mansonia titillans Theobald	2	0
Aedes hastatus Dyer	1	0
Culex amazonensis	1	1				1
Culex portesi Senevet and Abonnenc	1	0
Culex spissipes	1	1			1
Psorophora albipes	1	1			1
Psorophora lutzii Theobald	1	0
Total	83	42	9	4	27	1	1

Discussion

Here we examined the host range of blood-engorged mosquitoes collected around the ASSR in northeast Trinidad at locations most likely to bring humans under potential threat for enzootic arbovirus transmission. We were successful in determining the vertebrate host identities for 42 blood-fed mosquitoes representing eight species based on mtDNA gene sequence analysis. Although a very broad range of vertebrate species including humans were identified, the majority were nonhuman mammals, both domestic and wild. Our identification of domestic dogs as the most common host is not unprecedented (Molaei et al. 2007), and multiple studies have reported high seroprevalence for multiple arboviruses in dogs (Kile et al. 2005, Rocheleau et al. 2017, Thongyuan and Kittayapong 2017). Although dogs probably represent nonamplifying (or dead end) arbovirus hosts (Godsey et al. 1988), they do present strong potential as sentinel species for monitoring local enzootic arbovirus activity (Rocheleau et al. 2017).

As the predominant blood-fed species collected from the ASSR, Cx. declarator and Cx. nigripalpus are of particular interest to human health, as these two species obtained blood meals from birds, nonhuman mammals, and human hosts, and therefore pose significant risks as potential bridge vectors for enzootic arbovirus transmission in the ASSR area as well as other sylvan areas in Trinidad. That is, the broad host range we observed for these two species means they represent threats as bridge vectors for pathogens that typically circulate among nonhuman mammal species as well as those pathogens known to circulate between birds and nonhuman mammals. Of note, a novel Bunyavirus, Cumuto (CUMV), was recently isolated from Cx. declarator specimens collected along the ASSR boundary (Auguste et al. 2014), and although the vertebrate host is unknown, it would most likely circulate in nonhuman mammals. Furthermore, eastern equine encephalitis virus (EEEV) was previously isolated from Cx. nigripalpus specimens in a region northeast of ASSR (Downs et al. 1959) and is known to circulate across vertebrate groups but most generally among bird species. Both mosquito species are widely distributed across the neotropics (Gaffigan et al. 2018) and have been documented in the southern United States (Shaman et al. 2003, Darsie and Shroyer 2004). St. Louis encephalitis virus (SLEV) was repeated isolated in the area around ASSR during 1953–1962 from multiple mosquito species, including Cx. nigripalpus (Aitken et al. 1964). Of note, Cx. nigripalpus was the major mosquito vector for epizootic transmission of SLEV during an epidemic outbreak in Florida in 1990 (Shaman et al. 2003). More recently, serological studies have confirmed West Nile virus, as well as Venezuelan equine encephalitis virus (VEEV), EEEV, and SLEV as circulating in Trinidad (Thompson et al. 2012).

A previous investigation of arbovirus diversity in mosquitoes collected using CO₂ baited CDC light traps at sites in northeastern Trinidad during 2007–2009, including ASSR, reported that the majority (~56%) of the specimens collected were within the Culex subgenus Melanoconion, and included Culex portesi (~30%) and Culex pedroi (~13%; Auguste et al. 2010). Furthermore, ~92% of the arbovirus isolations were from Cx. portesi, primarily with Mucambo virus. However, here we collected only small numbers of blood-fed Cx. portesi (n = 6) and Cx. pedroi (n = 1) at sites around the ASSR where human/mosquito interactions would be most likely, but were only successful in determining the host in a single Cx. portesi specimen, which was human. Both species have been identified as among the most common potential enzootic vectors for multiple arboviruses in the neotropics across numerous studies and confirmed as vectors for EEEV and VEEV (Cx. pedroi) and western equine encephalitis virus Cx. portesi (reviewed in Hutchings et al. 2018).

In summary, our results suggest that based on blood-feeding activities in Cx. declarator and Cx. nigripalpus, these species may represent the greatest threat for enzootic arbovirus transmission in key contact areas around ASSR. Further research in these areas and others across Trinidad are needed to determine how best to integrate and interpret surveillance data on total mosquito collections versus our focus here on only blood-engorged specimens collected in high threat areas.